@

5802; ; Material Nonlinear Analysis - TEMP(LOAD) - Superelements vs. Non

-Superelements
If superelements are used in a nonlinear run with temperature dependent
material properties, the superelement properties are calculated using the
TEMP(INIT), rather than the TEMP(LOAD). This is a limitation since
superelements are linear, their material properties are not changed during
the analysis. This can lead to different answers when comparing
superelement models to non-superelement models.
Avoidance: In order to use the TEMP(LOAD) for the material properties, use
the following DMAP alter in Version 68:
COMPILE SELG, SOUIN=MSCSOU, NOLIST, NOREF
ALTER 11,11
SSG1 SLT,BGPDTS,CSTMS,SILS,EST,MPTS,ETT,EDT,MJJ,CASESX,
DIT,,,,/PJX,PTELEM,/LUSETS/1//'NLST' $
@
5803; 69; Miscellaneous - GINO - SYSTEM FATAL ERROR (bio_c.read)
UNIX versions of MSC/NASTRAN Version 68 and MSC/EMAS Version 3 may report
a SYSTEM FATAL ERROR (bio_c.read) in the LOG file but will complete the
job normally.
Avoidance: No avoidance is necessary, the error message can be ignored.
@
5804; 69; Miscellaneous - RC File - NASTRAN OLDQ4K, NSINIT ERROR 20, UFM 1
7
"NSINIT ERROR 20 *** SYNTAX ERROR IN CONTROL FILE"or "UFM 17
UNIDENTIFIED NASTRAN CARD KEYWORD" occurs if NASTRAN OLDQ4K is placed in a
runtime configuration file or the data deck. Refer to section 3.4 of the
MSC/NASTRAN Version 68 Installation and Operation Instructions for
information on the location and format of the runtime configuration files.
Avoidance: Two avoidances exist. Either OLDQ4K can not be the first
system cell on the NASTRAN command line or the equivalent system cell,
SYSTEM(173), must be used in place of the keyword.
@
5805; ; Nonlinear Transient Analysis - NLTRD, RESTART - SFM 4276 EC 1044
, UIM 4550
IF a model containing superelements is executed in solution 106, and then
restarted into solution 129 (using PARAM, SLOOPID), and the "METHOD" field
on the TSTEPNL entry is set to AUTO, then the job fails and the following
fatal message is issued:
SYSTEM FATAL MESSAGE 4276 ERROR CODE 1044 (BIOMSG: ERROR 1044 HAS
OCCURRED IN ROUTINE OPEN FILE INDEX (FILEX) = 0)
Immediately preceeding this fatal message, USER INFORMATION MESSAGE 4550,
NEW STIFFNESS MATRIX REQUIRED is issued. This error may happen even
though the nonlinear iteration module output indicates that all time-steps
have been successfully completed.
 Avoidance: Use the default ADAPT method.
@
5807; 70; DMAP - COMPILE - UFM 423
If a user attempts to compile a user dmap where the name of the dmap
begins with the letter "E", the job may fail with the message "USER FATAL
MESSAGE 423 (ATTDBS) THE FOLLOWING PHYSICAL FILE DOES NOT EXIST
/msc/msc68/emas3/sunos/del/ESS.MASTERA".
Avoidance: Change the main dmap and SOL name so that it does not begin
with "E". This apparent error is the result of an MSC/NASTRAN feature
that will run MSC/EMAS if a deck with a 'SOL 300' or 'SOL E*' entry is
submitted. If such a SOL entry is found, the script attempts to attach
the MSC/EMAS Solution Sequences. This delivery database does not exist in
a pure MSC/NASTRAN installation.

@
5808; 68.2; Differential Stiffness and Buckling - EIGRL, EIGB - Wrong Answe
rs or UFM 4683
If the a-set and g-set are the same (no SPCs, MPCs, or reduction) in any
superelement or the residual structure, then the differential stiffness
matix used in solution 105 (buckling) or the buckling option in solution
200 will not be correct. The error may lead to wrong answers or "USER
FATAL MESSAGE 4683, DIFF. STIFFNESS MATRIX NEEDED FOR EIGENVALUE
ANALYSIS".
DMAP avoidance:
COMPILE SEKDR, SOUIN=MSCSOU, NOLIST, NOREF $
ALTER 42, 42 $
EQUIVX KDGGX/KDAA/-1 $
@
5819; 69; Static Analysis - TEMP(LOAD), LSEQ - UFM 9058
If the LOADSET, LSEQ combination and TEMP(LOAD) are specified in
MSC/NASTRAN Version 68, then "USER FATAL MESSAGE 9058" may be issued in
SOL 101. This error occurs when there are unreferenced LSEQ entries.
Avoidance: Remove the unreferenced LSEQ entries or insert the following
alter:
compile sestatic souin=mscsou noref nolist
alter 59 $
PARAML CASEsx//'DTI'/nskip/7//S,N,TEMPLD $
@
5820; 68.2; Optimization/Design Sensitivity - ANALYSIS=SAERO - Wrong Answers
Applied loads, for example to simulate subcase dependent fuel loads, are
ignored and wrong answers are produced if a static aeroelastic design task
is submitted in solution 200.
Avoidance: Insert the following DMAP avoidance:
compile phase1dr souin=mscsou nolist noref
alter 81,81
alter 89
equivx casesx/casew/-1 $
@
5823; 68.1; Inertia Relief - Multiple Subcases, SUPORT1 entries - Wrong SPC
forces
If inertia relief is used in conjunction with multiple subcases each
referencing a different set of SUPORT DOFs, then the spcforces will be
incorrect. This is a Version 68 error since prior versions do not support
multiple inertia relief sets. This feature is activated with the use of
the SUPORT1 entries.
Avoidance: Insert the following DMAP alter avoidance.
compile sestatic souin=mscsou nolist noref
alter 100,100 $
CALL SUPER3 CASECC, ,UL, , , , , , , ,
CRX, , ,,,,
PCDB,XYCDB,POSTCDB,FORCE,EMAP,MAPS,EQEXINS,
PVTS,CASES,SLT,ETT,GOAT,,LOO,CRJ,LAO,UOO,UOX,GPLS,USET,
SILS,PJ,EDT,YS,GM,PSS,KFS,KSS,QR,,,,,,
DYNAMICS,BGPDTS,CSTMS,MPTS,DIT,EST,GEOM2S,GEOM3S,ECTS,
EPTS,INDTA,KELM,KDICT,GPECT,VELEM,,SILX,EQEXINX,ECTX,
BGPDTX,,GEOM4S,PG,,,,EPT,MPT,GEOM1,GEOM2,GEOM3,GEOM4,
PTELEM,,,
,,MEDGE,VIEWTB,ESTVU,GEOM1VU,SILVU,
CSTMVU,DEQATN,DEQIND,GDNTAB,OINT,PELSETS,
KSF,KMG,,,,,,,,,,,,/
UG,QG,CASEDR,XYCDBDR,PUG,GPSF, ,
OUG1,OPG1,OQG1,OEF1,OES1,OEE1,OGPS1,
OUG2,OPG2,OQG2,OEF2,OES2,OEE2,OGPS2,
OESM1,OESG1,OEEM1,OEEG1,GPS,ONRGY,OGPFB/
APP/APP/RSONLY/FALSE /FALSE /FALSE/DROPT/FALSE/
S,PFILE/S,CARDNO/0/FALSE/0/PEXIST/
ADPTINDX/ADPTEXIT $
$

@
5829; 69; Plotting - SEUPPLOT - Missing Plot Output
In SOL 103, if the SEUPPLOT command is used to generate assembly plots,
plots will be available for only the first half of the modes in the
solution.

Avoidance: Apply the following DMAP avoidance in Version 68:

compile semodes, souin=mscsou, nolist, noref
alter 1
$ disable scratch LTU logic - otherwise the datablocks will
$ not be there when needed for plotting
putsys(0,192)
$
compile sedrcvr, souin=mscsou, list, noref
alter 1
type db,phidlg $unused datablocks in SOL 103
$
alter 48 $ after sdr2
copy ug/ugcopy $
append ug,ugcopy/ugtwice $ get "double" set of solutions
$ SDR2 to get "double plot vectors"
 message //'seid='/seid $
SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,,ETT,OL2,BGPDP,
PJ1,QG,UGtwice,EST ,XYCDBDR,OINT,PELSETS,VIEWTB/
,,OUGVxxx,,,phidlg/APP1/S,N,NOSORT2xx/
NOCOMPS/////ACOUT/PREFDB/-1./-1./
ADPTINDX/ADPTEXIT/FALSE $
$
compile super3, souin=mscsou, list, noref
alter 227,227 $ replace PLTVEC
type db,phidlg
dbdict datablk=(phidlg,phidrg,pug),
select(name,size,seid,peid,zname,version) $
DBVIEW PhidlgF=Phidlg (WHERE SEID=* AND WILDCARD=TRUE) $
PLTVEC EMAP,DRLIST,PCDB,EQEXINX,SILX,SILSF,GPLSF,phidlgf/
PUGX,PCDBX/SEID/'SEID' $
@
5831; 68.2; DMAP - MATMOD, FBS, TRNSP - Wrong Answers, SFM

There is a special form of an identity matrix called a form 8.

Conventional identity matrices are form 6. This type of matrix can be
generated with MATMOD option 2, or may be generated internally in modules.
The new sparse options may have trouble with this form of matrix.
Troubles observed to date include:
Wrong answers from FBS when more than one pass is required. The symptoms
for the FBS problem found to date include a solution which has about 50%
more non-zero terms than it should.
A SYSTEM FATAL MESSAGE in the TRNSP module.
Avoidance: The error in FBS may be avoided by switching back to the non
sparse methods by setting SYSTEM(126)=1. It is recommended this be done
only locally around the offending module with PUTSYS statements. The
error in TRNSP may only be avoided by producing a form 6 identity matrix.
MATMOD option 1 outputs a conventional identity matrix.
@
5835; 68.0.2; Miscellaneous - RC File - Directory Does Not Exist
If a scratch directory is specified in an RC File that does not exist,
the run fails even though a directory that does exist is specified in a
subsequent RC File that is meant to take precedence over the first
specification.

Avoidance: Change the directory specification to a directory that exists.

@
5838; 69; Loads - TEMP(LOAD) - UWM 4012, UFM 4017
When doing a static analysis with temperature loads in Version 68, it is
necessary to specify the temperature for all grid points in the model. If
this is not done, the following messages will be issued, "USER WARNING
MESSAGE 4012, THERE IS NO ELEMENT, GRID POINT, OR DEFAULT TEMPERATURE DATA
FOR TEMPERATURE SET 102, WITH RESPECT TO ELEMENT ID = 0","USER FATAL
MESSAGE 4017. THERE IS NO TEMPERATURE DATA FOR ELEMENT 15 IN SET 2".
Unfortunately, the element ID and set numbers are incorrect.
Avoidance: Always include a TEMPD bulk data entry in any static analysis
with temperature loads.
@
5839; 68.2; Heat Transfer - heatconv, SPC - Cannot Convert Large Field Forma
t, Wrong Answers
The heatconv program does not correctly handle all large field SPC
entries. In particular, given the following large format SPC entry:
SPC* 1 1 1
2.00000000E+01
heatconv will incorrectly copy this to a small format entry as follows:
SPC 1 1 1 2.000000
Whenever the temperature fields, i.e., fields 5 & 8, exceed eight
characters in length, incorrect results will occur.

@
5841; 2001.0.1; Composites - PCOMP, MAT2 - UFM 4291
Certain MAT2 entries that are referenced from a PCOMP are not checked for
correctness. In particular, if 13 and/or 23 terms are non-zero and the 33
term is zero, the matrix is not positive definite. In this case
MSC/NASTRAN will create potentially incorrect PSHELL/MAT2 combinations
from the PCOMP/MAT2 inputs.
Avoidance: Enter a positive definite MAT2 entry.
@
5843; 68.2; P-Elements - RFORCE - Wrong Answers or SFM 4276
Although it is not documented, RFORCEs are not implemented for p-elements.
If RFORCES are applied to p-elements either wrong answers or a SYSTEM
FATAL MESSAGE 4276 results.
Avoidance: Do not apply RFORCEs to p-elements.
@
5844; 69; P-Elements - HEX20, PENTA15, TETRA10 - Missing Documentation
The CHEXA element with 20 grids, the CPENTA element with 15 grids, and the
CTETRA element with 10 grids may not be used as p-elements by referencing
the PVAL entry. If attempted, the following message is issued:
*** USER FATAL MESSAGE 6390, (GP0A),
ELEMENT ID = 1 TYPE HEXA20 OR PENTA15 AND/OR TETRA10 ARE NOT
SUPPORTED IN P-ELEMENT ANALYSIS.
USER ACTION: CHECK ELEMENT CONNECTION ENTRY OR SET ENTRY.
@
5848; 68.2; Optimization/Design Sensitivity - DSCREEN - Poor Convergence
The DSCREEN entry permits the user to override default screening
parameters for design optimization. This capability does not work for
responses TDISP,TVELO,TACCL,TSPC,TSTRE,TFORC,TRIM,STABDER and FLUTTER.
Avoidance: The only avoidance is to do a TABEDIT on the DSCREN datablock.
The following words correspond to the data locations for the responses
 TRS NSTR
TDISP 179 180
TVELO 182 183
TACCL 185 186
TSPC 188 189
TSTRE 191 192
TFORC 194 195
TRIM 245 246
STABDER 248 249
FLUTTER 251 252
A candidate DMAP for performing the TABEDIT is:
compile desopt souin=mscsou nolist noref
alter 41 $
call dbstore dti,dtindx,,,//1/2/'dball'/0 $
compile predom souin=mscsou nolist noref
alter 12 $
call dbfetch /dti,dtindx,,,/1/2/0/0/0 $
dtiin dti,dtindx/control,,,,,,,,, $
tabedit dscren,control,,,/ndscren/-1 $
equivx ndscren/dscren/always $
$
$ the following dti modifies word 195 as required
$ for a TFORC request and nstr
$
dti,control,0
dti,control,1,cr,1,dw,195,195,1
,1
dti,control,2,er,endrec
$
@
5849; 68.1; Matrix Methods - FBS - UWM 4398, Wrong Answers
MSC/NASTRAN Version 67.7 and 68 on the CRAY C90 computer may produce wrong
answers for models where FBS method 2 is selected. The error may be
evidenced by "USER WARNING MESSAGE 4398" which indicates "DEGREES OF
FREEDOM DO NOT PASS SYMMETRY CHECK". The error only occurs when there are
strings in the factor that are longer than 64 terms. Note that except in
superelement models, that FBS method 2 is rarely selected.
Avoidance: FBS Method 2 should be deselected by setting "FBSOPT=-1" in
the system RC file on Version 67.7 and 68 CRAY C90 computers.
@
5850; 68.1; Miscellaneous - IFP, XRCARD - Infinite CPU Loop, Overflow Input
Buffer
Some invalid bulk data input may result in an infinite cpu loop in the
routine XRCARD. Although the exact type of input that results in this
error is not known, the specifics of the input line that resulted in this
error is that the line was over 80 characters, contained many leading
blanks and included an embedded negative, "-", sign. The infinite cpu
loop has only been reproduced on the CRAY computer but may occur on other
systems.
Avoidance: There is no practical avoidance.
@
 5852; ; Documentation - DMAP Module Dictionary, MODTRL - Bad Input Matri
x
If the MODTRL module is used to increase the number of columns in a
matrix, the resulting matrix is
unusable in most modules, including matprn and ADD. The cause of the
error is that the matrix does
not have column headers for the added null columns.
Avoidance: The MATGEN and ADD module should be used to increase the
number of columns in a matrix.
An example follows:
matgen ,/nullab/7/rn/cn $ null matrix with rn rows and cn cols. user
must set rn and cn above.
add nullab,oldmat/newmat $ trailer will match first input
equivx newmat/oldmat $ oldmat is now 'bigger'
@
5858; 68.2; Optimization/Design Sensitivity - SPCFORCE and MPCFORCE - Wrong
Answers
If scr=yes is used in SOL 200 with dynamic optimization or sensitivity,
then MPCFORCE and SPCFORCE output does not include mass and damping
effects after the first design cycle. (Version 68 only)
Avoidance: Specify scr=no. An extensive DMAP alter is also available if
scr=yes is desired.
@
5860; 69; Aeroelastic Analysis - GI, CAERO1, CAERO2 - Wrong Answers
MSC/NASTRAN Version 68 on Sun computers may produce wrong answers for
aeroelastic flutter problems when aerodynamic bodies, i.e. CAERO1 or
CAERO2 elements, are present.
Avoidance: There is no avoidance.

@
5861; 69; Dynamics - PARAM,MODACC - Poor Displacement Results
The mode acceleration method is intended to improve element stress data
recovery in dynamic solutions using modal formulations. However an error
may occur in the total displacements when free-body motions are present.
This error will NOT affect strains, stresses or constraint forces. The
major problem is that the angular displacements or pointing directions
will be affected. The problem occurs because the "improved" relative
displacements are not orthogonal to the free-body modes and create small
additional displacements and rotations. The error is proportional to the
local frexibility of the SUPORT degrees of freedom. This error is
considered fixed in Version 69 by the addition of a new mode acceleration
option which is selected by setting PARAM,MODACC,1 instead of
PARAM,MODACC,0.
The avoidance is to use conventional dynamic recovery methods in a modal
formulation when free-body modes are present. The "Residual Vector" method
is an accurate alternative.
@
5862; ; Database - DBUNLOAD - UFM 3007, illegal input to subroutine PVT
If the dbunload is done on a long word machine such as Cray and dbload is
 attemted on a short word machine, then UFM 3007, illegal input to
subroutine PVT may be issued. This error occurs only if the initial run
which created the databases had parameter specifications in the Case
Control. This occurs since the diliminator for the parameter in the Case
Control is machine dependent.
Avoidance: There are two avoidances:
1. Remove the parameter specifications from the initial run which creates
the databases.
or.
2. Use the following alter in the dbload run.
ACQUIRE NDDL
ASSIGN DBLOAD='neutrala.cray',FORM=FORMATTED
DBLOAD FORMAT=NEUTRAL
$
$ avoidance is to add the following alter $
$
SOL DBTRANS
TIME 5
compile dbtrans souin=mscsou list noref
alter 114,114
CEND
BEGIN BULK
ENDDATA
@
5863; 68.2; P-Elements - GP4 - UFM 4276, CSTM DOES NOT EXIST (GP4)
If a model contains p-elements and there is insufficient memory in the GP4
module the message "USER FATAL MESSAGE 4276, CSTM DOES NOT EXIST (GP4) may
be output.

Avoidence: Request more memory and rerun the job.

@
5864; 68.2; Aeroelastic Analysis - DIAG 50 - Unsplined Rigid Stability Deriv
atives
For Version 68, static aeroelastic analyses (SOLs 144 and/or 200) produce
Unsplined Rigid Stability Derivatives. These data are presented only to
provide a check of the splining and are not used in the solution. For
antisymmetric models and models with no symmetry (SYMXZ = -1 or 0), the
numbers may be in error. For sideforces the error is likely to be simply
one of sign. For yawing and rolling moments, the error is less obvious.

@
5866; 68.2; Heat Transfer - RADMTX - Duplicate Continuation Entry with Multi
ple Cavities
For the multiple cavity radiation view factor calculation, each cavity
will have a RADLST and RADMTX entry associated with it. The punch file
which contains the MSC/NASTRAN calculated RADMTX entries for each cavity
contains duplicate RADMTX continuation identifiers. Thus, subsequent
restarts using the punch file fail.
 Avoidance: Since continuation identifiers are not necessary if the input
is in sequential order, blank out all the continuation identifiers in
field 10 and field 1 of the RADMTX entries.
@
5867; 68.2; Mulitpoint Constraints - SPLINE2 - Singular Matrix, Wrong Answer
s
The Quick Reference Guide indicates that the DTOR value input on the
SPLINE2 entry must be greater than 0.0.
However, this requirement is never checked. If the user leaves DTOR field
blank the run could give wrong answers or fail due to a singular spline
matrix.
Avoidance: Specify a value greater than 0.0 for DTOR.
@
5868; 68.1; Installation - Authorization - UFM 3060, BUCKLING not in Approve
d List
An authorization code for the workstation option is suppose to include
statics, normal modes and buckling. An error in the authorization code
generator resulted in workstation options not including buckling.

Avoidance: The error in the authorization code generator has been

corrected. Contact MSC for a new code.
@
5870; 68.2; Multipoint Constraint - SPOINT - UFM 2101A DOF, Wrong Answers.
If the user inadvertently specifies a degree of freedom greater than 1 for
an SPOINT in an MPC equation, then wrong answers may result or the message
"USER FATAL MESSAGE 2101A, GRID POINT xxx COMPONENT yyy ILLEGALLY DEFINED
IN SETS zzz" may be output referring to the wrong grid point/degree of
freedom.
Avoidance: Other than correcting the SPOINT degree of freedom, there is no
avoidance.
@
5871; 68.2; Restarts - RESTART PROJECT - Inefficient Restart
In Version 68, the PROJECT keyword is ignored on the RESTART statement
which will result in an inefficient restart. For example,
1. The initial run creates project A:
PROJECT='A'
ASSIGN MASTER='A.MASTER', DELETE
2. The subsequent run restarts from project A into project B:
PROJ='B'
ASSIGN MASTER='A.MASTER'
RESTART PROJECT='A',KEEP
Avoidance: For SOLs 101 through 200 insert the following DMAP alter:
compile ifpl souin=mscsou noref nolist
alter 80,80
type parm,,i,,projid=1,versid=1 $
projid=1 $ set the projid to the integer equivalent of the project you are
restarting from.
versid=1 $ set the versid to the version you are restarting from $
dbdict datablk=geom1q $
 alter 118,131
DBVIEW GEOM1R=GEOM1Q(WHERE version=versid and PROJECT=PROJID) $
DBVIEW GEOM2R=GEOM2(WHERE version=versid and PROJECT=PROJID) $
DBVIEW GEOM3R=GEOM3(WHERE version=versid and PROJECT=PROJID) $
DBVIEW GEOM4R=GEOM4(WHERE version=versid and PROJECT=PROJID) $
DBVIEW EPTR=EPT(WHERE version=versid and PROJECT=PROJID) $
DBVIEW MPTR=MPT(WHERE version=versid and PROJECT=PROJID) $
DBVIEW EDTR=EDT(WHERE version=versid and PROJECT=PROJID) $
DBVIEW DITR=DIT(WHERE version=versid and PROJECT=PROJID) $
DBVIEW DYNAMICR=DYNAMIC(WHERE version=versid and PROJECT=PROJID) $
DBVIEW AXICR=AXIC(WHERE version=versid and PROJECT=PROJID) $
DBVIEW PVTR=PVT(WHERE version=versid and PROJECT=PROJID) $
DBVIEW MATPOOLR=MATPOOL (WHERE version=versid and PROJECT=PROJID) $
DBVIEW caseccrr=caseccr (WHERE version=versid and PROJECT=PROJID) $
DBVIEW EDOMR=EDOM WHERE (version=versid and PROJECT=PROJID) $
compile phase0 souin=mscsou noref nolist
alter 281,297
type parm,,i,,projid=1,versid=1 $
projid=1 $ set the projid to the integer equivalent of the project you are
restarting from.
versid=1 $ set the versid to the version you are restarting from $
DBVIEW CASESR=CASES (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW GEOM1SR=GEOM1S (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW GEOM2SR=GEOM2S (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW GEOM3SR=GEOM3S (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW USET0R=USET0 (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW RMGR=RMG (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW YGR=YG (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW DYNAMCSR=DYNAMICS (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW EPTSR=EPTS (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW MPTSR=MPTS (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW MAPSR=MAPS (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW PVTSR=PVTS (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW BACKR=BACK (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW HARMR=HARM (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW KVALR=KVAL (WHERE version=versid AND PROJECT=PROJID) $
DBVIEW FORER=FORE (WHERE version=versid AND PROJECT=PROJID) $
alter 352,352
DBVIEW EQEXINSR=EQEXINS (WHERE version=versid AND PROJECT=PROJID) $
@
5876; 69; File Management Section - EXPAND, RESTART - SFM 4276 XIO, EXPAND
, SCR300
An attempt to expand a database fails when restart is attempted in the
same run. A sample File Management Section demonstrating the error
follows:
assign dbup1a='scr7a.dbup1a'
expand dbup1, logical=(dbup1a(100))
restart version=1 keep
The sample input leads to the following fatal message due to the scr300
file being improperly defined:
"SYSTEM FATAL MESSAGE 1030 (XIO)
BLOCK NUMBER MISMATCH IN READ
SYSTEM FATAL ERROR 4276.XIO ERROR CODE 4276"
 Avoidance:
1) run an "expand" job only (i.e. expand and endjob)
2) run the restart job

@
5877; 67.5.0.1; Miscellaneous - SCO/UNIX - UFM 3060
The SCO/UNIX Version 67.5 of MSC/NASTRAN may issue "USER FATAL MESSAGE
3060" indicating authorization code violation during long running job with
many module changes (e.g. nonlinear).
Avoidance: There is no avoidance. This problem will be fixed on a
rerelease version (MSC/NASTRAN v67.5.0.1 for SCO/UNIX).
@
5878; 69; Sequencer - SEQP - SFM 3008, Insufficient Memory
A small p-element model ( 1 Hexap and 2 Pentap ) fails in the sequencer,
module SEQP, with the following error message from subroutine seqos:

SYSTEM FATAL MESSAGE 3008 MODULE = SEQP

0INSUFFICIENT MEMORY AVAILABLE FOR SUBROUTINE SEQOS
Avoidance: Avoid the sequencer by setting PARAM,NEWSEQ,-1.
@
5880; 68.2; P-Elements - TEMP - UFM 4017
All p-version models which contain thermal loads via TEMP entries and more
than one type of solid element (HEXA, PENTA, TETRA), fail with the
following fatal message, "USER FATAL MESSAGE 4017. THERE IS NO TEMPERATURE
DATA FOR ELEMENT y IN SET x".
Avoidance: If the thermal loading is a uniform temperature increase
throughout the entire structure, use a TEMPD Entry. Otherwise, the user
must use an equivalent TEMPF Entry. Also, it may be possible to use only
one element type in the model.
@
5881; 69; Documentation - PARAM,BAILOUT - Missing Documentation
MSC/NASTRAN Version 68 supports multiple boundary conditions in the
structured solution sequences. However, PARAM BAILOUT cannot be changed
between subcases.
@
5889; 68.2; Optimization/Design Sensitivity - DRESP2 - Wrong Answers
If a DRESP2 entry refers to two or more DRESP1 entries and there are
DVPREL2 entries in the input file, then the sensitivities printed in the
columns of the DSCM2 matrix associated with the DRESP2 entry are in error.

Avoidance: There is no avoidance other than to recast the problem so that

it does not require the offending DRESP2,DRESP1,DVPREL2 combination.
@
5891; 70.5.1; Elements - PENTA - Possible Wrong Answers, UFM 4654
If a pyramid-shaped degenerate PENTA element is input to MSC/NASTRAN, USER
 FATAL MESSAGE 4654 is issued warning of two corner points with the same
coordinates but the run will continue. Pyramid-shaped PENTA elements are
not supported and should not be used.
@
5893; 69; Sequencer - SEQP, PARAM,NEWSEQ,6 PARAM,OLDSEQ,6 - SFM 4276, EC 3
002
It is possible that jobs that request param,newseq,6 prior to Version 69
or param,oldseq,6 in Version 69 and after may encounter SYSTEM FATAL
MESSAGE 4276, ERROR CODE 3002 in the SEQP module.
By default sequencing in the SEQP module was turned off in Version 69 and
beyond. PARAM,NEWSEQ was changed to PARAM,OLDSEQ to force the user to go
out of his way to request sequencing. Thus, this error is considered
corrected in Version 69 by the fact that sequencing is no longer
recommended.
@
5894; 2001; Transient Response Analysis - TIC - Initial Conditions Not Refle
cted in Results
An initial velocity, V0, TIC (transient initial condition) is not
properly reflected in the Velocity (first time derivative) results. The
discrepancy between the value specified and the value output occurs
because the program uses a central difference approach. The displacements
are correct.
@
5896; 68.2; Miscellaneous - NASTRAN SYSTEM(i) - Bad Processing Order
MSC/NASTRAN processes systems cells specified by the format "NASTRAN
SYSTEM(i)=value" before those specified by "SYSTEM(i)=value". Thus system
cells specified in the latter format will take precedence regardless of
where they are specified.
Avoidance: Always use the format "NASTRAN SYSTEM(i)=value" in the system,
architecture, user, and local RC files.
@
5900; GL; Material Nonlinear Analysis - RESTART, NLPARM - DIM 9005, UFM 20
44, SFM 3007
If thermal loads are applied in a SOL 106 cold start, and, if in the
restart, PARAM,LOOPID is not the last loop id for the subcase prior to
PARAM, SUBID then the program will issue various misleading fatal
messages. In one example the message "USER FATAL MESSAGE 2044 UNDEFINED
TEMPERATURE SET, SYSTEM FATAL MESSAGE 3007 MODULE=NLCOMB ILLEGAL INPUT TO
SUBROUTINE NLCPTM" was encountered.
Avoidance: Other than specifying the correct LOOPID, there is no
avoidance.
@
5901; 68.2; Installation - Multiple Auth Codes - Only One Code in Auth File
If authorization codes for MSC/EMAS and MSC/NASTRAN are entered at
different times during the same installation, only the last code entered
is placed in the authorize.dat file.
Avoidance: Enter all codes for all products at the same time in the
installation. The installation script allows you to enter a series of
codes. If one of the products is using an existing authorize.dat file,
modify the appropriate <install-dir>/conf/"rc file" to have the AUTH
 keyword point to the correct file.
@
5911; 68.2; Complex Eigenvalue Analysis - SPCForces - Wrong answers
In MSC/NASTRAN Version 68, the SPCForce results are incorrect in SOLs 107,
110, and 145 only. The SPCForces do not include the effects of the
imaginary part of the eigenvalue which results in "undamped" SPCforces.
Avoidance: Insert the following DMAP alter.
COMPILE CFORCE SOUIN=MSCSOU noLIST NOREF $
ALTER 47,47 $ replace DSAP
PARAML OL//'TRAILER'/1/S,N,NFREQS $ NUMBER OF Eigenvalues
TYPE PARM,,I,N,II=0,iip3,iip4 $ LOOP COUNTER
TYPE PARM,,CS,N,IOMEGAX,MOMEGX $
FILE qx1=APPEND $
DO WHILE (II<NFREQS) $
II = II + 1
iip3=6*(ii-1)+3 $ clama word 3
iip4=6*(ii-1)+4 $ clama word 4
PARAML OL//'DTI'/2/IIP3/S,N,omega $ real eigenvalue
PARAML OL//'DTI'/2/IIP4/S,N,omegai $ imag eigenvalue
IOMEGAX = CMPLX(omega,OMEGAi) $
MOMEGX = IOMEGAX**2 $
MATMOD KUx,,,,,/KI,/1/II $ COLUMN EXTRACTION
MATMOD BUx,,,,,/BI,/1/II $
MATMOD MUx,,,,,/MI,/1/II $
ADD5 KI,BI,MI,,/FI/1.0/IOMEGAX/MOMEGX $
APPEND FI,/qx1/2 $
ENDDO $ NFREQS
@
5914; 68.2; Nonlinear Transient Analysis - PARAM,G, PARAM,W3 - Incorrect St
ructural Damping
In MSC/NASTRAN Version 68, solution 129, the structural damping for linear
elements specified by parameters G and W3 is doubled. The structural
damping specified by GE and parameter W4 are correct. (V68 only)
Avoidance: Apply the following DMAP alter:
compile nltran, souin=mscsou, nolist, noref $
alter 224, 224 $
ADD5 BDD0,KDDNL,K4DDNL,,/BRDD//GW3/RW4 $
endalter $
@
5915; 69; Case Control - GPSDCON, ELSDCON - Missing Output
Element and grid point stress discontinuity output (ELSDCON and GPSDCON
Case Control commands) cannot be requested for a single subcase or a set
of subcases. This output is only generated if the request is made for all
subcases in Case Control.
Avoidance: The user must place these commands above the subcase level or
in each subcase.
@
5918; 69; Complex Eigenvalue Analysis - CEAD - Poor Message, UFM 4215
 When running a complex eigenvalue analysis, the message "USER FATAL ERROR
4215. COLUMN 1 OF DATA BLOCK SCRATCH IS NULL BUT ROW 1 IS NOT" may be
issued. However, this null column does not affect the solution.
Avoidance: There is no avoidance.
@
5920; 68.2; Transient Heat Transfer - MATT4,TABLEM4 - SFM 2090
Contrary to the documentation , the temperature dependent specific heat,
T(CP), on the MATT4 entry, cannot be specified with the TABLEM4 entry
(coefficient of a power series). If this definition is attempted, the
message "SYSTEM FATAL MESSAGE 2090, TABLE DICTIONARY ENTRY 201 MISSING" is
output.
Avoidance: Use the TABLEM1, TABLEM2 or TABLEM3 entry to specify
temperature dependent specific heat.
@
5921; 70; Nonlinear Analysis - TEMPP1, TEMPP3 - Wrong Answers
If the user specifies a temperature gradient using the TEMPP1 or TEMPP3
entry in a nonlinear static thermal stress analysis, then the TEMP(INIT)
case control command treats the thermal gradient as zero despite the value
specified by the user for the thermal gradient.
Avoidance: The user may avoid this problem if the thermal expansion
coefficient is not temperature dependent. In this case, the thermal load
from the thermal gradients depends only on the difference of the
thermal gradients. The user should, on the TEMPP1 entry for TEMP(INIT)
place a 0.0 or blank in the T' field, and on the TEMPP1 entry for
TEMP(LOAD) place T'(load) - T'(init) in the T' field.
@
5922; 69; Miscellaneous - ALPHA OSF/1 Only - UFM 1012 DBSET DBALL IS FULL
On the ALPHA OSF/1 release of MSC/NASTRAN when the system runs out of
allocated disk space the job may hang. When the process is killed, the
system sends the message that the job is finished and the output files
look as they should with no indication of a problem. This error has also
been observed in connection with SFM 4393 followed by SFM 4276 Error Code
3002.
Avoidance: Other than killing the job, there is no avoidance.
@
5923; ; Miscellaneous - DIAG 49, F04 Analyze - Missing Output, SWM 1007,
LOGIC ERROR 4
Large models that request DIAG 49 (f04 performance summary) may fail and
issue the following message in the f04 (Execution Summary) file. Models
without DIAG 49 have also less frequently encountered this error.

SYSTEM WARNING MESSAGE 1007 (F04SUM)

LOGIC ERROR 4 WITH FOLLOWING RECORD(S):
or
SYSTEM FATAL MESSAGE - bio_c.sysrd.read
A BIO operation on a GINO file has failed.
File descriptor: 7.
Result: 49152.
 The error occurs whenever the file size is smaller than the buffsize.
Avoidance: Other than running Version 67.5 there is no avoidance.
@
5924; 68.2; Nonlinear Transient Heat Transfer - TEMPBC - SFM 4276 EC 3021
If a TEMPBC entry of TYPE="STAT" and TYPE="TRAN" is referenced in the same
nonlinear transient run, then the job fails and the message "SYSTEM FATAL
ERROR 4276.WRITE ERROR CODE 3021", is issued.
Avoidance: Use SPC's to fix the constant temperature boundary condition.
Use TEMPBC TYPE="TRAN" to designate the time-varying boundary condition.
@
5927; 68.2; Matrix Methods - MCE1, RBEi, MPC - UWM 6137, INPUT MATRIX IS RAN
K DEFICIENT
Although it is not illegal to specify a DOF as independent on one rigid
element and dependent on another rigid element, if the definitions form a
"closed loop" such that DOF A is dependent on DOF B and DOF B is either
directly or indirectly dependent on DOF A, and sparse decomposition is
selected the message, "USER WARNING MESSAGE 6137, INPUT MATRIX IS RANK
DEFICIENT" is output.
Avoidance: This error is corrected in Version 67 and later systems for
standard decomposition. Thus, if the user forces standard decomposition
by setting the SPARSE keyword to 1, then USER FATAL MESSAGE 6202 is
produced which identifies the rigid elements in error.

@
5930; 68.2; Transient Heat Transfer - MATT4, RADMT - UWM 2013
In a transient thermal analysis (SOL 159), if QVECT load is used in
conjunction with temperature dependent absorptivity. This input leads to
the message "USER WARNING MESSAGE 2013, NO STRUCTURAL ELEMENTS EXIST.
FATAL ERROR" being issued. The QVECT load is directional load such as
solar load.
Avoidance: Defines a constant absorptivity value instead of temperature
dependent absorptivity value.

@
5931; ; DMAP - COMPILE - UFM 151 (LINKAR)
A DMAP compile error may be encountered when logical parameters are in a
call statement.
An avoidance is to replace the logical in the call statement with another
type, such as an integer or to delete the parameter from the subdmap call
altogether.
@
5933; GL; DMAP - DCMP, DECOMP - Wrong (Zero) Results
The DECOMP and DCMP modules have an optional set of parameters which
should give the scaled determinant (DET) and the scale factor (POWER) of
the input matrix. For the sparse methods, they are set to zero.
Avoidance: An avoidance is to select the nonsparse decomposition methods
 when a determinant is needed. That is set system cell 126 to 1 (a.k.a.
NASTRAN SPARSE keyword) for symmetric matrices and system cell 209 equal
to 0 (a.k.a. NASTRAN USPARSE keyword) for unsymmetric matrices.
@
5934; 68.1; Documentation - GMBC - UFM 315
The following message is issued when the LID field of the GMBC entry is 0
and it contains a continuation entry, i.e., ENTITY = FEFACE and METHOD =
LINEAR.
USER FATAL MESSAGE 315 (IFPDRV)
FORMAT ERROR ON BULK DATA ENTRY GMBC 0
SORTED CARD COUNT = **
Remark 4 under the Version 68 Quick Reference Guide's description of the
GMBC entry should be updated to include this condition and it should be
stated that not only is the METHOD field not required but will result in a
USER FATAL MESSAGE if included.
This error is essentially the same as error 5506, it has been corrected in
the 2nd printing of the Quick Reference Guide.
@
5935; 69; Miscellaneous - ADD5 - SFM 4276, EC 70020
An error in the bufferpool logic may result in "SYSTEM FATAL ERROR
4276.BPREAD ERROR CODE 70020" to be issued from the ADD5 module. This
error is very much problem and buffersize dependent.
Avoidance: Changing the buffersize may avoid the error but the sure
avoidance is to turn off bufferpooling prior to the ADD5 module by setting
SYSTEM(119) as follows:
ALTER xxx $ turn off buffpool just before ADD5
putsys (2,119)
ALTER yyy $ turn buffpool back on just after ADD5
putsys (3,119)
@
5941; 69; Optimization/Design Sensitivity - DOPTPRM - Missing Output
If the P1 parameter on the DOPTPRM bulk data entry is either set to the
default value of 0 or simply not specified and allowed to become the
default value of 0, and the P2 field is specifed, then requested data on
the P2 field is not printed for the last design cycle.
Avoidance: Set parameter P1 to 1 to obtain the output for the last design
cycle or apply the following dmap alter:
$
$ This alter will print the design optimization output
$ for the last design cycle regardless of what the value
$ of the 'p1' parameter in the doptprm entry is set to.
$
ksh: cray.err: permission denied
$ **** THIS ALTER IS FOR VERSION 68.2 ONLY ****
$
compile desopt nolist noref souin=mscsou
alter 'CALL PREDOM'
$ dummy matgen calls to satisfy dmap complier
file cntabrgp=save,ovrwrt/cvalrgp=save,ovrwrt/cvalox=save,ovrwrt/
propip=save,ovrwrt/propox=save,ovrwrt/r1tabrgp=save,ovrwrt/
r1valrgp=save,ovrwrt/r1valox=save,ovrwrt/rsp2rgp=save,ovrwrt/
r2valrgp=save,ovrwrt/r2valox=save,ovrwrt/optprmgp=save,ovrwrt/
objtbgp=save,ovrwrt/drstblgp=save,ovrwrt/frqrprgp=save,ovrwrt $
if ( desiter=-1 ) then
matgen ,/cntabrgp/1/ $
matgen ,/cvalrgp/1/ $
matgen ,/cvalox/1/ $
matgen ,/propip/1/ $
matgen ,/propox/1/ $
matgen ,/r1tabrgp/1/ $
matgen ,/r1valrgp/1/ $
matgen ,/r1valox/1/ $
matgen ,/rsp2rgp/1/ $
matgen ,/r2valrgp/1/ $
matgen ,/r2valox/1/ $
matgen ,/optprmgp/1/ $
matgen ,/objtbgp/1/ $
matgen ,/drstblgp/1/ $
matgen ,/frqrprgp/1/ $
endif $
$
alter 'IF ( CNVFLG = 2 ) THEN'
$
message //'dom10 call after hard convergence - descycle='/descycle $
$
type parm,,i,n,desp $
desp = descycle - 1 $
DOM10,
DESTAB,xprev,xinit,CNTABRGp, CVALRGp, CVALOx,
DVPTAB,PROPIp,PROPOx,R1TABRGp,R1VALRGp,R1VALOx,
RSP2RGp,R2VALRGp,R2VALOx,OPTPRMGp,OBJTbgp,DRSTBLgp,
tol1,fol1,frqrprgp//
desp/desp/objpv/OBJOUT/ $
$
alter 'DOM9,'
$
$ save dom10 inputs
equivx cntabrg/cntabrgp/always $
equivx cvalrg/cvalrgp/always $
equivx cvalo/cvalox/always $
equivx propi/propip/always $
equivx propo/propox/always $
equivx r1tabrg/r1tabrgp/always $
equivx r1valrg/r1valrgp/always $
equivx r1valo/r1valox/always $
equivx rsp2rg/rsp2rgp/always $
equivx r2valrg/r2valrgp/always $
equivx r2valo/r2valox/always $
equivx optprmg/optprmgp/always $
equivx objtbg/objtbgp/always $
equivx drstblg/drstblgp/always $
equivx frqrprg/frqpprg/always $
$
alter 'CNVFLG=1'
$
DOM10,
DESTAB,XINIT,XO,CNTABRG, CVALRG, CVALO,
 DVPTAB,PROPI,PROPO,R1TABRG,R1VALRG,R1VALO,
RSP2RG,R2VALRG,R2VALO,OPTPRMG,OBJTbg,DRSTBLg,
tol1,fol1,frqrprg//
DESCYCLE/DESMAX/OBJIN/OBJOUT/ $
$
endalter
@
5946; 68.2; Optimization/Design Sensitivity - DRESP1 - Uses Von Mises Only
Remark 1 for the DRESP1 entry indicates that, for stress/strain item codes
that have dual meanings, the selection is made based on the option
specified in case control. This pertains to Von Mises and Max Shear
selection. Due to a programming error, the remark is incorrect and Von
Mises results will be selected, regardless of the case control request, as
it was in Version 67.
Avoidance: Users desiring to use Max Shear in a design task will have to
synthesize it from basic stress results.
@
5948; 70.5; Documentation - QRG, USET, PARTN - SFM 3007
If a USETi bulk data entry lists a standard MSC/NASTRAN set, such as S or
M, the program may fail in the PARTN module with the message "SYSTEM FATAL
MESSAGE 3007, ILLEGAL INPUT TO SUBROUTINE".
Avoidance: This entry should only reference new sets defined on DEFUSET
bulk data entries.
@
5953; 68.2; Optimization/Design Sensitivity - DCONADD,DCONSTR - UFM 316
If the DCONADD entry references only a single DCONSTR entry on any
continuation entry, then the job fails and the following message is
issued:
USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA CARD DCONADD xx SORTED
CARD COUNT = yy
An example of a DCONADD entry that causes this error is as follows:
DCONADD, 1, 10, 11, 12, 13, 14, 15, 16, +DCON1
+DCON1, 17, 18, 19, 20, 21, 22, 23, 24, +DCON2
+DCON2, 25
$
Note that the second continuation entry contains a single DCONSTR entry
(25 in this case).
Avoidance: Add a dummy DCONSTR to the DCONADD continuation entry. The
limits should be set to a large value to ensure that the constraint
screening will remove this constraint prior to performing optimization. A
suggested lower bound (LALLOW) and upper bound (UALLOW) values would be
-1.0E20 and 1.0E20, respectively.
@
5954; 69; Grid Point Force - DEFORM, GPFORCE, GPFDR - SFM 6223 (GPFDR2), S
FM 4276, EC 3012
When the DEFORM command is used and ELAS elements exist in the deck, the
 following fatal message is issued from the GPFDR module.
*** SYSTEM FATAL MESSAGE 6223 (GPFDR2), RECORD POINTED AT BY INDEX IN
SCRATCH FILE IS WRONG.
USER ACTION: CONTACT MSC CLIENT SUPPORT.
CASECC NUMBER = 4
LOAD SET NO = 4
ELEMENT TYPE = 1
EXT ELEMENT ID = 101
*** SYSTEM FATAL ERROR 4276.OPEN ERROR CODE 3012
Avoidance: Do not request GPFORCE output for the subcase with DEFORM
loads.

@
5956; 68.2; Old Design Sensitivity - PARAM,POST,-2 - Missing Datablocks
In MSC/NASTRAN Version 68, if sensitivity output (SENSITY command) is
requested using SOL 103 in conjunction with param,post,-2, not all
datablocks are written to the output2 file. Note this error applies only
to the old design sensitivity method. It does not apply to the new
sensitivity calculation (SOL 200).
Avoidance: Insert the following DMAP alters.

compile super3 souin=mscsou nolist noref $

alter 90,90 $
gp1 GEOM1,GEOM2,,,/GPL,EQEXIN,GPDT,CSTM,BGPDT,SIL,/0/0/0 $
alter 208,208 $
alter 211,211 $
@
5960; 69; Buckling Analysis - EIGRL, SYSTEM(146) - SFM 5407, SFM 5401, SFM
4276
If SYSTEM(146) is set to save scratch space and reduce I/O in sparse
decomposition during a Lanczos eigenvalue buckling run, then the job
fails. Different fatal messages have been encountered including "SYSTEM
FATAL MESSAGE 5407 INERTIA (STURM SEQUENCE) COUNT DISAGREES WITH THE
NUMBER OF MODES ACTUALLY COMPUTED IN A (SUB-) INTERVAL", "SYSTEM FATAL
MESSAGE 5401 (REIGL) LANCZOS METHOD IS UNABLE TO FIND ALL THE EIGENVALUES
IN RANGE" AND "SYSTEM FATAL MESSAGE RECOVERY ERROR CODE 7777" in sparse
decomposition.
Avoidance: Do not use SYSTEM(146) during a Lanczos buckling run.
@
5961; 68.2; Heat transfer - SET1, RADCAV - Wrong Answers
If the SET identification pairs for the calculation of global view factors
exceeds 10 pairs, then an erroneous result is output for the global view
factors. The documentation states that only 10 pairs can be shown for one
run. This limit is expanded to 30 in Version 68.2.
@
5962; 70.5; Acoustic Analysis - REIGL, SUPORT, MERGE - UFM 4276, EC 3007
SUPORT entries do not work in acoustic analysis. USER FATAL MESSAGE
4276, ERROR CODE 3007 from subroutine merge called from the REIGL module
 is issued. Other eigenvalue extraction methods result in other fatal
messages.
Avoidance: Remove SUPORT entries.
@
5963; 69; Case Control - STRAIN(CORNER) - Missing Output
If a STRAIN(CORNER) request is made without a STRESS(CORNER) request, only
centroidal strains are output.
Avoidance: Request STRESS(CORNER,PLOT) in the Case Control Section.
@
5967; 70.5; DMAP - MPYAD - Wrong Answers
The MPYAD module performs the following matrix operation:
[A][B]+[C]=[D]
with the following specification: MPYAD A,B,C/D $
If [A] does not exist then the MPYAD module assumes [A] is null and is
supposed to make [D] equivalent to [C]. However, if [C] is complex then
[D] contains only the real part of [C].
Avoidance: Test for the presence of [A] before executing MPYAD:
PARAML A//'pres'////s,n,noamat $
if ( noamat>-1) then $
MPYAD A,B,C/D $
else $
EQUIVX C/D/-1 $
endif $
@
5974; 68.2; Miscellaneous - Accounting Files do not Include the Time Zone
The accounting files generated on some versions of AIX will not include
the time zone when LANG="". Since the various scripts must set this to
function properly, the only recourse is for the execution and accounting
summary scripts to account for this potential discrepancy.
Avoidance: The running summary written by the execution script, i.e.,
nast68 or emas3, will be incorrect. The accounting summary script can be
corrected by manually editing the accounting file to insert the time zone
parameter into each line as appropriate. The accounting summary script
will then report correct summary information.
@
5977; 69; Buckling Analysis - SPCD - Wrong Answers
If SPCD entries are used in SOL 105 and the SUBCASE used for the buckling
analysis does not contain a call for a LOAD set using the SPCDs, the SPCD
entries are ignored in the solution.
Avoidance: Either use different SPC sets for each SUBCASE, or include a
LOAD command in the buckling SUBCASE pointing to the SPCD set.
@
5980; 68.2; Executive - XEQUIV - SFM 4276 EC 7777
If during the equivalence of a local scratch datablock the PVA table runs
out of space, the expansion of the PVA table may corrupt open core. Jobs
 that encounter this error may fail in many unpredictable ways. The most
likely failure will be "SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777". The
job that encountered this error is an optimization analysis with image
superelements.
Avoidance: Increase or decrease the size of RAM with the INIT FMS
statement.
@
5981; ; Optimization/Design Sensitivity - DSAD, FRRD1 - Insufficient Tim
e, SFM 4276
If an optimization run encounters an "INSUFFICIENT TIME" type message
during the finite element part of the solution (the example encountered
was insufficient time to complete the frequency response solution, FRRD1)
it is possible the run will continue into the optimization process and
fail in the optimizer (the example fails in DSAD with SFM 4276, ERROR CODE
3001).
If an unexplained error occurs during optimization, check the output file
for other messages.
Avoidance: Specify sufficient time on the TIME statement in the executive
control section.
@
5982; 69; Eigenvalue Analysis - AGIV, AHOU - SDCOMP FATAL ERROR, UIM 5458
If an eigenvalue analysis specifies either the AGIV or the AHOU method,
then standard GIV and standard HOU are attempted first. If the mass
matrix is singular, an unnumbered fatal message "SDCOMP FATAL" is issued
even though the problem continues and the program switches to the modified
Givens or Householder method.
Avoidance: If the problem runs, ignore the FATAL message.
@
5983; 69; Optimization/Design Sensitivity - DOPTPRM, DOM10 - SFM 3001
An optimization run with P1=1 and P2=15 on the DOPTPRM entry fails and
issues "SYSTEM FATAL MESSAGE 3001" from the DOM10 module even though the
run works fine without these parameters. These are output control
parameters and should not effect the solution.
Avoidance: If P1 and P2 have been used on the DOPTPRM and the run fails,
try the run again without specifying P1 and P2.
@
5984; 68.2; Nonlinear Analysis - PARAM,NMLOOP , SE's - Missing Output
When using the nonlinear normal modes option in Sol 106, no modal analysis
is performed when there are superelements and PARAM,NMLOOP is specified in
the Case Control.
Avoidance: Place the PARAM,NMLOOP statement in the bulk data.
or
Use SUPER=ALL in the case control.
or
DMAP alter:
 compile nlstatic souin=mscsou nolist noref $
alter 480 $
seid= 0 $
pvt pvts,cases/pvtxx/ $
endalter $

@
5986; 68.2; Superelements - PARAM, EXTRCV - Wrong Answers
When using PARAM, EXTRCV with the RESTART FMS statement, wrong answers
result if the residual structure subcase qualifiers (LOAD, SPC, MPC, etc.)
differ from those specified in the residual structure subcase of the run
which generates the external superelement.
Avoidances:
1. Maintain the same residual structure subcase qualifiers in all runs.
2. Insert the following alter in the PARAM, EXTRCV run:
COMPILE IFPL souin=mscsou, noref nolist
alter 152,152

3. Use the multi-MASTER approach in the PARAM, EXTRCV run. Assuming that
the external superelement data base is called extse1.master, replace:
RESTART and
ASSIGN MASTER, if any, with the following:
nastran system(143)=1
assi extse1=extse1.master
dbloc logi=extse1
ASSIGN MASTER, if any, with the following:
nastran system(143)=1
assi extse1=extse1.master
dbloc logi=extse1
Insert PARAM, SERST, SEDR into the Case Control or Bulk Data.
@
5987; 68.2; Postprocessing - PARAM,POST,-2 - IDEAS Dataloader Abort
In MSC/NASTRAN Version 68, the structured solution sequences with
PARAM.CHECKOUT,YES and PARAM,POST,-2 do not write an END-OF-DATA on the
FORTRAN file.
Avoidance:
COMPILE PHASE0 SOUIN=MSCSOU NOREF NOLIST
ALTER 527 $ Before CALL PRTSUM
OUTPUT2 //-9/OUNIT1//OMAXR $ write END-OF-DATA
CEND
@
5992; 68.2; Grid Point Stress - GPSTRESS, SURFACE, QUAD4 - Missing Output
Certain setids (120 thru 138, 148 thru 159, 174 thru 181) may cause
incomplete output for GPSTRESS when used in combination with the SURFACE
 case control command. An example case control which results in the error
follows:
SET 21=120
STRFIELD=21
GPSTRESS=21
OUTPUT(POST)
SET 95=19,20,21,59,60,61
SURFACE 120 SET 95 NORMAL Z SYSTEM ELEMENT
BEGIN BULK
@
5993; 69; Documentation - PBEAM - Continuation Entries
The documentation for the PBEAM entry in the MSC/NASTRAN Version 68 Quick
Reference Guide states that the fields "SO" and "X/XB" on the second
continuation entry are required. A footnote should be included (as was in
the Version 66 User's Manual) that states that these fields are only
required when the second continuation entry is desired to request
intermediate stress output.

@
5995; ; Elements - QUAD4, TEMP(LOAD) - Wrong answers
Warped QUAD4 elements with thermal loads give wrong answers. This error
is similar to the thermal expansion error in elements with rigid offsets,
there is no expansion for the offset.
If offsets and orthotropic materials are present, the errors are
magnified.
@
5998; ; Documentation - Linear Static Analysis User's Guide - Wrong outp
ut
Incorrect output result files are included in Figures 11-2 and 11-3 of the
MSC/NASTRAN Linear Static Analysis User's Guide. Due to a missing carriage
control character, a pload1 entry for element # 2 is missing for the
inertia2.dat file that is delivered with the delivery tape. Furthermore,
due to error 5522, PLOAD1 yields incorrect results for the BAR, BEAM, or
BEND elements when used in conjunction with inertia relief. Thus, this
example (using PLOAD1) should not be emulated until error 5522 is
corrected.

Avoidance: Replace the PLOAD1 with equivalent grid point loads and
regenerate the output file.
@
6000; 2001; Transient Analysis - RESTART, PARAM,NOAP - Poor Answers
If TIC entries are used to set initial velocity values, the computed
initial velocity will be slightly different from the specified initial
velocity. On a case where v0 was set to 1.0, the output velocity was
computed as 0.9969. This model has a dt of 0.1
Avoidance: An avoidance is to run the model once to determine the c
omputed
initial velocity, and increase the input initial velocity enough so that
the computed intital velocity is the value desired on a second run.
@
6004; 68.2; SSSALTER Library - ZFREQ - Wrong Answers
 The ZFREQ alter delivered in the Version 68 SSSALTER library allows
frequency dependent springs and dampers to be defined. However, the
nominal spring stiffness is not updated in the data recovery phase.
Thus, element force output for these elements is incorrect. This is a
known limitation of the alter that should have been documented in the
ZFREQ.RDM file. The information contained in this error report is to be
included in the Version 68.2 ZFREQ.RDM file and the error is considered
corrected.
Avoidance: An avoidance is to place a stiff spring in series with the
mount, the forces in the stiff spring represent the correct element force
for the mount spring. The error introduced by this technique is
proportional to the ratio of the stiff spring rate to the mount spring
rate. For example, if the ratio is 1000:1, the overall mount spring rate
across both devices is 0.1% lower than the actual mount rate. The user
may modify his mount spring rate data to account for this effect. In most
instances the mount spring rates are not known to this level of accuracy,
so that the effect may be ignored.
To avoid numerical difficulties the stiff spring should be of the same
order of stiffness as the structural elements to which it attaches. In
most instances the structure to which the mount attaches is much stiffer
than the mount.

@
6007; 69; Documentation - PARAM, USETPRT PARAM, USETEL - K-set is not Outp
ut
The MSC/NASTRAN Quick Reference Guide incorrectly states that the
parameters usetprt and usetsel can be used to print k-set degrees of
freedom. That is, the k-set decimal equivalent is listed as an option in
table 1 of the usetprt parameter description.
@
6010; 68.2; Nonlinear Slide Line Contact - BCONP, BLSEG - Wrong Answers
An overhang slave node is not considered as "overhang" if the overhang
value is less than 5% of the master segment length. For models with large
master segments, this may lead to wrong answers. Typical symptoms include
the printout of a wrong slip code (e.g., SLIDE instead of OVERHANG) and
incorrect surface coordinates.
Avoidance: Use smaller master segments when modeling overhang situations.
@
6011; 70; Security - NAO - UFM 3060, Segmentation Fault, Signal Number 11
in Log File
Large jobs that use Network Authorization (NAO) may fail in various
moudules on the DEC/ALPHA OSF with the message, "Receiving signal number
11: Segmentation fault", in the MSC/NASTRAN log file. Similarly, if using
node lock authorization, and running a looping solution sequence like
nonlinear and a large problem, then User Fatal Message 1019 "Fortran Unit
Open Error" pointing to the authorization file may be issued. This message
is followed by USER FATAL MESSAGE 3060. The problem occurs because a
socket (file) is left open each time MSC/NASTRAN checks the authorization
code and when the system reaches the open sockets limit, the job fails as
indicated above.
Avoidance: Edit the system configuration table (sysconfigtab) and increase
 the value of the OPEN_MAX parameter. (This is a system administrator job
that may require assistance from DEC.) Then reboot the machine. To
determine the curren OPEN_MAX
value, execute the command: getconf OPEN_MAX
@
6012; 69; Dynamic Response Analysis - EPOINT - Single Precision Accuracy
If there are EPOINT entries in the bulk data file, the eigenvalue matrix
PHIA is expanded to the d-set matrix PHIDH in the GKAM module. On
short-word machines PHIA is double precision. At present PHIDH is single
precision. As this matrix is used to reduce load matrices and expand
modal solutions to physical DOFs this may reduce the precision of the
solution.
Avoidance: An avoidance is to discard the PHIDH matrix output by the GKAM
module and build an alternate matrix with DMAP. The equations defining
how PHIDH is formed from PHIA are given in the MSC/NASTRAN REFERENCE
MANUAL.
@
6013; 69; DMAP - DMIG - Wrong Answers
If a DMIG entry references a degree of freedom which does not exist, the
corresponding coefficients are assigned to another degree of freedom,
without any diagnostic messages.
In one example, an open set was used to define a series of scalar points.
A DMIG term referenced a point that was not defined on an SPOINT entry
which occurred between two points which were listed on SPOINT entries.
The coefficient was placed on the last DOF.
@
6015; 68.2; Documentation - Aeroelastic Users' Guide - UFM 2268 (FA1)
Example HA145I in the Version 68 Aeroelastic User's Guide contains an
alter that is in error. The alter inserts a MATPRN after line 50. It
should be after line 45. If submitted as documented, the run fails and
issues the message "USER FATAL MESSAGE 2268 (FA1) FMETHOD SET 0 NOT
FOUND".
The data deck delivered in the tpl has been made current. The
documentation will be updated to reflect the latest version when it is
updated.
@
6016; 68.2; P-Elements - HEXA, PENTA - Excessive Pivot Ratios at p=2 (UFM 90
50)
If a model consists of HEXAs or PENTAs at a p-level=2 or higher connected
end-to-end, the following message may occur:
DMAP FATAL MESSAGE 9050 (SEKRRS) - RUN TERMINATED DUE TO EXCESSIVE
PIVOT RATIOS
IN MATRIX KLL. USER PARAMETER BAILOUT MAY BE USED TO CONTINUE THE
RUN.
This fatal message is caused by the reduced integration used at a
p-level=2 and higher. Reduced integration is the current default.
Avoidance: SYSTEM(212)=1 may be set to work around the problem.
@
6017; 68.2; Nonlinear Slide Line Contact - BCONP, BLSEG - Extraneous Output
The following parameters are printed in the "LINE CONTACT REGION" output
 when using the slide line contact capability.
CONT BEHA FO CO
TYPE
These values should be ignored as they are not implemented in the program.
Avoidance: These values have no effect on the rest of the output and they
can be safely ignored.
@
6018; 69; Acoustic Analysis - SORTi - SORT2 Output in Place Of SORT1
In solution 108, a request for SORT1 output in an acoustic analysis
results in SORT2 output.
Avoidance: Add PARAM,CURVPLOT,1 to the bulk data.

@
6024; 68.2; Optimization/Design Sensitivity - DLINK - Wrong Answers
If the user inputs a nonzero value for C0 on the DLINK entry, it is
incorrectly processed by the optimization algorithm. Values used by the
optimizer for the dependent design variables will be incorrect. This will
directly affect constraint calculations from the upper and lower bounds
input on the associated DESVAR entry. These constraints may not be
critical and and the optimization results may be valid. The printed
summary values for the dependent design variables will be incorrect.
Avoidance: There is no avoidance.
@
6025; 68.2; Plotting - XYPLOT(TEMP) - Infinite Loop
If the SET1 entry references non-existing grid IDs, then the XYPLOT
command which reference this SET1 entry cannot identify the missing grid
point(s), which causes the program to go into an infinite loop.
Avoidance: Check all the grid points on the SET1 entry and make sure they
exist.
@
6030; 68.2; Optimization/Design Sensitivity - MPCFORCE, PARTN, DSAP - UFM 21
66, UFM 5901
In Version 68, if MPC forces are requested in an optimization run under
static analysis, the program fails at line 121 of subDMAP SEDISP with UFM
2166 stating that the number of columns in the partition vector is
incorrect. Under frequency response analysis, UFM 5901 will be issued
from the DSAP module.
Avoidance: Either remove the MPCFORCE output request or apply the
following DMAP alter to avoid the error:
$for statics approach:
compile sedisp, souin=mscsou, nolist, noref
alter 109
noqmg=ltoi(delta) $
$ For Frequency approach:
compile sedisp, souin=mscsou, nolist, noref
alter 238
if (not delta ) then $
 alter 239
endif $
@
6031; 68.2; Documentation - gentim2.dat - Data File Not Found
The MSC/NASTRAN V68 Installation and Operation Instructions implies that
the gentim2.dat file used to update timing constants is in the same
directory with the SSS.MASTERA and USS.MASTERA (ie /msc68/nast68/aix/del)
but it is not. The file is located in the /msc68/nast68/del directory.
The install guide should be clarified to tell users where the data file is
located.
@
6033; 68.2; Elements - TETRA(10), TEMP(LOAD) - Wrong Answers for TETRA(10) S
tresses
If the temperature at the grid points of a TETRA(10) element vary (i.e.
they are not constant), then the stresses for the element are incorrect.
Avoidance: The error is proportional to the variance of temperature
within the element; however, refining the mesh so that the temperature
within each element is near a constant is not considered a practical
avoidance. Stresses requested at the gauss points are correct.
@
6034; 3.2; Documentation - Installation and Operation Instructions - Timing
Constants
Section 3.6 of the MSC/NASTRAN Version 68 Installation and Operation
Instructions explains how to install the timing constants. Step 2 of the
documented procedure states:
"2. The timing data are generated or changed by running gentim2.dat
(delivered in the "del" library)..."
Later in this step you are instructed to change to the directory
install-dir/msc68/{prog}/{mach}/del
where the files *SS.MASTERA, etc are stored. In other words, there are
two "del" directories
1) install-dir/msc68/{prog}/{mach}/del
and
2) install-dir/msc68/{prog}/del
The first directory contains the *SS.* files and the second contains the
gentim* files. The installation procedure should be more clear.

Avoidance: Copy the file install-dir/msc68/{prog}/del/gentim2.dat to

install-dir/msc68/{prog}/{mach}/del
@
6035; 69; Nonlinear Analysis - NLITER, NLPARM, INTOUT=NO - UFM 4551
Nonlinear jobs have been found that converge successfully when INTOUT=YES
but fail to converge and issue "USER FATAL MESSAGE 4551 STOPPED PROBLEM
DUE TO FAILED CONVERGE" when INTOUT=NO. The exact cause of the problem is
under investigation.
@
 6036; 68.2; P-Elements - Punch Output not Supported
When stresses and strains are output in the punch file, they are output in
the print file (F06) format, with titles and labels, rather than a format
suitable for the punch file.
@
6038; 69; Elements - CTRIA3, TA1 - No Message, Trap Encountered
If a CTRIA3 element references a nonexistent material coordinate system
(field 7), the job should issue USER FATAL MESSAGE 2025 UNDEFINED
COORDINATE SYSTEM. The IBM RS/6000 versions of MSC/NASTRAN do not issue
this message. The job stops without a message being issued. The f04,
Execution Summary, indicates that the last module executed is either TA1
or PHASE1DR and the log file contains a "Trap Encountered" message.
@
6042; GL; Elements - PENTA(6), TEMP(LOAD) - Wrong Answers for Stresses
If the temperature at the grid points of a PENTA(6) element vary (i.e.
they are not the same at all points), then the stress output is wrong.
Both corner point (default) and Gauss location (selected on PSOLID)
options give wrong answers. The stress error is proportional to the
amount of temperature change within the element.
Avoidance: This is a limitation of the element. The only avoidance is to
refine the mesh so that the variation of temperature within the element is
minimal.

@
6044; 68.2; Optimization/Design Sensitivity - Incorrect Sensitivities for Su
perelements
Incorrect design sensitivities are computed for superelement models in
which a superelement has
design variables but no design responses. This error occurs only in
statics.
Avoidance: The avoidance is to specify WEIGHT=ALL (for all superelements)
as a design response.

@
6045; 2001; Installation - Timing Constants - UWM 6080
The text of "USER WARNING MESSAGE 6080 (SETKRN), THE TIMING CONSTANTS DATA
BLOCK TIMEBLK NOT FOUND ON THE DELIVERY DATABASE" refers the user to "AM
7.6". This reference should changed to the a more generic term, such as
machine dependent documentation.
@
6046; 68.2; Optimization/Design Sensitivity - PARAM,DESPCH - Incorrect Actio
n
Setting PARAM,DESPCH to a negative value will not prevent DESVAR and GRID
Entries from being written to the PUNCH file.
Avoidance: The following DMAP alter may be used to avoid the problem:
compile exitopt nolist noref souin=mscsou
alter 24 $ just before executing module DOM11
 type parm,,i,n,desitert
if (despch < 0) then
desitert = desiter
desiter = 0
endif
alter 25 $ right after execution of module DOM11
if (despch < 0) then
desiter = desitert
endif
$$$ end of alter
@
6047; 68.2; SSSALTER Library - RESFLEXR - Wrong Answers
The alter to add residual vectors into a dynamic solution delivered in the
sssalter directory with Version 68 does not handle SUPORT entries properly
for free-free structures. If a SUPORT entry is present, residual vectors
are not generated for the residual structure.
Avoidance: Call MSC for an updated version of the alter.
@
6050; 68.2; P-Elements - GMLOAD in convective coordinate systems
If a GMLOAD is applied in a convective coordinate system defined on a
GMCURV or GMSURF where the CIDIN is other than basic coordinate system,
then the following message is output from SSG1:
USER FATAL MESSAGE 2025, UNDEFINED COORDINATE SYSTEM 2
FATAL ERROR
Avoidance: Possible avoidances include defining the GMCURV or GMSURF in
the basic coordinate system, and applying the load in a non-convective
system, or using PLOAD4.
@
6051; 69; Documentation - MAT1 - Programs Action not Clear
Remark 3 under the documentation for the MAT1 entry in the Quick Reference
Guide states that the relationship between E, G and NU "must be
satisfied". It should be clarified that satisfying the relationship is
merely a recommendation, rather than a requirement. The values furnished
are used as indicated in Table 1. and if the relationship is not satisfied
the program proceeds but issues USER WARNING MESSAGE 2251A and 2251B.
Furthermore, the relationship indicates ">" when it should indicate "<".
@
6052; 69; Acoustic Analysis - ACMG - Job Aborted, No Message, Receiving Si
gnal Number 10
An acoustic job that specifies DIFF on the ACMODL entry aborted in the
ACMG module with no message issued in the f06 file. A message may appear
in the log file indicating the job has aborted but the message will be
different depending on the computer. One such message is "Receiving
signal number 10: Bus error"
Avoidance: The cause of the error is under investigation. Selecting the
IDENT method on the ACMODL entry will probably allow the problem to
execute to complete but unless the fluid and structural mesh are
compatible will not yield the correct results.
@
6058; 68.2; P-Elements - Convective output coordinate system
 If a convective coordinate system is specified for the output coordinate
system of a GMSURF, GMCURV, FEFACE, FEEDGE, or GRID, the following message
is output from GP1:
*** USER FATAL MESSAGE 2104, UNDEFINED COORDINATE SYSTEM 13
Avoidance: There is no avoidance.
@
6059; ; Nonlinear Static Analysis - Restarts - Missing Displacements
In the nonlinear static solution sequences, displacement output is missing
from the restart run, when the following conditions are met:
1) A new subcase is added which requests intermediate output (INTOUT=YES
on NLPARM).
2) The restart run regresses to a subcase prior to the last executed
subcase (using PARAM,LOOPID and PARAM,SUBID). In other words,
if the cold start finished with subcase 4, the new restart subcase 5 would
start with the last LOOPID of subcase 3 or prior and a SUBID of 5
(effectively skipping subcase 4).

When these conditions are met, the .F06 and the punch file contains ONLY
the first intermediate displacements for the new subcase, and no other
displacement output for that subcase. Other forms of output are
unaffected.
Avoidance: Avoid one of the above conditions.
@
6065; 70.7; Composites - PCOMP, MID4, TEMP(LOAD) - Incorrect Term Output for
MAT2
If unsymetric thermal properties and thermal loads are present on a
composite plate that is using the PSHELL/MAT2 instead of the PCOMP/MATi
method of composite property description, then the PSHELL's MID4 MAT2 must
have a ID greater than 400,000,000 and the thermal properties in that MAT2
must be written in a special nonstandard format (see Remark 13 on the MAT2
description of the MSC/NASTRAN Quick Reference Guide). This is because
the MSC/NASTRAN composite formulation is different than the noncomposite
formulation for unsymmetric thermal loading. The MID4 MAT2 ID is the flag
that tells MSC/NASTRAN whether the PSHELL/MAT2s are composite or not
composite.
PSHELL/MAT2s that are internally generated from PCOMPs and which are
written to the .f06 file or to the .pch file have the correct material IDs
on them, and if used without change, will produce correct results.
However, if the MID4 MAT2 ID is changed to a number less than 400,000,000
and unsymmetric thermal material properties and loads are present, or if
user supplied MAT2s are used and do not follow the above rules, incorrect
results will occur. Note that a double field format is required on the
MAT2 in order to have an ID this large. The MAT2s echoed in the .pch file
are already in a double field format and thus may be used without change.
The MAT2s echoed in the .f06 file however, are in a nonstandard print
format and must be manually changed to a standard MSC/NASTRAN double field
format.
Avoidance:
Use a MID4 MAT2 ID > 400,000,000 and the proper thermal property format
 for composites.
@
6067; 68.2; P-Elements - Punch File Output not Supported for Error Tables
When the error table is output in the punch file, it is output in the
print file (F06) format, with titles and labels, rather than a format
suitable for the punch file.
@
6068; 68.2; P-Elements - GMBC - UIM 6353
If a zero displacement value is entered on the GMBC entry, the following
message is issued for each grid point associated with the GMEDGE or GMFACE
entry:
USER INFORMATION MESSAGE 6353 (EGENDD)
ENFORCED DISPLACEMENT VECTOR IS NULL FOR COMPONENT=** OF EDGE ID= *** ON
GMBC LOAD ID= 0 AND SPC ID= **
Avoidance: Although there is no way to avoid the information message from
being output, there is no other adverse affect on the run and the message
may be ignored.
@
6070; 68.2; Aeroelastic Analysis - PARAM,LFREQ, OPPHIPA, HFREQ - UFM 5423
A flutter or dynamic aeroelastic run that uses PARAM LFREQ or HFREQ to
eliminate modes and PARAM OPPHIPA to print the normal mode shapes
including aerodynamic grid points, fails with the message "USER FATAL
MESSAGE 5423 ATTEMPT TO MERGE INCOMPATIBLE MATRICES".
Avoidance: There is no avoidance other than do not use PARAM LFREQ or
HFREQ in combination with OPPHIPA.
@
6072; 69; Optimization/Design Sensitivity - DEQATN - UFM 224 (PRERMM),
If the PI(1) function on the DEQATN entry is used with an integer, it
results in a mixed mode arithmetic operation error and the following
message is issued "USER FATAL MESSAGE 224 (PRERMM) MIXED MODE USAGE BY THE
ARITHMETIC OPERATOR * IS INVALID DURING DEQATN PROCESSING".
Avoidance: Use a real number instead of an integer. For example, PI(1.0)
is acceptable and produces the desired result. The user could also define
pi on a DTABLE entry as a constant and not rely on the DEQATN function.
@
6079; 69; Dynamics - Data recovery with PARAM,DDRMM,0 (default) - Missing
output
By default (PARAM,DDRMM,0) in dynamic data recovery for modal frequency or
modal transient analysis (SOLs 11, 12, 111, 112, 146, and 200) any FORCE,
STRESS, SPCF (if PARAM,DYNSPCF,OLD is specified), or STRAIN request in a
subcase subsequent to the first subcase is not honored. For example, if
the subcase structure shown below is specified then the forces for element
20, 40 and 60 are not output.
set 1=100,200
set 2=20,40,60
subcase 10
FORCE = 1
Subcase 20
FORCE = 2
 Avoidance: Including PARAM,DDRMM,-1 avoids the problem but this forces the
output to be in SORT1 format.
Another avoidance is to add a new subcase above the existing subcase
structure for the FORCE(PLOT)=ALL and STRESS(PLOT)=ALL as shown below.
With this avoidance the output is in SORT2 format.
set 2=20,40,60
Subcase 1
FORCE(PLOT)=ALL
subcase 10
FORCE = 1
Subcase 20
FORCE = 2

@
6080; 70.5; Eigenvalue Analysis - REIGL - SFM 5407, SFM 5401
Large Lanczos buckling runs may fail in the REIGL module on the CRAY C-90
release of MSC/NASTRAN with the following messages:

SYSTEM FATAL MESSAGE 5407.

INERTIA (STURM SEQUENCE) COUNT DISAGREES WITH THE NUMBER OF MODES ACTUALLY
COMPUTED IN AN (SUB-)INTERVAL

SYSTEM FATAL MESSAGE 5401 (REIGL).

USER FATAL MESSAGE 4405,

NO EIGENVECTORS COMPUTED FOR COMPONENT MODE SYNTHESIS OR SYSTEM SOLUTION.
Avoidance 1: An avoidance is to set SHFSCL on the EIGRL entry near the
absolute value of the lowest expected root.
Avoidance 2 (preferred over Avoidance 1): include sss alter br.v69 for
Nastran version 69.0.2 or br.v691 for Nastran version 69.1. Leave the
SHFSCL entry on the EIGRL card blank.
@
6086; 68.2; Superelements - PARAM, ERROR, 0 - Misleading Messages, Incorrect
Action
PARAM, ERROR, 0 does not work in SOLs 101 through 200 in Version 68. If
an error occurs in a superelement, or the residual structure, the program
will incorrectly continue processing the same superelement rather than
skip to the next superelement, or if the residual structure is being
process, terminate the run.
Avoidance: Insert the following DMAP alter:
compile phase1dr souin=mscsou noref nolist
alter 100
if ( not skipse ) then $
alter 101
endif $
alter 103
if ( not skipse ) then $
 alter 104
endif $
@
6087; 68.2; Heat Transfer Analysis - heatconv, CHBDY large field format - Wr
ong Conversion
The heat transfer converter program, heatconv, cannot process CHBDY large
field format with ambient points correctly. Specifically, the GA1, GA2,
GA3, GA4 points on the CHBDY continuation entry are not copied to the new
Version 68 bulk data CONV entry.
Avoidance: Use small field CHBDY format.

@
6093; 68.2; Optimization/Design Sensitivity - BNDGRID, RBEi, RBAR - UFM 2101
A, UFM 3001
When a shape optimization job is run which uses BNDGRID entries to define
shape basis vectors, if the same
component of a grid referenced on a BNDGRID entry is also referenced by a
rigid element, then the message "USER FATAL MESSAGE 2101A" is correctly
issued stating that a grid component can not be defined in both S-set and
M-set. This is a user input error.
For versions prior to 68.2, if a shape optimization job uses Analytic
Boundary Shape and also includes MPC or rigid element entries in the
primary model, error message "USER FATAL MESSAGE 3001 MODULE = UPARTN THE
INPUT DATA BLOCK NAMED USET0 AT POSITION 1 DOES NOT EXIST." may be
output.
Avoidance: Replace the offending RBE's with extremely stiff elements.
@
6095; 2001; Material Nonlinear Analysis - CQUAD4, CTRIA3, PSHELL, MID3 - No
Warning Message
Material nonlinear analysis does not support the transverse shear
stiffness in a CQUAD4 or CTRIA3. This is documented in the MSC/NASTRAN
nonlinear seminar notes. However, if the MID3 references a MATS1 entry,
no warning message is issued stating that material nonlinearities are not
supported for this stiffness.
Avoidance: There is no avoidance other than realizing the fact that the
transverse shear stiffness remains linear for CQUAD4 and CTRIA3.
@
6100; ; Aeroelastic Analysis - PK Flutter Analysis - Extra Eigenvectors
If a user requests flutter eigenvectors in a pk flutter analysis and there
are real roots, an eigenvector is printed for each real root. This is
confusing to our users since only half of the real roots are provided in
the flutter summary.
Avoidance: Ignore the first half of the flutter eigenvectors that have
real eigenvalues.
@
6101; ; Aeroelastic Analysis - ADR, AERO - Incorrect Frequency Labels.
If the AEROF request is used in a flutter analysis, the frequency used to
annotate the forces is incorrect. The frequency is computed based on the
 reduced frequency of the flutter analysis. In dimensionalizing this
reduced frequency, the velocity from the AERO bulk data entry (VELOCITY)
is used rather than the velocity from the flutter analysis (VFLUT).
Avoidance: The correct frequency can be computed based on:
f = f * (VFLUT / VELOCITY)
correct printed
@
6111; 68.2; Dynamic Response Analysis - PARAM,DYNSPCF,OLD - Wrong Answers
If PARAM,DYNSPCF,OLD is used in a modal solution and the matrix method of
data recovery is used (the default), incorrect SPC Force output occurs.
Avoidance: Either specify PARAM,DDRMM,-1 or include the following DMAP
alter:
compile sedrcvr, souin=mscsou, nolist, noref
alter 163
TYPE PARM,,CHAR8,Y,DYNSPCF='NEW' $
if (dynspcf='old') then $
DDRMM CASEDR,UH1,OL1,OUGV2,oqg2,OES2,OEF2,XYCDBDR/
OUGV2Y,oqg2y,OES2Y,OEF2Y, $
equivx oqg2y/oqg2/always $
else $
alter 164
endif $
@
6112; 2001; Documentation - MATT8 - MSC/NASTRAN V68 Reference Manual
The documentation in Chapter 5 of the Version 68 MSC/NASTRAN Reference
Manual referencing the MATT8 entry is incorrect. The MATT8 entry does not
exist and it is not possible to define temperature dependent orthotropic
material properties.
@
6116; 68.2; P-Elements - RBE, MPC - UFM 5289
The external id field in the message "USER FATAL MESSAGE 5289, COMPONENT
WITH EXTERNAL ID = ***, DOF = **, APPEARS MORE THAN ONCE ON THE SAME MPC
OR RIGID ELEMENT" is limited to eight characters. Thus, if the message
attempts to identify an internally generated degree of freedom from a
p-element "********" are output.
Avoidance: There is no avoidance.
@
6118; 68.2; Heat transfer - CHBDYE, TA1 - SFM 4276 EC 201
The message "SYSTEM FATAL ERROR 4276.WRITE ERROR CODE 201" occurs when a
CHBDYE element references a non-existing conduction element ID.
Avoidance: Make sure all the CHBDYE have a corresponding conduction
element ID

@
 6123; 68.2; Optimization/Design Sensitivity - DOPTPRM, P2=8 - Missing Respon
se Output
The design optimization print produced when P2 is set to 8 on the DOPTPRM
Bulk Data entry does not
include transient responses.
Avoidance: Insert the following DMAP alter in V68.2:
compile desopt noref nolist
alter 'dom10',''
dbview xxxx=oli where (solapp='mtran') $
DOM10,
DESTAB,XINIT,XO,CNTABRG, CVALRG, CVALO,
DVPTAB,PROPI,PROPO,R1TABRG,R1VALRG,R1VALO,
RSP2RG,R2VALRG,R2VALO,OPTPRMG,OBJTbg,DRSTBLg,
xxxx,fol1,frqrprg//
DESCYCLE/DESMAX/OBJIN/OBJOUT/ $

@
6124; 68.2; Elements - PARAM,COUPMASS,1 CBAR, EMG - SFM 4276, EC 7777
A run that contains BAR elements that have zero area specified on the PBAR
entry but non zero values for K1 and K2 the area factors for shear, may
fail and issue the message "SYSTEM FATAL ERROR 4276, RECOVR ERROR CODE
7777" in the EMG module.
Avoidance: Either remove the area factor for shears or input a non-zero
value for the area of the PBAR.
@
6126; 69.1; Superelements - PLOADX1 - Wrong Answers
The PLOADX1 bulk data entry is not supported when using superelements. No
load is generated for a PLOADX1 entry that references elements that belong
to an upstream superelement,
Avoidance: Place any elements loaded by PLOADX1 entries in the residual
structure.
@
6127; 69; Eigenvalue Analysis - EIGR, GIV, MGIV, HOU, MHOU - Wrong Answers
If any transformation method (i.e. GIV, MGIV, HOU or MHOU) is used to
extract eigenvalues from a two degree of freedom problem the eigenvectors
may be reversed.
Avoidance: Select the lanczos method of eigenvalue extraction.
@
6129; 69; Elements - CTETRA - Poor Answers, Distorted Element Not Detected
Severely distorted TETRA elements may not always be identified, this may
lead to poor answers.
Avoidance: Other than avoiding severely distorted elements, there is no
avoidance.
@
6137; 69; Restart/Superelement - SOL 103/129 - Inefficient Restart
 When performing a restart from SOL 103 into SOl 129, if the model has
multilevel superelements, the restart logic performs dynamic reduction on
the last level of superelements before the residual structure.
Avoidance: The following DMAP alter for Version 68 bypasses the automatic
restart logic which is causing this error. The user must beware that
bypassing the automatic restart logic may result in wrong answers if the
model has been changed.
compile phase0 souin=mscsou noref nolist
alter 336,336 $ v68
alter 267,267 $ v675
endalter

@
6138; 68; OUTPUT2 - RCOUT2 - BUS ERROR (CORE DUMP)
In Version 67.5, if the RCOUT2 program is used to read a formatted output2
file on the HP machine, the job may fail with following error messages:

BUS ERROR (CORE DUMP)

This error only occurs on the HP Version 67.5 system when reading an
output2 that is written in a neutral format.
Avoidance: Use the Version 68 RCOUT2 program to read the Version 67.5
output2 file.
@
6142; 68.2; P-Elements - EMG - UFM 4659
For some complicated geometric surfaces, such as a sphere which has
singularities at the poles, element edges which are mapped directly to the
surface may be distorted enough to cause the the fatal message:
USER FATAL MESSAGE 4659 (EMGPMD) DURING INTEGRATION FOR ELEMENT WITH
ID = ***
BAD GEOMETRY HAS BEEN DETECTED.
Avoidance: There is no current workaround.
@
6147; 68.1 (w/ ZA; Matrix Methods - FBS-Method 2 and Unsymmetric DECOMP (Non-s
parse)
Errors have been found in the Vector Math Kernels for Hitachi VOS3. The
errors will only occur when sparse is deselected (i.e., when NASTRAN
SPARSE=1,USPARSE=0 is specified). There are two separate errors -- one in
FBS Method 2 and one in Unsymmetric Decomp. Both errors can be corrected
using the following ZAP:
//... JOB ...
//*
//* CORRECT ERRORS IN NON-SPARSE VECTOR KERNELS
//*
//ZAPKVCT EXEC PGM=AMASPZAP
//SYSPRINT DD SYSOUT=*
//SYSLIB DD DISP=SHR,DSN=...NAST68.NASTRAN ..NASTRAN LIB..
 //SYSIN DD *
NAME NASTKVCT $PMDINF
VER 0000 C8C9,E3C1
REP 0018 0000,0002,0000,0016,0000,005F
NAME NASTKVCT VOS3VCT2
VER 20CA 47B0,F094
VER 2844 0840,4242,0842,4442
REP 20CA 47D0
REP 2844 0840,4442,0842,4642
NAME NASTKVCT VOS3VC1A
VER 3450 0C40,4442
VER 36E2 2362
VER 373C 2A8A
VER 3742 2B20
VER 3832 2A8A
VER 3838 2B20
VER 3BE2 2362
VER 3C12 2A03
VER 3E72 2362
VER 3EA0 2A03
VER 3FFC 2C44,4C0E
VER 4064 084C,4442,084E,4642
REP 3450 0C40,4242
REP 36E2 2862
REP 373C 2B8A
REP 3742 2A20
REP 3832 2B8A
REP 3838 2A20
REP 3BE2 2862
REP 3C12 2B03
REP 3E72 2862
REP 3EA0 2B03
REP 3FFC 2C44,4C18
REP 4064 084C,4042,084E,4242
/*
//
There is one important thing to note about this ZAP -- one of the routines
being modified is the $PMDINF routine (to change the Vector Kernels module
date to indicate that the ZAP has been applied). Note that this routine
name starts with a dollar sign ('$'). This is actually character code
X'5B' and may be represented as a Yen symbol in some VOS3 systems. It is
very important that this character be checked.

@
6148; 68.2; Optimization/Design Sensitivity - DMAP FM 9015 (FEA) USET TABLE
NOT PRESENT
Shape optimization jobs that contain an auxilary model to define the
boundary shapes but do not include a static subcase may fail with the
message:

DMAP INFORMATION MESSAGE 9052 (FEA) - STATIC ANALYSIS COMPLETED.

DESIGN CYCLE NUMBER= 1
DMAP FATAL MESSAGE 9015 (FEA) THE USET TABLE IS NOT PRESENT FOR:
ANALYSIS = FREQ
MPC = -1
SPC = -1
AUXMID = 1
Avoidance: Apply the following DMAP alter:
compile sedisp souin=mscsou list noref
alter 206
diagon( 20 ) $
alter 207
diagoff( 20 ) $
endalter
compile descon souin=mscsou list noref
alter 108
diagon( 20 ) $
alter 109
diagoff( 20 ) $
endalter
compile desopt souin=mscsou list noref
alter 166
statcc = true $
modecc = false $
buckcc = false $
dfrqcc = false $
mfrqcc = false $
mtrncc = false $
saercc = false $
dvrgcc = false $
flutcc = false $
alter 172,172
CALL BNDSHP CASEXX,EDOM,EQEXIN,GPDT,BGPDT,CSTM,SIL,
USET,CASEVEC,AMLIST,UG,GEOM4/
CASENEW,CVEC,DVIDS,USET0,YG,rmg/
AMBS/GMBS/LUSET $
alter 174
MDCASE CASEXXu/,xcasem,xcaseb,xcasef,xcasemt,
xcasea,xcasedv,xcasef,,,,xcasead,,//
S,N,MODECC/S,N,BUCKCC/S,N,DFRQCC/S,N,MFRQCC/
S,N,MTRNCC/S,N,SAERCC/S,N,DVRGCC/S,N,FLUTCC $
alter 193
bndshp = true $
endalter
compile bndshp souin=mscsou nolist noref
alter 1,1
SUBDMAP BNDSHP CASESTAT,EDOM,EQEXIN,GPDT,BGPDT,CSTM,SIL,
USET,CASEVEC,AMLIST,UG,GEOM4/
CASENEW,CVEC,DVIDS,USET0,YG,rmg/
AMBS/GMBS/LUSET $
alter 35,35
type parm,nddl,i,n, mpc,suport $
mpc = 0 $
suport = 0 $
GP4 CASENEW,GEOM4,EQEXIN,SIL,GPDT,BGPDT,CSTM,,,/
RMG,,USET0,/
LUSET/S,N,NOMSET/0/S,N,NOSSET/0/S,N,NORSET/
NSKIPX/0/0 $
endalter
compile fea souin=mscsou nolist noref
alter 120
type parm,nddl,logical,n, bndshp $
bndshp = false $
$ message //' bndshp in fea = '/bndshp $
alter 204,204
 if ( ((dfrqcc or mfrqcc) and
(feaopt='analysis' or feaopt='datarecovery'))
or
((itol(-dfrqdsa) or itol(-mfrqdsa)) and
feaopt='sensitivity' )
) then $
alter 363,363
if ( (mtrncc and
(feaopt='analysis' or feaopt='datarecovery'))
or
(itol(-mtrndsa) and feaopt='sensitivity')
) then $
alter 453,453
if ( (saercc and
(feaopt='analysis' or feaopt='datarecovery'))
or
((itol(-saerodsa) or itol(-statdsa)) and
feaopt='sensitivity')
) then $
endalter

@
6152; 68.2; P-Elements - GMCORD, GMCURV, FEEDGE - Wrong Answers
When a GMCORD coordinate system is specified on a GMCURV, the 2 and
3-directions are incorrect. The 2-direction is opposite the normal to the
associated GMSURF, and the 3-direction is the cross product of the tangent
to the edge and the normal to the face. The 3-direction should be the
normal, and the 2-direction should be the cross product of the normal to
the surface and tangent to the curve.

@
6155; 69; Miscellaneous - XCSA - Poor Performance
The documentation states that an alphabetic character must be used to
start the ID entry if one is used. If a non-alphabetic character is used,
the program will issue UFM 300. On some client machines, it appears that
excessive time is taken before this message is issued.
Avoidance: Do not start any words on the ID entry with a non-alphabetic
character.
@
6156; 68.2; Elements - CTRIAi - Poor Answers, No Skew Check
There is no skew check performed on triangular plate elements in
MSC/NASTRAN. Excessively skewed elements may result in poor answers.
Avoidance: Many pre-processors have skew checks for triangular plate
elements. A skew check is in MSC/NASTRAN Version 68.2. The test in
Version 68.2 is based upon the three vertex angles. A warning message is
issued if any vertex angle is less than "T3SKEW", where "T3SKEW" is a user
controllable NASTRAN keyword with a default value of ten degrees.
@
6157; ; Performance - DPD - Excessive Time in DPD Module
On large-size models, the main function of the DPD module is to generate
the usetd table from the uset table. It has been observed that the time
to perform this operation on large models can exceed the time to generate
 the uset table and process all rigid elements in the GP4 module.
Avoidance: An avoidance on sol 103 models which do not have EPOINTs is to
write a DMAP to branch over the DPD module, and equiv uset to usetd,
dynamics to eed, and set noue to -1.
@
6159; 69; Documentation - MPCFORCE - Misleading Documentation
The MSC/NASTRAN Version 68 Quick Reference Guide description of the
MPCFORCES command, remark 1, states that the MPCFORCE requests are honored
in the Structured Solution Sequences only. This remark implies that
MPCFORCE output is available in all of the SSS. It is not available in the
nonlinear solutions SOLs 106, 129, 153, or 159. This fact is documented in
Table 3 of the Case Control section of the Quick Reference Guide.
@
6166; 69.1; Eigenvalue Analysis - GIV - Wrong Answers, UIM 4158
Using the GIVENS method to calculate modes of a structure which has a non
positive definite mass matrix, should result in the following error
message:
"USER FATAL MESSAGE 4646---- THE MASS MATRIX IS NOT POSITIVE
DEFINITE,USING THE GIVENS METHOD. USE MGIV INSTEAD"
However, the run may proceed and incorrect answers may be produced. Since
MSC/NASTRAN automatically omits massless degrees of freedom, this error
only occurs with models involving nearly massless mechanism. The user is
warned of the poor answers by a very large value for the factor diagonal
ratio in the following message:
USER INFORMATION MESSAGE 4158 ---STATISTICS FOR CHOLESKY DECOMPOSITION OF
DATA BLOCK MXX FOLLOW
MAXIMUM RATION OF MATRIX DIAGONAL TO FACTOR DIAGONAL =
The user may also be warned of the poor results by the issuance of a
machine dependent message indicating a floating point overflow.
Avoidance: Use MGIV or LANCZOS method of extracting eigenvalues for the
structures of these nature.
@
6167; ; Material Nonlinear Analysis - NLITER - SFM 3007 NLITER (NLQNVC)
A material nonlinear problem using the Quasi-Newton (QN) method for a
nearly elastic-perfectly plastic material (i.e., strain/work hardening
slope, H<<1) may fail with following fatal message:
*** SYSTEM FATAL MESSAGE 3007 MODULE = NLITER
0ILLEGAL INPUT TO SUBROUTINE NLQNVC
The error may occur since the same data point (involved in the QN vector
generation) may be used in two successive iterations.
Avoidance: Change the number of load steps (i.e. number of increments
NINC on NLPARM entry).
@
6168; 68.2; DBC, TRANS, ACCESS - Error in makefile.
The access and util makefiles, <install-dir>/msc68/access/makefile and
<install-dir>/msc68/util/makefile have two errors:
 1) The implicit rules for ".ff.f" are wrong
The error is corrected by changing the following rules (in both
files):
.ff.f:
$(CPP) -P $< $@
to:
.ff.f:
$(CPP) -P $(CPFLAGS) $< $@
2) The install target does not work (the architecture was already
included in the EXE macro).
The error is corrected by removing the macro "$(ARCH)/" from the
pathname of the
install target and all dependent targets.
@
6171; ; Plotting - CGAP - UWM 979
XYPLOTs of gap elements work incorrectly. For instance the command
"XYPLOT stress /899(2)" for gap element 899 causes the following error
message:
0*** USER WARNING MESSAGE 979, AN XY-OUTPUT REQUEST FOR POINT OR ELEMENT
ID 899
-STRE- CURVE IS BEING PASSED OVER. THE ID COULD NOT BE FOUND IN DATA
BLOCK 105
Three plots are created with this command. The first plot is the correct
plot, the second and third are empty plots.
@
6181; 68.2; Optimization/Design Sensitivity - VECPLOT - UFM 5423
If acoustic analysis is requested in SOL 200 with ANALYSIS=MFREQ, DFREQ,
or MTRAN and also ANALYSIS=STATICS, then the run fails in the VECPLOT
module during static analysis data recovery.
Avoidance: Insert the following DMAP avoidance:
compile sedrcvr,souin=mscsou,noref,nolist $
øalter 28,28
IF ( NOT(PEXIST) AND NOT(NLHEAT) AND (NSILS*6)<>LUSETS ) THEN $
@
6184; 70.5.2; Nonlinear Static Analysis - NLPARM - Possible Divergence
If the stiffness update method, KMETHOD field, is specified incorrectly on
the NLPARM entry, a fatal message is not issued and the default AUTO
method is used. Note that the double quotes surrounding the legal choices
in the MSC/NASTRAN Version 68 Quick Reference Guide should not be included
when specifying the KMETHOD field. If the double quote is used, it will
be considered as part of the string specified.
Avoidance: Specify a correct method in the KMETHOD field.
@
6185; 68.2; Installation - MSC/NASTRAN will not run on CRAY UNICOS V9
The nast68 script does not properly identify the model when running under
CRAY Unicos V9, thus MSC/NASTRAN will not run. The script may be
corrected by making the following changes:
 Around line 325, change:
osraw=`uname -m | awk '{print $1}'`
test "$osraw" = "CRAY" && {
to:
uname -m | grep '^CRAY' > /dev/null 2>&1 && {
and a few lines later, change:
sysmod=`uname -m | awk '{print $2}'`
to:
sysmod="Y-MP"
/bin/crayc90 && sysmod="C90"
This changed script will continue to work on UNICOS V7 and V8.

@
6186; 68.2; Optimization/Design Sensitivity - DSAH - SFM 4276 GREAD EC
If a design optimization task contains multiple subcases that include
design constraints on dynamic responses (either frequency or transient)
and there are no active constraints for the first dynamic subcase, the
above fatal error occurs.
Avoidance: A possible avoidance is to force retention of a response using
a DSCREEN bulk data entry with a threshold set to a large negative number.
@
6192; 69.0.2; Eigenvalue Analysis - EIGRL, REIGL - Poor Message
If ND, the number of desired roots, is specified on the EIGRL entry to be
greater than the total number of eigenvalues and , the Eigenvalue Analysis
Summary "TERMINATION MESSAGE" may indicate "NOT ALL EIGENVALUES FOUND IN
RANGE" even though all possible eigenvalues have been found.
Avoidance: Either specify a frequency range or specify an appropriate
value for ND.
@
6196; 68.2; Optimization/Design Sensitivity - DOM9 - SFM 3002 EOR RSP2RG
In a design optimization task, if the objective is a synthetic (DRESP2)
response and there are no constraints on synthetic responses, the above
error may occur.
Avoidance: An avoidance is to formulate and constrain a DRESP2 response.
For example the DRESP2 = 1.0 * DRESP1. In order to force retention of the
constraint, a DSCREEN entry with RTYPE = EQUA and TRS = -1000. should be
used.
@
6201; 68.2; Postprocessing - Heat Transfer, PARAM POST,-2 - Missing Output
In heat transfer analysis solutions 153 and 159, the temperature and flux
results are not properly written to the OUTPUT2 file under when using
PARAM, POST, -2.
 Avoidance: Insert the following DMAP alter:
øcompile sedrcvr souin=mscsou noref nolist
alter 66,66
alter 68,68
IF ( POST=-2 AND GETSYS(NH,56)<>0 ) THEN $
MATMOD ougv1,,,,,/tougv1,/13 $
alter 112,112
IF ( POST=-2 AND GETSYS(NH,56)<>0 ) THEN $
@
6202; 68.2; Installation - MSC/N Will not Install on HP-UX 10.00
A utility program on HP-UX 10.00 returns unexpected information. As a
result, the installation and execution scripts do not work.
Avoidance: The only work-around is to replace the HP-UX 10.00 version's
expr program, found in /bin/expr (a.k.a. /usr/bin/expr) with the HP-UX 9.0
version (found in /bin/expr).
@
6206; 69; Miscellaneous - IBM Hardware/Software Error
The following documentation was received from IBM:
Today we are informing you of an Engineering Change (EC) package that
provides updates for selected systems which you have received. The
package contains two changes that apply to a limited number of the subject
machines.
The EC package will eliminate the remote possibility that these models
could store data to memory incorrectly. While this condition could
potentially cause a data error without generating a system error message,
it can occur only under an extremely rare set of circumstances.
The EC package also includes an enhanced cooling system for the Model 590.
This features a new fan and baffle to improve airflow and cooling in the
processor complex.
The EC package with both changes is available to you at no charge.
Our records show that the package applies to at least one of the machines
shipped to you. You can quickly determine whether any machine can benefit
from the changes by making a simple inquiry from a system-attached
console, as shown on the Attachment.
If you decide that you wish to have the EC package installed, contact your
IBM Service Center at 1-800-IBM-SERV (1-800-426-7378) and request ECA
#190.

Attachment to IBM Letter of January 9, 1995

To Determine if a particular machine within the RS/6000 Models 58H, 590,

990, 99E or 99F could benefit from the Engineering Change (EC) package,
your Systems Administrator can issue an AIX command at a system-attached
console. The procedure should take about 30 seconds to complete and will
not affect other processes that are running on your system.
 From the AIX Prompt, issue the command:
lscfg -l sysplanar0 -v | pg
On the display, read the line which says:
Device Specific.(Z1)........nnYYnn
If YY is:
41 or 42, then the machine, as manufactured, already includes the EC
package.
or
Anything else, then the machine is a candidate for the EC package.
If you call the IBM Service number to request the EC package, your Service
Representative will perform this procedure to verify the status of your
machines before any repair activity begins, and will exchange parts only
if required.

@
6221; 69; Miscellaneous - CORD2i, GP1 - Missing Error Messages
On the DEC/OSF1 (and possibly other machines), when a large number (e.g.,
1.E+35) is used to define the vectors of a secondary coordinate system
(e.g., CORD2S), the job may fail without issuing an error message in the
F06 file. A message such as "abnormal time termination" may be observed
in the LOG file. On other machines, SFM 2005, UFM 316, or SFE 4276 EC
7777 may be issued in the F06 file.
Avoidance: Use smaller numbers to define the vectors.
@
6224; 69; Transient Response - DMAP, TRD1 - SFM 3008
The transient response solutions in MSC/NASTRAN do not support complex
arithmetic. If a complex matrix is introduced by DMAP, the program
attempts to process the problem, but may fail in the module TRD1 with a
message similar to the following: "INSUFFICIENT CORE FOR SPARSE FBS IN
SPFBS USER ACTION: INCREASE CORE BY xxx WORDS. *** SYSTEM FATAL MESSAGE
3008 MODULE = TRD1 INSUFFICIENT MEMORY AVAILABLE FOR SUBROUTINE."
Avoidance: Do not use complex matrices in a transient solution.
@
6226; 68.2; Optimization/Design Sensitivity - SDSC - SFM 4276, Attempt to Re
ad Past EOR
In a dynamic response sensitivity analysis, the above message can occur
when the are multiple superelements and one of the superelements has no
active responses.
Avoidance: A possible avoidance is to explicitly specify each
superelement subcase in the case control and not place a design condition
for the superelements that are known not to be active in the sensitivity
analysis. A less satisfactory fix is to delete the SDSC and DSTAP2
modules in the EXITOPT subdmap. The sensitivity matrix would then be
printed without the IDENTIFICATION OF COLUMNS IN THE DESIGN SENSITIVITY
MATRX table.
@
6228; 69; Optimization/Design Sensitivity - DVPRELi - DSCM Matrix Row Defi
cient
The number of rows in the DSCM matrix should be equal to the number of
independent design variables plus the number of DVPREL2 entries. If there
are no DVPREL2 entries and the trailing independent design variables are
not referenced by any DVPREL1 entries, the number of rows in DSCM matrix
will be deficient by the number of unreferenced trailing design variables.

Avoidance: An avoidance is to rearrange the ID's of independent design

variables so that the unreferenced ID's are not the highest numbers.
@
6229; 68.2; Heat Transfer - CHBDY, Forced Convection Heat Flow - Inconsisten
t Calculations
The forced convection heat flow is calculated by upwinding shape
functions in the solver and by the average fluid temperature in the data
recovery. This inconsistency results in an unbalanced energy between the
forced convection heat flow output and the mass flow convection
(mdot*Cp*Delta T). This error may lead to poor answers for the
"FORCED-CONVECTION" and "HEAT FLOW INTO HBDY ELEMENTS" results.
Avoidance: There is no avoidance.
@
6230; 68.2; Heat Transfer - ECNVMD, NCNVMD - Wrong Answers
If the Prandtl number convection exponents for heating (EXPPI) and cooling
(EXPPO) are not the same, then MSC/NASTRAN incorrectly uses EXPPI for
cooling and EXPPO for heating calculations. In addition, the shape
functions of forced convection boundary conditions (CONVM) are incorrectly
assigned Peclet numbers less
than 0.01 which have negative mass flow rates.
Avoidance: Exchange the values of EXPPI and EXPPO and do not use negative
mass flow rates when Peclet numbers are less than 0.01.

@
6233; 69; Aerodynamic Flutter - SOL 145 - Poor Performance
The solution approach used in SOL 145 is different in Version 68 as
compared to Version 67.5. Invariant matrices are created in Version 68
for aerodynamic optimization. These matrices are larger than the matrices
used in the Version 67.5 approach. Therefore, the performance (e.g., in
the FBS module) may be slower in Version 68 as compared to Version 67.5.
Due to these larger matrices, the disk space requirement may also be
larger.
Avoidance: There is no avoidance.
@
6234; 70; Optimization/Design Sensitivity - DEQATN Size Increased to 32000
Characters
MSC/NASTRAN Version 70 has increased the allowable number of characters
for the DEQATN entry from 12,500 to 32,000.
@
6235; 68.2; Static - PLOAD1, CBEAM, CBEND - Infinite Loop
If a problem contains PLOAD1, CBEAM, and CBEND entries, then MSC/NASTRAN
may become trapped in an infinite loop. The error only occurs when no
PLOADs are applied to any CBEND elements.
Avoidance: Apply a zero dummy load on a CBEND element.
@
6240; 68.2; Static Analysis - QSETi, SSG3 - SFM 3007, UWM 4698, UWM 5221, UF
M 5225
In static analysis, if QSETi entries are defined for SPOINTs and ASETi
entries are defined for physical degrees of freedom then the run will fail
in SOL 61 and 101 with SFM 3007 in module SSG3 and subroutine SSG3A. In
SOL 24, the run will issue USER WARNING MESSAGEs 4698 and 5221 and then
fail with USER FATAL MESSAGE 5225.
Avoidance: Insert the following avoidance for SOL 101 or remove the QSETi
entries:
compile selrrs souin=mscsou noref nolist
alterø 5,5
PARAML USET//'USET'//////'q'/S,N,NOqSET/
't'/S,N,NOtSET $
IF ( FIXEDB <> -2 ) THEN $ A-SET PARTITIONING ( Q + T )
EQUIVX PA/PT/NOQSET $
IF ( NOQSET>=0 AND NOTSET>=0 ) UPARTN USET,PA/PT,,,/'A'/'T'/'Q'/1 $
ENDIF $
EQUIVX PT/PL/NORSET $
alter 8,8
UPARTN USET,PT/PL,PR,,/'T'/'L'/'R'/1 $

@
6241; 69; Random Response Analysis - STRESS, BEAM - Wrong Answers
If stresses are requested for BEAM elements in a random response analysis,
and the STRESS output is requested as real/imaginary values, then the PSD
output for the RMS stresses at point F of end A are incorrect.
Avoidance: Request the stresses in magnitude and phase by using the
STRESS(PHASE) Case Control command.
@
6242; 69; Optimization/Design Sensitivity - DSAD - Datablock Does not Exis
t, Poor Message
The message "INPUT DATA BLOCK NAMED OUGVS1 AT POSITION 12 DOES NOT EXIST.
DSAD" is produced when the user requests responses at a set of grid points
that do not exist. If some of the grid points exist and some of them
don't, a more meaningful message is produced. A similar nonexistent
datablock error is produced for other response types; e.g., the OES1
datablock is missing if none of the requested elements exist.
Avoidance: Do not request responses for grids that do not exist.
@
6244; 68.2; Post Processing - PARAM,POST,-1; PATVER,2.5 - Extra Data Block L
AMA
If PARAM,POST,-1 and PARAM,PATVER,2.5 are used in SOL 103, MSC/NASTRAN
Version 68 incorrectly writes the LAMA (OLB) datablock to the output2 file
for PATRAN. PATNAS cannot recognize this datablock and issues an error
message.
Avoidance: Use the following DMAP alter:
COMPILE SEDRCVR SOUIN=MSCSOU NOLIST NOREF
ALTER 92 $
IF (POST=-1 AND PATVER=2.5) THEN $
OUTPUT2 OUGV1//OTAPE2/OUNIT2//OMAXR $
ELSE $
ALTER 93 $
ENDIF $
@
6249; 69; Miscellaneous - PARAM,NMLOOP - UWM 4500, DMAP FATAL MESSAGE 9015
If "PARAM,NMLOOP" is included in the Bulk Data section for a Structured
Solution Sequence input file other than SOL 106 or 153, then the job may
fail with one or more of the following error messages:
USER WARNING MESSAGE 4500, STIFFNESS MATRIX PURGED. GRID POINT SINGULARITY
NOT TEST
DMAP FATAL MESSAGE 9015 (SEBUCKL) - THE USET TABLE IS NOT PRESENT
Note that this parameter is not used except in SOL 106 and 153.
Avoidance: Remove the offending parameter as this parameter is only used
for SOL 106 and 153.
@
6252; 69; Documentation - SPCFORCE - PLOT Option is Available
The SPCFORCE output request supports the plot option even though it is not
listed in the MSC/NASTRAN Quick Reference Guide.

@
6253; 68.2; Complex Eigenvalue Analysis - CEAD, CLAN - Incorrect Eigenvector
s
The complex lanczos method of eigenvalue extraction may produce
non-orthogonal eivenvectors for models with multiple roots at the same
eigenvalue. This error is under investigation, currently it has been
experienced on short-word machines only. The CRAY produced correct
results.
Avoidance: Use the Hessenberg method.
@
6255; 69; Matrix Methods - SOL 307 - SFM 3001
In a run using the default sparse solver if there is insufficient memory,
then the program should revert to the active column method of
decomposition and continue. In this condition "USER WARNING MESSAGE
6136" is output. However, for SOL 307, after outputting this message, the
program issues the following fatal message, "SYSTEM FATAL MESSAGE 3001
MODULE = CEAD THE EXECUTIVE FILE 0 DOES NOT EXIST. PROGRAMMER
INFORMATION: SUBROUTINE GOPEN".
 Avoidance: Increase the memory using the 'mem' command line keyword or
turn off the sparse solver with the by setting the SPARSE keyword to 1.
@
6257; 69; Miscellaneous - File Not Found, UFM 423
Prior to Version 69, filenames were limited to 80 characters. Filenames
that were longer were trunctated by MSC/NASTRAN, this generally resulted
in file not found errors.
Version 69 increases the allowable filename length limit to 256
characters.
@
6258; ; Documentation - DMAP Module Dictionary - MATMOD Option 16
The first line of the example DMAP for MATMOD option 16 in the MSC/NASTRAN
DMAP Module Dictionary (page 196) is incorrect.
Avoidance: In the DMAP example, replace the following line
MATMOD EQEXIN/INTEXT/9/0/LUSET $
with
MATGEN EQEXIN/INTEXT/9/0/LUSET $
@
6260; ; Nonlinear Heat Transfer - NOLINi, ADAPT - Wrong Answers
The ADAPT strategy and the PW convergence criterion leads to an erroneous
solution if NOLIN cards are enforcing square wave type of power input
loading that is based on specified temperatures at certain grid locations
of the model. Particularly, the iteration strategy fails when the
temperature of the curve is nearly constant. This phenomenon occurs for
the ablation problem for which NOLIN entries are used to model the heat
transfer through out the medium.
Avoidance: Set the ADJUST of the TSTEPNL entry to zero, so that the
automatic time step adjustment is disabled. Also set the convergence
criterion to U, and set the EPSU to 1.0-5 instead of the 1.0-3 default
value. This avoidance yields correct answers but the actual CPU time to
reach convergence will be much longer than using SOL 89 for Version 67.5.
****New Avoidance****
Set the ADJUST equal to zero , and set the convergence criterion to P
only.
@
6261; 70; Miscellaneous - ADAPT,OUTPUT,OUTRCV - UFM 300
If free field format is being used with the ADAPT, OUTPUT, or OUTRCV bulk
data entries, and the first comma does not occur in columns 1-8, the fatal
message, "USER FATAL MESSAGE 300, DATA ERROR IN FIELD UNDERLINED", occurs.
Avoidance: Eliminate spaces between the entry name and the first comma.
@
6264; ; Thermal Stress Analysis - TEMPBC - Wrong Answers
The TEMPBC bulk data entry is used to specify time-dependent temperatures
(TYPE=TRAN) at grid points for transient thermal analysis (SOL 159).
Internally in MSC/NASTRAN, this time dependent load is accomplished using
 SLOAD and the CELAS2 bulk data entries. For thermal analysis, these
generated CELAS2's produce correct temperatures. However, if a TEMPBC
entry with the TRAN option is included in a thermal stress analysis (SOL
101, SOL 106), then the thermal stress calculations are wrong because
extra forces are introduced in the problem.
Avoidance: Remove TEMPBC entries with TYPE=TRAN.
@
6266; 68.2; Miscellaneous - nastran/emas script - Unexpected Result for post
Command
In the nast68/emas3 script, the post and app=yes operations may interfere
with one another. If a client attempts to compress the output files from
MSC/NASTRAN or MSC/EMAS and uses app=yes, the obvious command is:
nast68 myjob app=yes post='gzip $out.out'
Unfortunately, the append commands are run after the post command, so the
file myjob.out does not yet exist.
If the user were to change the command to
nast68 myjob app=yes post='gzip $out.f06 $out.f04 $out.log'
the append operation will fail because the three output files have been
compressed and renamed.
Avoidance: A work-around that provides the effect of the initial command
is:
nast68 myjob app=no post='cat $out.f06 $out.f04 $out.log | gzip -c >
$out.out.gz'
NOTE:
The versioning operation performed by old=yes will _not_ save files with a
".gz" suffix.
@
6267; 68.2; Static Analysis - SUPORT - UFM 3055
In MSC/NASTRAN Version 68, if multiple SUPORT sets are used in SOL 101, it
is possible to get "USER FATAL MESSAGE 3055 ATTEMPT TO MULTIPLY
INCOMPATIBLE MATRICES". This error occurs when there are a different
number of loading conditions for the various SUPORT points.

Avoidance: If multiple SUPORT sets are used, insert the following DMAP
alter for SOL 101
compile sestatic, souin=mscsou, nolist, noref
alter 69
type parm,,i,n,storit=0 $
call dbstore crx,,,,//suport/0/'DBALL'/s,storit $
$
compile sedisp, souin=mscsou, nolist, noref
alter 14
type parm,,i,n,gotit=0
purgex /crx,,,, $
call dbfetch /crx,,,,/suport/0/0/0/s,gotit $
 $
@
6269; 68.2; SSSALTERS - DDAMF.FOR - Missing Subroutines, DDAM Will not Compi
le
The ddamf.for FORTRAN program delivered with the SSSALTER library for the
ddam related alters is missing two subroutines. This is a Version 67.5
plus error.
Avoidance: The updated ddamf.for will be delivered with Version 68.2. If
you need a copy of this updated program, you can contact your local
technical support representatives.
@
6270; 70.5; Buckling Analysis - EIGRL - SFM 5299, SFM 5401, UFM 4405
In a buckling analysis, if the buckling load is reversible (e.g., a
torsional buckling load applied to a cylinder), then the job may fail with
the following error messages:
*** SYSTEM FATAL MESSAGE 5299.UNRECOVERABLE TERMINATION FROM LANCZOS
ITERATION
*** SYSTEM FATAL MESSAGE 5401 (REIGL). LANCZOS METHOD IS UNABLE TO FIND
ALL EIGENVALUES IN RANGE.
*** USER FATAL MESSAGE 4405,
NO EIGENVECTORS COMPUTED FOR COMPONENT MODE SYNTHESIS OR SYSTEM
SOLUTION.
In this case, the Lanczos shift logic finds many negative legitimate
roots. The number of these negative roots will increase as the model size
increases.
Avoidance 1: If you are interested in only a few modes, which is the case
for most buckling analysis, use the SINV method. If you wish to use the
Lanczos method for buckling and you encounter the above problem, specify a
shift scale value on the EIGRL entry.
Avoidance 2: (preferred over Avoidance 1 for Nastran versions 69 and
69.1): Include alter bra.v691 from the SSS alter library. This alter
computes an estimate for the lowest buckling mode then passes this
estimate to REIGL where it is used as a shifting scale. With this alter
REIGL solves buckling problems much better.
@
6271; 69; Aeroelastic Analysis - AESTATRS - Wrong Answers
In a static aeroelastic analysis, if two dynamic pressures differ by less
than 5.0e-4, the analysis will be repeated with the results for the first
dynamic pressure duplicated rather than new results being generated for
the second.
Avoidance: The following DMAP avoidance gives differing results whenever
the dynamic pressures differ by more than 5.0e-7:
compile aestatrs souin=mscsou nolist noref
alter 30,30
IQ=NINT(1000000.*Q) $
@
 6272; 69; Documentation - DMIG, M2GG, K2GG, B2GG - No Check for Symmetry
User-provided matrices, such as M2GG, K2GG, and B2GG, are not checked by
the program for symmetry. If non-symmetric matrices are entered for any
matrix which should be symmetric, the program may fail in a later
operation, or give incorrect answers.
Avoidance: Use the symmetric option on the DMIG entry.
@
6274; 68.2; Heat Transfer - RADSET - Wrong Answers for Multiple Large Caviti
es
For multiple radiation enclosures, the contents of the radiation matrix
for the first cavity is overwritten if the element size of the second
cavity is greater than 200. This error also occurs for large sizes of
third, fourth and subsequent cavities.
Avoidance: Use a single cavity or keep the sizes of the second, third, and
subsequent cavities small.

@
6282; 69; Buckling - Thermal Loads - Wrong Answers
If multiple buckling analyses are executed in a single solution 105 run,
and thermal loads are present, then wrong answers (eigenvalues,
eigenvectors etc.) are generated for the second and greater buckling
subcases. The error occurs because MSC/NASTRAN uses incorrect temperature
information when generating the differential stiffness matrix for the
second and greater buckling subcases. The solutions in the static subcases
are correct.
Avoidance: Specify all loads in both the static and buckling subcases,
and include the following dmap alter:
compile sekdr souin=mscsou nolist noref $
alter 25, 25 $
if (seid = 0) then $
case cases,/case1/'tran'/colno $
else $
equivx cases/case1/-1 $
endif $
call setq case1//seid/peid/s,mtemp/s,k2gg/s,m2gg/
s,b2gg/s,mpc/s,spc/s,load/s,deform/s,templd/
s,p2g/s,dyrd/s,meth/s,mfluid $
message //'load '/load/'templd '/templd $
endalter $

Example case control to avoid error:

SPC = 1
SUBCASE 1
LOAD = 21
TEMP(LOAD)=11
SUBCASE 2
LOAD=22
TEMP(LOAD)=12
SUBCASE 3
 STATSUB=1
LOAD = 21 $ AVOIDANCE ERROR 5977
TEMP(LOAD)=11 $ AVOIDANCE ERROR 6282
METHOD = 1
SUBCASE 4
STATSUB=2
LOAD = 22 $ AVOIDANCE ERROR 5977
TEMP(LOAD)=12 $ AVOIDANCE ERROR 6282
METHOD = 1
BEGIN BULK
ETC...
@
6285; ; Nonlinear Transient Analysis - MATi, TRD1 - SFM 4276 EC 105
In a nonlinear transient problem (SOL 129), the job fails in TRD1 and the
misleading message "SYSTEM FATAL ERROR 4276.PACK ERROR CODE 105" is output
whenever all the mass properties are omitted. Therefore, the message
occurs whenever the user forgets to specify the mass density on their MATi
bulk data entries and doesn't have any other concentrated masses (CONMi,
CMASSi) specified.

Avoidance: Specify the mass density on MATi bulk data entries or include
a concentrated mass.

@
6289; 69; Static Analysis - GPSDCON, ELSDCON - SFM 4276 EC 3002
When attempting to calculate Grid Point Stress Dicontinuities (GPSDCON) or
Element Stress Dicontinuities (ELSDCON) on a model containing HEXA, PENTA
and TETRA solid elements it is possible to get the following fatal message
"SYSTEM FATAL ERROR 4276.FREAD ERROR CODE 3002" if the calculations are
done over a volume containing both types of elements.
Avoidance: Perform two runs and define the Volume set in the OUTPUT(POST)
section to be PENTA or HEXA only. Also, grouping the element types by
element id number allows the STDCON module to run.
@
6291; 69; Miscellaneous - SORT2, PUNCH - Missing SUBCASE ID
If punched output is requested in SORT2 format (e.g.,
STRESS(SORT2,PUNCH)=all), then the punch output does not contain the
appropriate subcase labels. This error is considered fixed in Version 69
since the subcase id may be obtained by setting system(210)=1.
Avoidance: Either request SORT1 output or specify system(210)=1.
@
6292; 69; Nonlinear Transient Heat Transfer - 20 Node HEXA - Infinite CPU
Loop
If a 20 node HEXA element is used in a transient thermal analysis (SOL
159), the solution will not converge because the weighting of the
capacitance matrix is not correct. The job may go into an infinite CPU
loop and will have to be manually killed by the user.
Avoidance: Use 8 noded HEXA instead.

@
 6294; ; Nonlinear Transient Analysis - TSTEPNL - SFM 4276 EC 1044
Solution 129 runs may fail with the message, "SYSTEM FATAL ERROR 4276.OPEN
ERROR CODE 1044" in the NLTRD2 module. The error is still under
investigation but the error only seems to appear when the NO field is not
set to 1 or the timestep field (NDT) is not 1.E**.
Avoidance: Specify the NO field to 1 or NDT to 1.E** (i.e., .001, .01,
.1, 1. ...).
@
6296; 69.0.2; Nonlinear Analysis - DBC, PARAM,POST,0 - DBCMSG ERROR 23 FROM GR
IDBK/ELMBLK
If in a Version 68.2 nonlinear transient analysis with PARAM,POST,0 the
following warning messages are encountered the graphic database, xdb, will
be corrupt.
0*** SYSTEM WARNING MESSAGE 6065 (DBC)
*** DBCMSG ERROR MESSAGE 23 FROM SUBROUTINE GRIDBK
OUTPUT LIST SYNCHRONIZATION PROBLEM FOR TOL VALUE =
1.2879E+00, POSITION = 52
0*** SYSTEM WARNING MESSAGE 6065 (DBC)
*** DBCMSG ERROR MESSAGE 23 FROM SUBROUTINE ELMBLK
OUTPUT LIST SYNCHRONIZATION PROBLEM FOR TOL VALUE =
1.2879E+00, POSITION = 52
Avoidance: The following DMAP avoidance may avoid the error in Version
68.2. The error has been corrected in Version 69.0.2; however most Unix
workstations were released as Version 69.0.0 and Version 69.0.2 must be
specially requested.
COMPILE NLTRAN SOUIN=MSCSOU,LIST,NOREF
ALTER 392,392
ALTER 406
IF (POST=0) DBC OESNLXD ,,,,,,,,,,,,,,,,,,,//
'OESNL'////////////////////
-1/DBCPATH/S,N,CP/'NLST'//GEOMU/LOADU/POSTU/
DBCDIAG/DBCCONV/DBCOVWRT $
@
6307; 68.2; Heat transfer - SSG1, QVECT, RADBND - SFM 3007
If the QVECT (direction heat flux) is used in conjunction with RADBND
(wave-length dependent radiation) for more than one element , then you
SYSTEM FATAL MESSAGE 3001 will be encountered in the SSG1 module. The
exact message is:
SYSTEM FATAL MESSAGE 3007 MODULE = SSG1
ILLEGAL INPUT TO SUBROUTINE HMAT
FATAL ERROR
Avoidance: Instead of using the QVECT entry to specify heat flux, use the
QHBDY entry to define the heat load.
@
6308; ; Grid Point Stress Discontinuities - GPSTRESS, GPSTR1 - Poor Accu
racy
Even though the stress discontinuities are exactly zero in the normal-x,
normal-y and shear-xy directions the major and minor directions may
 indicate small, non-zero stress discontinuities.
@
6310; 68.2; Buckling - ESE, GPFDR - Missing Output, Infinite CPU Loop, SWM ?
???, UWM 2354
Element strain energy and grid point force does not work in Version 68,
SOL 105 buckling runs. A request for ESE or GPFORCE may produce an
infinite cpu loop or SYSTEM FATAL MESSAGE 4276 with a traceback in the log
file. The job may eventually fail due to lack of disk space because the
following warning message is issued repeatedly:
USER WARNING MESSAGE 4506--- TOTAL STRAIN ENERGY IS ZERO. THE PERCENT OF
TOTAL STRAIN ENERGY IS ALSO SET TO ZERO.
USER INFORMATION MESSAGE 4507--- IN SUBCASE ****
THERE ARE *** QUAD4 ELEMENTS HAVING STRAIN ENERGY WHICH IS LESS THAN
.0010 PERCENT
OF THE TOTAL STRAIN ENERGY OF ALL ELEMENTS.
The following message is another symptom of this error.
SYSTEM WARNING MESSAGE ???? ***, GPFDR FINDS A SYMMETRY SEQUENCE LENGTH
19 WHILE ATTEMPTING TO COMPUTE GRID POINT FORCES AND ENERGIES. THERE IS
AN INSUFFICIENT NUMBER OF VECTORS AVAILABLE 0
USER WARNING MESSAGE 2354. GPFDR MODULE IS UNABLE TO CONTINUE AND HAS
BEEN TERMINATED DUE TO ERROR MESSAGE PRINTED ABOVE OR BELOW THIS MESSAGE.
THIS ERROR OCCURRED IN GPFDR CODE WHERE THE VARIABLE -NERROR- WAS SET =
9
Avoidance: Insert the following DMAP alter:
compile sedrcvr souin=mscsou noref nolist
alter 264
øIF ( APP1='bkl1' ) APP2='reig' $ø

@
6311; 70; Optimization/Design Sensitivity - SOL 200, GDR, CMS, QSET - Dyna
mic Reduction
A request for dynamic reduction ( CMS or GDR ) on upstream superelements
is ignored in SOL 200. The error only occurs when a statics subcase is
also included along with the modes subcase.
Avoidance: Specify the METHOD and/or DYNRED Case Control commands in the
static subcases.
@
6312; 68.2; Heat Transfer - CHBDYE - UWM 2154
A CHBDYE which is defined as side 4 of a PENTA element does not work. If
attempted the message "USER WARNING MESSAGE 2154, A ZERO AREA OR ILLEGAL
CONNECTION FOR CHBDYE ELEMENT NUMBER xxxxx" is generated.
Avoidance: If possible reorder the grids so the CHBDYE can be defined for
a side other than side 4.

@
6313; 69; DBC/TRANS/ACCESS - CHBDYP, PHBDY, Heat Transfer - SFM T10 (XMIT)
 TRANS will not work on heat transfer decks that contain CHBDYP elements
with
PHBDY property cards. The following message is written out and no neutral
file is created:
*** SYSTEM FATAL MESSAGE T10 (XMIT)
A DATABLOCK WAS ENCOUNTERED THAT HAD MORE FIELDS THAN WERE DEFINED
IN THE DATA DEFINITION LANGUAGE.
DATABLOCK NAME = PHBDY
DATABLOCK LENGTH = 7
DATA DEFINTION LENGTH = 4
Avoidance: The only avoidance is to run MSC/NASTRAN on the machine that
you intend to perform post processing on (or other compatible machine) so
that TRANS is not required. Note that machines that use the same UNIX
IEEE structure machines are compatible and TRANS is not required. The
list of compatible machines is documented in Chapter 6 of the Installation
and Operation Instructions. IBM RS/6000, SGI, CONVEX, HP 9000, SUN SPARC
SUNOS, and SUN SPARC Solaris are compatible machines.
@
6316; 70; Frequency Response/Random Analysis - RANDOM - BIOMSG: ERROR 3001
IN FREAD
If the RANDPS entry references subcases that do not exist the following
uninformitive message is issued:
BIOMSG: ERROR 3001 HAS OCCURRED IN ROUTINE FREAD
Avoidance: Correct the RANDPS entry that is in error.
@
6317; 68.2; Optimization/Design Sensitivity - DRESP1 - UFM 6293
"USER FATAL MESSAGE 6293 (DSAD16) GRID NO. *** REFERENCED ON A DRESP1
ENTRY COULD NOT BE FOUND" may be issued in error in a frequency response
optimization or design sensitivity analysis if there are multiple
frequencies and multiple grid points appearing on DRESP1 entries, but not
every response is designed for every frequency.
Avoidance: The first check should be to make sure the message is in
error. If this appears to be the case, it is necessary to add a
constraint for every response at every frequency. Very generous limits
could be set on these added constraints so that they play no role in the
design process.
@
6319; ; DMAP - INPUTT2 - Poor Message
If a DMAP alter using INPUTT2 is used, and a parameter is used to specify
the INPUTT2, the parameter must be assigned a value before execution of
the INPUTT2 module. If no value is assigned, the module will fail with a
FORTRAN read error. The message may be different for each computer type,
and appears only in the log file. The .f06 and .f04 files end abruptly and
no error message is printed.
Avoidance: Be sure the unit number is defined before INPUTT2 is used.
@
6322; GL; Superelements - SEMA - SFM 4252, Inefficient restart
On a restart run, if the user changes PARAM,AUTOSPC for the residual
 structure, then SYSTEM FATAL MESSAGE 4252 may be issued or all
superelements will be recomputed. This error was found on a sol 103
restart to sol 129. In a sol 129 run, the PARAM,AUTOSPC default is "NO"
in the residual and it should not be changed. Therefore, single point
constraints for singularities that were eliminated by PARAM,AUTOSPC,YES in
the sol 103 run must be explicitly defined in sol 129 for the residual.
Avoidances:
1) Specify PARAM, SPCGEN, 1 in the cold start run to punch SPC entries
based on the
the AUTOSPC operation. Include the puched SPC entries in the restart
Bulk Data section.
2) Use the following alter
compile ifpl souin=mscsou,nolist,noref
$alter 89,89 $ v67.5 removes autospc restart check
alter 150,150 $ v68, autospc
compile phase0 souin=mscsou,nolist,noref
$alter 231,231 $ v675 autospc
alter 304,304 $ v68 autospc

@
6326; 68.2; Optimization/Design Sensitivity - PARAM,AUTOSPC; MERGE, PARTN -
Various SFM
In SOL 200 with PARAM, AUTOSPC, YES, if the model is a fluid/structure
model then the run may fail with various matrix partition or merge errors
indicating incompatible matrices. The error occurs after the first design
cycle if the AUTOSPC operation produces a different set of constraints
then in the previous cycle. (Version 68 only)
Avoidance: Insert the following DMAP alter:
compile dpd1 souin=mscsou noref nolist
alter 29
delete /vafs,,,, $
@
6328; 69; Authorization - UFM 3060
If an MSC/NASTRAN Version 67.5 based authorization file is merged with a
Version 68 authorization file, it is possible that Version 67.5 runs will
fail with USER FATAL MESSAGE 3060 indicating that OPTION NOT IN APPROVED
LIST even though the file contains a valid Version 67.5 authorization
code. This problem only occurs when Version 68 authorization codes
precede the Version 67.5 authorization code.
Avoidance: Version 68 authorization codes should be appended to the end
of the authorization file.
@
6329; ; Optimization/Design Sensitivity - PCOMP - Wrong Answers
If the ply orientation angle is selected as as a design quantity and the
initial ply orientation angle specified by the DESVAR/DVPREL1 combination
is 0.0, incorrect sensitivities can result. The error occurs because the
finite difference perturbation used to produce the sensitivity is too
small when the initial ply angle is zero.
Avoidance: Provide an initial design with a non-zero ply orientation
angle.
@
6332; 68.2; Eigenvalue Analysis - EIGRL, REIGL, FBSMEM, SYSTEM(146) - SFM 54
01, UFM 4405
If the NASTRAN keyword FBSMEM (a.k.a. SYSTEM(146)) is set to too large a
value such that there is insufficient memory to honor the request, then
LANCZOS stops in the REIGL module and issues the messages:
SYSTEM FATAL MESSAGE 5401 (REIGL). LANCZOS METHOD IS UNABLE TO FIND ALL
EIGENVALUES IN RANGE.
ACCEPTED EIGENVALUES AND ADDITIONAL ERROR MESSAGES MAY BE LISTED ABOVE.
USER ACTION: RERUN WITH ANOTHER METHOD OR ANOTHER SETTING ON EIGRL ENTRY.
USER FATAL MESSAGE 4405
NO EIGENVECTORS COMPUTED FOR COMPONENT MODE SYNTHESIS OR SYSTEM SOLUTION
Avoidance: Large values of FBSMEM are not necessary in Lanczos eigenvalue
analysis. A value of 4 is the largest recommended value. Larger values
are only a significant benefit to frequency response and transient
solution sequences. Either increase the mem request or lower the FBSMEM
request.
@
6338; 68.2; Heat Transfer - TRIAX6, QVOL - Wrong Answers
TRIAX6 elements with QVOL heat loading produce wrong answers on long word
machines such as CRAY Y-MP and CRAY C-90.
Avoidance: There is no avoidance.
@
6339; 68.2; Optimization/Design Sensitivity - DESGLB - UFM 6545 (DOPR3)
The following fatal error message is issued for superelement optimization
when a global response is selected from case control using the DESGLB
entry. This error only occurs when a weight or volume response is selected
in a non-statics case.
USER FATAL MESSAGE 6545 (DOPR3)
A DESGLB = 10 COMMAND APPEARS AT THE SUBCASE LEVEL.
USER ACTION: EITHER REPLACE THE DESGLB COMMAND WITH A DESSUB COMMAND OR
PLACE THE DESGLB REQUEST ABOVE THE FIRST SUBCASE.

Avoidance: The only avoidance is to specify a dummy statics subcase and

to select the global (weight or volume) response with a DESSUB command
from the statics subcase.
@
6348; 69; Optimization/Design Sensitivity - DVPREL1, PBEND - UFM 7052, Mis
leading Message
If an invalid property type is referenced on a DVPREL1 entry, the message
"USER FATAL MESSAGE 7052, DOMPTC. ILLEGAL PROPERTY ENTRY NAME *****" is
printed. In most cases this message would be caused by a typographical
error on the DVPREL1 entry and would be easy to spot. However, the
message is misleading for the legal PBEND property entry. As is
documented in remark 3 under the DVPREL1 entry, the PBEND entry is not
supported on the DVPREL1 entry. The message should read "USER FATAL
MESSAGE 7052, DOMPTC. ILLEGAL PROPERTY NAME ***** REFERENCED ON DVPREL1
 ENTRY". Note that UFM 7052 is not documented in the Reference Manual and
needs to be added.
@
6353; 68.2; Optimization/Design Sensitivity - DRESP1 - UFM 5502
If a frequency sensitivity or optimization is performed at multiple
frequencies and there are multiple
element types and only a subset of the output frequencies are invoked on
DRESP1 entries, then the following message is issued:
USER FATAL MESSAGE 5502 (DSAD17)
ELEM NO. **** REFERENCED ON A DRESP1 ENTRY COULD NOT BE FOUND.
USER ACTION : USE OF UNREFERENCED ELEMENTS IS NOT RECOMMENDED AND
MAY INDICATE A USER INPUT ERROR,
RESULTS FOR OTHER RESPONSES ARE NOT AFFECTED.
One message is output for each element in the model.
Avoidance: Only perform the analysis at frequencies for which design
results are required.
@
6355; 2001; Miscellaneous - IFPL, MAT1, MAT8 - SIGNAL NUMBER 11
If the same ID is used a MAT1 and MAT8 entry, the job should fail with the
following message:
USER FATAL MESSAGE 2017, MATERIAL PROPERTY IDENTIFICATION x WAS
DUPLICATED.
On the Intel SCO/UNIX, the job terminates without any error message in the
f06 file. A symptom of this error is the appearance of the following
message in the log file:
SIGNAL NUMBER 11

Avoidance: Use unique ID's for each MATi entry.

@
6359; 68.2; Nonlinear Analysis - Superelements, OLOAD, CLOAD - Incorrect OLO
AD Output
In nonlinear superelement analysis the OLOAD resultant for the residual
includes loads from the upstream superelements. This error has been
corrected in SOL 106 in MSC/NASTRAN Version 68.2.
Avoidance:
compile dmap=NLSTATIC souin=mscsou nolist noref
alter 290,290 $
VECPLOT PGRES,BGPDTH,EQEXINS,CSTMS,CASESX2,,,/PGB/
GRDPNT/0/+1/'OLOAD' $
@
6360; ; Miscellaneous - NFS Mount - Job Hangs
Submitting a job from a machine that is nfs mounted may result in the job
hanging. The exact cause of this specific problem is under investigation
but is often the result of incompatibilities in the operating systems.
Avoidance: Prior to MSC/NASTRAN V69, set system cell SYSTEM(207)=1;
 after that, set the keyword "sysfield=lock=no". This feature is designed
to block file locking for those situations where, despite the correct
programming of the file locking calls, the system hangs or aborts.
@
6361; 68.2; Postprocessing - PARAM,POST,-1/PARAM,OUGCORD,BASIC - Missing out
put
If the neutral format option is requested when using "param,post,-1" and
"param,ougcord,basic", then the ougv1pat datablock will not be written to
the output2 file. This is due to the fact that the datablock ougv1pat is
not defined in the NDDL.
Avoidance: Add the following DMAP alter avoidance:
compile sedrcvr souin=mscsou nolist noref
alter 74,74 $
OUTPUT2 OUGV1PAT//OTAPE2/OUNIT2//OMAXR/'oug' $ V68
@
6364; 68.2; Single and Multipoint Constraints - MPCFORCE - Wrong Answers for
Superelements
In static analysis in the structured solutions the MPCFORCE output is
wrong for upstream superelements when rigid elements are on the boundary
between the residual and the superelement or completely interior to the
superelement.
Avoidance: Insert the following DMAP alter:
ø
compile sedisp souin=mscsou noref nolist
alter 121,121 $ PARTN PG,COL,ROW/,,,PMI/1 $
upartn uset,pg/pmx,,,/'G'/'M'/'S'/1 $
if ( seid=0 ) then $
PARTN Pmx,COL,/,,,PMI/1 $
else $
equivx Pmx/pmi/-1 $
endif $
@
6365; ; Material Nonlinear Analysis - QUADX, TRIAX - Poor Answers
The radial degree of freedom should be constrained for all grid points on
the axis of symmetry for hyperelastic axisymmetric problems else poor
answers may result.
@
6366; 2001; Elements - CROD - Wrong Answers
When a CROD or CONROD that has torsional properties only, nonzero axial
stresses are printed if the attached grids move.

Avoidance: Use a torsion only CBAR or ignore the nonzero stresses.

@
6367; 68.2; Optimization/Design Sensitivity - OFREQ, ANALYSIS=DFREQ - SFE 42
76, EC 3001
If a OFREQ Case Control command is present in a direct frequency response
optimization/design sensitivity analysis the following message is issued:
 SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 (DSAD)
Avoidance: Insert the following DMAP alter:
$
compile fea souin=mscsou list noref
alter 297,298 $
CALL SUPER3,
CASECC, ,ULF,FOL1,PPFX,PSF1, , , ,DLT,
, , ,,,,
PCDB,XYCDB,POSTCDB,,EMAP,MAPS,EQEXINS,PVTS,CASES,
SLT,ETT,GOAT,GOAQ,,,,UFOO,,GPLS,USET,SILS,
PJ,EDT,YS,GM,,KFS,KSS,QR,,,,MAR,MEA,
DYNAMICS,BGPDTS,CSTMS,MPTS,DIT,EST,GEOM2S,GEOM3S,ECTS,
EPTS,INDTA,,,GPECT,,,SILX,EQEXINX,
ECTX,BGPDTX,,GEOM4S,PG,,,,EPT,MPT,
GEOM1,GEOM2,GEOM3,GEOM4,,,,CASERS,,,,,,GPSNTS,PVT,,,,,,
KSF,KMG,MMG,MSF,BMG,BSF,K4MG,K4SF,
K2MG,M2MG,B2MG,K2SF,M2SF,B2SF/
UPF,QPF,CASEDR,XYCDBDR,PUG,,oli,
OUPF1,OPPF1,OQPF1,OFEF1,OFES1,OFEE1,,
OUPF2,OPPF2,OQPF2,OFEF2,OFES2,OFEE2,,
,,,,,,/
$ /APP1 /CYCLIC/NLHEAT/AERO
APP/APP /RSONLY/FALSE /FALSE /FALSE/DROPT/FS/
S,PFILE/S,CARDNO/DESCYCL1/FALSE/DMRESD/FALSE/0/FALSE $
$
IF ( FEAOPT='ANALYSIS') EQUIVX OLI/FOL1/ALWAYS $
$
ENDALTER $
@
6368; ; p-Elements - FEEDGE - Undetermined Displacement Coordinate Syste
ms on Edges
If two faces with different displacement coordinate systems intersect at
an edge that has no explicitly defined displacement coordinate system,
then the coordinate system used for the edge cannot be easily determined.
Avoidance: In this situation, specify the displacement coordinate system
for the edge explicitly.
@
6374; 69; Grid Point Force - TEMP(LOAD), SUBCOM, GPFORCE - SFM 6220
If a SUBCOM involves a temperature load, and gpforce is requested as an
output, then the job will fail with the following error messages:
SYSTEM FATAL MESSAGE 6220 (GPFDR2), ERROR SEARCHING INDEX IN SCRT9 FILE
Avoidance: Replace the SUBCOM with an appropriate SUBCASE.
@
6377; 68.2; Documentation - Quick Reference Guide - Superelements, SESET
The Gi description for the SESET entry is incorrect. The GRID ID limit of
less than 1,000,000 is a pre Version 68 limit. The new limit is less than
100,000,000.
@
6378; ; Element - GENEL - UFM 316
 If the [S] matrix is left out (to be generated automatically by
MSC/NASTRAN) when using the GENEL element, then {ud} must contain six and
only six degrees of freedom. This fact is documented in the MSC/NASTRAN
Quick Reference Guide under the GENEL entry description. However, if less
than six degrees of freedom is defined when using the Z option, then the
following unclear error messages is issued:
USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA CARD GENEL x SORTED
CARD COUNT = y
On the other hand, if the K option is used, no error message is issued.
The job most likely fails downstream with other error messages.
Avoidance: If the [S] matrix is to be generated automatically by
MSC/NASTRAN, define six and only six degrees of freedom for {ud}.
@
6380; ; Nonlinear Analysis - RESTART, LOOPID, SUBID - Missing Output
If PARAM,LOOPID and PARAM,SUBID are used to restart a nonlinear analysis
and SUBID references a SUBCASE that is not the next sequential subcase
following the subcase associated with LOOPID, then missing output occurs.

Avoidance: If the user wants to restart an analysis and continue at the

beginning of a new subcase, then the subcase(s) should be added
sequentially following the subcase associated with LOOPID.

@
6383; ; P-Elements - MODES - UFM 6500 (ADAPT)
The modes command does not appear to work with p-elements and yields an
incorrect fatal message. The following Case Control:
ADAPT = 150
VUGRID(PLOT)=ALL
ADACT = ALL
$
METHOD=151
SPC=999
DISP=ALL
SUBCASE 1
MODES=1
SUBCASE 2
STRESS=ALL
resulted in the following fatal message:
USER FATAL MESSAGE 6500 (ADAPT)
MULTIPLE SUBCASES WITH DIFFERENT METHOD COMMANDS AND/OR SPC/MPC/SUPORT
COMMANDS HAVE BEEN SPECIFIED
WITH P-ELEMENTS IN NORMAL MODES ANALYSIS.
USER INFORMATION: ONLY ONE BOUNDARY CONDITION AND ONE METHOD COMMAND PER
RUN IS ALLOWED.
Note that the method command and constraint requests are above the subcase
level.
Avoidance: There is no avoidance.
@
6386; 69; Documentation - DMAP Module Dictionary - UFM 2143
If the MPL listing for the module show that the parameters are of the BCD
type, then they must be of length 8 as shown by the following example.
type parm,,char8,n,major,setj,comp='comp' $
major='g' $
setj='a' $
vec uset/vgacomp1/major/setj/comp $
If other lengths are used, e.g., char4 instead of char8, then the job may
fail with the following error
message:
USER FATAL MESSAGE 2143, MODULE VEC UNABLE TO IDENTIFY SET OR
SUBSET DESCRIPTOR G A
This fact is not documented in the DMAP Module Dictionary.
Avoidance: Define the BCD parameter type as char8.
@
6387; 68.2; Nonlinear Analysis - NLTRD2 PARAM,TSTATIC - SFM 4276, EC 7777, I
nefficient
An error in the bisection logic may cause nonlinear analysis to overshoot
the load factor and the subcase may be repeated. If this error is
encountered in a transient nonlinear analysis using PARAM,TSTATIC then
(depending on machine) the job may issue SYSTEM FATAL MESSAGE 4276 ERROR
CODE 7777 in the NLTRD2 module or may exit the NLTRD2 module without any
message.

@
6388; ; p-Elements - GMSPC, GMBC, CORDiR, CORDiC - Wrong Answers
When constraints, either GMSPCs or GMBCs, are defined in a cylindrical or
spherical displacement coordinate system on an edge or face, the
constraints can only be approximately satisfied. It has been observed
that in some cases this leads to non-physical singularities.
Avoidance: If possible, the basic coordinate system should be used.
@
6390; 68.2; Installation - Installation Script Does Not Work on AIX Post 4.1
.2
The installation script for MSC/NASTRAN, EMAS and MN4Pro Version 68 based
systems will not work on AIX Version 4.1.2. (The current version is 4.1.2,
and this error was reported on a pre-release version that has not yet been
named).
The uname command returns different information than in previous releases.
As a result, the scripts do not recognize the system as AIX.
The correction requires changing the architecture selection code that
examines the uname output. Currently the code uses the string "AIX" in
case statements. The case selector must be changed to "AIX*". This
construct is used in the following files:
 bin/nast68 or bin/emas3, bin/util68, msc68/access/access,
msc68/util/util, bin/mscdoc1
A typical correction would be to change the line:
AIX )
to
AIX* )
@
6391; 69; DMAP - Matrix K4AA Has an Improper Trailer in Solution 129
In a SOL 129 run, the K4AA matrix for the superelement is FORM=1 (square)
instead of FORM=6 (symmetric). This may cause DMAP fatal messages in user
DMAPs that expect the matrix to be symmetric. No adverse affect in the
delivered solution sequences is known.
Avoidance: In Version 68, the following alter changes the trailer to be
FORM=6:

compile semrb, souin=mscsou, nolist, noref

alter 13
modtrl xaa////6 $
@
6396; 69; Documentation - SPLINE1, SPLINE2 - Missing Documentation
The description of the SETG field of the SPLINE1 entry in the MSC/NASTRAN
Version 68 Quick Reference Guide refers to remark 4 which does not exist.
Even though the SETG field on the SPLINE2 entry serves the same function,
the description does not refer to any remark at all.
Avoidance: The remark's original intent was to inform the user that SETi
entries are bulk data entries and as such they are described in the same
section.

@
6397; 68.2; Lagrange Multipliers - SYSTEM(198), SYSTEM(205) - SFM 5407, 540
1
In Version 68.1, if the Lagrange Multiplier Technique is used on the SGI,
the job may fail with the following (or possibly other) error messages:
SYSTEM FATAL MESSAGE 5407.
INERTIA (STURM SEQUENCE) COUNT DISAGREES WITH THE NUMBER OF MODES ACTUALLY
COMPUTED IN AN (SUB-)INTERVAL
SYSTEM FATAL MESSAGE 5401 (REIGL). LANCZOS METHOD IS UNABLE TO FIND ALL
EIGENVALUES IN RANGE.
This error is most likely to occur when using one of the Lagrange
Multiplier alters from the sssalter library (e.g., lmt11.v68).
Avoidance: Set SYSTEM(198) and SYSTEM(205) to 1 at the top of the input
file with the following NASTRAN statements:
 NASTRAN SYSTEM(198)=1
NASTRAN SYSTEM(205)=1
@
6402; 69; Data Recovery - ESE - SFM 6212
If ESE=n references a SET n then the run fails with SFM 6212.
Avoidance: Use ESE=ALL.

@
6404; ; Restart - SCR=YES, DBVIEW - Datablocks Deleted
If a restart run uses the DBVIEW statement in conjunction with SCR=YES, it
is possible that the DBVIEWed datablocks may be deleted from the database.
This is in spite of a RESTART, KEEP FMS statement.
Avoidance: On restart runs, do not use SCR=YES.
@
6405; 69; Postprocessing - RCOUT2 - SFM 1208 (LDSTR) on RCOUT2
Using RCOUT2 to convert a neutral binary output file that was ftp'd from
an IBM MVS/XA system results in SYSTEM FATAL MESSAGE 1208 (LDSTR). This
error occurs because a hex 71 is mistakenly written from MSC/NASTRAN, and
it is not an acceptable character in EBCDIC.
The error will be fixed in future versions. The client can fix
MSC/NASTRAN at their site using the following superzap:
NAME NASTRAN UNSTR
VER 031C BD71
REP 031C BD5F
IDRDATA FXUNSTR

@
6406; 68.2; Nonlinear Analysis - PARAM,TESTNEG,2 - Differential Stiffness no
t Ignored
In nonlinear static analysis PARAM,TESTNEG,2 option is suppose to allow
the user the option to ignore the affects of differential stiffness. This
parameter does not work.
Avoidance: A DMAP avoidance will be supplied shortly.

@
6413; 68.2; Miscellaneous - script, SMEM - SMEM Incorrectly Set
The runtime script incorrectly sets smem=100 as a default value.
Unfortunately, this overrides the system supplied default value. The
script should not enter a value for smem unless the user wants to change
the default value. The 68.0 and 68.1 scripts may be corrected as follows:
Change the line:
smem="100"
to:
 smem=""
Change the line:
echo "SMEM=${smem}"
to:
test -n "$smem" && echo "SMEM=${smem}"
@
6416; 68.2; Static Analysis - SOL 101, SSG1 - UFM 2025 Does Not Terminate Ru
n
For any structured solution sequence which computes static loads, if UFM
2025 (UNDEFINED COORDINATE SYSTEMS) is issued more than once by the SSG1
module then the job will continue through DCMP and attempt a static
solution instead of terminating the run. This error is caused by load
entries such as GRAV, FORCE, MOMENT that reference a coordinate system
that does not exist.
Avoidance: Insert following DMAP alter:

COMPILE SELG SOUIN=MSCSOU NOREF NOLIST

ALTER 12,12 $ v68
IF ( NOGO<0 ) THEN $
@
6418; 69; Optimization/Design Sensitivity - SFM 5502 (DSAC02) vs UFM 5502
(DSAD17)
The message number 5502 is used in different routines for different
reasons. The Version 68 Reference Guide documents only one of these
messages.
Avoidance: No avoidance is necessary, unique numbers will be assigned in
Version 69 and the documentation updated to reflect the new numbers.
@
6419; 68.2; Miscellaneous - DBUNLOAD, SMEM=0 - FORTRAN BIO Fatal Message
If a DBUNLOAD is requested and a FORTRAN unit number is not explicitly
specified, then the default unit number is 51. This number causes a
conflict when SMEM=0 on systems with FORTRAN BIO. For Version 67 this
error only occurs on Hitachi OSF machines when ESS is used.

Avoidance: Specify a FORTRAN unit number on the DBUNLOAD entry that is

less than 50.
This was fixed in v68.2 for both Hitachi and NEC
@
6420; ; Nonlinear Transient Analysis - SCR=YES, NLTRD - SFE 4276, EC 103
9
In a nonlinear transient analysis, if the AUTO method is selected and the
job is submitted with scr=yes, then the job may fail with the following
error message: (HP only)
0*** SYSTEM FATAL ERROR 4276.SKPFIL ERROR CODE 1039
Avoidance: There are two avoidances:
 1. Submit the job with scr=no
or insert the following dmap alter avoidance
2. Compile nltran souin=mscsou nolist noref
alter 1 $
putsys(0,192) $
@
6421; 68.2; Elements - QUAD4, STRESS(CORNER) - Poor Default
In MSC/NASTRAN Version 68 there are three different options (CUBIC, BILIN
and SGAGE) for the calculation of QUAD4 corner stresses. Any of these
options may be explicitly requested on the STRESS entry. The default
option obtained by simply selecting STRESS(CORNER) is the CUBIC option.
It has been determined that the BILIN option produces better results for a
wider range of problem, thus beginning with Version 68.2 the default
option is changed to BILIN.
Avoidance: In versions prior to Version 68.2 the user should explicitly
specify STRESS(BILIN) to obtain the BILIN option.
@
6422; 69; Optimization/Design Sensitivity - DOM6 - SFM 3001
If there are no retained type 1 responses, a request for design
sensitivity output (PARAM OPTEXIT = 4 or 7) produces the following fatal
message:
SYSTEM FATAL MESSAGE 3001 MODULE = DOM6
THE INPUT DATA BLOCK NAMED DSCMG AT POSITION 6 DOES NOT EXIST.
Avoidance: Force the retention on at least one type one response. This
can be done by creating a dummy invariant response that is constrained at
a level that will require its retention but that will not affect the
design task when OPTEXIT=7.
@
6423; 68.2; Miscellaneous - DBMGR, DBSTORE, and DBFETCH, DIAG 20 - Excessive
Output
DIAG 20 activates a DIAG which produces excessive output if placed in the
Executive Control. DIAG 20 is also used to print diagnostic messages in
DBMGR, DBSTORE, and DBFETCH. A separate DIAG (DIAG 47) has been added in
Version 68.2 so that the diagnostic messages from DBMGR, DBSTORE and
DBFETCH may be obtained with out excessive output of DIAG 20.
@
6424; ; Nonlinear Analysis - CGAP - UFM 2025, Poor Error Message/Documen
tation
If a gap element connects coincident grid it is a documented requirement
that a coordinate system must be specified else USER FATAL MESSAGE 2025
will be issued. Although this is a user input error, the fatal message
should be improved to include more descriptive text such as:
"USER FATAL MESSAGE ----, COINCIDENT GRIDS WITH NO CID FOR THE GAP ELEMENT
----."
@
6426; 69; DMAP - MATMOD - Poor Message, SFM 4276 EC 105
If the first datablock for the MATMOD module does not exist, then SYSTEM
FATAL ERROR 4276.PACK ERROR CODE 105 may result. This has not been
 observed in the solutions provided with MSC/NASTRAN. It has only been
observed when a user DMAP is used.
Avoidance: Make sure that the first datablock exists.
@
6430; 69; Optimization/Design Sensitivity - CORDi, DOM11 - SFM 3002
"SYSTEM FATAL MESSAGE 3002 MODULE = DOM11" may occur in an optimization
run that has many coordinate systems.
Avoidance: The error may be avoided by providing a set of dummy GRID
entries in the deck. Let NGRID be the total number of GRID entries
including dummy ones and NCS be the total number of coordinate systems
defined in the deck. Then, NGRID should be determined as follows:
NGRID > NCS * 13 / 8
@
6431; 69; Optimization/Design Sensitivity - GPSTRESS - UFM 1126
In solution 200, if a GPSTRESS request is made in two subcases with
different analysis types, for example analysis=statics and analysis=modes,
then the following fatal message is encountered:
USER FATAL MESSAGE 1126 (GNFIST)
DMAP MODULE SUPER3 IS ATTEMPTING TO OUTPUT DATABLOCK 206 NAME =
GPSF , WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE GPSF =OVRWRT
Avoidance: Remove the GPSTRESS request from one of the subcases.
@
6433; 69; Grid Point Stress - PARAM CURV, PARAM,S1M - SFM XXXX
In a normal mode analysis (e.g., SOL 63, 103, etc.), if the stress output
is requested in a material coordinate system, then the following error
messages may be issued:
SYSTEM FATAL MESSAGE XXXX. SUBROUTINE CURV2
HAS RETURNED WITH ERROR CONDITION -2
LOC CODE = 600 IN SUBROUTINE CURV2
FILE NUM = 301
In addition to printing the regular stress output, the run will print the
transformed stresses for the first mode only.
Avoidance: Add modes=x in the Case Control Section. The value "x" is an
integer greater than or equal
to the number of modes in the model.
@
6436; 68.2; Optimization/Design Sensitivity - DOPTPRM, APRCOD=2 or 3 - SFM 4
276, EC 7777
"SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777" indicating a divide by zero is
possible in the DOM9 module when either the default approximation method,
Mixed Method, or type 3, Convex Linearization, is used. The approximation
method is set by specifying APRCOD on the DOPTPRM entry.
Avoidance: If this error is encountered set APRCOD equal to 1, Direct
Linearization. Note that Direct Linearization is recommended for shape
optimization problems.
@
6447; 68.2; Complex Eigenvalue Analysis - CEAD, HESS - Poor Answers
The epsilon to determine convergence in the Hessenberg Method of complex
eigenvalue extraction may be too large for some problems and result in
poor answers.
Avoidance: There is no avoidance. The epsilon is squared in Version 68.2
to correct this error, that is where in prior versions the epsilon was of
the order of 1.0E-7 in Version 68.2 the epsilon will be of the order of
1.0E-14.

@
6454; ; Dynamic Response Analysis - SDAMP, FRRD1 - UFM 2089
The SDAMP case control command selects modal damping and thus should be
ignored in direct dynamic solutions. When it is present in a direct
solution, it results in the following misleading diagnostic messages:

USER FATAL MESSAGE 2089 MODULE = FRRD1

TABLE 4 IS UNDEFINED.
USER FATAL MESSAGE 2089 MODULE = FRRD1
TABLE 5 IS UNDEFINED.
Avoidance: Remove the SDAMP case control command.

@
6457; 69; Optimization/Design Sensitivity - DSVG4, APD, BUFFSIZE - SFM 427
6
When a nonstandard buffsize is used, it is possible that a run will fail
and issue a SYSTEM FATAL MESSAGE 4276. To date this error has only been
experienced on HP systems in the DSVG4 and APD modules. The DSVG4 module
is no longer executed in the Version 68.2 solution sequences.
Avoidance: To avoid the error in general specify a standard buffsize. To
avoid the error in the DSVG4 module apply the following DMAP avoidance:
compile respsen souin=mscsou nolist noref
alter 171,171

PARAML UGy//'TRAILER'/1/S,N,NCOL $ no. of cols reqd. in SPAW1

PARAML UGx//'TRAILER'/1/S,N,nrow $ no. of rows reqd. in SPAW1
matgen ,/spaw1/4/ncol/nrow/0/1/nrow/1/1/nrow $
diagonal spaw1/spawn/'whole'/0. $
mpyad ugx,spawn,ugy/ugsn $

@
6458; 69; Optimization/Design Sensitivity - DSVG1 - UFM 5165
The message: "USER FATAL MESSAGE 5165 (DSVG1D-1) UNDEFINED INPUT FILE =
108, NAME = (NONE) , ERROR CODE = -2, MODULE DSVG1 IS TERMINATED" has
been encountered in a static - normal modes design optimization run with
PARAM,OPTEXIT,7. (Version 68 and 68.2 only)
 Avoidance: Insert the following DMAP alter:
compile desopt souin=mscsou noref nolist
alter 257,257 $ version 68
alter 269,269 $ version 68.2
delete /fol1,ppf1,,, $
alter 291,291 $ version 68
alter 305,305 $ version 68.2
delete /fol1,ppf1,,, $
@
6460; 70.5; Superelements - PARAML - UFM 1127 (GNFIST) DMAP MODULE PARAML
If the split database approach is used in the SSS in Version 68 or later
releases, if any upstream superelement databases are "OFFLINE" (not
available), it is possible that MSC/NASTRAN may fail in module PARAML with
USER FATAL MESSAGE 1127 stating that the USET is not present. This only
occurs if the program is allowed to attempt to process the superelements
whose databases are OFFLINE.
Avoidance: As long as there are no changes to the previously processed
superelements, the SEALL command may be used pointing to only the
superelements to be processed in the current run. If the previously
processed superelements are changed, their databases must be ONLINE in
order for MSC/NASTRAN to include the effects of the changes.
@
6461; 69; Heat Transfer Analysis - CQUAD8 - Wrong Answers
The capacitance matrix for the higher order CQUAD8 element uses a lumped
capacitance matrix instead of a consistent capacitance formulation.
Therefore, the weighting on the QUAD8 capacitance is not correct.
Avoidance: Use the CQUAD4 element instead.
@
6463; 69; Data Recovery - CURV3 - SFM XXXX SUBROUTINE CURV3
The following error message may occurs when PARAM,CURV,1 and PARAM,S1M,1
are used.
SYSTEM FATAL MESSAGE XXXX. SUBROUTINE CURV3
HAS RETURNED WITH ERROR CONDITION -3
LOC CODE = 150 IN SUBROUTINE CURV3
FILE NUM = 302
Although the fatal message is printed, the run continues and gives the
correct results.
The reason for this error is not known at this time.
Avoidance: Ignore the error message.
@
6464; 70; Miscellaneous - SPCF=set id, DISP=ALL = Missing Output
If a deck contains a request for DISP=ALL and a SPCFORCE output request
that refers to a set id, then the displacement output may also be
restricted to the set. This error does not occur all the time. The exact
cause is under investigation.
Avoidance: Request SPCFORCE=ALL instead of a set id.
@
 6466; ; Miscellaneous - UFM 316
"USER FATAL MESSAGE 316" which indicates a format error may be issued
when the large-field input format is used when the format of the entry
appears correct. The error message may be caused because the number of
large-field entries is odd. Section 3.7.1 of the Reference Manual states
that the number of large-field format entries must be even. The program
should check and issue a specific fatal message if this rule is violated,
or better yet, accept an even or odd number of entries.
Avoidance: Add a blank wide field continuation line if you have an odd
number of large-field lines.

@
6468; 69; Eigenvalue Analysis - REIGL - Poor Diagnostic Message, UFM 5401
If the user sets f2 on the eigrl entry less than the first frequency of
the component, the design intent
is that the run is to continue, and set neigv=-1. The dmap writer then
can decide to continue or write his own fatal error. Lanczos (a.k.a. the
REIGL module) does all this, but also puts out the following misleading
message:
*** SYSTEM FATAL MESSAGE 5401 (REIGL). LANCZOS METHOD IS UNABLE TO FIND
ALL EIGENVALUES IN RANGE.
ACCEPTED EIGENVALUES AND ADDITIONAL ERROR MESSAGES MAY BE LISTED
ABOVE.
USER ACTION: RERUN WITH ANOTHER METHOD OR ANOTHER SETTING ON EIGRL
ENTRY.

@
6469; 69; Nonlinear Transient Analysis - NLTRD2, AUTOSPC - UFM 6137, SFM 3
007
When using component modes synthesis on superelements in a nonlinear
transient analysis with adaptive time stepping (METHOD=ADAPT on the
TSTEPNL entry), you must select the exact number of generalized
coordinates to match the number of modes; otherwise, your job may fail
with one of the following error messages:
*** USER FATAL MESSAGE 6137 (DFMNRS/D)
INPUT MATRIX IS RANK DEFICIENT, RANK = x

SDCOMP INFORMATION MESSAGE--COLUMN YYY OF DATA BLOCK SCRATCH IS NULL.

SYSTEM FATAL MESSAGE 3007
ILLEGAL INPUT TO SUBROUTINE NLTRD2
Avoidance:
1. Specify NASTRAN SYSTEM(69)=16.
2. First determine the number of modes in the range of interest for each
component by specifying a small frequency range for SINV on the
EIGR entry. Then specify the number of modes desired not the frequency
range of interest.
3. Select another METHOD on the TSTEPNL entry.
@
6476; 2001; Plotting - XYPLOT - Stress Plots Default to Max Shear
The MSC/NASTRAN Quick Reference Guide states that von Mises stresses are
the default, but if von Mises stresses (item code 9 for QUAD4) are
requested through a xyplot request, but there is no corresponding case
control request, then max. shear is calculated instead of von Mises. This
happens because the stress request in case control is defaulted to von
Mises in IFP, but the corresponding default is not set in SDR2 or DDRMM
for implied stress recovery using the xyplot section.
Avoidance: The user can request STRESS(PLOT)= n in the case control to
force von Mises.
@
6477; 69; p-Elements - Missing Mass Matrix for p-elements
If the first p-element processed (sorted by element type, then ID) has
zero density, then the mass of all elements are ignored.
Avoidance: If a zero mass element is desired, then it must be given a
higher ID. Another alternative is to assign a tiny density, like 1.e-30.
@
6478; 68.1; Executive - XCSA - Unrecoverable Library Error, Traceback
If a CEND is missing in the input deck the CRAY Version 68.0 of
MSC/NASTRAN hangs in module XCSA with the following messages:
lib-1212 /msc/msc68/exe/unicos/analysis.Y-MP: UNRECOVERABLE library error
An internal WRITE tried to write beyond the end of an internal file.
apparent state: unit 6 named out//infinite_loop.f06
last format: (A4)
lately writing sequential formatted internal
SIGNAL - Signal SIGABRT(6) caught, job aborting.
See signal(2), CRI manual SR-2012 for further information.
TB001 - BEGINNING OF TRACEBACK
- $TRBK WAS CALLED BY RECOVR AT 525143a (LINE NUMBER 33)
- RECOVR WAS CALLED BY sysrcv AT 525363a (LINE NUMBER 112)
- sysrcv WAS CALLED BY __handlr AT 1053431a
etc.
@
6481; ; p-Elements - EMGPMD - UFM 4659
P-element models with cylindrical or spherical coordinate systems
generated using the Aries evaluator may encounter the message "USER FATAL
MESSAGE 4659 (EMGPMD) DURING INTEGRATION FOR ELEMENT WITH ID =". This
error is demonstrated by tpl deck pfms03.dat.
Avoidance: Other than removing cylindrical and spherical coordinate
systems which is probably not practical, there is no guaranteed avoidance.
Changing the meshing or geometry of the model slightly may avoid the
error.
@
6482; 69; p-Elements - ADAPT, DATAREC, OUTRCV, GP1 - UFM 2104
If the p-Element Case Control commands ADAPT, DATAREC, or OUTRCV reference
nonexistent bulk data entries, then the job may fail in the GP1 module
with the message "FATAL MESSAGE 2104, UNDEFINED COORDINATE SYSTEM"
 Avoidance: Add the appropriate bulk data entries.
@
6483; 69; Postprocessing - OUTPUT2, OGS1, OGPS - UWM 6286
If a param,post,-1 is used in sol106 then following warning message is
issued while writing the
OGS1 datablock.
USER WARNING MESSAGE 6286 (ZREAD)
THE NDDL DESCRIPTION OF DATA BLOCK=OGS1 , RECORD=IDENT ,DOES NOT HAVE
AN EITHER/OR TO MATCH OPT= 10.
ITEMS ARE ASSUMED TO BE UNDEFINED.
USER INFORMATION: PROBABLE CAUSE IS WRONG NDDL DESCRIPTION CODED, OR
,
WRONG GENERIC DATABLOCK USED (E.G. OES INSTEAD OF
OEF).
@
6484; 70.5.1; SSSALTER - PARAM,DRMH,1 - SFM 4276, EC 3021
Single precision machines, i.e., Cray, some SGI, and NEC, fail when trying
to run the files "CBAR1G.DAT" and "cbar9g.dat" in the sssalter directory.
This data deck requests data recovery using client developed modules,
drmh1 and drmh3, which have not been ported to single precision (long
word) machines. SUN systems may experience problems running the DRMH3
module only. The jobs may end in the DRMH3 module with the message "time:
command terminated abnormally" issued to the log file.
Note that in versions past v68, the problem test decks have been renamed
to alter1g.dat, 2g, 3g, and 9g respectively.
@
6485; 70; DMAP - TRNSP - Poor Message, CALLING PEXIT FROM DUMMY
The following error message may be output from the TRNSP module when the
DMAP forces single precision to be used by setting system cell(55) in the
DMAP (i.e., PARAM //'SYSTEM'/S,N,IPREC/55/1 ):
CALLING PEXIT FROM DUMMY
Avoidance: A user should avoid forcing the precision. MSC recommends
letting the precision default to machine precision. The avoidance is to
remove this PARAM module from the dmap.
@
6487; 69; Optimization/Design Sensitivity - DSAL, MODES - SFM 4276 GETSTR
E.C. 1409
If the MODES case control command is used in SOL 200, module DSAL will
issue the message "SYSTEM FATAL MESSAGE 4276 ERROR CODE 1409".
Avoidance: Do not use the MODES command in SOL 200.
@
6488; 69; Executive - COMPILER,LIST,REF - Missing Output, Fails to trap sy
ntax error
When COMPILE,LIST is input instead of COMPILER,LIST, the job runs failing
to trap the syntax error and the listing is not produced.
@
6490; 69; Optimization/Design Sensitivity - DESOBJ - Misleading Subcase Id
entifier
The subcase identifier for the design objective produced when the P2
DOPTPRM parameter is set to 1 lists the first subcase when the objective
function is weight or volume.
Avoidance: Weight and volume responses are not a function of the subcase
and the misleading subcase identifier should be ignored. Version 69
corrects this error by indicating a subcase id of zero.
@
6491; ; DMAP - COMPILE, DIAG 4,14, COMPILER - Extra Output
Starting in Version 68, if the COMPILE NDDL command is used to punch out
the NDDL and both the SOL and DIAG 4,14 (or COMPILER LIST,REF) statements
are specified in the Executive Control Section, then NDDL will be written
to the punch file twice.
Avoidance: There are two avoidances:
1. Remove the SOL statement.
2. Remove either 4 or 14 from the DIAG (or REF or LIST from the
COMPILER) statement.

@
6492; 69; Design Sensitivity - DSVG1, MATi - UFM 5171 (DSVG1D-7) Unexpect
ed EOR
The following message occurs when mass sensitivities are being calculated
and some of the materials associated with the designed elements have a
density of zero and others do not.
UFM 5171 (DSVG1D-7) Unexpected EOR in File 101.
SYSTEM FATAL MESSAGE 3003 MODULE = DSVG1
ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK MDICT (FILE
106) IN SUBROUTINE DSVG1D
Avoidance: Assign a small, non-zero density in place of the zero values.
@
6494; 69; Eigenvalue Analysis - SDRX - Infinite Loop
In SOLs 39, 63, 103, and 200 with ANALYSIS=MODES, if PLOAD1 entries are
present and they are applied to CBAR or CBEAM elements then the run will
go into an infinite loop in the SDRX module. (Version 68.2 only)
Avoidance: Remove PLOAD1 Bulk Data entries or remove the STRESS and FORCE
Case Control command requests.
@
6504; 2001; Rigid Elements - GP4, RBE3 - Poor Message
If the components specified on the "UM" set reference the same grid point
as the REFGRID, and the "UM" components is an incomplete subset of the
REFC components, then the job may fail with the following error message in
versions prior to Version 68.
0*** USER FATAL MESSAGE 2332, DEPENDENT MPC COMPONENT HAS BEEN SPECIFIED
TWICE. EXTERNAL ID = xx DOF = yy
 In Version 68 and above, depending on the solution sequence, the job may
terminate with no messages in GP4 or it may fatal with the following
unclear error message:
FATAL ERROR
A sample of an RBE3 that can cause this error is as follows:

RBE3, 10,, 16, 1234, 1., 123, 11,

26,
,33, 4.7, 12, 17, 31
,UM, 16, 14

Avoidance: Remove the offending "UM" grid points from the continuation
entry.
@
6508; 69; Optimization/Design Sensitivity - ANALYSIS=BUCK - UFM 4683
Shape optimization with the analytic boundary shapes approach and buckling
analysis (ANALYSIS=BUCK) in solution 200 produces the following error
message:
USER FATAL MESSAGE 4683, KDIF MATRIX NEEDED FOR EIGENVALUE ANALYSIS
Avoidance: The error is avoidable if the user generates basis vectors
using either the direct input of shapes method or the DVGRID (i.e.,
manual) approach. Using geometric boundary shapes will also produce the
same error.
@
6509; 68.2.3; p-Elements - PLOAD4 - Wrong Answers
If a PLOAD4 is used to apply a pressure load to a p-element that has the
connectivity inverted, i.e. that the right hand rule applied to the
numbering of the first face produces a vector pointing out of the element,
then the pressure is applied in the wrong direction and wrong answers
result.
Avoidance: There is no simple avoidance.

@
6514; 69; Optimization/Design Sensitivity - RESPSEN - Poor Performance
The RESPSEN subDMAP uses the UG datablock rather than UG1. The result is
that much of the efficiency that can be gained from load case and/or
frequency deletion is lost. Further, there is no current requirement for
static spcforces in this subDMAP.
Avoidance: A new RESPSEN subDMAP is available on request.
@
6515; 69; Optimization/Design Sensitivity - DOPR1B, DLINK - UWM 7001 (DOPR
1B)
The message "USER WARNING MESSAGE 7001 (DOPR1B) DLINK GIVES WRONG XINIT
VALUE" is produced in error whenever NMSG = INT (( ndv - 4 )/ 7 ) - ndvi
+1 is a positive number. For example, if there are 30 design variables,
but only one independent design variable, 3 incorrect messages are
 produced. The user input value for the design variable value is
incorrect. The value computed based on the DLINK entry is correct. Thus,
the error does not produce wrong answers.
Avoidance: No avoidance is necessary since the error does not produce
wrong answers but only an incorrect warning message.
@
6517; 69; p-Elements - GMCORD, DBUNLOAD, OUTPUT2 - Bad CSTM Trailer
Data decks that contain only convective coordinate systems (GMCORD) and no
CORDij type data entries have bad CSTM trailers. This does not cause a
problem on homogeneous platforms, however, difficulty may be encountered
during the neutral OUTPUT2 or DBUNLOAD operations.
Avoidance: Include a dummy CORDij entry in the data deck.
@
6521; ; Nonlinear Analysis - TABLES1 - UFM 2076, Poor Message
In nonlinear analysis, USER FATAL MESSAGE 2076 is issued if the MATS1
entry is type PLASTIC and the slopes of the first two segments of the
stress-strain curve on the referenced TABLES1 entry are the same. The
message:

USER FATAL MESSAGE 2076, ELEMENT 0 REFERENCES TABLES1 4 WHICH IS THE

INCORRECT FORM FOR NONLINEAR ELASTIC ANALYSIS
does not clearly indicate the problem and incorrectly references element
0.
@
6525; 69; Nonlinear Analysis - CGAP - Incorrect SPCF Output
When the GAP element has an offset moments may be created by the shear
force times the axial offset and the axial force times the lateral
offset). These moments are balanced by forces of single point constraint,
however, the SPC Force output is incorrect in both magnitude and sign.
Avoidance: Remove the offset.
@
6527; 69; Documentation - Quick Ref. Guide, Ref. Manual - PARAM,ACOUT,RMS
PARAM,ACOUT,RMS only applies to element output. The current documentation
in the MSC/NASTRAN Version 68 Quick Reference Guide erroneously implies
that all the pressure output will be RMS.
Avoidance: In the MSC/NASTRAN Quick Reference Guide, in the definition
for parameter ACOUT, the following:
ACOUT specifies the type of sound pressure level output to be
used with the DISPLACEMENT or PRESSURE Case Control command in
coupled fluid-structural analysis (see the MSC/NASTRAN Reference
Manual, Chapter 15). The default requests the peak pressure
level. ACOUT = RMS requests RMS pressure level and is only
available in Solution Sequences 108 and 111.
To obtain sound pressure level in units of dB and dBA, a peak
reference pressure must be specified with PARAM, PREFDB. The
dB level is defined as:
 Should be replaced with:
ACOUT specifies whether peak or root-mean-square (RMS) output
is to be used with the FORCE Case Control command in coupled
fluid-structural analysis (see the MSC/NASTRAN Reference
Manual, Chapter 15). The default requests peak.
ACOUT = RMS requests RMS and is only available in Solution
Sequences 108 and 111.
In order for decibel information (dB and dBA) from the
DISPLACEMENT, PRESSURE, or FORCE Case Control commands to make sense, a
peak reference pressure must be specified with PARAM, PREFDB.
The dB level is defined as:

In the MSC/NASTRAN Reference Manual, Chapter 15.4.2 Coupled

Fluid-Structural Analysis, Output and Plotting, the following:
The DISPLACEMENT or PRESSURE Case Control command is used to
request:
1. Pressure and peak sound pressure level at the fluid
points.
2. RMS sound pressure level is requested by the
PARAM,ACOUT,RMS Bulk Data entry. This output is
available in Modal and Direct Frequency Response analysis
only (SOLs 111, 108, and 200).
3. Sound pressure level in dB and dBA - a peak reference
pressure must be specified on the PARAM, PREFDB Bulk
Data entry. The dB level is defined as
dB = 20 log(P/PREFDB)
(for air PREFDB = 20 x 10-6 Pascal).
should be replaced with:
The DISPLACEMENT or PRESSURE Case Control command is used to
request:
1. Peak and root-mean-square (RMS) sound pressure
level at the fluid points.
2. Sound pressure level in dB and dBA calculated from
the RMS sound pressure. Note, a RMS reference pressure
must be specified on the PARAM, PREFDB Bulk Data entry.
The dB level is defined as dB = 20 log(P/PREFDB).
For example in air, a common RMS reference pressure is
PREFDB = 20 x 10-6 Pascal.
The FORCE Case Control command is used to request complex fluid
accelerations, velocities and pressure levels at the fluid
elements. This information is peak by default but can be changed
to RMS form if PARAM,ACOUT,RMS is used. This output is available
in Modal and Direct Frequency Response analysis only (SOLs 111,
108, and 200).

@
6528; 68.2; Composites - SDR2, SDRCOMP, STRESS(CORNER) - Wrong Answers, SFM
4276
If a STRESS(CORNER) request is made for a deck with composite elements
wrong answers result. The run may also issue SYSTEM FATAL MESSAGE 4276 EC
7777 or stop with a "Terminated Abnormally" message in the log file.
Avoidance: The only avoidance is to either not request STRESS(CORNER) at
all, execute STRESS(CORNER) for noncomposite elements in a cold start and
then recover the stresses in the layered composite elements in a restart
run. In Version 68.2 this error has been addressed by treating a
STRESS(CORNER)=ALL request as STRESS(CORNER) for noncomposite elements
only and outputting center stress for the composite elements.
@
6529; 68.2.3; Elements - TETRA, TEMPi, SSG1 - Wrong Answers
Incorrect stresses result for the TETRA(10) element with curved edges and
thermal loads. The error is due to incorrect thermal loads, and thus even
the displacements are poor. This is a result of under-integration, and
the magnitude of the error depends upon the amount of element geometry
distortion.
Avoidance: There is no known avoidance short of modelling with straight
edges.
@
6534; 69; Acoustic Analysis - RESTART, GEOM2 - SWM 2072
The message "SYSTEM WARNING MESSAGE 2072, CARD TYPE 7408 NOT FOUND ON
DATA BLOCK" may be issued in an acoustic analysis restart. The warning
message does not affect the run and may be ignored.
Avoidance: No avoidance is necessary.

@
6535; 69; Acoustic Analysis - RESTART - Inefficient Restart
Changing the PANEL definition in the restart of an acoustic analysis
causes the modes to be unnecessarily re-calculated.
Avoidance: Apply the following DMAP alter to MSC/NASTRAN Version 68.1
COMPILE IFPL SOUIN=MSCSOU NOLIST NOREF
ALTER 146,146
@
6537; 68.2.1; Dynamics - STRAIN - Missing output
STRAIN output is missing in SOLs 110, 111, 112, 145, 146, and 200,
with ANALYSIS=MFREQ, MTRAN, and FLUTTER if PARAM, DDRMM is 0, which is the
default. This error occurs in MSC/NASTRAN Version 68.2.0 only. This
version of MSC/NASTRAN is included with MSC/Aries 7.0.0 on the IBM/RS6000,
HP 9000/700, SUN SUNOS and SUN SOLARIS systems. Version 68.2.0 has not
been delivered to non MSC/Aries users. Version 68.2.1 corrects this error
and is delivered to non MSC/Aries clients. (Version 68.2. only)
Avoidance: Insert PARAM, DDRMM, -1 or the following DMAP alter:
compile sedrcvr noref nolist
alter 174
øDDRMM CASEDR,UH1,OL1,,,OSTR2,,XYCDBDR/,,OSTR2Y,,//3 $
@
6538; 69; Documentation - Illegal data on EPOINT entry
The MSC/NASTRAN Version 68 Quick Reference guide does not document a limit
for the ID of an EPOINT. The limit is 1000000.

@
6539; 70; P-Elements - SDRP - SFM 4276 EC 7777 SDRP module
The job will stop in the SDRP module and the message "SYSTEM FATAL ERROR
4276.RECOVR ERROR CODE 7777" will be output when multiple output
requests are used on the output bulk data entry. The error only occurs on
the Convex Version 68 systems. The error only occurs when the output
requested is changed between sets.(i.e. when output requested for set 3 is
different than output requested for set 2. For example:
OUTPUT 1
ELSET=3,DISP=PRINT,STRESS=NONE,STRAIN=NONE,ERROR=NONE,
PVAL=NONE,FIRST=NO,BY=0
ELSET=2,DISP=NONE,STRESS=PRINT,STRAIN=NONE,ERROR=NONE,
PVAL=NONE,FIRST=NO,BY=0
ELSET=999999,DISP=NONE,STRESS=NONE,STRAIN=NONE,ERROR=PRINT,
PVAL=PRINT,FIRST=YES,BY=1

@
6540; 69; Nonlinear Transient Analysis - RESTART, EPOINT - SFM 3007, UMERG
E, CALCV
A solution 129 restart from solution 106 fails with the following message
whenever EPOINT bulk data entries are used:
SYSTEM FATAL MESSAGE 3007 MODULE = UMERGE
ILLEGAL INPUT TO SUBROUTINE CALCV

Avoidance: Apply the following DMAP Alter:

COMPILE NLTRAN SOUIN=MSCSOU NOLIST NOREF
ALTER 112,112
UMERGE USETD,PLX,/IFD/'D'/'A'/'E' $
ENDALTER
@
6544; 69; Optimization/Design Sensitivity - SDSC, DSCREEN - SFM 3001
The message "SYSTEM FATAL MESSAGE 3001 MODULE = SDSC THE INPUT DATA BLOCK
NAMED DSCMCOL DOES NOT EXIST" is produced in a sensitivity analysis when
there are no active responses; i.e., all the responses have been screened
out becuase they do not exceed the threshold required for sensitivity
computations.
Avoidance: Use the following DMAP avoidance (for V68.2.3):
compile exitopt list noref souin=mscsou
alter 'putsys'
dbstatus R1TABRG,RSP2RG//S,N,NORSP1/S,N,NORSP2 $
if (NORSP1>0 OR NORSP2>0) THEN $
 alter 'optexit'(7)
endif $ NORSP1>0 OR NORSP2>0
@
6547; 70; Heat Transfer Analysis - CHBDYG, RMG2 - SFM 4276 EC 211
If a connection order error is made on the CHBDYG, AREA4 element, the
message
SYSTEM FATAL ERROR 4276.BLDPK ERROR CODE 211
may be output which does not give the user any clude to the input error.
Avoidance: There is no simple avoidance other than manually checking
your CHBDYG,AREA4 elements for the correct connectivity.

@
6551; 69.1; Elements - HEXA, TEMP(LOAD) - Wrong answers
The stress output for the HEXA element has an error. This error occurs
only for grid locations (not at the center location). The error occurs
only when the following conditions are all true:
1. An eight node HEXA element with bubble functions is used.
2. The temperatures are not the same at all grid points.
3. The stress output coordinate system axes do not align with the element
coordinate system axes.
Avoidance: The error may be averted by avoiding any one of the
aforementioned conditions.
@
6552; 69; Heat Transfer Analysis - VIEW - Missing Output, UFM 6660, Not Up
ward Compatible
Using the old VIEW module produces a punch file in which the RADMiX
ambient element (CHBDYP,POINT) is missing. If the punch file is included
in a subsequent job, MSC/NASTRAN fails and the following message is
issued:
USER FATAL MESSAGE 6660 (VIEW)
THE RADIATION MATRIX FOR RADIATION CAVITY ID= 11 DEFINED ON RADLST OR
RADMTX BULK DATA ENTRIES HAS WRONG FORMAT OR DATA. USER INFORMATION: THE
RADIATION MATRIX MUST BE SPECIFIED COMPLETELY.

Avoidance: Use the VIEW3D entry to avoid the problem.

@
6554; 69; Heat Transfer Analysis - PARAM,NDAMP - Phase Change Analysis, Po
or Convergence
PARAM,NDAMP only allows a maximum value of .1. This is inadequate for the
highly nonlinear phase change problem. The value of .1 is equivalent to a
beta of .55 using the Crank-Nicholson method. The phase change problem
requires a beta value of 1.0. Convergence is extremely difficult for the
phase change problem with the current limitation on PARAM,NDAMP.
Avoidance: There is no avoidance.
@
6556; 69; Postprocessing - RFxxD66, POST=-1 - Missing Geometry Datablocks
Geometry datablocks will not be written to the OUTPUT2 file with RFxxD66
when PARAM,POST,-1 and PARAM,OGEOM,YES is specified. This is contrary to
the instructions for RFxxD66 listed in the alter.
Avoidance: For RF63D66 insert the following below the line: IF ( POST=-1
) OTAPE2=-1 $
IF (POST=-1 AND OGEOM='YES') THEN $
OUNIT1=OUNIT2
OUTPUT2 CSTM,GPL,GPDT,EPT,MPT//OTAPE2/OUNIT1//OMAXR $
OTAPE2=0
OUTPUT2 GEOM2,GEOM3,GEOM4,,//OTAPE2/OUNIT1//OMAXR $
ENDIF
Similar changes are required for the other RFxxD66 rfalters.
@
6557; 68.2.3; P-Elements - TETRA, TEMP(LOAD) - Wrong Answers with TEMP Loads
Poor accuracy occurs for TETRA p-elements with thermal loads and curved
edges. Since the
poor accuracy results from an error in the integration of the thermal
loads, all results are poor. The error is much less if the edge p-value
is three or greater (for a quadratic curved edge, p will be at least two).
@
6558; 69; Matrix Methods - SMPYAD, DBVIEW - Wrong Answers
If SMPYAD is asked to do a matrix product of the form A(transpose)*B*C,
and matrix C is a DBVIEW version of the datablock with the same name as
matrix A, then SMPYAD substitutes matrix A for matrix C and performs
A(transpose)*B*A instead, which will lead to wrong answers. For example:
type db phg,mgg
type parm,nddl,i,n,meth
meth=6
dbview phg2=phg(where meth=3) $ Points to phg, but where method=3
not 6!
smpyad phg,mgg,phg2,,,/phimphi1/3////1 $ Does phg*mgg*phg, not
phg*mgg*phg2!!!!!!
Avoidance: The simplest avoidance is to set nastran system cell 129 to 1
prior to the smpyad (e.g. insert PUTSYS(129,1) into the DMAP. The system
cell may be reset to zero immediately after the smpyad by inserting
PUTSYS(129,0).
@
6570; 69; Nonlinear Transient Heat Transfer - RESTART - UFM 2166
The enforced temperature datablock, YSNL, is not stored in the database.
This causes restarts in solution 159 to fail with the message:
USER FATAL MESSAGE 2166, MATRIX TO BE PARTITIONED IS OF SIZE xx ROWS BY yy
COLUMNS.
ROW PARTITION SIZE IS xx COLUMN PARTITION SIZE IS zz (INCOMPATIBLE).

Avoidance: Insert the following alter in both the cold start and all
 subsequent restarts:
compile nltcsh souin=mscsou noref nolist
alter 583 $ v68
$alter 585 $ v68.2
call dbfetch /ysnl,,,,/0/0/0/0/0 $
alter 618 $ v68
$alter 620 $ v68.2
call dbstore ysnl,,,,//0/0/' '/0 $

@
6571; 68.2; Miscellaneous - SSG1 - SFM 4276 EC 7777
In MSC/NASTRAN version 68.0 or 68.1 it is possible to get the message
"SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777", in the module SSG1 because of
an error in SSG1 which causes overlapping memory regions. This error is
therefore data dependent and is most likely to occur on the Hewlett
Packard 9000, SGI R8000 or the DEC/Alpha OSF/1 systems. This error has
been fixed in the V68.2 system.
Avoidance: It may be possible to avoid this error if the model were
changed slightly.
@
6572; ; p-Elements - GMCURVs with MSC/Aries Geometry Evaluator
For certain cases, when the end GRID or the secant point is mapped back to
the GMCURV, the MSC/Aries geometric evaluator (i.e., group MSCGRP2 on the
GMCURV or GMSURF entry) returns the following message:
SYSTEM FATAL MESSAGE 6498 (Aries B-spline Evaluator, Get Curve Parameter)
GEOMETRY API MESSAGE FOLLOWS.
USER INFORMATION: Numerical Error in Evaluator.
Curve: 150.
Point: ( 1.000000e+01 1.288252e+02 1.402565e+02).
Avoidance: The problem is under investigation and it is unclear
specifically when this error occurs or how to avoid the problem.
@
6573; 69; Documentation - Reference Manual, PLOTPS - Obsolete, Superseded
Documentation
Section 13.4 of the MSC/NASTRAN Reference Manual is obsolete. It has been
superseded by the Utilities section of the Installation Guide for the each
computer type.
@
6574; 69; Heat Transfer Analysis - CHBDYE - Wrong Answers
If a CHBDYE (Boundary element) is used where side 5 represents an edge of
a shell element such as a CQUAD4, CQUAD8, TRIA3, or TRIA6 element, the
surface area calculated will be incorrect and wrong answers are produced.
Avoidance: Use the CHBDYP,LINE option to represent the convection or
radiation off the edge of the shell element.
@
6575; 2001; Nonlinear Analysis - QUAD4, ZOFF - Wrong Answers
In Nonlinear Statics Analysis the output element forces from the QUAD4
 elements are incorrect when the uniform offset (ZOFF field in the CQUAD4
bulk data) is non-zero. The force values are actually calculated in the
non-offset element coordinate system while the stresses are calculated
correctly in the offset reference system. All other solutions calculate
both outputs in the reference system.
Avoidance: Use the stress output and ignore the element force output.
@
6577; 2001; Eigenvalue Analysis - DPD, EIGR, EIGB, EIGC - UFM 2075 and UFM 2
079
When the POINT method of normalizing eigenvectors is selected on EIGR,
EIGB or EIGC entries, and the specified grid/dof does not exist, the
following mis-leading error messages is issued:
1) If the grid does not exist:
SYSTEM OR USER DMAP FATAL MESSAGE 2075, IMPROPER VALUE xx FOR FIRST
PARAMETER IN DMAP INSTRUCTION SDR2
2) If the component is invalid (in the case of SPOINT for example):
USER FATAL MESSAGE 2079, GRID OR SCALAR POINT 0 HAS AN UNDEFINED
COORDINATE REFERENCED ON A NOLINI CARD.
Note that these messages are issued even if the EIGR, EIGB or EIGC at
fault is not selected by a METHOD
entry in the Case Control section.
Avoidance: Remove or change the offending entry.
@
6578; ; Heat Transfer - VIEW3D, CHBDYG,AREA8, CHBDYG,AREA6 - Wrong Answe
rs
The boundary surface elements CHBDYG type AREA8 and AREA6 were implemented
in Version 68 to complement higher-order conduction elements such as the
CQUAD8, HEX20, and TRIA6. Using these surface elements in View factor
calculations results in an incorrect view factor if the surfaces are not
in-plane or on a curve surface. Therefore in order to model surfaces on
curvilinear surfaces such as cylinder it is best to use AREA4 and AREA3
elements.
Avoidance: Use CHBDYG,AREA4 instead of AREA8 to model curvilinear surface
for view factor calculation.
Similarly, use CHBDYG,AREA3 instead of AREA6 to model curvilinear surface
for view factor calculation.

@
6579; 69; Dynamic Response - SDR2, OFREQUENCY, SET=ALL - SFM 4276 EC 7777
In a frequency response run if the OFREQUENCY references a set where "all"
frequencies are specified, then the job may fail with the message "SYSTEM
FATAL ERROR 4276.RECOVR ERROR CODE 7777".
Avoidance: Remove the ofreq=xxx case control command.
@
6580; 69; Nonlinear Transient Heat Transfer - VIEW - SFM 4276 EC 7777
An integer overflow, SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777, has been
 encountered on the VAX VMS system while calculating the matrix density of
RECM in the VIEW module.
Avoidance: The only way to avoid the error is to reduce total number of
HBDYi elements for each radiation cavity.
@
6582; 69.1; Postprocessing - DBC - XDBs are Too Large
XDB files generated on a NEC SuperUX system are excessively large (by a
factor of 8). These files are still readable from RECEIVE, the
corresponding NTRL files are the correct size.
Avoidance: The corresponding NTRL files are the correct size, so storing
the XDB files in this form resolves the problem.
@
6590; 69; DMAP - ALTER, Missing Blank - No Warning Message
When the following DMAP statement:
alter 449,449 $
was mistyped as:

alter449,449 $
the alter was ignored and no warning message was issued.

@
6596; 70; Documentation - PARAM,USETPRT PARAM,USETSEL - Missing Documentat
ion
In the Version 68 MSC/NASTRAN Quick Reference Guide, the documentation for
PARAM,USETPRT and USETSEL should include the following set names and their
decimal equivalents:
U1 2147483648
U2 1073741824
U3 536870912
U4 268435456
U5 134217728
U6 67108864

@
6598; GL; Miscellaneous - PARAM - UFM 300, Improper Output
MSC/NASTRAN has a restriction on the magnitude of an integer parameter of
nine (9) digits. If an integer value with more than 9 digits is used for
a parameter value (e.g. PARAM*,USETSEL,2147483648) then the parameter does
not function properly. If the value of the parameter is hard coded
directly in the DMAP (e.g. TABPRT USET,EQEXIN//'USET'/0/2147483648) the
message " USER FATAL MESSAGE 300 FROM CRCARD ABOVE" is issued.
Avoidance: There is no avoidance.
@
6599; 69; Frequency Response Analysis - FRRD1, DECOMP - UFM 6137
When singular matrices are encountered in a frequency response analysis,
 MSC/NASTRAN Version 68 would fail in sparse decomposition with the
message:
USER FATAL MESSAGE 6137 (DFMNRS/D)
INPUT MATRIX IS RANK DEFICIENT, RANK = 2
USER ACTION: CHECK MODEL
The most common cause of the singular systems is a free body excited at
zero frequency. Non-sparse decomposition, which was the default prior to
Version 68, behaves differently. Non-Sparse decomposition would output
messages that warn the user that a small term was being added to the
diagonal so that decomposition could continue. An example of the messages
are:
USER WARNING MESSAGE---ROW 3 OF LOW. TRI. FACTOR HAS DIAGONAL TERM =
0( OR .LT. 0 IF CHOLESKY )
USER INFORMATION MESSAGE--- 1.0E-10 SUBSTITUTED FOR DIAG. TERM OF LOW.
TRI. FACTOR AT ROW 3
USER INFORMATION MESSAGE 4158---STATISTICS FOR SYMMETRIC DECOMPOSITION OF
DATA BLOCK SCRATCH FOLLOW
MAXIMUM RATIO OF MATRIX DIAGONAL TO FACTOR DIAGONAL = 1.0E+10
AT ROW NUMBER 3

The solution for this frequency would then have a constant displacement on
the order of 1.E+10 and a zero acceleration. Sparse decomposition has
been enhanced in Version 69 to behave similarly to non-sparse
decomposition. Thus, frequency response analysis at zero frequency with
rigid bodies should no longer fail with USER FATAL MESSAGE 6137.
Note that DMAP has also been added to issue a warning message and discard
the solution for singular systems in frequency response. However, with
the aforementioned change to add a small term to the diagonal we do not
expect this DMAP to be activated under normal circumstances. Future
versions may rely entirely on the DMAP action of discarding the solution
rather than then adding a small term to the diagonal in decomposition.
@
6601; 69; Buckling - TEMP(INIT) - Wrong Answers
In a linear buckling analysis, if TEMP(INIT) references a temperature that
is non-zero, then the eigenvalue calculation will be incorrect, which in
turn will lead to a wrong buckling load. The results of the static
subcase are correct.
Avoidance: Use one of the following avoidances:

1. Set TEMP(INIT) to zero and adjust the applied temperature to yield the
same delta temperature
or
2. Set the initial temperature using the TREF field on the MATi entry
instead of using the TEMP(INIT) Case
Control Command.
@
6603; 70; DMAP - SELA, SEMA, SEDR, PLTVEC, SEP4 - SFM 4252, UIM 4570
Starting in Version 67.7, the superelement modules, SELA, SEMA, SEDR,
PLTVEC, and SEP4, only recognize a qualifier name of SEID for the
qualifier which defines the data block family in the DBVIEW WHERE clauses.
For example, in subDMAP SELA, if we modify the DBVIEW PAUP statement for
 a family of user- defined data blocks, MYDB, and modify the 2nd parameter
in the SELA module,
DBVIEW PAUP = MYDB (WHERE MYQUAL=*) $
DBVIEW MAPUPL = MAPS (WHERE SEDWN=TEMP AND SEID=* AND PEID=*) $
SELA PJC,SLIST,EMAP,EQEXINS,PAUP,MAPUPL,GDNTAB/
PG/
SEID/'MYQUAL'/0/S,N,NOPGS $
then the SELA module will not find any members in the family PAUP.
Avoidance: Modify the NDDL so that the PATH statement for the data block
MYDB contains SEID as a qualifier. Also, modify the DBVIEW statement
WHERE clause to include SEID=*. For example,
NDDL statements:
PATH MYPATH SEID ...
DATABLK MYDB PATH=MYPATH ...
DMAP statements:

DBVIEW PAUP = MYDB (WHERE SEID=*) $

DBVIEW MAPUPL = MAPS (WHERE SEDWN=TEMP AND SEID=* AND PEID=*) $
SELA PJC,SLIST,EMAP,EQEXINS,PAUP,MAPUPL,GDNTAB/
PG/
SEID/'SEID'/0/S,N,NOPGS $

@
6610; 69; Heat Transfer Analysis - ROD, BAR, BEAM - Inconsistent Capacitan
ce Matrix
The rod element was changed in Version 68 to have a consistent heat
capacitance matrix. However, all the other one dimensional heat conduction
elements such as the beam and bar elements have a lumped representation.
This difference between the one dimensional element formulations may
result in slight differences between a model comparing RODs to BARs and/or
BEAMs. For a typical transient thermal analysis comparing the result
between ROD and BAR elements, the difference in temperature output is
generally less than 0.2%.
Avoidance: Use bar or beam element.
@
6613; 69; Heat Transfer Analysis - CHBDYE, SSG1 - Incorrect Loading, Wron
g Answers
If a CHBDYE surface type element is used to model an edge with mid-side
nodes and a heat flux loading is applied to this edge, then the answers
will be incorrect. This error occurs because the SSG1 module does not
know there is loading on the midside nodes.
Avoidance: Do not reference elements with mid=side nodes on the CHBDYE
entry. For example, instead of using the higher-order CQUAD8 elements
that have mid-side nodes, use the lower order CQUAD4 elements.
@
6614; 69; Acoustic Analysis - ACSRCE - Wrong Answers
The ACSRCE which defines the power versus frequency curve for an acoustic
 source results in an incorrect acoustic fluid power input. The power is
off by a factor of i*omega, the radian frequency.
Avoidance: Use the RLOAD data entry instead with Qdot, (volume
acceleration). Qdot is calculated as the square-root of 8 times pi times
the speed of sound in the fluid times power divided by the density of the
fluid:
Qdot = sqrt(8*pi*c*power/rho)

@
6618; 69; Documentation - String-based alter - UFM 300
When using the string-based alter feature in Version 68.2, the dmap
strings should be enclosed by the ' single-quotation string on both ends
(e.g., 'string1'). In the Version 68.2 Release Notes, the dmap strings
are shown to be enclosed with the ` single-quotation string on the left
end and the inverted single-quotation string on the right end (e.g.,
`string1') throughout chapter 5. An example of this documentation error
is shown on page 40 of the Release Notes as

ALTER `CALL.*PHASE1B' $
The correct format for the above string-based alter is
ALTER 'CALL.*PHASE1B' $
Note that the V68.2 Release notes can not be corrected. Since this
documentation error does not exist anywhere else it is considered fixed in
Version 69.
@
6620; 69; Installation - No Authorization Code File - Poor Message
During the installation of MSC/NASTRAN V68.2 the user is asked to run a
test problem after installation. If the user requests to run a test
problem and there is no authorization code file, then the job aborts with:
'The test appears to have failed for an unknown reason.'
If a bad authorization code is installed, then the installation test
problem determines the correct error.
But if no authorization code has been installed, then the noted message is
issued.

Avoidance: Install a correct authorization code.

@
6626; 69; Optimization/Design Sensitivity - DOPTPRM, GSCAL - Poor Default
The GSCAL parameter on the DOPTPRM entry is used to avoid a divide by zero
when calculating constraints. It has been found that the default value of
0.001 is too large when constraints are being calculated for property
values that can be many orders of magnitude smaller than this. Some
models with metric units have displayed this problem. The problem is
characterized by the design task finishing prematurely.
Avoidance: Set the value of GSCAL on the DOPTPRM entry to a value that is
several orders of magnitude
smaller than the expected property values. Along with this, it is always
a good practice to avoid constraints that have a lower or upper bound of
 0.0
@
6628; 69; Heat Transfer Analysis - QHBDY, IFP - UFM 316
A negative heat flux can not be applied using the QHBDY entry. For
example the entry:
QHBDY 1 LINE -500.0 6.28 6 7

results in the message:

USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA CARD QHBDY 1 SORTED
CARD COUNT = 82
Although the use of a negative value is rare in heat transfer, it should
be allowed.
Avoidance: Use the QBDY entry to apply the negative heat flux value.
@
6631; 69.0.2; Optimization - Superelements, DRESP2 - UFM 5609
If an optimization model contains superelements, UFM 5609 may be
incorrectly issued if different response types are combined using a
DRESP2. For example, this error occurs if a superelement is defined in a
frequency response optimization model, and both FRDISP and FRFORC
responses from DRESP1 entries are combined on a DRESP2 entry.
Avoidance: If different response items need to be combined, use a
residual-structure only model or use external superelements.
@
6632; 69; Documentation - SYSTEM(198), SYSTEM(205) - Missing Documentation
Version 68.1 of MSC/NASTRAN introduced a very significant performance
improvement to sparse decomposition. This enhancement is called the
high-rank update. The capability is controlled by two system cells,
SYSTEM(205) and SYSTEM(198). The optimum value of these system cells is
machine dependent. In Version 68.1 the optimum value of these two system
cells are set in the delivered system rc file. In Version 68.2 the
optimum values are set in the code. In either version the values may be
overwitten by the user but this is generally not recommended.
SYSTEM(205) sets the value for the rank of update. The value used for any
run may be determined from user information message 4157 output in the f04
execution summary. SYSTEM(198) sets a lower bound on the number of rows
to be eliminated in a front, hence the alias name MINFRONT. The meaning
of system cell 198 has not changed from 68.0 to 68.2. In version 68.0
this feature was only helpful on a few very large problems on CRAY
computers. In Version 68.1 and 68.2 it becomes more critical and can
help performance on many other machines. Normally, it is best to set
system cells 198 and 205 to the same value, this is done be default.

@
6643; 69; Optimization/Design Sensitivity - PARAM,OPTEXIT, DSAL - SFM 4276
, EC 3021
The message "SYSTEM FATAL MESSAGE 4276, ERROR CODE 3021" occurs in the
 DSAL module when a converged optimum solution is achieved in a modal
frequency response design task and OPTEXIT=7 is used to obtain
sensitivities of the final design.
Avoidance: Remove the OPTEXIT=7 request. A new run that uses the final
design and has OPTEXIT=4 will provide the desired information.
@
6656; 69; Optimization/Design Sensitivity - DEQATN - Segmentation Fault, T
rap Encountered
A DEQATN bulk data entry may not end with a semicolon(;). If it does, the
run fails while processing the input in a machine-dependent manner. This
error has appeared as a segmentation fault on SGI machines and a "Trap
Encountered" on RS/6000 machines.
Avoidance: Do not use a semicolon as the last character on a DEQATN
entry.
@
6668; 68.2; DMAP - OUTPUT2 - FORTRAN I/O ERROR 910
On the HP, if a user dmap contains an output2 request and there is no
corresponding assign statement, then the job may terminate without any
error messages in the f06 file. The following error messages may appear
in the log file.
*** FORTRAN I/O ERROR 910: ACCESS PAST END OF RECORD ATTEMPTED
FILE: ftn11, UNIT: 11
Avoidance: Explicitly define the output2 file using the assign statement
in the FMS section.
@
6676; ; Plotting - SPCFORCE, MPCFORCE, XYPLOT - Missing Output
If XYPLOTs of SPCFORCEs or MPCFORCEs are requested, the items must be
requested in the case control before the OUTPUT(XYPLOT) section. For
example, if SPCFORCE plots are to be requested for points 1 and 5, the
following output request (or a similar one) would be required in the case
control:
SET 99 = 1,5
SPCFORCE = 99
@
6678; ; Nonlinear Analysis - TEMP(LOAD) - Wrong Nonlinear Strains with T
emperature Loads
In nonlinear solution sequences, the output entitled "Nonlinear Stresses"
prints mechanical strains which do not include the thermal strains. The
true strains are the sum of the mechanical and thermal strains. The only
exception to this limitation are the hyperelastic elements which output
the true strain.
Avoidance: There is no avoidance.
@
6682; 69; DMAP - MATMOD - MATMOD with option P1=2 and P5<0.0.
Option P1=2 of the MATMOD module is intended to filter the matrix and set
all terms in the matrix with magnitude smaller than P5 to 0.0. However,
if P5 is set to a negative value, no filtering will take place.
Avoidance: Use a positive value for P5.
@
 6684; 69; Optimization/Design Sensitivity - DOPTPRM - UFM 316 (IFPDRV)
On the DOPTPRM data entry, a P1 value higher than 4 or 5, depending on the
computer, causes the FATAL message:
USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA CARD DOPTPRM
Avoidance: Select a value for P1 that is equal to or less than 4.
@
6701; 69; Transient Heat Transfer - TETRA - Wrong Answers
The 10-noded tetra for the transient thermal analysis (SOL 159) will
obtain wrong answer.

Avoidance: Use lower order tetra elements.

@
6705; 69; Optimization/Design Sensitivity - MPCFORCE - Missing Output, Wro
ng Answers
Running SOL 200 with SCR=NO may result in wrong or missing MPCFORCE
output. ANALYSIS=STATICS results in missing MPCFORCE output for all but
the first cycle. Other approaches result in wrong but plausible answers
for all but the first design cycle.
Avoidance: Submit job with SCR=YES.
@
6711; 69; Miscellaneous - INCLUDE - UFM 307 (IFPDRV)
If an include statement immediately follows a line which ends in a blank
comma field, as in the example below, MSC/NASTRAN may fail with the
message "USER FATAL MESSAGE 307 (IFPDRV) ILLEGAL NAME FOR BULK DATA ENTRY
INCLUDE, USER FATAL MESSAGE 9002 (IFP) - ERROR ENCOUNTERED IN MODULE
IFP".with the two lines below:
EIGRL, 1, -0.001, , 2, 0, 2, , $
include 'struct.inc' $

Avoidance: The error can be avoided by separating the lines or removing

the trailing commas.
@
6718; ; P-Elements - GRMAX, GNSTRT, GNMAX, VGSTRT - Keyword Settings Ign
ored
The NASTRAN keywords GRMAX, GNSTRT, GNMAX, VGSTRT,VGMAX,VESTRT,VEMAX, that
control p-element parameters may be ignored.
Avoidance: Use the actual system cells 175, 178, 179, 180, 181, 182, and
183.
@
6721; 69; Miscellaneous - Network - Error Writing File, Poor Message
The following message may occur when MSC/NASTRAN cannot access the disk
where the .f06 file is to be written in a reasonable amount of time
 because of network problems.
The sequential WRITE request cannot be completed because an AIX errno
value of 28 was received while writing the sequential file ./csr5504s.f06.
The program will continue if ERR= or IOSTAT= was coded in the WRITE
statement. 1525-099 Program is stopping because errors have occurred in an
I/O request and ERR= or IOSTAT= was not coded in the I/O statement.
Avoidance: Either re-run the job when network traffic is reduced or
direct the output files to a local disk, not through a network.
@
6722; 69; Transient Heat Transfer - TRIA6 element using CHBDYE - Wrong Ans
wers
The heat capacitance for the edge of the TRIA6 element is not correct.
This results in wrong answers whenever the CHBDYE (boundary element) is
used with SIDE other than side 1.
Avodiance: Use TRIA3 element with CHBDYE or CHBDYP,LINE for the edge
convection.
@
6724; 69; DMAP - Linking with SOL or LINK statement - UFM 149 and SFE 4276
in LINKAR
If UFM 149 is issued then, in some cases, SFE 4276 will be issued and the
run does not terminate gracefully. The exact symptoms are under
investigation. If this occurs during the creation of a delivery data
base, then the data base will be corrupted.
Avoidance: Other than correcting the DMAP error indicated by UFM 149,
there is no avoidance.
@
6725; 68.2; Matrix Methods - Sparse Decomposition, DECOMP - UFM
A large model that has run successfully on other machines experiences
negative terms on the factor when decomposing KOO on the MSC/NASTRAN
Version 68.1 on the SGI R8000 computer.
Avoidance: If decomposition errors are encountered that do not appear to
be due to modeling problems, then it may be due to the High Rank Update
feature introduced in Version 68.1. Disable this feature by setting
NASTRAN SYSTEM(198)=1 SYSTEM(205)=1 or switch to Version 68.2 where the
error has been corrected.
@
6726; 68.2; Miscellaneous - MPYAD - SFM 4276
Large client model failed in the MPYAD module on the SGI R8000 computer.
Avoidance: The problem was avoided by disabling the sparse option for
MPYAD. This was done by NASTRAN SPARSE=24, however, a significant penalty
in CPU resulted.
@
6727; 69; Cyclic Symmetry - CYSUP - UFM 5423
In MSC/NASTRAN Version 68.2 it is possible to get the message "USER FATAL
MESSAGE 5423 --- ATTEMPT TO ADD INCOMPATIBLE MATRICES, (SADD5)" in a
cyclic symmetry (SOL 114) run using the CYSUP entry. Version 67.5 produces
incorrect answers under the same circumstances.
 Avoidance: For V67.5
COMPILE STATCYC SOUIN=MSCSOU NOREF NOLIST
ALTER 8,8
NOA=LTOI(NOGSET=NOASET) $
CEND
Avoidance: For V68
compile statcyc noref nolist
alter 'noa=ltoi',''
noa=ltoi(nogset=noaset) $
@
6728; 69; heat transfer - apply boundary condition on both side on CHBDYE
or CHBDYG aqa_02
heat transfer - apply boundary condition on both side on CHBDYE or CHBDYG
-
SEE CSR 7462 for the fix that have corrected this error in version 69
Wrong summary for the H E A T F L O W I N T O H B D Y E L E M E N
T S
The user has the option using same boundary element to specify front side
and
the backside convection and radiation. The temperature vectors calculated
from
using this option is correct. However, the summary table only show the
energy
on the front side of the element
H E A T F L O W I N T O H B D Y E L E M E N T S (CHBDY)
ELEMENT-ID APPLIED-LOAD FREE-CONVECTION
FORCED-CONVECTION RADIATION TOTAL
13 4.000000E+00 0.000000E+00
0.000000E+00 0.000000E+00 4.000000E+00
Avodiance: use different CHBDYE or CHBDYG ID for specifying the front side
and
the backside.
@
6731; 69; Cyclic Symmetry - SPCFORCE, CYSUP - Missing Output
The forces of single point constraints in the static cyclic symmetry
solution sequence SOL 81 do not include the reactions at the degrees of
freedom specified on the CYSUP Bulk Data entry.

Avoidance: Use SOL 114 or insert the following DMAP alter:

compile sol81 souin=mscsou noref nolist
alter 818,818
@
6733; ; P-Elements - TEMP(LOAD), PVAL,ADAPT - Wrong Answers
If thermal loads are applied to p-elements and the PVAL entry (referenced
by the ADAPT entry) defines the polynomial order to be 1, then the
solution may converge to wrong results. Wrong results definitely occur
for models with only one element.
 Avoidance: Set POLYi field on the PVAL entry to 3 or more.
@
6738; 69; Superelement Analysis - SEMA - Poor Performance
If the residual structure of a large superelement model contains a
relatively large percentage of the total grid points, then the total run
time may be significantly larger then the same model without
superelements. For example, in one client's superelement model with over
100,000 grid points in the residual structure with 16 superelements (each
with about 1500 grid points) the total run time increased by a factor of
7. The increased time required was due to the SEMA (superelement
assembly) module. SEMA is intended to work on a collector superelement (or
residual structure) that is much smaller than the total model size.
Avoidance: Try to keep the residual structure and collector superelement
size to less than 10 percent of the model. The smaller these components
are, the more efficient the assembly process is. A DMAP alter avoidance
is also available from your MSC Technical Support representative.
@
6739; 69; Dynamic Analysis - ASETi, OMITi - Performance Degradation
If a static reduction is requested in a dynamic solution and there is no
mass on the omitted degrees of freedom, the program will still perform the
operations to reduce the mass matrix. This operation is not necessary and
may be skipped by a DMAP alter.
Avoidance:
In the SSS, include the following alter for V68:
compile semr2, souin=mscsou, list, noref
alter 19
$
$ check for null moo and moa
$
matmod moo,,,,,/tempaa,/12/s,n,nulmoo/1 $
matmod moa,,,,,/tempbb,/12/s,n,nulmoa/1 $
paraml moo//'trailer'/1/s,n,nomoo $
paraml moa//'trailer'/1/s,n,nomoa $
if (nulmoo=nomoo and nulmoa=nomoa)then $
equivx maa1/maa/always $
else $
alter 23
endif $
@
6744; 70.5; Optimization/Design Sensitivity - MTRXIN,GMA - User Fatal Messag
e 5423
The message "USER FATAL MESSAGE 5423" may occur in the MTRXIN module when
an attempt is made to read M2PP, B2PP and K2PP matrices a second time.
This error occurs in optimization analyses with K2PP, B2PP and TF inputs.
This error was fixed in Version 69.0.2 and 69.1 but occurred in Version
70. The error is fixed in Version 70.5.
Avoidance:
Use the following DMAP to skip MTRXIN for subsequent times:
 COMPILE GMA SOUIN=MSCSOU NOREF NOLIST $
ALTER 2 $
TYPE DB,ZUZR01,ZUZR02,ZUZR03 $
TYPE PARM,,I,N,NOM2PP,NOB2PP,NOK2PP $
ALTER 9,9 $
PARAML ZUZR02//'PRESENCE'////S,N,NOZUZ2 $
PARAML ZUZR03//'PRESENCE'////S,N,NOZUZ3 $
IF ( NOZUZ2<0 AND NOZUZ3<0 ) THEN $
MTRXIN CASES,MATPOOL,EQDYN,,TFPOOL/K2PP,M2PP,B2PP/LUSETD/
NOK2PP/NOM2PP/NOB2PP $
COPY M2PP/ZUZR01 $
COPY B2PP/ZUZR02 $
COPY K2PP/ZUZR03 $
ELSE $
COPY ZUZR01/M2PP $
COPY ZUZR02/B2PP $
COPY ZUZR03/K2PP $
NOM2PP = 0 $
NOB2PP = 0 $
NOK2PP = 0 $
ENDIF $

@
6745; 69; P-Elements - GMBC, SPCADD - Wrong Answers
GMBC entries referenced by SPCADD bulk data entries are ignored.
Avoidance: Other than not using the SPCADD entry, there is no avoidance.

@
6753; GL; Nonlinear Static - RESTART, PARAM, NMLOOP - Wrong Answers
If PARAM,NMLOOP is used to restart from SOL 106 (nonlinear statics) into a
linear solution, the data recovery output is performed using linear
element properties. Ths design causes displacements and element results
to be incorrect. However, the solution is performed using the nonlinear
equilibrium position (allowing EIGENVALUES to be correct). Only data
recovery is in error, hence displacements and element results are wrong.
This design does not affect a restart from a nonlinear run to a nonlinear
run.
Avoidance: There is no reason to perform a RESTART from solution 106 to
101, perform the desired data recovery in solution 106.
@
6755; 69; Documentation - Inertia Relief, SUPORT, PARAM,INREL - Wrong Answ
ers
In addition to a SUPORT entry, PARAM,INREL,-1 is required to perform
inertia relief in solution 101. This requirement is not documented under
the description of the SUPORT entry. It is documented under the SUPORT
entry description the SUPORT entry "does not perform the same function as
a constraint". In fact if PARAM,INREL,-1 is not included the SUPORT entry
will act like a constraint.
@
6759; 69; SSSALTER Library - RFLAG,LMT - Fatal Error
The SSSALTER Library files RFLAG and LMT will cause fatal errors when used
with SOLS 107, 108, and 109 if the stiffness matrix is singular.
 There are two possible avoidances:
1. Modify the dmap alter with the following change:
Change the DMAP Alter from:
GETLLL=(STATICS OR NORSET>-1 OR NO3SET>-1 OR
(MODACC>=0 AND (SOLTYP='MODAL' OR APP='REIG'))) $
to:
GETLLL=(STATICS OR NORSET>-1 OR
(MODACC>=0 AND (SOLTYP='MODAL' OR APP='REIG'))) $
Removing the NO3SET. This matrix does not need to be decomposed for
direct analysis. The alter causes a branch over this unneeded operation.

2. Modify the model by adding enough SUPORT entries to stabilize KLL.

These SUPORTs will not modify the answers.

@
6761; 69; P-Elements - GMLOAD - GMLOAD Entry Does Not Check if FEFACE Exis
ts
The GMLOAD entry allows the following options in the entity field:
FEEDGE, GMCURVE, FEFACE, and GMSURF
If there are no FEFACE, GMCURV or FEFACE entries in the model, no error
message is issued, and the user is not given a clear idea of what is
wrong.
@
6768; 69; Documentation - Reference Manual, CURV, RF24D40 - Curve Projecti
on Limits
The description for the CURV module processing in the Reference Manual
under RF24D40 (page 15.8-19) Remark 4.a. states:
- -
If |n.i|>= 0.4 the projections of the y-axis on the surface of the element
is taken as the reference axis.

This statement should read:

- -
If |n.i| > 0.4 the projections of the y axis....

@
6772; 69; P-Elements - TEMP(LOAD), CHEXA,PSOLID,PVAL - Wrong Answers
A skewed p-hexa element with p=1 (on the PVAL entry) produces incorrect
results with thermal loading. If the p-hexa is not skewed or if p>1, the
answers are correct. Incorrect results may also occur if the p-element is
next to an h-element, since the p-element will be reduced to p=1 on that
face to be compatible with the h-element.
Avoidance: Turn off the bubble functions using the PSOLID entry (5th
 field) or set p>1 on the PVAL entry.
@
6785; 70.5; Buckling Analysis - EIGRL - SFM 5401
In a buckling analysis, if only the V1 and ND fields are specified on the
EIGRL entry, then the job may fail with the following error messages:
SYSTEM FATAL MESSAGE 5401, LANCZOS METHOD IS UNABLE TO FIND ALL
EIGENVALUES IN RANGE.
A symptom of this error is the continuing shifting of Lanczos when ND
roots have been found.
Avoidance 1: Specify the V2 field in addition to the V1 and ND fields.
Setting SHFSCL on the EIGRL entry near the absolute value of the lowest
expected root may also avoid the error and improve efficiency. Another
alternative is to use the SINV method if only a few roots are desired
(which is true for most buckling analysis).
Avoidance 2 (preferred over Avoidance 1): A new alter has been added to
the sssalter library that can be used with Version 69.0.2. (the alter can
not be used with prior releases) that solves this problem. The alter is
named br.v69 for Version 69.0.2 or br.v691 for Version 69.1. Leave the
SHFSCL entry on the EIGRL card blank. The alter will be made available on
the web and on the 69.1 delivery.
@
6786; 69; Optimization/Design Sensitivity - FBS, QSETi - SFM 2198 and 2199
If QSETi entries are specified in solution 200 for the purposes of dynamic
reduction and the a-set contains only the q-set degrees of freedom and
there is also a static subcase, then the run will fail with SYSTEM FATAL
MESSAGE 2198 and 2199. (Version 68.2 only)
Avoidance: There are three possible avoidances:
1. Remove the q-set and the dynamic reduction requests and, if not
already done so, then replace with a Lanczos eigenvalue analysis
(see EIGRL entry).
2. Specify ASETi entries for one or more physical degrees of freedom.
3. Insert the following DMAP alter:
compile fea noref nolist
alter 'FBS'(,-1),'FBS' $
DBSTATUS UL,USET,LLL//S,N,NOUL/S,N,NOUSET/S,N,NOLLL $
IF ( NOUL=-1 AND NOLLL>0 ) FBS LLL,,PL/UL $
@
6788; 2001; Transient Response Analysis - XYPLOT - UWM 982, Missing Output
If an xyplot is requested for the velocity of a grid point which does not
exist and there is no VELO=
request in the CASE control for printed data, then no displacement,
velocity, or acceleration data will be printed or plotted and the
following message will be issued:
*** USER WARNING MESSAGE 982. FORMAT OF OUGV1 INCOMPATIBLE WITH SDR3
DESIGN. CODE = 165
 Avoidance: Review the output requests versus the grid and element ids in
the model, insuring that all requests reference proper ids. Another
avoidance is to add a VELO(PLOT)=ALL command to the case control. This
command will produce very large files on large size models, however.
@
6789; ; Nonlinear Transient Analysis - PARAM,LGDISP, PARAM,G - Incorrect
Damping
There is an error in the way element damping is implemented in nonlinear
transient analysis (SOL 99 and 129). This error can result in answers
with apparently very wrong damping when large displacement effects are
requested with PARAM LGDISP. These solution sequences create a "B"
damping matrix, from which damping forces are computed. This matrix is
updated whenever the stiffness is updated. The problem is that many
structures have stiffness ratios (and hence damping ratios) in
perpendicular directions that are orders of magnitude different. A tiny
misalignment of the damping can be intolerable.
Avoidance: There is no avoidance.
@
6790; 69; Nonlinear Statics - MATGEN, SCE1 - SFM 3007, UFM MATGEN-4, Poor
Message
In nonlinear static analysis, SOLs 66 and 106, if the model has no l-set,
a poor message is issued.
In SOL 66, SFM 3007 is issued for module SCE1 at line 858 in Version 68.
In SOL 106, " USER FATAL MESSAGE MATGEN-4, ILLEGAL VALUE FOR PARAMETER 2 =
-1" is issued for module MATGEN at line 143 in Version 68.
Avoidance: Insert the following DMAP alter in SOL 106 to obtain the
correct message, UFM 4402.
compile nlstatic souin=mscsou noref nolist
$alter 117 $ V68
alter 119 $ V68.2
IF ( NOLSET=-1 ) CALL ERRPH2 //SUBDMAP/4402 $
@
6794; 69; Postprocessing - Heat Transfer, PARAM,POST,-1 - Missing Databloc
ks
Executing solutions 153, heat transfer, or 159, nonlinear transient heat
transfer, with PARAM,POST,-1 does not produce the geometry data blocks
GEOM1, GEOM2 etc. in the OUTPUT2 file. This error only occurs in
MSC/NASTRAN Version 68.2.
Avoidance: Apply the following alter:
compile super3 noref nolist
alter 217,217
alter 220,220
@
6795; 69; Miscellaneous - SGI R8000 Only - Command Terminated Abnormally
Jobs submitted with scr=yes on MSC/NASTRAN Version 68.2 on the SGI R8000
computer may in rare cases not write the "END OF JOB" to the f06 file but
instead issue a "Command Terminated Abnormally" message in the log file.
In the single case experienced the job otherwise completed processing and
 the error was easily avoided by running with scr=no.
@
6797; ; DBC/TRANS/ACCESS - V6 & Later for IBM does not support READONLY
Version 6 and later releases for MSC/ACCESS for IBM Mainframes running
MVS/XA does not allow for current access of XDB databases via the READONLY
request. As per the information from the NASTRAN Product Manager, there
will be no more versions of MSC/NASTRAN for the IBM MVS platforms so this
error will not be corrected.
@
6799; 69; Optimization/Design Sensitivity - DOM2 not available in V68
Prior to Version 68, design variable values could be passed into the DOM2
module and an EPT datablock based on the values and the property to design
variable relations would be produced. This capability was
removed from DOM2. The functionality has been re-instated in the DOM9
module in Version 69.
@
6800; 2001; Transient Analysis - DPD - Poor Message, UFM 2071
In solution 109, if an initial condition, ic, is specified without the
presence of a DLOAD command in the case control, the following fatal
message is produced:
USER FATAL MESSAGE 2071 MODULE = DPD
DYNAMIC LOAD SET 1 REFERENCES UNDEFINED TABLE 1
This message occurred despite the fact that no DLOAD was selected, and
the TLOADi,1 entry in the bulk data was a TLOAD2, which can not reference
a table.
Avoidance: Input a DLOAD command in the case control which references a
TLOADi entry which in turn
references a DAREA entry with a zero coefficient.
In general, all dynamic load input must be complete, even when it can not
be selected. For example, if there is incomplete frequency response load
data in a transient analysis input file, they will cause fatal errors in
the DPD module, even though they can not be selected in that context.
@
6801; 70; Optimization/Design Sensitivity - DRESP1, SUPER3 - UFM 9058
In solution 200, if a DRESP1 entry is provided for an analysis type that
is not present in a case control ANALYSIS command (e.g. RTYPE = DISP, but
there is no STATICS of SAERO subcase) then "USER FATAL MESSAGE 9058
(SUPER3) THE SOLUTION FOR THE RESIDUAL STRUCTURE DOES NOT EXIST" may
occur.
Avoidance: Do not use unreferenced DRESP1 entries.
@
6802; 69; Optimization/Sensitivity - DOM9 - Quits No Message
If an optimization problem has a near mechanism and param,bailout is set
to force it through static decomposition without a fatal message, then the
job may fail on DEC ALPHA systems in the DOM9 module with no message.
Other machines will produce meaningless answers.
Avoidance: There is no avoidance.
@
6808; ; Documentation - CBEND, MSC/NASTRAN Reference Manual - Incorrect
Equations
There are typographical errors in equation 7 of section 5.2.3 of the
Reference Manual. R-sub-R should be R-sub-B and in the integral for Z, "R"
in the expression R/R-sub-b should be "r".
Avoidance: The correct equations may be found in the old MSC/NASTRAN
User's Manual
@
6809; 70; Miscellaneous - XSORT - UFM 1126 (GNFIST)
If the input data contains two "BEGIN BULK" entries, then the following
misleading message may be produced:

USER FATAL MESSAGE 1126 (GNFIST)

DMAP MODULE XSORT IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME = IBULK,
WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE IBULK =OVRWRT
@
6810; 69; Optimization/Design Sensitivity - DESGLB - UFM 6545 (DOPR3)
The message "USER FATAL MESSAGE 6545 A DESGLB= XXX COMMAND APPEARS AT THE
SUBCASE LEVEL" is issued in error when the DESGLB command appears above
the subcase level and there is more than one "super" discipline in case
control. There are three super disciplines in MSC/NASTRAN design
sensitivity and optimization: Frequency (ANALYSIS=MFREQ or DFREQ),
Transient (ANALYSIS=MTRAN) and everything else
(ANALYSIS=STATICS,MODES,BUCK,FLUTTER or SAERO).
Avoidance: Remove the DESGLB command and invoke the constraint with a
DESSUB command from the first subcase.
This will force the evaluation of the constraint at the subcase level even
though there are no subcase
associated responses involved in the constraint evaluation.
@
6812; 69; Eigenvalue Analysis - READ, EIGR - UWM 5402
In normal modes or buckling analyses which are defined by the EIGR Bulk
Data entry, if the reduced stiffness matrix is less than 0.0001% dense,
and the stiffness is not null, then UWM 5402 is issued and no eigenvalue
analysis is performed.
Avoidance: Use the EIGRL Bulk Data entry or insert the following DMAP
alter:
compile moders souin=mscsou noref nolist
alter 100 $ Version 68 and 68.2
modtrl kxx///////1 $ set density to .0001
@
6813; 70.5; Restarts - SOL 144, TRIM, Aero User's Guide - UFM 5423
The RESTART example ha144a_rst in Chapter 4 of the Aeroelastic Analysis
User's Guide deletes two of the three subcases from ha144a.dat, the cold
start run. This causes the RESTART job to fail when it tries to use the
UOOX matrix from the cold start with the new GOAT matrix.
 Avoidance: Add a dummy FORCE Bulk Data entry to the Bulk Data section and
it need not be referenced by the LOAD command. In general lines should not
be deleted from the Case Control, subcases and output requests should be
added.

@
6814; ; p-Elements - GMCURV, GMSURF - Poor Answers
When p-elements are connected to h-elements the polynomial order at the
interface should be p=1.
If GMCURV or GMSURF data is used along the interface surface the code
ignores the different connections and allows the polynomial order to grow.
Severe stress concentrations will be calculated at the connected points.
Avoidance: Since the interfaces must be straight there is no need for
geometry, remove the gmcurv or gmsurf from the interface.
@
6815; 68.2; Optimization/Design Sensitivity - PARAM,POST - Missing Results f
or PATRAN
If a SOL 200 analysis is run with PARAM,POST,-1, only the last design
cycle results are available for postprocessing in MSC/PATRAN using the
output2 file, and there is no any model information.
Avoidance: The error has been fixed in MSC/NASTRAN Version 68.2, the user
should switch to this version.
@
6818; 69; P-Elements - FEEDGE - UFM 2104
When point entries are used with the FEEDGE entry, the order is not
supposed to matter; however, some runs may fail and issue the message
"USER FATAL MESSAGE 2104, UNDEFINED COORDINATE SYSTEM".
Avoidance: There is no simple avoidance.
@
6821; 69; Acoustic Analysis - ACMODL, ACMG - Trap Encountered, Traceback
In acoustic/fluid analysis, if DIFF is used on the ACMODL entry and some
fluid grids on the fluid/structure interface are not captured by the FSTOL
tolarance, then the job may fail without any message being written to the
f06 file. The log file may indicate a traceback as follows:
Trap encountered, Traceback:
Offset 17dc in procedure acmg_
Offset 1f0 in procedure xsem05_
Offset 140 in procedure mainal
Offset 0 in procedure start

Avoidance: Set a value for FSTOL that captures all of the fluid grids at
the fluid/structure interface.
@
6823; 2001; Optimization/Design Sensitivity - PCOMP, IFP - SFM 3003
The auxiliary model technique of shape optimization (solution 200) does
not work when PCOMP entries are present. The following message is
received:
 SYSTEM FATAL MESSAGE 3003 MODULE = IFP
ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK IEPTX
(FILE 202) IN SUBROUTINE IFP6
Avoidance: The avoidance is to use one of the other three shape
optimization methods.
@
6826; 69; Inertia Relief - PARAM,INREL/ SUPORT - UFM 6137, UFM 3005
If the inertia relief feature is used in SOL 101, and the model does not
contain a mass matrix, then the job will fail with the following unclear
messages:
*** USER FATAL MESSAGE 6137 (UDSSDD)
INPUT MATRIX IS RANK DEFICIENT, RANK = 0
USER ACTION: CHECK MODEL
*** USER FATAL MESSAGE 3005 MODULE = SOLVE
In SOL 24, the job will terminate without any messages. A symptom of this
error is printing of the parameters after echo of the input file.

Avoidance: Include the appropriate mass required for inertia relief.

@
6829; 69; Miscellaneous - Network Job Submittal - Job Not Submitted
If the current working directory is auto-mounted, and there is a delay in
starting execution of an MSC/NASTRAN or MSC/EMAS job (e.g., submitting job
with aft= or que=), then there is a chance that the job will not be
submitted. Depending on your network setup, you may receive the following
message:
/tmp_mnt/net/<node>/<somedir> : not found
from the daemon.
This error is encountered when it attempts to change the working directory
to the current working directory as determined by the nast682 or emas32
script. The error can occur on any system where the "pwd" command does
not return the canonical working directory. For example,
% cd /net/<node>/<somedir>
% pwd
/tmp_mnt/net/<node>/<somedir>
If sufficient time passes, the system will release the mount of
'/tmp_mnt/net/<node>/home/<somedir>' and the subsequent reference to this
directory will fail. When this happens, neither the F04, F06, nor the LOG
file will be created since the job is never submitted. Note this error
only occurs if you are submitting a job over the network and the current
directory is auto-mounted.
Avoidance: There are two possible avoidances.
1. Do not delay the submittal time for the job.
2. Leave the current directory window open without changing directory.
Sadly, this may not work on SunOS.
@
6831; 69; Installation - Remote Installation - Poor Message
If the "Remote Installation" option of the script is chosen, AND the
user's shell on the remote system is csh AND the remote system is not
properly configured so that test(1) is in the PATH during an rsh command,
the installation will fail indicating the CD or media could not be found
on the remote system.

Avoidance: The installation cannot use the specified remote system until
the PATH is corrected. The script has been modified to explicitly test
for this user error so that a meaningful message is output.

@
6833; ; Documentation - DMAP Module Dictionary - Examples Result in UFM
The example in Section 2.1 of the DMAP Module Dictionary has several
errors in it. The following lines of the input file cause fatal errors if
used as is:
QVOL 123 1.E+2 1 THRU 9
GRID 5 0 36O.O 240.0 0.0 0 3456
MAT1 1 30.0E3 1. 6.5-6 5.37+2 +MAT1
*MAT1 16.99235344 11.814744801 1.E04
CORD2R, 3, 0, 0.,0., 0., 0., 0., 1., +CORD1
+CORD 1. 0. 1.

Avoidance: Change the above input to:

QVOL 123 1.E+2 1 THRU 9
GRID 5 0 360.0 240.0 0.0 0 3456
MAT1 1 30.0E3 1. 6.5-6 5.37+2
+MAT1
*MAT1 16.99235344 11.814744801 1.E04
CORD2R, 3, 0, 0.,0., 0., 0., 0., 1., +CORD1
+CORD1 1. 0. 1.
@
6836; 68.2.3; Miscellaneous - OUTPUT2 - UFM 4130 (OUTPN2)
The CRAY Y-MP Version of MSC/NASTRAN may fail on a CRAY J-90 in the
OUTPUT2 module with the message "USER FATAL MESSAGE 4130 (OUTPN2) MODULE
OUTPUT2 - ILLEGAL TAPE CODE HEADER = NASTRAN FORT TAPE ID CODE -".
Avoidance: MSC has released a new version of MSC/NASTRAN specifically for
the CRAY J-90 computer that not only corrects this error but runs
approximately 30% faster than the Y-MP version. All clients with J-90s
should receive this new version.
@
6839; 69; Dynamic Analysis - FREQ, TSTEP, SDAMPING - Missing Documentation
, Messages
By design intent, solutions 107 through 112 do not support changes of
freq=, tstep=, method=, cmethod=, sdamping commands in subcases. For
 example, if in sol 111 there are two subcases, and each has a different
freq= command, only that in the first subcase will be honored. All
similar commands in the second and subsequent subcases will be ignored.
The output for all subcases will be for the frequency set selected in the
first subcase. Note that solutions 108 and 111 do correctly support
changes in the DLOAD command between subcases.
Avoidance: There is no avoidance.
@
6849; 69; Aeroelastic Response - PARAM,DDRMM,-1 - Wrong Answers
When a transient analysis is performed in SOL 146 with PARAM,DDRMM,-1
wrong answers are produced.
Avoidance: Use PARAM,DDRMM,0 (i.e., the default), use SOL 76 or apply the
following DMAP avoidance for Version 68.2:
compile seaero nolist noref
alter 'call mfreqrs'
app1=app $
@
6850; 69; Optimization/Design Sensitivity - PARAM,OPTEXIT,-4 - Output Ove
rwritten
If "param,optexit,-4" is used in a design optimization run (SOL 200), then
the OUTPUT2 results for table DSCMCOL will be overwritten by the matrix
DSCM2. This error is due to assigning both the output2 and output4 files
to the same unit.
Avoidance: Insert the following dmap avoidance.
compile exitopt souin=mscsou nolist noref $
alter 20,21 $ v68.02
output4 dscm2,,,,//-1/12 $
output4 ,,,,,//-2/12 $
@
6853; 70.5; Frequency Response/Random Analysis - RANDOM - Infinite CPU Loop
If a zero hertz excitation frequency is specified in a run which has a
free-free model and is using modal frequency response to perform a random
analysis, the run may encounter an infinite loop. This problem appears to
be due to the zero frequency solution being a numeric infinity. If the
same model is used in a direct frequency response, the run will fail
during the decomposition due to the 0.0 hertz static solution being
singular. In the modal solution, a matrix decomposition is not performed
(in the uncoupled solution), therefore, a solution is obtained for the 0.0
hz excitation.
Avoidance: Do not request a 0.0 excitation frequency in a modal frequency
response analysis.
@
6854; ; Sequencer - PARAM,SUPER, SEQP - SFM 4276, EC 7777, Traceback SEQ
MR
If a superelement model uses PARAM,SUPER,2 without PARAM,MPCX,sid (where
sid is the MPC set to be used), the run may fail with no messages in the
sequencer.
 Avoidance: Include PARAM,MPCX in the run if PARAM,SUPER,2 is used.
@
6855; 69; SSSALTERS - SEP1, lmt11.v682 - SFM 6357
In version 68.2, if the lagrange multiplier solution dependent sssalters
lmt*.v682 are used, then the job may fail with the following error
messages:
SYSTEM FATAL MESSAGE 6357 - NO GRID POINTS DEFINED FOR SEP1
Avoidance: Use the solution-independent lagrange multiplier sssalter
rflag.v682 instead of the lmt*.v682.
The rflag.v682 alter combines the function of all the lmt*.v682 alters.
@
6859; 69; DMAP - MODTRL, PARTN - Infinte CPU Loop
If an element of a matrix is given an index for a row number larger than
the number of rows in the matrix, as defined by the matrix trailer, and
partn operates on this matrix, partn will go into an endless loop.
The MTRXIN and MODTRL modules, and perhaps inputt4, can generate such
matrices when used injudiciously.

Avoidance: One way to avoid this error is to first put the matrix through
the ADD module. The module will throw away all terms beyond the
boundaries of the matrix as defined by the matrix trailer.
@
6860; 69; Postprocessing - PARAM,POST,-2 PARAM,OUNIT2 - Missing EOF, Datal
oader Abort
In MSC/NASTRAN Version 68.2.3 if PARAM,CHECKOUT,YES and PARAM,POST,-2 are
used the end of file is written to unit 11. If PARAM,OUNIT2 is used, the
eof will go to the OUNIT2 file, but the data is written to unit 11.
Avoidance: Please see error 5987.
@
6866; 69; Miscellaneous - Cray J90, T90 - UFM 4311
If you are running MSC/NASTRAN V68.2 or MSC/EMAS 3.2 on a Cray T90 or J90
which has more than 16 CPUs, then your job may fail with the following
error messages:
USER FATAL MESSAGE 4311 (XSEM)
TIME EXPIRED PRIOR TO LINK NS01, CALLING MODULE - 19 IFPL

Note that this error may occur even if you are not using more than one CPU
as long as the T90 or J90 has more than 16 CPUs. The error is due to one
of the timing kernels on the T90 and J90 returning an erroneous number.
Avoidance: A fix for this kernel error is available from Cray Research.
Even though this is not MSC's error, the error is marked as fixed in
Version 69 to remove it from the list.

@
6867; 69; Plotting - PLOTPS/NEUTRL - No .ps file from .neu
According to the documentation, the plotps utility should be able to
generate a postscript formatted plot from the MSC/NASTRAN plot file (.plt)
or a neutral (.neu) format plot file. (The neutral file format is
 necessary for transferring data between different machine types).
However, when the neutrl utility translates a plot (.plt) file into a
neutral (.neu) format, the plotps utility cannot translate the neutral
file and fails with the following message "USER WARNING MESSAGE (PLTFIL)
No frames found in file." The plotps utility works fine for translating
standard plot (.plt) files.
Correction:
The error exists in the neutrl utility, which is provided in source form.
In subroutine RDBIN, change line
240 from:
ibyte = 0
to
ibyte = 1
Note: A neutral (.neu) plot file is unnecessary unless the user wishes to
use one of the unsupported plot utilities given with a prior version of
MSC/NASTRAN. The PostScript formatted file (.ps) is already a neutral
format and is transferable across machine types.
@
6868; 69; Data Recovery - CBAR, SDR2 - SFM 4276, EC 7777, Command Terminat
ed Abnormally
Extremely small stresses for bar elements may lead to extremely large
margin of safety output. These large numbers may cause MSC/NASTRAN to
abort in the SDR2 module. This error has been observed on the DEC ALPHA
OSF/1 and the HP/9000 systems. The DEC OSF/1 issues a "command terminated
abnormally" message in the log file, while the HP/9000 issued a "SYSTEM
FATAL MESSAGE 4276, ERROR CODE 7777" followed by a traceback in the log
file. The traceback in the log file indicates that the job was in the
"numcvt" routine at the time of failure.
Avoidance: The error may be avoided by not setting PARAM,POST,0 for
MSC/XL data recovery. The error may also be avoided by removing the
stress limits on the MATi entries which results in the margins of safety
being calculated.
@
6875; 69; Utilities - Incorrect Ownership Statement
The MSC ownership statement in the utilities source
(<ins-dir>/msc682/util/*) is incorrect. The correct ownership statement
is:

THIS PROGRAM IS CONFIDENTIAL AND A TRADE SECRET OF THE MACNEAL-

SCHWENDLER CORPORATION. THE RECEIPT OR POSSESSION OF THIS PROGRAM
DOES NOT CONVEY ANY RIGHTS TO DISCLOSE ITS CONTENTS, SELL, LEASE, OR
OTHERWISE TRANSFER IT TO ANY THIRD PARTY, IN WHOLE OR IN PART, WITH-
OUT THE SPECIFIC WRITTEN CONSENT OF MSC.
This statement permits the user to modify these routines for use with
MSC/NASTRAN and MSC/EMAS.
@
6876; 69; Matrix Methods - ADD5 - Inefficient Processing
NO EXTERNAL ERROR REPORT NECESSARY
If the scaling factor (for example, alpha) for any matrix in the ADD5
statement is zero and there are additional matrices input to the module,
 the module will scale each term of the associated matrix by 0.0 and add it
to the additional matrices.
Avoidance: No avoidance is necessary as this error is simply an
efficiency issue and wrong answers are not produced. DMAP may be used to
check if any scaling factors are 0.0 and the associated matrices may be
excluded from the addition.
@
6877; 69; Composites - Stress and Failure Indices - Mislabeled and/or Extr
aneous Output
In SOL 106, composite stress and failure index output is not labeled with
the load step as is found in other nonlinear output. Thus, PATRAN is not
able to distinguish composite output at different load steps.
In addition, in any solution that supports composite output, even if no
failure index output is requested, the header labels for failure index
output are always printed.
Avoidance: There is no avoidance.
@
6879; 70; Acoustic Analysis - DISP, ACCE - Extra, Incorrect Output.
If a fluid structure analysis is performed and accelerations are requested
for structural points and displacements are requested for the fluid
points, additional (incorrect) output will occur beyond the requested
fluid point output. Various machines may behave slightly differently.
Some machines may fail with the message "SYSTEM FATAL MESSAGE 4276, ERROR
CODE 7777" indicating a numeric overflow, other machines may output "NaNQ"
or "**********" indicating a numeric overflow.
Avoidance: Do not request structural accelerations in the same run
requesting fluid displacements. Perform a restart and keep the requests
separate.
@
6880; 69; Differential Stiffness - QUAD4, PARAM,LGDISP - SFM 4276, EC 777
7
Nonlinear analysis with QUAD4 elements and PARAM,LGDISP or buckling
analysis may fail in MSC/NASTRAN Version 68 on long word machines such as
CRAY and NEC. The run may issue SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777
indicating a numeric overflow. Other messages may be written to the log
file.
Avoidance: The error can be avoided by setting SYSTEM(170)=1 which select
the old quad4 differential stiffness formulation.

@
6881; 70.5; Miscellaneous - TA1 - UFM 2010, Missing Properties
If a model has a large number of property entries and insufficient memory
is allocated, the program may issue USER FATAL MESSAGE 2010 stating that
properties which are in the input file do not exist. This occurs in DMAP
module TA1.
Avoidance: Increase the amount of memory for the run or increase the
BUFFSIZE on the NASTRAN statement.
@
6882; 69; Miscellaneous - SFM 4276, EC 923, Insufficient Disk Space, Exces
sive Output
NO EXTERNAL ERROR REPORT UNTIL I GET A DECK AND A BETTER DESCRIPTION AS TO
WHEN THIS HAPPENS.
If the files used by an MSC/NASTRAN run exceed the available space on a
disk, the program will provide SFM 4276 EC 923, indicating that system
resources were insufficient. However, the message is followed by a
"core-dump" which will be several thousand lines long. This additional
output has no meaning to the user.
@
6883; 69; Elements - HEXA(8), TEMP(LOAD), PSOLID - Wrong Answers
If an eight node HEXA uses full 27 point integration (IN=3 on the PSOLID
entry), incorrect thermal loads are computed resulting in wrong answers.
Avoidance: The avoidance is to use the default integration method.
@
6887; 69; Frequency Response Analysis - PARALLEL - Wrong Answers
MSC/NASTRAN Version 68 may produce wrong answers for frequency response
analysis whenever PARALLEL=n is selected, where n is greater than 0.
Avoidance: The error only affects parallel sparse complex decomposition.
Parallel may be selected for other operations while being deselected for
this operation by:
instead of parallel=1, the workaround is parallel=262145
instead of parallel=2, the workaround is parallel=262146
instead of parallel=3, the workaround is parallel=262147
instead of parallel=4, the workaround is parallel=262148

@
6888; 69; SSSALTER - mfreqa.v68 and mfreqa.v682 - Error in Scaling of Resu
lts
In the sssalter mfreqea.v68 and mfreqea.v682 the scaled results of the
modal contributions are incorrectly scaled. Only as many modal
contributions as there are excitation frequencies are scaled and the
others are not printed.
Resolution/Avoidance:
Change the following line :
mpyad onec,totalt,/scalit $
to be
mpyad oner,totalt,/scalit $
In the alter for V68.2, also change the following line:
mpyad kinetc,onec,/totalk/1 $
to be:
mpyad kinetc,oner,/totalk/1 $
@
6893; ; Complex Eigenvalues, Matrix Methods - CEAD, UDCOMP - No Message,
Traceback
If sparse unsymmetric decomp spills the run may go into an infinite cpu
loop or fail with the following traceback being issued to the log file but
no other message:
Traceback:
Offset 0x00000010 in procedure txuycd_
Offset 0x00000010 in procedure udspcd_
Offset 0x00000164 in procedure udsncd_
Offset 0x000031dc in procedure udsfa_
Offset 0x00000440 in procedure udcomp_
Offset 0x000000e8 in procedure factru_
Offset 0x00000120 in procedure classd_
Offset 0x00000e88 in procedure clasdd_
Offset 0x00000c24 in procedure cead_
Offset 0x0000006c in procedure xsem11_
Offset 0x00000994 in procedure mainal

Avoidance: Increase memory to avoid spill. In Version 70.7 the pivot

threshold has been changed from 1.e-6 to 1.e-10 to make this error less
likely. Furthermore, it is now possilble to override the pivot threshold
with system cell 91. The threshold is set to 10^(sys91)
@
6894; 70; Miscellaneous - No Output - Intel Windows NT 3.1
During testing of MSC/NASTRAN Version 68.2 some jobs produced no output on
the Windows NT 3.51 system. The error could not be debugged because when
the command prompt window was closed and re-opened the error could not be
reproduced. Thus, this report is being written to assist users who
possibly encounter the same problem.
@
6902; 69; SSSALTER - Possible error using alter2g.v682
If the DMAP alters for coupled analysis are used with the OUTPUT2
(MATPOOL) option and a component has no damping matrix, the component
matrices may not be stored properly on the database.
Avoidance: In the alter, change NOM2GG to NOB2GG in the following DMAP
statement
mtrxin ,,baap,EQEXINS,SILS,/baaext,,/LUSETS/
S,N,NOM2GG///1 $
@
6903; 69; p-Elements - FEEDGE, GMLOAD, CORDi - Wrong Answers
If a GMLOAD is applied to an FEEDGE or GMCURV which have grids with
displacement coordinate systems other than the Basic coordinate system,
the loads are not properly applied.
Avoidance: Set the grids' displacement coordinate system to the Basic
coordinate system.
@
6904; 69; Nonlinear Analysis - MATS1, QUAD4, TRIA3 - Wrong Answers
 There are two errors in element force output for nonlinear QUAD4 and TRIA3
elements. The first error results in over-estimated force components.
This error exist in Version 68 systems only. This error occurs when
geometric nonlinear (LGDISP=1) is turned on and the element is in the
nonlinear material portion of the stress-strain curve (MATS1). Forces may
be over-estimated by more than 40%. The second error results in
under-estimated force components, on the order of 4%, when param,lgdisp is
the default, -1. This day one error is directly linked to the default
value of param,nlayers,4.
Avoidance: There is no avoidance for the over-estimated element forces
when param,lgdisp is 1. The under-estimated element force error may be
reduced to .7% by setting param,nlayers to 12.

@
6905; 69; Miscellaneous - NOTIFY - Command Garbled
A syntax error in the "handlenotify" shell function in the MSC/NASTRAN
Version 68.2.3 and EMAS Version 3.2.3 script produces a sed error:
sed: command garbled:

Avoidance: The problem is in the line that reads:

ncmd=`echo "$ncmd" | sed "s^{}^\'$msg\'"`
Replace this line with:
ncmd=`echo "$ncmd" | sed "s^{}^\'$msg\'^g"`

@
6906; 69; Documentation - HP-UX Installation Guide - Illegal Option
The documented example in the "Installation and Operation Guide" for the
ncmd command line option is incorrect for HP-UX computers. The example
should read:
ncmd=echo {} | mailx -s {} `whoami`

@
6910; 70.5; Documentation - OUTPUT4, INPUTT4 - Illegal Unit Number for INPUT
T4 -11
If the sparse option is used to write a matrix using OUTPUT4 (IUNIT<0),
then the program will fail when reading the matrix using INPUTT4 if the
sparse input option is used (IUNIT<0).
Avoidance: Do not use the sparse option when reading or outputting a
matrix with inputt4 or output4. Use a positive value for IUNIT. .
@
6913; 70.5.2; Superelements - Can Not Perform Dynamic Data Recovery for Extern
al S.E.
Data recovery is not supported for external superelements in dynamic
solution sequences.
Avoidance: There is no avoidance. Version 70.5.2 has been enhanced to
 provide this support, please see PARAM,EXTDR.
@
6914; 68.2.3; Installation - LANG - awk: UNIX command not found
MSC/NASTRAN Version 68.2 may get the following message, "awk: UNIX command
not found", or some other UNIX level fatal message on SGI machines outside
the US. Non US language AIX machines are unable to understand the 'mem'
command line parameter. Other machines may exhibit different problems.
The problems are all caused by the fact that the LANG environmental
variable is set to null in the Version 68.2 NASTRAN script and not
explicitly to 'C'.
Avoidance: All non US Language machines should change the MSC/NASTRAN
script to specify C explicitly:
# Set the "C" locale
LANG="C"; export LANG
This error is corrected in the Version 68.2.3 R2 CD.
@
6916; 69; Acoustic Analysis - HEXAPR - Incorrect Phase Output
The "COMPLEX ACCELERATION VELOCITY AND PRESSURE LEVEL" output is incorrect
for models that include damping. "REAL/IMAGINARY" output will always
indicate a zero value for the imaginary part of the acceleration and
pressure output and the real part of the velocity output. The nonzero
part indicates the magnitude. Likewise "MAGNITUDE/PHASE" output gives a
correct magnitude but an incorrect phase angle. An example of the
incorrect output is shown below.
C O M P L E X A C C E L E R A T I O N S V E L O C I T I E S A N D
P R E S S U R E L E V E L S
(REAL/IMAGINARY)
EL-TYPE X-ACCELERATION Y-ACCELERATION Z-ACCELERATION X-VELOCITY
Y-VELOCITY Z-VELOCITY PRES(DB)
HEXPR -1.215000E-09 3.504862E+03 -1.262103E-09 .0
.0 .0 3.E+01
.0 .0 .0 9.668665E-11
-2.789081E+02 1.004350E-10

Avoidance: There is no avoidance.

@
6921; 69; Matrix Methods - MPYAD - Negative Time Estimates, Poor Performan
ce
It is possible to get a negative time estimate for MPYAD. This error may
lead to poor method selection and thus poor performance.
Avoidance: There is no avoidance.
@
6928; 69; Optimization/Design Sensitivity - DESOBJ - Wrong Answers
In Versions 68, weight and volume cannot both be selected as design
responses in the same solution 200
 job. The problem will run to completion but one of these responses will
have zero value (i.e., it will be ignored in the design process). This
leads to incorrect answers.
The avoidance is to either:
1.) Use only weight or only volume in the design model
2.) Include the following DMAP alter (version 68, 68.2):
compile desinit souin=mscsou nolist noref
alter 66
wgtvol = 3 $
endalter
@
6932; 2001; Optimization/Design Sensitivity - PACABS, ELTPRT - UFM 5165, SF
M 3002
If the B field of the PACABS (acoustic absorber element) data entry is
specified as a design variable in an optimization analysis, then the
execution will result in the following fatal message:
USER FATAL MESSAGE 5165 (DSVG1D-1)
UNDEFINED INPUT FILE = 106, NAME = (NONE) , ERROR CODE = -6,
MODULE DSVG1 IS TERMINATED.
Avoidance: Use the following alter for 68.2.3:
compile pslgdv list noref souin=mscsou
alter 239, 239
IF ( DODFREQ>0 OR DOMFREQ>0 OR DOMTRAN>0 ) EMG,
EST,CSTMS,MPTS,DIT,,,,,,,GPSNTS,EQEXINS,SILS/
kelmx,kdictx,,,BELM,BDICT//0/s,n,nobgg $
In addition to the above error, optimization of the PACABS properties with
the objective to minimize weight fails at module ELTPRT in RESPSEN with
the following error message:
SYSTEM FATAL MESSAGE 3002 MODULE = ELTPRT
0EOF ENCOUNTERED WHILE READING DATA BLOCK ESTDV2F (FILE 105) IN
SUBROUTINE SKPREC

Avoidance: Minimize a DRESP2 which references the PACABS properties

instead of minimizing weight.

@
6933; 69; Optimization/Design Sensitivity - PMIN - Wrong Answers
If the user specifies a value of PMIN that is less than -0.99e35, the
value could potentially be incorrectly reset to 1.0e-20. This will happen
for I12 on a PBAR or PBEAM and for orientation on the PCOMP entry.
Avoidance: Specify PMIN values that are greater than -.99e35
@
6934; 69; Optimization/Design Sensitivity - DVPREL2 - Incorrect UFM 7618
When USER FATAL MESSAGE 7168 is encountered, the DVPREL2 ID referenced is
in error.
USER FATAL MESSAGE 7168 (DOPR1Z)
EQID XXX REFERENCED ON DVPREL2 YYY IS NOT PRESENT ON A DEQATN ENTRY.
 Avoidance: A nonexistent EQID has been referenced on a DVPREL2 entry. The
burden is on the user to determine which DVPREL2 makes the incorrect
reference.
@
6936; 69; Documentation - umsec6.doc - File Does Not Exist
The "readme" file in the nast68/doc directory states that the directory
contains the file UMSEC6.DOC which includes a list of MSC/NASTRAN
diagnostic messages. This file is no longer included with the delivery.
The MSC/NASTRAN online encyclopedia, delivered with all Version 68
systems, contains this information (and much more) in an easily accessable
format. Thus the online encyclopedia is a far superior method for
accessing information. The readme file will be updated for Version 69.
@
6937; 69; DMAP - COMPILE - UFM
Multiple COMPILE statements for the same subdmap may result in a user
fatal message. For example the following DMAPs should be equivalent but
the first may lead to a UFM:
COMPILE A COMPILE A
ALTER 1 ALTER 1
COMPILE B ALTER 3
ALTER 2 COMPILE B
COMPILE A ALTER 2
ALTER 3
Avoidance: Combine the alter so that a single COMPILE statement may be
used for the subdmap. Re-arranging the COMPILE statements such that the
COMPILE statements for each subdmap follow each other directly should also
avoid the problem.
@
6939; 69; Nonlinear Analysis - PARAM,NMLOOP, DBSTATUS - SFM 4276 EC 7777
Nonlinear analysis jobs with PARAM,NMLOOP may fail on the DEC ALPHA OSF/1
Version of MSC/NASTRAN.
Avoidance: Insert following DMAP alter:
compile nlstatic noref nolist
$alter '/s,n,fact','' $ V68.2
alter 486,486
PARAML ESTNL//'DTI'/0/6/S,N,FACT $ LOAD FACTOR
@
6941; 69; SSSALTER - MATMOD, ZFREQ - SFM BLDPACK ERROR
The Zfreq alters in the sssalter library require the presence of DMI
matrices for both frequency-dependent stiffness and damping if the
frequency-dependence part of the alter is to take effect.
If a zero term is input for any of these effects a BLDPACK error occurs in
MATMOD option 28, when
attempting to make a diagonal matrix from a vector.
Avoidance: Put in a very small number instead of a 0.0 value.
@
6943; 69; Optimization/Design Sensitivity - SHPCAS, TEMP(INIT) - UIM 9052,
SFM 3001
 Using TEMPERATURE(INITIAL) in a solution 200 shape optimization run with
an auxiliary model results in the messages:
USER INFORMATION MESSAGE 9052 (FEA) - STATIC ANALYSIS COMPLETED. DESIGN
CYCLE NUMBER= 1
SYSTEM FATAL MESSAGE 3001 MODULE = SHPCAS
THE INPUT DATA BLOCK NAMED YGBNDR AT POSITION 2 DOES NOT EXIST.
Avoidance: Specify TREF on the MAT1 entry or apply the following DMAP
alter:
compile bndshp nolist noref souin=mscsou
alter 'auxmid'(2)
dbview ugx=ug (where mtemp=*) $
alter 'partn','partn'
PARTN UGx,,AXMVEC/UGA,,,/1 $
alter 'equivx','equivx'
EQUIVX UGx/UGA/ALWAYS $

@
6947; 69; Miscellaneous - INPUTT4 - SFM 4276 EC 7777
When reading a double-precision datablock on an SGI R8000 computer using
INPUTT4, the run will fatal in the INPUTT4 module with the following
message in the .f06 file:
RECEIVED SIGNAL 0, ABORTING ...
BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR
FILE INDEX (FILEX) = 0
SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777
Avoidance: An avoidance is to have the datablock output in the neutral
format (OUTPUT4 in the ASCII format) and edit the datablock header to
change the data type (the 4th word in record 1) from 2 (double precision)
to 1 (single precision).
@
6949; 69; Documentation - RS/6000 IOI DBset ASSIGN Description format is i
ncomplete.
In Section 4.3 of the MSC/NASTRAN Installation And Operating Instructions,
the description of the DBset ASSIGN statement is incomplete. The
reference to the "action" options is not specified. Please see the
MSC/NASTRAN Quick Reference Guide for a complete description of the ASSIGN
statement.
@
6951; 70.5; Nonlinear Analysis - P2G - Wrong Answers, Null Load
If P2G is used to import loads onto the residual structure (with or
without superelements) in SOL 106, their contribution to the overall
loading is ignored.
Avoidance: Insert the following DMAP alter:
compile nlstatic
alter 'ssg1 '
MTRXIN CASESx2,MATPOOLS,EQEXINS,,/
 P2J,,/LUSETS/S,N,NOP2J///2 $
ADD Pgres,P2J/Pgres1/CP1/CP2 $
equivx pgres1/pgres/-1 $
endalter
or use a LOADSET callout above the SUBCASE level and a CLOAD within each
residual subcase to include the P2G contribution as shown:
LOADSET = 10
P2G = ExtLds
SUBCASE 1
LOAD = 20
CLOAD = 30
BEGIN BULK
LSEQ, 10,9999,20
CLOAD,30,1.,1.,9999

@
6952; ; P-Elements/Superelements - GP4 - UFM 2101, SFM 3062
The output coordinate specification is not allowed for the FEEDGE or
FEFACE entries when an FEEDGE or FEFACE is specified on the superelement
boundry. Do not use global coordinates on superelement boundries in
p-element substructures else the following fatal message will be
encountered:
*** USER FATAL MESSAGE 2101 (GP4)
GRID POINT 101000048 COMPONENT 1 ILLEGALLY DEFINED IN SETS UM
UB

@
6953; 69; Optimization/Design Sensitivity - DRESP1 - Poor Message, SFM 712
7
The following message is misleading in that it should be a USER FATAL
MESSAGE and it does not indicate that the error is on a DRESP1 entry.
SYSTEM FATAL MESSAGE 7127 (DOPR3B)
AN ELEMENT WHICH REFERENCES PTYPE 52 20 PID 34 IS NOT DEFINED IN THE
ANALYSIS MODEL.

@
6956; 69; Nonlinear Analysis - NOLINi - Wrong Answers
NOLINi loads produce wrong answers in nonlinear transient analysis if the
loads are a function of velocity and the TSTEPNL entry uses methods "AUTO"
or "TSTEP".
Avoidance: Use the ADAPT option on the TSTEPNL entry.
@
6957; 69; Documentation - NOLIN2
Remark 4 under the MSC/NASTRAN Version 68 Quick Reference Guide
description of the NOLIN2 entry should state "GK-CK" not "CK-CK".
@
6960; 69; Optimization/Design Sensitivity - SUBSEQ, SDR2 - SFM 4276, EC 30
01, SWM 2092
If zero values are used for the SUBSEQ command for a SUBCCOM subcase in a
solution 200 analysis the following messages will be encountered from the
SDR2 Module:
SYSTEM WARNING MESSAGE 2092 MODULE = SDR2
SDR2 FINDS A SYMMETRY SEQUENCE LENGTH = X
AND AN INSUFFICIENT NUMBER OF VECTORS AVAILABLE = Y WHILE ATTEMPTING TO
COMPUTE STRESSES AND FORCES.
ALL FURTHER STRESS AND FORCE COMPUTATION TERMINATED.
SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0
Avoidance: The avoidance is to not use 0.0 coefficients in a SUBSEQ
request when performing design sensitivity and optimization.
@
6963; 69; Documentation - SUBCOM, SUBSEQ - Unexpected Answers
The documentation for the SUBSEQ command is misleading. Unless it is read
carefully, the user might expect the first number to apply to the first
SUBCASE and so on. Actually, the way it is applied is that the last number
is applied to the most recent SUBCASE, the next to last number is applied
to the SUBCASE preceding it, and so on. If scaling factors are not
included for all preceding SUBCASEs, unexpected results may result.
Avoidance: Include scaling factors for all preceding SUBCASEs.
@
6964; 69; Utilities - Intel Windows/NT Only
If the utilities are executed on the Intel Windows/NT Version 68.2 release
of MSC/NASTRAN without any argument the command line input format is not
output to the console. This command line description is instead output the
FOR000 file. The command line description is also documented in section 6
of Installation and Operation Instructions.
@
6965; 69; Miscellaneous - Include File Not Found
MSC/NASTRAN on Windows NT can be run one of two ways. One is via command
line and second is by clicking on an icon. If nastran is run via the
second method and the input data deck contains an include file, the
include file will not be found.
Avoidance: 1. Run nastran using the command line.
2. Define the environmental variable in the MSCN4NT1.INI file
specifying the location of the
data deck. This also requires the same environmental symbol
be included while specifying the
include file. This method is described in the section 4.1 of
the Installation and Operation
instructions.
3. Add a complete path name to the include file.
@
6966; 69; Nonlinear Transient Heat Transfer - HBDY, FLUX, SET - Wrong Answ
ers
 If a set definition is used for a FLUX request, then wrong (zero) values
in the radiation column of the
heat flow into HBDY elements may be produced. This error causes all
results in the radiation column to be incorrect. This error does not
affect the temperature results.
Avoidance: A FLUX=all request avoids the problem.
@
6968; 69; Miscellaneous - Bad File Names - Run Never Starts
When using MSC/NASTRAN in either Windows95 or Windows/NT, the program will
fail if any embedded blanks exist in either the filenames or directory
names used.
Avoidance: Do not include any embedded blanks in file or directory names.
@
6969; 69; Installation - MSC/NASTRAN Will Not Install on Windows NT Versio
n 4.0
For Windows NT Version 4.0, which is the Windows 95 interface on top of
Windows NT 3.5, the installation of MSC/NASTRAN will not work. There is a
problem with the installation software (Installshield).
Avoidance: Do not upgrade your operating system to 4.0.
@
6971; 69; Loads - RFORCE - Wrong Answers
On one client model, RFORCE entries generated null loading vectors in
MSC/NASTRAN Version 68.2. Correct answers were produced in MSC/NASTRAN
Version 68.0 and V68.1. This error results in incorrect results.
Avoidance: We have been unable to duplicate this problem on any other
input files. For the client input deck, setting METHOD to 2 on the RFORCE
entry resulted in the correct loads being generated. Our recommendation is
whenever you have an RFORCE in your input file, include a SUBCASE which
applies the RFORCE only and verify the OLOAD resultant is correct for that
SUBCASE. If the OLOAD resultant for that SUBCASE is null set method to two
on the RFORCE entry and please send the deck to MSC.
@
6972; 69; Installation - MSCID - UFM 3060
The MSC Authorization program (an icon) has, on some configurations
(usually laptops with docking stations), shown a "0" for "MSC id". This is
not correct. If you get a "0" while using the authorization program during
installation or running from the icon, do not use this number to get an
authorization code from MSC.
To get the proper ID, run a MSC/NASTRAN job and the 3060 error message in
the .f06 file will contain the correct ID. Use the ID in the .f06 file
when calling MSC for an authorization code.
@
6976; 69; Acoustic Analysis - EMA, PSOLID - Poor Message
When executing a fluid/structure analysis, if the FCTN field is not set to
FLUID on the PSOLID data entry for a fluid element the execution will fail
with the message:
SYSTEM FATAL ERROR 4276.BLDPK ERROR CODE 211
 Avoidance: Other than correcting the model, there is no avoidance.
@
6980; 69; Heat Transfer Analysis - QVECT - Wrong Answers for HBDY Output
The QVECT load defines thermal flux from a distant source into a face of a
CHBDYi boundary element. If a coordinate system(CE) for the thermal vector
flux is rotated with respect to the basic coordinate system, the data
recovery for the APPLIED LOAD under the HEAT FLOW INTO HBDY ELEMENTS is
not correct.
However, the actual temperature results are correct. The actual loading
can be seen using the OLOAD=all output. There is no error if the
coordinate system is merely translated, not rotated, with respect to the
basic coordinate system.
Avoidance: Ignore the incorrect output.
@
6981; 69; Nonlinear Transient Heat Transfer - RESTART, PARAM,SLOOPID - Wro
ng Answers
If PARAM,SLOOPID is specified for a nonlinear transient heat transfer (SOL
159) analysis restarting from a steady state thermal analysis (SOL 153),
the initial temperature from the SOL 153 run is zero.
Avoidance: Use the following DMAP ALTER.(V68.2 ALTER)
$ DMAP alter to repair SLOOPID restart from sol 153
$ MSC/NASTRAN V68.2
$
compile nltcsh,souin=mscsou,nolist,noref
$
$ bypass the restart deletion of estnl in case that a
$ TEMPBC entry is added to the bulk with the sol 159 input
$
alter 39 $
type db zuzr01 $
dbview xestnl = estnl (where nloop=sloopid) $
copy xestnl/zuzr01 $
alter 281,281 $
alter 300,300 $
copy zuzr01/estnl $
$
$ save the sol 153 temperatures to database
$
alter 310 $ V68.2
equivx ulx/ulnt/noeset $
cend
@
6982; ; Heat Transfer Analysis - PARAM,LGDISP - Solution Does Not Conver
ge
PARAM,LGDISP is intended only for the nonlinear structural analysis.
However, if it is used in heat transfer analysis (SOL 153 or 159) the
solution will not converge.
Avoidance: Delete the PARAM,LGDISP from any heat transfer analysis.
@
6984; 69; DMAP - MALTER - UFM 806
 If the input file for a restart run contains a DMAP alter using MALTER
commands and has no "compile" commands, the run will fail with User Fatal
Message 806 stating that the MALTER statements were not found in the
source files.
Avoidance: This error may be avoided by either:
1) Inserting ACQUIRE NDDL in the FMS section of the restart deck.
or
2) Inserting a COMPILE command for a subdmap in the deck.
@
6986; 69; Dynamics - DPD, SEQP - Poor Efficiency on Convex
Models which have been resequenced cause the DPD module to run up to 5
times slower than models which have not been resequenced on MSC/NASTRAN
Version 68.2 on the Convex. This also affects reading large numbers of
FORCE bulk data entries.
Avoidance: The combination of the sparse solver and the seqp module is
almost always more than the sparse solver without resequencing. Thus,
sequencing should be simply turned off. Remove resequencing by:
1. Remove any PARAM, NEWSEQ entry and replace it with PARAM,NEWSEQ,-1.
2. Remove any SEQGP entries which may be in the bulk data file.
This avoidance will not help the reading of the FORCE bulk data entries.
@
6990; 69; Acoustic Analysis - PARAM,G - Wrong Answers
If the structural damping (either PARAM,G or "ge" damping on the MATi
entry) is the only change in a restart to a fluid-structure acoustic
analysis, the change is ignored and the answers are incorrect, i.e., they
remain the same as before the change.
Avoidance: Execute the analysis as a cold start.
@
6991; 69; Heat Transfer - CHBDYG, CTRIAX6 - Wrong Answers
Currently there is a design limitation using CHBDYG,REV to represent the
higher order CTRIAX6 element because only two grid points are allowed on
the REV option. Therefore, the weighting on the mid-side node is
misrepresented if the two node REV option is used.
Avoidance: Use CHBDYE to represent the CTRIAX6 element with midside nodes.
Please note that without the midside node, the CHBDYG,REV can be used.
Please note that the REV option G1,G2,G3
Where G1 and G2 are the edge grids and the G3 is the midpoint grid
@
6992; 69; Heat Transfer Analysis - OLOAD, CTRIAX6 - Wrong Answers for LOAD
VECTOR
Applying a heat flux to the edge of the midside node of a CTRIAX6 element
results in wrong answers for the LOAD VECTOR output at the midpoint. This
error is only in the data recovery. The temperature results are not
affected.

Avoidance: Ignore the OLOAD output for the midside node of the CTRIAX6
element
@
6994; 2001; Miscellaneous - ASET1 - UFM 2049
If an ASET1 entry references a non-existing grid as an end point for the
"thru" option, then the following fatal message is issued: "USER FATAL
MESSAGE 2049, UNDEFINED GRID POINT ** HAS A COORDINATE REFERENCED ON A
ASET CARD." Besides the fact that this message should not be encountered
at all, this message is poor for the following reasons:
(1) The missing grid point specified in the message will always be the
first grid point in the "thru" range, not the actual missing grid point.
(2) The word "coordinate" is incompatible with other documentation which
uses the word "component".
(3) It uses the obsolete word "CARD"
(4) It repeats the message twice.
(5) If the user explicitly specified the grids on the ASET1 entry rather
than using the "thru" option, they would have gotten User Warning Message
2049A instead of a fatal message.
Avoidance: Don't specify "thru" ranges on the ASET1 entry that have a
missing end point.
@
6996; ; Optimization/Design Sensitivity - Q4TAPER - UWM 5491, UFM 4297
If an auxiliary model is used to generate the shape basis vector for a
shape optimization analysis and in the subsequent dblocate optimization
run an element exceeds the maximum allowable taper value of .5 then EMG
issues UWM 5491 that may identify elements with bad geometry that do not
exist in the model. The run may also issue USER FATAL MESSAGE 4297
indicating that an element has an interior angle greater than 180 degrees.
An example of the warning message follows:
USER WARNING MESSAGE 5491 (EQD4D)
BAD GEOMETRY FOR QUAD4 ELEMENT ID = 2600 THE TAPER OF 0.582639 EXCEEDS THE
MAXIMUM VALUE OF 0.500000
USER ACTION: THE TOLERANCE MAY BE ADJUSTED WITH Q4TAPER KEYWORD ON THE
NASTRAN STATEMENT.

Avoidance: EMG is reporting the internal identification numbers for the

perturbed elements. These element id's are incremented based on the base
line finite elements. The model should be modified to a finer mesh or the
Q4TAPER keyword should be set on the NASTRAN statement to a larger value.

@
6997; 70.5; Acoustic Analysis - Superelements, Reference Manual - Wrong answ
er
The Version 68 MSC/NASTRAN Reference Manual Section 15.4.2 states that the
"grid points at the fluid-structure interface may be assigned to the
residual structure only". This statement is not restrictive enough. The
MSC/NASTRAN Common Questions and Answers Guide states the restriction
correctly; "for coupling to occur, all fluid elements that are in contact
with the structural elements must be in the residual structure, and all
structural elements that are in contact with the fluid elements must also
belong to the residual structure.
@
6998; 69; Heat Transfer Analysis - QBDY2 - Wrong Answers
 QBDY2 (Boundary Heat Flux Loading) loading on CQUAD8, CTRIA6, CHEXA20,
CPENTA15, CTETRA10 elements results in incorrect data recovery for the
element. The temperature vectors are correct. The heat flow into the HBDY
elements is wrong.
Avoidance: Ignore the Heat Flow into HBDY element for the Applied Heat
load. All other output is unaffected by this error. If the user wants to
see the correct applied load under the Heat Flow into the HBDY element,
they must use the QBDY3 bulk data entry to apply the loading.
@
7001; 70.5; Nonlinear Transient Heat Transfer - QRG, PARAM,AUTOSPC - UFM, Fa
iled to Converge
PARAM,AUTOSPC,YES will automatically constrain floating degrees of
freedom. The floating grid is used in heat transfer to model ambient
temperature as a function of time .Heat transfer analysis uses these
floating grids to apply temperature as a function of time through the
TEMPBC,TRAN. Use of this PARAM,AUTOSPC,YES in nonlinear heat transfer
results in a run that fails to converge. This limitation is documented in
the Version 68 Release notes but needs to be documented in the Quick
Reference Guide.
@
7002; ; Sequencer - GP2 - UFM 2007
If an attempt is made to use the punched grid and element output of a
p-element run in the bulk data the following error will occur and the run
will fail in the SEQP module or GP2 modules:
*** USER FATAL MESSAGE 2007 MODULE = GP2
ELEMENT ******** REFERENCES UNDEFINED GRID POINT 111001009

USER FATAL MESSAGE 2007, ELEMENT******** REFERENCES UNDEFINED GRID

POINT********.
This message occurs since grid ids larger than 100000000 are not valid on
the grid card and the vu-grids have ids greater than 100000000. .
Avoidance: Following system cells should be set in a p-element run to
generate
smaller vu-grid ids and vu-element ids.
System cell values are beginning and maximum for vu-grids and
vu-elements

nastran system(175)=9000, $ GRMAX = 100000000

system(178)=100, $ GNSTRT = 101000000
system(179)=150000, $ GNMAX = 200000000
system(180)=150001, $ VGSTRT = 201001001
system(181)=2000000, $ VGMAX = 300000000
system(182)=150, $ VESTRT = 1001001
system(183)=2000000 $ VEMAX = 2000000

@
7006; 69; Documentation - DRESP1, DVPREL2 - Incorrect RTYPE, Missing Docum
entation
The Version 68 MSC/NASTRAN Quick Reference Guide description of the DRESP1
 entry indicates the RTYPE for SPCFORCE as FRSPC and TSPC, it should be
FRSPCF and TSPCF. Remark 3 under the description for the DVPREL1 entry
for PTYPE="PBEND" should also be included for the DVPREL2 entry.
@
7007; 69; Installation - NASTRAN FOR WINDOWS - FILE TRANSFER ERROR NO. -3
If long file names have been disabled by the following registry entry the
NASTRAN FOR WINDOWS VERSION 68.2 Installation will fail with the message
"FILE TRANSFER ERROR NO. -3".
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\FileSystem\Win31FileS
stem
Avoidance: Reset the registry entry and restart your computer.
@
7010; 2001; Axisymmetric Fluid Elements - CAXIF4, VECPLOT - UFM 5423
Models with CAXIF4 axisymmetric fluid elements fail in the VECPLOT module
with the message:
USER FATAL MESSAGE 5423 - ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES.
Avoidance:
Apply the following alter:
compile selg nolist noref
alter 63,63
endalter
@
7011; 69; DMAP - VECPLOT - Excessive CPU
The VECPLOT module can consume a significant amount of CPU when processing
RESULTANTS and MAXIMUMS. This information may be of no interest to the
analyst yet the only way to turn it off is thru DMAP. MSC/NASTRAN Version
69 has been enhanced to give the analyst options to avoid these
calculations. PARAM,PRTMAXIM (default=no) and PARAM,PRTRESLT
(default=yes) have been added to control the calculation of MAXIMUMS and
RESULTANTS respectively.

@
7012; 70.5; Executive - DIAG 14,64 - UFM 69
The presence of both DIAG 14 and DIAG 64 in the same run will cause the
following message "USER FATAL MESSAGE 69, SUBDMAP STATEMENT PARAMETER APP
NOT DEFINED". This error is only known to occur in the structured solution
sequences (SOLs > 101)
Avoidance: Remove Diag 14 and/or Diag 64.
@
7013; ; Miscellaneous - COMIFP, PUNCH - Produces Incorrect Bulk Data Ent
ries
The comifp module has been modified for Version 69 to support the part
superelement concept. It can take ifp-like files from the sep2 module and
punch bulk data files which may then be used as part files. At present,
the RBE2 and RBE3 elements are punched incorrectly.
Avoidance: An avoidance is to separate the elements not properly
 processed by hand.

@
7014; 69; Heat Transfer Analysis - MAT4, CONVM, EMG - FATAL ERROR, Bad Dia
gnostic Message
If the user fails to define a dynamic viscosity in field 7 of the MAT4
entry for a forced convection analysis the job fails in the EMG module
with no indication of the cause of the error.
Avoidance: Make sure the dynamic viscosity is input on the MAT4 card if
the problem has CONVM bulk data entries.
@
7015; 69; Material Nonlinear Analysis - TEMP(LOAD), CLOAD - Wrong Answers
In Version 68.2 and SOL 106 only, thermal loads applied to upstream
superelements results in incorrect stresses and forces.
Avoidance: Use SOL 66 or apply the following alter to SOL 106:
compile super3 noref nolist
alter 'IF ( NOCASRS1>-1 AND SEID=0 )',''
IF ( (NOCASRS1>-1 AND SEID=0) or app='nlst' ) THEN $

@
7021; 69; Optimization/Design Sensitivity - DSCM2 - Wrong Ordering
In design sensitivity analysis, there is a table printed just prior to the
DSCM2 (design sensitivities matrix) which specifies the ordering of the
columns in the matrix. If some of the design responses specified on the
DRESP1 entry are the eigenvalues, then the table claims that the
eigenvalue responses are in ascending order by DRESP1 entry ID. However,
the matrix is actually in ascending order by mode number.
Avoidance: The only avoidance is to be aware that the output is in
ascending order by mode number.
@
7024; ; IDENT option on ACMODL entry
If INTER=IDENT on the ACMODL data entry is used in acoustic analysis,
then only two fluid grids per structural grids at the fluid/structure
inter face. This precludes modeling structures such as t-sections
stiffeners in tanks or t-sections immersed in fluid.
@
7040; 69; Documentation - DRESP1 - UFM 316
The documentation for the DRESP1 entry in the MSC/NASTRAN Version 68 Quick
Reference Guide states that the SPC force component is chosen with the
RTYPE=FRSPC or RTYPE=TSPC. However, if either of these RTYPEs are used,
the following fatal message is issued, "USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA CARD DRESP1 331 SORTED CARD COUNT = ***". The
documentation should say the the SPC force component is chosen with
RTYPE=FRSPCF or RTYPE=TSPCF.
Avoidance: Use RTYPE=FRSPCF or RTYPE=TSPCF not FRSPC or TSPC when the SPC
force component is desired as a response type.
@
7041; 69; Nonlinear Analysis - EMG - Poor Performance
NO EXTERNAL ERROR REPORT
When there are no linear elements in a model, the EMG module is
unnecessarily executed.
This appears to be a minor performance improvement, I see no need to
publish to clients.
@
7044; 69.1; Nonlinear Transient Heat Transfer - QVECT, DLOAD - Wrong Answer
s
Scaling by other than 1. of a vector flux QVECT load by a DLOAD bulk data
entry produces wrong results in the "APPLIED-LOAD" column of the "Heat
Flow into HBDY Elements" table. The actual load vector generated is
correct and therefore the temperatures are correct.
Avoidance: Use only scale factors 1 on the DLOAD entry.
@
7045; 69; Case Control - SUBCASE - Poor Message
The message " *** USER FATAL MESSAGE 609 (IFP1D) SUBCASE DELIMITER CARDS
MUST HAVE A UNIQUE IDENTIFYING INTEGER " which does not accurately
describe the input error is issued for the following situation:
1. When the subcase id exceeds 9999999.
2. In solution 200 when the subcase id of the preceding subcase is higher
than the id of subsequent subcase.
@
7047; 69; Aeroelastic Analysis - Unwanted Output in the Flutter Summary
In the pk flutter solution, all combinations of velocity, Mach number and
density are used. In certain situations (e.g., if analyses are only
required at 50 velocity-density combinations) it would be preferable to
only do ordered combinations.
Avoidance: The current avoidance is to put each required combination in
its own subcase. For the
example given above, this would require 50 subcases, but it would reduce
the number of flutter
analyses from 2500 to 50.
@
7048; 69; Submittal Script - node - Error With node Keyword
When the "node" keyword is used along with the "user" keyword ( i.e.
nastran ..... node=pelee user=ssp) then
job will not get submitted to remote node and the following message will
be displayed on the console:
nast68: (W) node=pelee specified, setting app=yes bat=no scr=yes
nast68: (E) node=pelee: no execute permission on remote node.
Avoidance: The following line in the nastran shell script should be
changed. The shell script is
$MSC_INSTALLDIR/bin/nastran.
 rx="$remcmd $tuser $node"
should be changed to
rx="$remcmd $node $tuser"

@
7050; 69; Multipoint Constraint - MPCFORCE Output - Labeled as SPCF in .pc
h file
MPCFORCE output when printed to a punch file, is mis-labeled as SPCF. The
data is correct however. Confusion may arise when both SPCFORCE and
MPCFORCE are printed to the punch file. The user must use care to read the
correct results. The error is considered fixed in Version 69; however,
setting of SYSTEM(210)=1 is required.
Avoidance: Print and punch the data so that the output can be compared
and correlated.

@
7053; 69; Documentation - DVPREL1 - UFM 2042 Misleading
The DVPREL1 entry is used to define a relationship between an analysis
model property and a design variable. However, not all analysis model
properties are valid choices. For example it is not valid to link the MID
field on a property entry (which is an integer) to a design variable
(which is always real). If this is done the MATi entry in the analysis
model is overwritten with a nonexistent MATi and then the following
misleading message is issued, "USER FATAL MESSAGE 2042, MISSING MATERIAL
TABLE******** FOR ELEMENT xx". Invalid choices for the analysis property
chosen on the DVPREL1 should be screened out prior to the start of
optimization.
Avoidance: Choose only valid analysis properties on the DVPREL1.
@
7055; 69; Documentation - Composites, Buckling, PSHELL, PCOMP - Missing Do
cumentation
The following remark should be added to the PSHELL bulk data entry
description in the MSC/NASTRAN Quick Reference Guide:
The use of MID4 will cause incorrect results in buckling analysis and
differential stiffness.

The following remark should be added to the PCOMP bulk data entry
documentation in the MSC/NASTRAN Quick Reference Guide:
An unsymmetric layup or the use of ZO to create an unsymmetric layup will
cause incorrect results in buckling analysis and differential stiffness.
@
7059; 2001; Miscellaneous - DMIG - Psuedo Infinite Loop, Device Full
If characters are input in the NCOL field of a DMIG header entry and the
matrix is rectangular (IFO=9), then the MTRXIN module may be caught in a
pseudo infinite loop which may be finally stopped by a "device full"
message. The MTRXIN module interprets the characters in NCOL as an
 integer and this is normally a huge number. The IFP module should stop
the execution with "User Fatal Message 327" in case that the input for
NCOL is not an integer.
@
7065; 69; P-Elements - Wrong Adaptivity Analysis - Poor Answers
If a data deck contains different TYPEs of P-elements (e.g. a HEXA and a
PENTA), the adaptivity error analysis likely to be in-accurate leading to
poor answers.
Avoidance: There is no avoidance.
@
7073; 69; Heat Transfer - PARAM,POST,-2 - SDRC Cannot Read TEMP Vector
SDRC is not able to read the temperature vector in the output2 file
produced with PARAM,POST,-2.
Avoidance: Use EQUIVX instead.

@
7076; ; Timing Estimate - DECOMP - UIM 4157
The CPU time estimates issued in USER INFORMATION MESSAGE 4157 for sparse
decomposition can be overstated by a factor of 7 to 15 times on some
machines. This is due to the fact that the timing estimates do not take
into account the high rank performance update enhancement which was
introduced beginning in Version 68.1. By overstating the CPU time
estimates the job may fail with a message that it does not have enough
time to complete Decomposition.
Avoidance: Include an extra large value for time on the TIME executive
control entry or remove the TIME entry.
@
7085; 70.5.2; Cyclic Symmetric - Dihedral, LOADCYH, DSYM - UFM 4342
A cyclic symmetry analysis using dihedral symmetry and centrifugal loads
may issue the following message:
USER FATAL MESSAGE 4342,
THE LOAD BULK DATA CARD GIVEN BELOW IS INCONSISTENT WITH SPECIFIED
ANALYSIS HARMONICS OR SEGMENTS OR SYMMETRY TYPE.
LOADCYH 20 0CSTAR 2000004
The message occurs when a SS (Symmetric=Symmetric) boundary condition is
specified on the DSYM case control command and the LOADCYH bulk data
requests an 'RLOAD' for the HTYPE field. In this condition the code
generates too many load vectors.
Avoidance: One avoidance is to change the HTYPE field to 'C' on the
LOADCYH bulk data entry. A better avoidance is to eliminate the DSYM
request in the case control.
@
7088; 70; Documentation - PSHELL, MID4, ZOFFS - MID4 and ZOFF may be combi
ned
Remark 6 under the description of the PSHELL entry in the Version 68 Quick
Reference Guide states that "Only one of the options MID4 or ZOFFS may be
 used". This is a guideline to prevent users from offsetting the element
twice. Depending on the intent of the user these affects may be combined
and correct answers achieved.
@
7093; 70.5; Complex Eigenvalues - CLAN - Wrong Answers
For one model, the complex lanczos eigenvalue method in solution 107
produced erroneous eigenvalues on the CRAY computer. The same model
produced consistent real eigenvalues across all machines so there does not
appear to be anything wrong with the model. MSC/NASTRAN Version 70.5
introduces the new Block Complex Lanczos method. This method correctly
solves this problem on CRAY and all other machines. This method must be
selected in Version 70.5 by setting system(108)=4 on the NASTRAN entry or
rc file.
Avoidance: Until the error is better understood, use SOL 110, modal
complex eigenvalues, with the Hessenberg Method.
@
7100; 69; Static Analysis - SPC - UFM 5423
A DMAP error exists in multiple boundary condition with non-zero SPC in
SOL 101. The error is triggered
when the same boundary conditions are not grouped together, i.e., the same
SPC is selected again following
a SUBCASE with a different SPC. The error is detected by the SSG2 module
with a UFM 5423 "attempt to multiply incompatible matrices." This error
is related to error 5977 which is detected when SPCD is used instead of
non-zero SPC.
Avoidance:
1. Shuffle the Subcases so that the same boundary conditions are grouped
together (to appear consecutive Subcases).
2. Use SOL 106 with unsorted SUBCASE set-up and run it as a linear problem
without any nonlinearity.

@
7108; 69; Acoustic Analysis - SDR1, MPYAD - Insufficient Time
Acoustic Analysis in MSC/NASTRAN does not support the matrix method of
data recovery. Thus, the data recovery performed by the modal
displacement method for small sets of data is very inefficient. This
deficiency is complicated by poor timing estimates in MPYAD. MPYAD, which
is called by SDR1 during the data recovery, may yield a time estimate that
overflows the machine. Version 69 will support data recovery in the DDRMM
module, i.e., the matrix method.
Avoidance: There is no simple avoidance. Perform modal reduction on the
fluid and limit the number of frequencies in the analysis. The following
alter can be applied to recover pressure only by the matrix method.
compile semfreq souin=mscsou noref nolist
alter 74
lamx ,,lamas/mlamas/-1 $
lamx ,,lamaf/mlamaf/-1 $
matgen ,/cp0/6/5/2/2/1 $
matgen ,/cp1/6/3/0/1/1/1 $
$
 partn mlamas,cp0,/,,emats0,/1 $
merge ,,,emats0,,cp1,/emats1/1 $
trnsp emats1/emats $
$
partn mlamaf,cp0,/,,ematf0,/1 $
merge ,,,ematf0,,cp1,/ematf1/1 $
trnsp ematf1/ematf $
$
append emats,ematf/emat $
lamx emat,/lamax $
ofp lamax/ $
ALTER 112,112
IF (DDRMM>=0) APP1='MMREIG'
alter 117
equivx lamax/lama/-1 $
$
compile super3 souin=mscsou
alter 24,24
IF ( APP1='MMREIG' ) THEN $
compile sedrcvr souin=mscsou noref
alter 61
if ( app1='mmreig' ) then $
equivx oes1/qes1/-1 $
equivx oef1/qef1/-1 $
delete /ougv1,oqg1,oes1,oef1, $
DDRMM CASEdr,UH1,OL1,qugv1,qQG1,qES1,qEF1,XYCDBdr/
OUGV1,OQG1,OES1,OEF1, $
EQUIVX OQG1/QQG1/ALWAYS $
EQUIVX OUGV1/QUGV1/ALWAYS $
endif $
alter 159,159
IF ( sort2 ) THEN $ SORT2
@
7109; 70.5; DMAP - MATMOD OPTION 17 - Wrong Answers
MATMOD option 17 generates a g-size partitioning vector from a
user-defined set of grid and/or scalar points. If the 3rd parameter
(SETFLG) is set to 1, it is supposed to look for the PARTN case control
command and put a value of 1.0 in the vector for grid points listed in the
set defined by the PARTN case control command. If there is no PARTN case
control command, it is supposed to use the sets defined by the 2nd
parameter (UBIT). However, the above operation returned a vector with all
values set to 1.0 instead. Depending on subsequent operations, this can
either lead to fatal errors or wrong answers.
Avoidance: Include the appropriate PARTN case control command.
@
7110; 69; SSSALTERS - evds.v682 - Fails to Converge
When using the eigenvector sensitivity dmap alter (evds.v682) delivered
with the sssalter library in Version 68.2, and the model does not contain
any elements which references a mass property entry, then the job may fail
to converge.
Avoidance: Replace the following dmap statement in evds.v682 subdmap fox:
TYPE PARM,,I,N,NULLDM $
 with

TYPE PARM,,I,N,NULLDM,NZWD $

@
7118; 69; Documentation - ADUMi, PDUMi - Undocumented Limits, User Modifia
ble NASTRAN
The number of additional fields "NP" on the PDUMi entry is limited to 24
words. This limit is not documented in the description of th NP field of
the ADUMi bulk data entry description.
@
7127; 69; Optimization/Design Sensitivity - PARAM,SNORM - UFM 2025 in modu
le TA1
If in SOL 200, parameter SNORM is greater than or equal to 0.0 and if
a coordinate system is specified in the MCID field of the connectivity
entry in the run then the following fatal message will be issued.
*** USER FATAL MESSAGE 2025 MODULE = TA1
UNDEFINED COORDINATE SYSTEM 3
This error might also occur in SOL 101 and 103 if p-element adaptive
analysis is performed.
Avoidance: Two avoidances exists:
1. Remove the SNORM parameter.
2. Add the following alter.
compile update souin=mscsou nolist noref
alter 51,51 $
TA1 ,,ECT,EPTX,BGPDTSCR,SIL,,CSTMSCR,/
ESTJNK,,,GPECT0,/
LUSETS/-1/S,N,NOSIMP/1/S,N,NOGENL/0 $

@
7136; 69; Restarts - MAT2, IFPL, RESTART - DEC Alpha Only Aborts on Restar
t
There is an error in the nddl for the MCSID field of the MAT2 entry. The
nddl types the field as real when it should be an integer. If this field
is changed in a RESTART the execution will fail. The message
"time: command terminated abnormally" is issued in the log file.

Avoidance: The run should be executed as a cold start.

@
7137; 69; Postprocessing - PARAM,POST,-2, SOL 112 - Extra Output
In SOL 112 with PARAM, POST, -2, the geometry tables are written twice to
the OUTPUT2 file. However, this does not seem to negatively affect the
IDEAS translator.
 Avoidance: There is no avoidance.
@
7138; 70; Documentation - Common Questions and Answers - Incorrect Problem
Descriptions
The problem description on page 103 of the Common Questions and Answers
,2nd Edition, does not match figure 40 on the same page. The acceleration
should be negative, as is reflected in fig. 40 and the 3 in field 5 of the
RLOAD2 entry should be moved to field 6.
The title for figure 39 on page 101 of the Common Questions and Answers
book refers to point loads when in fact the example discusses pressure
loads (PLOAD4s).
@
7139; 69; Executive System - COMPILE - Segmentation or Access Violation
NO EXTERNAL ERROR REPORT IS NECESSARY
MAR VERIFIES THIS EVERY TIME HE REBUILDS THE DELIVERY DATABASES.
An attempt to compile the NDDL and more than 75 subDMAPs in a single run
causes a segmentation or access violation. The limit in Version 69 has
been expanded to allow for at least 100 subdmaps.
Avoidance: Split the job into two runs.
@
7141; 70.7.2; Nonlinear Transient Analysis - CBEAM - UFM 4551 (NL2CON)
It has been found that in rare cases when then the orientation vector for
a beam element is exactly normal to the axis of the beam, the Cray may
exhibit a round-off error that leads to convergence difficulties and
ultimately USER FATAL MESSAGE 4551, STOPPED PROBLEM DUE TO FIVE
CONSECUTIVE DIVERGING TIME STEPS.

@
7142; 69; Optimization/Design Sensitivity - DOPR3 - SFM 4276
If the user provides constraints for multiple DRESP1 entries in a single
flutter subcase, then SYSTEM FATAL MESSAGE 4276 GREAD ERROR CODE
3001 will occur in the DOPR3 module.
Avoidance: If it is not possible to define all the flutter responses on a
single DRESP1 entry, multiple
flutter subcases are required, each with a single referenced DRESP1 entry.
@
7144; ; Nonlinear Transient Analysis - OTIME - Missing Output
In nonlinear transient analysis (SOLs 129 and 159), the solution vectors
(SDISP, SVELO, and SACCEL) should
be output for each subcase independently. However, SDISP, SVEL, or SACCEL
does not appear in the output unless the request is specified above the
subcase.
Avoidance: Specify the output request for SDISP, SVELO, or SACCEL above
the subcase level.
@
 7147; 69; Utilities - OPTCONV - Fails On Windows NT
If the OPTCONV program delivered with the Windows NT version of NASTRAN is
used to convert an old optimization deck to the current format, the
following message is obtained in a pop-up window: "The instruction at
004147cc referenced memory at 0d4212b9 The memory could not be written
Click on OK to terminate the application". Therefore, it is not possible
to convert any old optimization decks on Windows NT.
Avoidance: Convert the decks by hand or have them converted on a
different machine variation.
@
7156; 69; P-Elements - PARAM,USETPRT,0 - Incorrect Output
PARAM,USETPRT,0 produces incorrect output for P-Elements.
Avoidance: Use PARAM,USETPRT,1 instead.

@
7163; 69; Nonlinear Static Analysis - ASETi, OMITi, AUTOSPCR - UWM 5225, W
rong Answers
In nonlinear static analysis, any nonlinear degree of freedom in the o-set
(via ASETi or OMITi entries)
may result in a singularity in the o-set linear stiffness matrix (UWM
5225). Also, if PARAM,AUTOSPCR is used, all the nonlinear degrees of
freedom will be constrained and the resulting solution will be incorrect.

Avoidance: Include all the active nonlinear degrees of freedom in the

A-set.
@
7165; 69; Optimization/Design Sensitivities - CTRIAX6 - Wrong Answers
When CTRIAX6 elements without mid-side nodes are used in an optimization
job which includes specification
of temperature loads, the thermal displacement/stress sensitivities may
show wrong signs which may result in
strange convergence behaviors.
Avoidance: Use CTRIAX6 elements with midside nodes.
@
7167; 69; Nonlinear Heat Transfer - UIM 6194, UIM 6193, UFM 4551
Nonlinear Heat Transfer jobs with more than one subcase may fail to
converge and issue UIM 6194, UIM 6193 and USER FATAL MESSAGE 4551. This
error was experienced in a job where the second subcase was the same as
the first. The job should have converged immediately in the second
subcase since there was no change.
USER INFORMATION MESSAGE 6194,
STOPPED ITERATIONS DUE TO REACHING MAXIMUM ITERATION LIMIT WITHOUT
CONVERGENCE ***
USER INFORMATION MESSAGE 6193,
MAXIMUM NUMBER OF BISECTIONS OR MINIMUM LOAD STEP HAS BEEN REACHED.
USER FATAL MESSAGE 4551,
 STOPPED PROBLEM DUE TO FAILED CONVERGENCE
@
7170; 69; Heat Transfer - CHBDYP, LINE with midpoint node option
Currently there is a design limitation using CHBDYP,LINE to represent
thhigher order CQUAD8 or CTRIA6 for the edges with midpoint option.
Therefore, the weighting on the mid-side node is misrepresented if the two
node LINE option is used.
Avoidance: Use CHBDYE to represent the CQUAD8 or CTRIA3 element with
midsidenodes. Please note that without the midside node, the CHBDYP,LINE
can be used.
@
7171; 2001; Documentation - Reference Manual - Large Field description
There is an error in the Reference Manual documentation for the
description of the Small/Large/Free field format of the bulk data entries
as follows:
pp 3.7.1-2 Bottom of the page, Large Field example. The first field of
the parent card should be TYPE* not TYPE. The continuation of the parent
card should be QED123 not QED124.
pp 3.7.1-3 Top of the page, Large field with large field continuation
example. This is shown as all small fields. There should only be 4
(double wide) fields between the first and last fields.
@
7173; 69; P-elements - CHEXA - UFM 9050
CHEX p-Elements can have singularities in the stiffness for the default
integration options. This error only occurs in a model with a single layer
of elements. In particular the default reduced integration order with
rectangular elements with p values of [2,2,3] in the three directions may
be singular. This error may lead to the following fatal message:
USER FATAL MESSAGE 9050 (SEKRRS) - RUN TERMINATED DUE TO EXCESSIVE PIVOT
RATIOS
IN MATRIX KLL. USER PARAMETER BAILOUT MAY BE USED TO CONTINUE THE RUN.
Avoidance: Specify a different p value.

@
7175; 70.5; Documentation - CALL DBFETCH - DMAP Module Dictionary
The description of the DBFETCH command in the DMAP Module Dictionary
should enhanced to state "if you are using DBFETCH in a Sol 100 (USER
DMAP) to get data blocks from a previous run, include a RESTART FMS
statement to specify the version number on which the data blocks were
written".
@
7176; 69; Direct Frequency Response - CHACAB - Wrong Answers
The damping terms for the acoustic absorber element, CHACAB, are "lost" in
solution 108 resulting in null damping from this element.
 Avoidance: Place equivalent CDAMP or CVISC elements across the structural
boundary points.
@
7179; 69; Documentation - QRG - Inconsistent E,G, NU with MAT1
Table 1 under the MAT1 description in the Version 68 MSC/NASTRAN Quick
Reference Guide indicates that "E" and "NU" are used to define the
properties for "Membrane and Bending" of the CQUADi, CTRIAi and CCONEAX
elements. This should be clarified to state "should be clarified to
"Membrane, Bending and Inplane Shear".
@
7182; 69; Post Processing - SOL 103 and 105, OUTPUT2 - Missing Output
When multiple subcases are used in SOL 103 or 105 and PARAM,POST,-1/-2 is
chosen, the output2 file is written over with each subcase and only the
last subcase remains in the output2 file for post processing.
Avoidance: Request different output2 units in each subcase using
param,ounit2 or the following DMAP alters may be used for some post
processors (PATRAN3 v5)
For SOL 103:

COMPILE SUPER3 SOUIN=MSCSOU NOLIST NOREF

$
ALTER 'IF ( POST=-1 AND OUNITX\<\>OUN','IF ( ( ((DROPT=0 OR DROPT=4)'
PARAML CASECC//'DTI'/1/1//S,N,SUBC1 $
PARAML CASEDR//'DTI'/1/1//S,N,SUBCNO $
IF ( POST=-1 AND SUBCNO=SUBC1 AND OUNITX<>OUNIT2 AND NOT(PEXIST) AND
NOT((APP='NLST' OR APP='TRANRESP') AND PATVER>=3.) AND
(((DROPT=0 OR DROPT=4) AND APP1<>'BKL1') OR APP='BKL0')
) OTAPE2=-1 $
IF ( ( ((DROPT=0 OR DROPT=4) AND APP1<>'BKL1') OR APP='BKL0') AND
(OGEOM='YES' OR OGEOM=' ') AND SUBCNO=SUBC1 AND
((POST=-1 AND PATVER>=3. AND
NOT(APP='NLST' OR APP='TRANRESP')) OR
POST=-2 OR POST=-4) ) THEN $
$
ALTER 'IF ( ((DROPT=0 OR APP1=','IF ( ((DROPT=0 OR APP1='
PARAML CASECC//'TRAILER'/1/S,N,NSUBC $
PARAML CASECC//'DTI'/NSUBC/1//S,N,LASTNO $
IF ( (SUBCNO=LASTNO OR APP<>'REIG ') AND
(((DROPT=0 OR APP1='BKL1') OR
(PEXIST AND ADPTEXIT)) AND POST<0 AND POST>-4 AND
NOT(POST=-1 AND PATVER>=3.0 AND SCRSPEC>-1))
) OUTPUT2 //-9/OUNIT2//OMAXR $
ENDALTER $
For SOL 105:
compile super3
alter '-9'(2),''
compile sebuckl
alter 'dbdict1'
type parm,nddl,i,y,ounit2 $
type parm,,i,y,omaxr $
OUTPUT2 //-9/OUNIT2//OMAXR $
@
7184; 69; Postprocessing - PARAM,OUGCORD,BASIC, PATVER,3 - Incorrect Defor
med Shape
PATRAN 3 will display an incorrect deformed shape when PARAM OUGCORD is
set to BASIC and the gridpoint output coordinate systems (field 7, CD) is
anything other than the basic (coordinate system ID 0). SubDMAP SEDRCVR
transforms the displacement vector into the basic coordinate system based
on the value of the parameter OUGCORD and PATVER. In Version 67.5 the DMAP
IF test includes the value of the parameter PATVER which prevents
transformation if PATVER=3.0 (default for PATVER is 3.0).
66 ELSE IF ( NOT(PATVER>=3.) AND OUGCORD='BASIC' ) THEN $
67 VECPLOT UGVS,BGPDP,EQEXINS,CSTMS,CASEDR,/UGVSB//0/1 $
The reason for preventing this transformation is that PATRAN 3 understands
the CD field for the GRID entry. In Version 68 the DMAP test for PATVER is
not present:
70 ELSE IF ( OUGCORDX='BASIC' ) THEN $
71 VECPLOT UG,BGPDP,EQEXINS,CSTMS,CASEDR,,MEDGE,SIP/UGB/
/0/1/' PARAM,'/'OUGCORD,'/' BASIC '/ALTSHAPE $
This causes the displacement vector to be transformed into the basic
coordinate system for post processing by PATRAN 3. However, PATRAN 3 is
expecting the output in the output coordinate system of the gridpoints.
If the transformed displacement datablock is used by PATRAN 3, the
displaced shape is wrong.
The description in the Quick Reference Guide, Section 6 (parameters) also
needs to be updated under the section 'POST,POST,-1: Results Data Blocks'.
The PATVER description needs to include the interaction between parameter
OUGCORD and PATVER.
Avoidance: Do not specify the value of parameter OUGCORD as BASIC if you
are using PATRAN 3. PATRAN 3 expects the displacements to be expressed in
the CD value of the gridpoint (analysis coordinate system in PATRAN 3.)
You can transform displacements to different coordinate systems inside P3.
@
7185; 66; Response Spectra - SDR2 - Wrong Answers

The response spectra summation process for the NRL and SRSS methods
handle the CBAR and CBEAM elements
differently. Both the NRL and SRSS methods involve taking the square root
of the sum of the squares of the value
of interest, in this case stress. For the BEAM element the summation is
made for the combined bending and axial
stress, while for the BAR element the bending and axial stress is summed
independently to coincide with the BAR's
output which prints the axial stress separately. This leads to incorrect
results for the output of the maximum and
minimum stress.
Avoidance: The CBEAM element should be used for Response Spectra
Analysis whenever stress results are
desired.
@
7186; 70.6; Documentation - INCLUDE - Include File Not Found
The Installation and Operations manual Section 4.5 states that the default
directory for an include file is the input file's directory. The default
directory is actually the current directory.
 Avoidance: Place the file to be included in the current directory or
specify the complete path name of the file to be included.
@
7191; 69; Optimization/Design Sensitivity - TEMP(LOAD), TEMP(MATERIAL), SH
PCAS - SFM 3001
When a TEMP(LOAD) and/or TEMP(MATERIAL) entry is used in the primary
Subcase Control section of a shape optimization job using the analytic
shape boundary shape approach in MSC/NASTRAN Version 68.2, the job will
fail in module SHPCAS with the message "SYSTEM FATAL MESSAGE 3001".
Avoidance: Apply the following DMAP alter.
compile bndshp nolist noref souin=mscsou
alter 'auxmid'(2)
dbview ugx=ug (where mtemp=*) $
alter 'partn','partn'
PARTN UGx,,AXMVEC/UGA,,,/1 $
alter 'equivx','equivx'
EQUIVX UGx/UGA/ALWAYS $

@
7193; ; Miscellaneous - Wrong continuation format - UFM 313 or Wrong Ans
wers
If column 1 of the continuation entry is blank, then the continuation
entry is interpreted as a continuation of the last sorted bulk data entry.
The error is likely to result in USER FATAL MESSAGE 313 ILLEGAL NUMBER OF
WORDS ON CONTINUATION" or in rare cases possibly even wrong answers. An
example follows:
.
Begin Bulk
.
.
tic,100,101,1,0.,0.
$
grid,1000,,0.,2.5,2.5
$
grid,101,,0.,0.,0.
grid,102,,10.,0.,0.
$ 2 3 4 5 6 7 8 9 10
cbar 1 1 101 102 1000
+000015
+000015 0.25 0. 0. 0.25 0. 0.
$
.
.
enddata

will be interpreted as

BEGIN BULK
.
.
CBAR 1 1 101 102 1000
 +000015
.
GRID 101 0. 0. 0.
GRID 102 10. 0. 0.
.
GRID 1000 0. 2.5 2.5
.
.
TIC 100 101 1 0. 0.
+000015
+000015 0.25 0. 0. 0.25 0. 0.
.
.
ENDDATA
Avoidance: Correct the offending continuation entry.

@
7196; 2001; Miscellaneous - GP2, GENEL - Poor Message, SFM 3003
When at least one entire line is missing from a GENEL element, GP2 issues
the message:
SFM 3003 ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK
GEOM2S (FILE 101) IN SUBROUTINE GP2.
Avoidance: Correct any offending GENEL elements. At a minimum the
Reference Manual documentation should cite incorrect GENEL input as a
possible cause of the error message.
@
7199; ; Nonlinear Transient Analysis - PARAM,LGDISP - Poor Answers
In MSC/NASTRAN Version 68 nonlinear transient analysis, solutions 99 and
129, were improved to correct error 2665 which stated that in nonlinear
transient analysis the damping matrix is not updated. The damping forces
for the large displacement solution now use the current "updated" damping
matrix, which may have been calculated at an unconverged solution. This
approach was intended to provide better directionality to the damping
forces when they rotate. However, every time the damping matrix is updated
for convergence, a change will occur in the force balance which will cause
an erroneous step in the converged accelerations. ( It causes very large
errors. - DNH)
@
7201; 70; Element - CELASi - Forces/Stresses are Calculated in Single Prec
ision
Scalar spring and damper element's (ELASi and DAMPi, 1=1,2..4) stress and
force outputs are calculated with only single-precision arithmetic. If a
scalar element connects two points with small differences in
displacements, relative to the total displacements, some numerical
roundoff will occur. Since the stiffness and damping matrices are
calculated in double precision the overall results will not be affected.
The problem also does not occur on large-word, 64-bit computers such as
CRAY and NEC.
Avoidance: More precise results for the ELASi elements may be obtained
from the GPFORCE output request.
@
7202; 69; Case Control - MPCFORCE=NONE - Extra Output
 The mpcf=none command does not turn off mpcforce output in a subcase when
it has been requested above the subcase level.
Avoidance: Define set 1 = [x], where x is a nonexistent grid or scalar
point and use mpcf=1 instead.
@
7206; 69; Heat Transfer Analysis - CHBDYE,TUBE PHBDY - Wrong Answers
The convection area used by the CHBDYE for the structure TUBE element is
incorrect whenever D2 is left blank. In Version 69 the code is changed so
that if D2 is left blank then D2 is set to D1.
Avoidance: Use CHBDYP,TUBE option to represent the tube convection area or
explicitly set D2.
@
7207; 69; Restart - SPCFORCE - Missing Output
SPC force output request is ignored during automatic restarts in dynamic
structured solution sequences. See related Error 3119 which was fixed for
SOL 101 only.
Avoidance: Include the entry PARAM, SERST, SEDR in the bulk data section
of the restart deck.
@
7213; ; DMAP - FBS, SYSTEM(72)=1 - Wrong Answers
Left handed fbs solves the equation of the form:
t t
[X] [L][U]=[B]
In addition left handed fbs has an option of solving
T
[X] L U = [B]
This option is only documented in the MSC/NASTRAN Numerical Methods Guide
and is selected by setting SYSTEM(72)=1. This option produces wrong
answres when B is rectangular. The left hand fbs method is not used in
the delivered solution sequences for non symmetric matrices. Thus, only
user DMAPs may encounter this error.
Avoidance: Use right hand fbs.
@
7215; ; Cyclic Symmetry - SECSET1, VAXW, GIV, MGIV - UWM 3034, UFM 4645
Massless degrees of freedom that are placed in the c-set are not properly
removed before entering the READ Module in Cyclic Symmetry component mode
synthesis. Attempts to remove these degrees of freedom by explicitly
constraining them will also be ignored. This may lead to "USER WARNING
MESSAGE 3034, ORTHOGONALITY TEST FAILED" followed by a fatal message,
"USER FATAL MESSAGE 4645---- THE SHIFTED STIFFNESS MATRIX IS NOT POSITIVE
DEFINITE IN MGIVENS METHOD" or "USER FATAL MESSAGE 4646---- THE MASS
MATRIX IS NOT POSITIVE DEFINITE, USING THE GIVENS METHOD. USE MGIV
INSTEAD".
Avoidance: Do not put the degrees of freedom in the c-set, specify
Lanczos instead of a transformation method or add a small amount of mass
to the massless degrees of freedom.
@
 7219; 69; Miscellaneous - IFP - Infinite CPU Loop
The following continuation entries while legal when using the NASTRAN MESH
option result in an infinite cpu loop in IFP when the NASTRAN MESH command
is removed.
SPC1 10 10 11
= *1 *-1 *1
DESVAR 20 ABC 1.
=2 *1 = *1.
Avoidance: Add the NASTRAN MESH command or correct the continuation
entries in error.
@
7220; 69; Postprocessing - PARAM,POST - Missing Output
If the ASSIGN statement is not used or if the complete path is not
specified on the ASSIGN statement for an OUTPUT2 file on the Windows NT
Version of MSC/NASTRAN, the file is not written to the jid directory but
is instead written to the "bin" directory.
Avoidance: Specify a complete path on the ASSIGN statement.

@
7231; 69; Optimization/Design Sensitivity - XSORT - Incomplete Sorted Outp
ut
The XSORT Module may encounter errors for large input files, when it is
executed more than once during a run. The messages encountered are
various, but the sorted file will be incomplete in every case. The error
occurs only on large files, and only on the second and subsequent bulk
data files. The errors occur when the bulk data to be sorted does not fit
into memory. In Version 68 this error occurs only in design optimization
analysis.
Avoidance: Increase the memory request so that the file will fit into
memory, or reduce the size of the file.
@
7232; 69; Miscellaneous - IFP, RBEi- UFM 4276 EC 7777
When the RBE3 bulk data entries contains illegal data in the REFC field,
USER FATAL MESSAGE 316 should be issued to identify the offending entry.
USER FATAL MESSAGE 4276 ERROR CODE 7777 is issued instead, which does not
identify the illegal data.
Avoidance: Other than locating the input in error and correcting it,
there is no avoidance.
@
7233; 69; Matrix Methods - FBS, MPYAD - UFM 4547
FBS may set the density of GOAT incorrectly for some large models. The
density should be between 0 and 10000 (i.e., 100%). However, it sometimes
overflows this field and is printed in the f04 as *******. In this case
the run will fail in the MPYAD module with the message "USER FATAL MESSAGE
4547"
Avoidance: Use the following alter to re-set the trailer to a value of 80%
(or any value they desire).
 compile sekr, souin=mscsou, nolist, noref
alter 106 $ the line which creates GOAT in his run
modtrl goat///////8000 $
@
7234; 69; Loads - PLOAD1 - Missing Documentation, Unexpected Answers
If on the TYPE field of the PLOAD1 entry, the element coordinate system
direction (e.g. TYPE=FYE) option is selected, then the projection (i.e.
SCALE=FRPR or LEPR) option is ignored and the result is the same as the
SCALE=FR (or LE) option.
Avoidance: There is no avoidance.
@
7237; 70.5; Documentation - DMAP Module Dictionary, OUTPUT2, INPUTT2 - SFE 4
276 EC 1401
MSC/NASTRAN Version 68 may not be able to handle an OUTPUT2 file of a
matrix generated from an external client written computer program that
follows MSC's documentation description of proper OUTPUT2 format. It is
not the INPUTT2 step that is crashing, it is any subsequent operation on
the datablock, such as TABPT, OUTPUT4, ADD, MATPRN or the like.
Avoidance: Use V67.5 to convert the OUTPUT2 file to an OUTPUT4, and then
read in the OUTPUT4 file into V68.

@
7238; 69; p-Elements - DATAREC, OUTPUT - UWM 6514 (SDRPOP)
When there are any OUTPUT bulk data entries, but no DATAREC case control
keywords, the following
warning is produced:
*** USER WARNING MESSAGE 6514 (SDRPOP)
THE DATAREC= 0 CASE CONTROL COMMAND REFERENCES UNDEFINED OUTPUT
BULK DATA ENTRY.
USER INFORMATION: THE DATAREC COMMAND WILL BE IGNORED.
USER INFORMATION: THERE WILL BE NO ERROR-ESTIMATE OUTPUT FOR THE
ADAPTIVITY ANALYSIS.
Avoidance: To avoid the warnings, the OUTPUT entries should be removed.
@
7242; ; p-Elements - GMCORD - Inefficient Processing
If a GMCORD coordinate system is used for the material, load, or
displacement coordinate system, execution may be inefficient since the
coordinate system is processed independently for each integration point.
Avoidance: Other than not using GMCORDs there is no avoidance.
@
7243; 70.5; Documentation - TABPT, BGPDTS - DMAP Module Dictionary
According to the DMAP module dictionary "TABPT may occasionally
misidentify real numbers or character values on the IBM/MVS-like
computers". This problem is not limited to the IBM/MVS-like computers.
 Avoidance: Use TABPRT instead of TABPT.
@
7244; 69; Optimization/Design Sensitivity - ANALYSIS=FLUTTER, DMIG - Wrong
Answers
For flutter design sensitivity and optimization, the sensitivity results
will be in error if there
are multiple subcases and a second or subsequent subcase includes a
different M2PP, K2PP or B2PP request in the case control. The sensitivity
analysis is based on the generalized matrices from the first subcase and
will therefore be incorrect for the subsequent subcases.
Avoidance: There is no avoidance.
@
7253; 69; Nonlinear Analysis - MPCFORCE - Missing Output
MPCforces are not output in nonlinear analysis, solution 106.
Avoidance: Insert the following DMAP alter in Version 68.2:
compile nlstatic
alter 'ADD DELCVECT,CVECT0'(,-1)
add5 qqgni,qgni,,,/qmgni//-1. $
call dbstore qmgni,,,,//nloop/0/' '/0 $
alter 'DBVIEW CVECTH'
call dbfetch /qmgni,,,,/solcuri/0/0/0/0 $
call dbstore qmgni,,,,//solcur/0/' '/0 $
alter 'APPEND QGNI,/QGNR'
$
file qmgni=ovrwrt $
$
call dbfetch /qmgni,,,,/nloop/0/0/0/0 $
file qmgnr=append $
APPEND QmGNI,/QmGNR/2 $ APPEND MPC FORCE VECTORS
alter 'EQUIVX QGNR/QGN'
call dbstore qmgnr,,,,//0/0/' '/0 $
compile super3
alter 'ADD PG,PJ/DPJ'
call dbfetch /qmgnr,,,,/0/0/0/0/0 $
equivx qmgnr/qmg/-1 $
COMPILE SEDISP
ALTER 'MPYAD PSS,CRX'(2)
MPYAD PG,CRX,/PGX $
$
ALTER 'IF ( NOQMG>-1 ','' $
IF ( NOQMG>-1 AND NOMSET>-1 AND NOT(DELTA) ) THEN $
$
ALTER 'UPARTN USET,PG','' $
IF ( APP <> 'NLST' ) EQUIVX PG/PGX/ALWAYS $
UPARTN USET,PGX/PM0,,,/'G'/'M'/'S'/1 $
$
COMPILE CFORCE
ALTER 'MPYAD KXY,UY,PX','' $
IF ( APP='STATICS' or APP='NLST' OR APP='BKL0' OR
(APP='CYC' AND APP1='STATICS') ) MPYAD KXY,UY,PX/QX///-1 $
@
7260; 69.1; Nonlinear Transient Heat Transfer - DELAY, CHBDY - Incorrect Loa
d Output
 If the DELAY bulk data entry is used to define an offset in time for the
applied load. The actual heat load is applied correctly, and the
resultant temperature value is correct. However in the data recovery for
the heat flux of the HBDY element the load is applied incorrectly.
Avoidance: Ignore the applied load output for the HBDY elements.
@
7266; 69; DMAP - MALTER - Extra output
The MALTER statement produces a listing of the altered subDMAP. This
printout cannot be suppressed even
if COMPILER NOLIST is specified. However, COMPILER REF and COMPILER DECK
produce the correct results.
Avoidance: There is no avoidance.
@
7270; 69; Acoustic Analysis - PARAM,WTMASS - Wrong Answers
PARAM,WTMASS is applied to both the fluid and the structure, even though
the mass of the fluid is in different units.
Avoidance:

a. use mass density for density values and ensure that all mass
input values are in mass units.
b. or insert the following DMAP alter in V68.2:
compile semg noref nolist
alter 'MJJX,/-1/WTMASS','' $
type parm,,cs,n,cwtmass,mwtmass $
IF (NOMGG>=0) then $
dbstatus VGFS//S,N,NOVGFS $
IF ( NOVGFS>0 and wtmass<>1.0 ) THEN $
EMA GPECT,MDICT,MELM,BGPDTS,SILS,CSTMS,MDICTP,MELMP/
MJJX1,/-1 $ MASS
matgen ,/ig/6/lusets/0/lusets $ gx1 matrix
cwtmass=cmplx(wtmass) $
add ig,vgfs/wtmass/cwtmass//3 $
matmod wtmass,,,,,/wtmassg,/28 $ gxg
mpyad mjjx1,wtmassg,/mjjx $
ELSE $
EMA GPECT,MDICT,MELM,BGPDTS,SILS,CSTMS,MDICTP,MELMP/
MJJX,/-1/WTMASS $ MASS
ENDIF $
endIF $ NOMGG>=0

@
7276; ; Composites - PARAM,NOCOMP - Inconsistent Stresses
Homogeneous stresses, output when using 0 or -1 for PARAM,NOCOMPS, for
PCOMP-based elements are calculated inconsistently between linear and
non-linear solutions. Since homogeneous stresses are based upon smeared
laminate properties, both calculations yield only approximate the lamina
stresses (see also current error list number 1818) but both solutions
should use the same method. The two methods currently being used are:
For linear analysis:
 S(z) = G1 * E - z * G2 * C (1)
while for nonlinear analysis:
S(z) = G1 * ( E - z * C ) (2)
where:
C = curvature (chi)
E = reference plane strain (epsilon)
G = modulus matrix
S = stress
z = distance from reference plane

@
7284; 69; Miscellaneous - DIAG - DIAGs Out of Range Not Checked
I don't think this needs to be an external error report.
Please read text below for error.

@
7288; 69; Nonlinear Buckling - MPC, RBE*, RBAR - Wrong Answers
If MPC's or rigid elements are used in a nonlinear buckling analysis, then
there is an error in the nonlinear buckling load factor (alpha computed by
DMAP in SOL 66 or 106). This error causes the critical buckling factor
(UIM 9040) to be estimated incorrectly. The remainder of the solution
(eigenvalues and displacements) is OK.
Avoidance: Use the following DMAP alter:
compile nlstatic noref nolist $
alter 'MPYAD DULV,DPL0,/NUM3/1','' $
mpyad dugni,dpgni,/num3/1 $

@
7289; 69; Optimization/Design Sensitivity - CELASi, TEMP(LOAD), SSG1 - UFM
3032
USER FATAL MESSAGE 3032 may occur in optimization solution sequence when
temperature loads are selected.
The error will occur when the following three conditions are true:

1) A structural model contains scalar elements such as CELAS1.

2) The grid points connected to these scalar elements are subjected to a
temperature load.
3) Only these scalar elements are designed and no other structural
elements are designed.
Note that scalar elements do not support temperature loading.
Avoidance:
1) Make a dummy design variable and related it to a non scalar structural
element in
the model. Then impose the same lower and upper limits on the dummy design
variables.
 2) Replace the scalar elements with non scalar elements, i.e., use CROD
instead of CELAS1.
@
7290; ; Nonlinear Analysis - QUAD4, TRIA3 - Update Documentation
The existing documentation , such as MSC/NASTRAN V68 Reference Manual,
Section 15.3.9 Small Strain Elements,
needs to be updated to indicate that the QUAD4 and the TRIA3 act
differently than the other "small strain" elements in that they satisfy
equilibrium in the deformed configuration.
This difference can have an increase in accuracy of internal loads in
nonlinear analysis if elements strains are significant.

@
7291; 69; Eigenvalue Analysis - REIGL, EIGRL - Poor Performance,
The REIGL module, when solving for normal modes of a vibration problem,
under certain rare circumstances will perform many negative shifts and
ultimately fail to find the modes requested by the user. Although one
negative shift often proves useful in picking up rigid body modes,
additional negative shifts are unnecessary and wasteful of computer
resources.
One way to duplicate the problem is to specify on the EIGRL card,
F1=blank, F2=very large, ND=blank. If the F2 value is several orders of
magnitude greater than the largest eigenvalue in the model, the shift
logic will usually fail with the multiple negative shifts.
@
7294; 2001; Miscellaneous - SUPORT1 - UFM 316
If a value of 0 (zero) is entered for the component number on a SUPORT1
entry USER FATAL MESSAGE 316 will result.
USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA CARD SUPORT1 ***
Avoidance: If you are referencing SPOINTs on the SUPORT1 entry, then use
1 (one) or blank.
@
7295; 70.5; Documentation - SYSTEM(196), SCRSAVE - No Documentation on Lancz
os System Cell
There is no documentation in the Quick Reference Guide on the use of
SYSTEM(196) (a.k.a. SCRSAVE). This system cell control reuse of scratch
files when segment logic is invoked. System(196) is useful only when
multiple frequency segments are requested on a Lanczos run. (Multiple
frequency segments can be requested via the NUMS field in the EIGRL entry
and by system(197).) Each frequency segment requires a minimum of three
scratch files. When multiple frequency segments are used on a single
processor computer then each frequency segment is solved serially. In
this case, it makes sense to let segment #2 use the scratch files which
were used by segment #1 since work for segment #1 has been completed
(otherwise it wouldn't be working on #2). Similarly, when work for
segment #2 is finished, segment #3 should be able to use #2's scratch
files. System(196)=1 allows such file reuse.
Note that when multiple processors at the Lanczos segment level is
selected, i.e. system(107) (a.k.a. parallel) is set greater than 1,
 system(196)=1 may cause much grief. Parallel processing may be retained
for other operations but deselected specifically for Lanczos segments by
adding 32768 to the number of processors. Note also that MSC/NASTRAN
Version 70 has parallelism at the segment level disabled. Therefore
system(196)=1 is a safe default.
@
7298; 70.7; Hyperelastic Elements - QUAD4, MATHP, PLPLANE - UFM 4297
Small hyperelastic QUAD4 elements may incorrectly fail with the message:
USER FATAL MESSAGE 4297 (EQDGED) QUAD ELEMENT WITH ID = XX
HAS AN INTERIOR ANGLE POSSIBLY GREATER THAN OR EQUAL TO 180 DEGREES.
event though the geometry is actually good. The error is caused by a bad
test on elements with lengths less than 1.0E-3.
Avoidance: The only avoidance is to use different units to make the
distances larger.
@
7300; ; Nonlinear Analysis - CGAP - Poor Results
If the adaptive GAP element indicates an update of an element stiffness
AFTER the program has converged on a solution, the program FAILS to
re-iterate upon that newly updated stiffness matrix hence the results are
based upon the GAP stiffness prior to the update.
This error has been encountered when the adaptive GAP gets the following
message:
*** NEW STIFFNESS MATRIX IS REQUIRED BY GAP FOR PENALTY VALUE
ADJUSTMENT ***
PENALTY VALUES ARE ADJUSTED BY YYY FOR ELID = xx
AFTER the program has converged on the solution.
Avoidance: Increase the number of increments, NINC on the nlparm entry, or
adjust the value of the stiffness on the offending PGAP or both.
@
7304; 69; Nonlinear Analysis - Creep - UIM 6193, UIM 6194, UFM 4551, Faile
d Convergence
Glenn;
What exactly was corrected for this error? All I know is that a creep
problem didn't converge in Version 68. Is there an error in all creep
problems? Does it have to do with specific elements, multiple subcases,
DT or what?
Regards Kevin Kilroy
@
7305; 70; Documentation - OTIME - Extra Output, OTIME Not Effective in Non
linear Transient
The OTIME case control command is used to limit output in transient
analysis. This command is not effective for nonlinear transient stress
output.
@
7306; 70; Online Documentation - Common Questions and Answers
The line "NASTRAN SCRATCH(MEM=n)" on page 267 of MSC/NASTRAN Common
 Questions and Answers (Second Edition) in the MSC/NASTRAN Encyclopedia,
page 144 in the printed text, should be: "INIT SCRATCH(MEM=n)"
@
7307; 2001; Dynamic Analysis - UWM 4684 LOADSET ID NOT FOUND
In a dynamic response analysis, when the bulk data contains LOAD entries
that are not referenced by the CASE control and if STRESS output is
requested, then the following warning message is issued.
USER WARNING MESSAGE NUMBER 4684 LOADSET ID NOT FOUND FOR SUBCASE 1
Avoidance: Remove the non referenced LOAD entries from the bulk data.
@
7312; 69; Optimization/Design Sensitivity - DOM11, GEOM1 - SFM 3001
When only scalar points (no GRID entries) are defined in an optimization
model and the SEQP module is not executed to resequence the model, then
there is no GEOM1 generated. The run subsequently fails in the DOM11
module when it tries to make a copy of GEOM1 with the message "SYSTEM
FATAL MESSAGE 3001 MODULE = DOM11 GEOM1 AT POSITION 12 DOES NOT EXIST".
Avoidance : Include a dummy GRID entry.
@
7314; 69; Performance - VECPLOT - Poor Performance
On problems with more than 65,000 dofs, the vecplot module is inefficient,
and may cost as much as the
rest of the calculations. It is used in several places, primarily as a
means of providing checks such as
applied load resultants.
Avoidance: If you observe that the amount of time being spent in this
module is excessive you may insert an alter to delete this module call.
The alter for all Version 68.2 structured solutions is:
compile selg nolist noref $
alter 63,63 $
@
7320; 69; Heat Transfer Analysis - CHBDYE, VIEW - SFM 3007
The message "SYSTEM FATAL MESSAGE 3007 MODULE=VIEW, ILLEGAL INPUT TO
SUBROUTINE VNASIN" will occur if using the backside on the CHBDYP LINE
element, for example:
CHBDYP 101 11 LINE 111 2 3
+
+ 1111 0 -1.0 1.0 0.0

Avodiance: Specifying the VIEWID on the frontside resolves this problem.

@
7326; 70; Documentation - FORCE(PLOT) - Missing Documentation
The MSC/NASTRAN Quick Reference Guide does not document the FORCE command
(PLOT) option. This option may be used if element force calculations are
desired for post-processing without printing them.
@
7331; ; Postprocessing - PBCOMP - Formatted OUTPUT2 on DEC Alpha dies Wr
iting EPT
If an EPT table containing entries for more than two PBCOMPs is written to
a formatted OUTPUT2 file on a DEC Alpha system, then the job will
terminate with either no message or a "time: command terminated
abnormally" message in the .log file.
Avoidance include: 1) Using PARAM, OGEOM, NO.
2) Writing to a formatted OUTPUT2 file.
3) Altering the OUT2GEOM subdmap and removing EPT from
the OUTPUT2 statement.
4) Running on a different platform
@
7333; 69; Frequency Response and Random Analysis - TETRA, PSDF - UFM 3052
Random response analysis that requests PSDF xyplots of TETRA stress
results may result in the following fatal message:
USER WARNING MESSAGE 3052, A RANDOM REQUEST FOR CURVE TYPE - STRE -,
POINT - 1
COMPONENT - 12 -, SPECIFIES TOO LARGE A COMPONENT ID. THE LAST
COMPONENT WILL BE USED.
The XYPLOT summary will then indicate that the correct plot is used, but
wrong results will be presented on the XYPLOT summary. In fact, if 6
plots are selected, with items codes 6,7,8,9,10 and 11, the summary will
say they all have the same RMS value, even though they are asking for
normalx, normaly, normalz, tauxy, tauxz, and tauyz, which cannot be the
same. Running a similar problem with a HEXA instead of TETRAs, produces
good results.
@
7335; 69; Optimization/Design Sensitivity - PBEAM - Wrong Answers
Errors occur in design optimization of beam elements if the following two
conditions are met:
1.) The beam section is constant (e.g., end b data is not supplied on
the PBEAM entry)
and
2.) End B data is referenced on a DVPREL1 entry.
By not supplying end B data, the code is mislead to believe the section is
constant, even if this is in conflict with information supplied in the
design model.
Avoidance: Supply end B data in the case of a tapered beam, and all
necessary intermediate stations in
the case of a variable beam section.
@
7338; 69; SSSALTER - resflexa.v682
An error in the resflexa.v682 alter in sssalter directory may result in
loadings applied to the residual structure not being included in the
residual vector calculation.
Avoidance: Add the following line after the line containing 'type db,pss
$'
 TYPE DB,PL
@
7339; ; Nonlinear Transient Analysis - NLTRD2, ADAPT - SFM 4276 EC 1044
Nonlinear transient analysis fails in the module NLTRD2 with the message
"SYSTEM FATAL ERROR 4276.OPEN ERROR CODE 1044". This error happens
when the adapt method is used if
1. There are no nonlinear elements in the model.
2. Stress output is requested.
3. The NO field of the TSTEPNL entry is greater than 2.
Avoidance: Use the default (1) for NO field of the TSTEPNL entry.
@
7340; ; DMAP - LTU - SFM 3001
User DMAP runs submitted with scr=yes may fail with SYSTEM FATAL MESSAGE
3001. This error may result if the LTU logic deletes a datablock even
though it is still needed.
Avoidance: Run the job using SCR=NO or create a copy of the datablock and
use the copy in the DMAP alter.
@
7342; 70; Optimization/Design Sensitivity - STATSUB - Wrong Answers
Wrong buckling sensitivities may occur when the following conditions are
true:
1. DSNOKD is nonzero,
2. The Subcase Control command STATSUB does not refer to the first subcase
and
3. The referred subcase does not have retained responses of its own.
Avoidance: Define an artificial response with very large upper and lower
limits and force retention of the response in the referred subcase using
DRESP1, DCONSTR and DSCREEN entries.
@
7356; 2001; P-Elements - DATAREC - UFM 6352
If a model contains output requests for p-elements (i.e., a DATAREC and
OUTPUT entry), but no p-elements are defined in the model, UFM 6352 may
occur.
Avoidance: Either create a p-element in the model or remove the p-element
output requests.
@
7358; 69.1; Installation - User Modifiable NASTRAN - CRAY C90 Only
The CD for User Modifiable MSC/NASTRAN Version 68.2 contains errors in the
unicos and unicosc90 directories.
1) The unicos*/undynamic* files are misnamed. They are renamed as
follows:
mv unicos/undynamic_7X.o unicos/undynamic_70.o
mv unicos/undynamic_8X.o unicos/undynamic_80.o
mv unicosc90/undynamic_7X.o unicosc90/undynamic_70.o
mv unicosc90/undynamic_8X.o unicosc90/undynamic_80.o
 2) The unicosc90/umnld script is incorrect. All references to
"unicos" must be changed to "unicosc90".

@
7365; ; Heat Transfer Analysis - VIEW3D, CHBDYE - No VIEW Factor Calcula
tion
The CHBDYE element references the conduction element for the geometry. If
the user attempts to calculate the view factor exchange between two plates
by using a line element such as a bar or rod and the CHBDYE element which
references the rod element for the surface information, then no view
factor calculation will be performed on the CHBDYE line element. .
Avoidance: Use the CHBDYP,LINE to model the surfaces instead of the CHBDYE
element.
@
7367; 70; Documentation - Design Sensitivity/Optimization, RBEi, MPC's, DV
GRID - UFM 6499
It is a limitation of MSC/NASTRAN, that the sensitivities are incorrect
for shape optimization runs which have DVGRID entries which reference grid
points on MPC or RBEi entries. Often, the sensitivities are 0.0 and
therefore the following message is produced "ONE OR MORE VIOLATED
CONSTRAINTS HAS A ZERO GRADIENT **** NO PROGRESS IS POSSIBLE. OPTIMIZATION
TERMINATED *** USER FATAL MESSAGE 6499 (DOM9) THE DOT OPTIMIZATION
ALGORITHM HAS ENCOUNTERED AN ERROR THAT MAKES FURTHER PROGRESS IMPOSSIBLE.
SEE MESSAGES ABOVE". This limitation should be documented under the
DVGRID entry.
Avoidance: It may be possible to replace the RBEi or MPC relation with an
element which has an extremely large stiffness.
@
7376; ; p-Elements - CORD3G - SFM 6333 (EQNINI)
When a convective coordinate system (cord3g) is specified as a stress
output coordinate system in OUTRCV, OFP crashes
with SFM 6333 (EQNINI)
Avoidance: Other than not referencing a cord3g entry, there is no
avoidance.

@
7380; 69; Documentation - QUAD4, PARAM,S1M, PARAM,CURV,1 - User Output Coo
rd. Systems
QUAD4 corner stresses do not support output in a user defined coordinate
system using PARAM,CURV,1 and PARAM,S1M,1 and a MCSID coordinate system on
a MAT1 or MAT2 entry. The MAT1 and MAT2 Quick Reference guide
documentation needs to be updated as follows:
The MAT1 remark 7 needs to changed to:
7. MCSID must be nonzero if PARAM,CURV is specified to calculate stresses
or strains at grid points on plate and shell elements only. This feature
does not support QUAD4 corner stresses.
 The MAT2 remark 6 needs to changed to:
6. MCSID must be nonzero if PARAM,CURV is specified to calculate stresses
or strains at grid points. This feature does not support QUAD4 corner
stresses.
@
7391; 69; Heat Transfer Analysis - RADCAV - SFM 4276, Infinite Loop
In SOL 153, if global view factors are specified (i.e. the continuation
entry on the RADCAV is used), then the run may go into an infinite loop or
issue SYSTEM FATAL MESSAGE 4276 for some set definitions.
Avoidance: Do not use global view factors or redefine which elements are
contained in sets selected by the RADCAV entry.
@
7397; 69; Documentation - QRG, APP COUPLED - Incorrect Reference to Therma
l User's Guide
The MSC/NASTRAN Version 68 Quick reference guide makes the following
refererence under the description of the APP Executive Control Command:
APP COUPLED Indicates that a coupled analysis combining heat transfer
and structural analysis
in SOL 153. See the MSC/NASTRAN Thermal Analysis Users
Guide.
In fact there is no description of the coupled thermal analysis in the
Version 68 Thermal Analysis User's Guide.
@
7403; 70.7; Optimization - PCOMP - UFM in SDRCOMP, Corner Stress Request
The presence of composites in SOL 200 when corner stresses are requested
will cause the following unnumbered fatal message in the SDRCOMP module:
USER FATAL MESSAGE. INCONSISTENT DATA DETECTED IN THE OES1A LAMINATE
STRAIN AND
OEF1A ELEMENT FORCE DATABLOCKS
In V70.5 and later, an error message similar to the following may be
output instead from the SDRCOMP module:
*** SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0
Avoidance: Remvoe the CORNER, BILIN, SGAGE or CUBIC option from your
STRESS or FORCE output requests. However if noncomposite load responses
are not used, the following DMAP alters will allow noncomposite corner
stress responses and composite center stress and failure index responses
to be used. But it will cause non-composite force responses to fail.
For V68.2:
compile sdr2stat noref nolist
alter 'oes1as'
SDR2 CASEDS,CSTMS,MPTS,DIT,EQEXINS,sils,ETTDC,EDT,BGPDTS,,
QGX,UGX,ESTDC,,,,,gpsnts/
,,,,OEF1AS,/'STATICS'//-2 $
alter 'sdrcomp',''
SDRCOMP CASEDS,MPTS,EPTS,ETTDC,ESTDC,OES1AS,OEF1AS,DIT/
ES1CS,EFITS,OEF1AA/0/1 $
alter 'sdrcomp'(2),''
 SDRCOMP CASEDS,MPTS,EPTS,ETTDC,ESTDC,OES1AS,OEF1AS,DIT/
EA1CS,EFITS,OEF1AA/1/1 $
For v70:
compile sdr2stat noref nolist
alter 'oes1as'
SDR2 CASEDS,CSTMS,MPTS,DIT,EQEXINS,,ETTDC,EDT,BGPDTS,,
QGX,UGX,ESTDC,,OINT,PELSETS,VIEWTB,GPSNTS,DEQATN,DEQIND,/
,,,,oef1as,/'STATICS'//-2/////ACOUT/PREFDB///
1/TRUE/FALSE $ forces
alter 'sdrcomp',''
SDRCOMP CASEDS,MPTS,EPTS,ETTDC,ESTDC,OES1AS,oef1as,DIT,BGPDTS/
ES1CS,EFITS,OEF1AA/0/1 $
alter 'sdrcomp'(2),''
SDRCOMP CASEDS,MPTS,EPTS,ETTDC,ESTDC,OES1AS,oef1as,DIT,BGPDTS/
ES1CS,EFITS,OEF1AA/0/1 $

@
7404; 69; p-elements - GP4 - SFM 3008, Insufficient Memory, Excessive Runt
ime
When converting a large h-element model directly to a p-element model, the
run may fail in GP4 with the following error messages:
SYSTEM FATAL MESSAGE 3008 MODULE = GP4
0INSUFFICIENT MEMORY AVAILABLE FOR SUBROUTINE CSTRAN
0FATAL ERROR
The runtime in GP4 will also be excessive. This error is most likely to
occur when there are many grid points which reference non-basic output
coordinate systems.
Avoidance: Other than increasing the memory on the mem keyword, there is
no other known avoidances. The amount of memory required for a large
problem can be substantial in this case.
@
7409; 69; Heat Transfer Analysis - SOL 101 - SFM 4276. FREAD EC 3002
Solution 101 heat transfer (app heat) superelement runs will fail in
solution 101 with the following message:
SYSTEM FATAL ERROR 4276.FREAD ERROR CODE 3002
Avoidance: Other than no running superelements there is no avoidance.

@
7410; 69; Eigenvalue Analysis - EIGRL - Wrong Answers
If the following EIGRL format
EIGRL,SID,V1,,ND
is used, and the first flexible mode is substantially higher than the
rigid body modes, then modes below V1 may also be included in the
eigenvalue table printout.
Avoidance: Either increase V1 or include a shift scale that is close to
 the estimated first desired mode.
@
7414; ; Optimization/Design Sensitivity - Poorly Labeled Output
It is difficult to relate the output in solution 200 (DISPLACEMENT,
STRESSES, SENSITIVITY, ETC ) to the design cycle. The output headers do
not include the design cycle number for which the output is printed.
Avoidance: Look for the "USER INFORMATION MESSAGE 9051" , which contains
the design cycle information preceding the output.
@
7425; ; DMAP - CALL - Incorrect PARAM Value Passed Back
If a parameter is passed into a subDMAP on the CALL statement, it is
possible that the program may pass a value of 0.0 back to the calling
subDMAP for that parameter. This has only been observed if the following
conditions are met:
1) The parameter appears on the CALL statement without the 'S' feature
2) The parameter is defined as being able to be changed in the bulk data
or case control ( 'Y' option on
the TYPE statement in the called subDMAP')
3) The same name is used for the parameter in the calling and called
subDMAPs
4) The value of the parameter is changed in the called subDMAP
Avoidance: Change the name of the parameter in the called subDMAP or use
the 'N' option on the TYPE statement in the called subDMAP.
@
7428; 70; Documentation/Elements - CGAP, CORDCi, CORDSi - Wrong Answers
If a cylindrical (CORD1C/CORD2C) or spherical (CORD1S/CORD2S) coordinate
system is specified for a gap element (CID field of CGAP entry), and one
of the connected GRIDS is defined at the center of that coordinate system,
wrong answers result. Wrong answers result, even if the GRID at the
center of the coordinate system is defined in a cartesian system.
Avoidance: Displace the GRID at the center by as little as 1.E-40 in the
direction of the desired gap axis, or define the v vector and leave the
CID field blank.
@
7452; 69; Postprocessing - Optimization, TRANS - Missing Datablock CHISTM
from xdb
MSC/NASTRAN Version 68.2 on CRAY computers fails to write the CHISTM
datablock to the .xdb file. An incorrect xdb will show the following
when the resulting xdb file is written to neutral format with the TRANS
utility:

.
.
DATA BLOCK PROJECT SE ID PATH SUBCASE SET ID APPROACH
CHIST 1 0 1 0 0 0
CQD4 1 0 1 0 0 0
CQD4 1 0 1 0 0 0
CQD4 1 0 1 0 0 0
CTAB 1 0 1 0 0 0
 CTAB2 1 0 1 0 0 0
DCYCT 1 0 1 0 0 0
DHIST 1 0 1 0 0 0
DINFO 1 0 1 0 0 0
DISPR 1 0 1 10 0 1
DISPR 1 0 1 10 0 1
DISPR 1 0 1 10 0 1
DOINFO 0 0 0 0 0 0
DOINFO 1 0 1 0 0 0
DVECT 1 0 1 0 0 0
.
.
.
.
Note that the correct xdb has a CHISTM datablock and two DINFOs, but the
incorrect xdb has no CHISTM and three DINFOs.
@
7454; 69; Nonlinear Analysis - PLSOLID, PLPLANE, TEMP(LOAD) - SFM 4276, EC
7777
Hyperelastic elements with temperature loads do not work on MSC/NASTRAN
Version 68 Alpha OSF/1 system. All
jobs with temperature loads and hyperelastic elements will fail with the
message "SYSTEM FATAL MESSAGE 4276, ERROR CODE 7777" in the NLITER Module.
Avoidance: Other than running on another machine there is no avoidance.
@
7460; 69; Optimization/Design Sensitivity - PCOMP, SUBCOM - SFM 3007
The message, "SYSTEM FATAL MESSAGE 3007 MODULE = DSAL" is produced if:
(1)there are active DRESP1 responses associated with composites
(CSTRESS,CSTRAIN,CFAILURE), (2) SUBCOM's are used and (3) One or more of
the subcases used in the SUBCOM has no active responses.
Avoidance: An avoidance is to add responses in the subcases referenced on
the subcom and force their
retention using DSCREEN.
@
7462; 69; Heat Transfer - CHBDYG, RADBC - Wrong Answers
If a negative EID is specified on the RADBC entry to indicate that the
boundary condition is to be applied to the back face of the element, then
the load applied is incorrect and wrong answers result.

Avoidance: Define another CHBDYG or CHBDYE on the backside and use it to

apply the load.

@
7463; 70.5; Stress Discontinuities - STDCON - Inefficient Processing
The STDCON module can be very inefficient. For some problems,
enhancements in Version 70.5 have improved the performance of this module
by several orders of magnitude.
@
7464; 69; P-Elements - CQUAD8, CTRIA6 - SFM 6398 (GPOI) Preliminary edge t
able
 If there are h-element CQUAD8s and/or CTRIA6s next to p-elements, and
there are midsides on the common edges, the following message may result:
SYSTEM FATAL MESSAGE 6398, (GP0I),
PRELIMINARY EDGE TABLE CREATED IN GP0B IS INVALID.
USER ACTION: CONTACT CLIENT SUPPORT AT MSC.
The error message should be more specific. The error is considered
corrected in Version 69 by issuing the following message:
USER FATAL MESSAGE 6544, (GP0I),
THE EDGE WITH GRIDS ID = **, ** ON P-ELEMENT ID = ** IS SHARED BY
AN H-ELEMENT WITH MID SIDE GRID.

@
7469; 2001.0.1; Frequency Response Analysis - Random - SFM 4276 EC 3001, Poor
Message
In a random analysis, the J and K fields (fields 3 and 4) on the RANDPS
entry should reference the subcase identification number, as documented in
the Quick Reference Guide. However, if the J and K fields reference the
subcase sequence number instead, the job will fail with the following
unclear messages.
BIOMSG: ERROR 3001 HAS OCCURRED IN ROUTINE FREAD , FILE INDEX (FILEX) =
0
SYSTEM FATAL ERROR 4276.FREAD ERROR CODE 3001 PID= 0
Avoidance: Modify the J and K fields on the RANDPS entry to reference the
appropriate subcase identification numbers.
@
7470; 69; Geometric Nonlinear Analysis - ELAS3, ELAS4 - Wrong Answers
CELAS3 and CELAS4 elements produce wrong answers in SOL 64.
Avoidance: Use CELAS1 or CELAS2 elements.

@
7476; 69; Heat Transfer - EMG, FTUBE, MATT4 - Quits in EMG
If the data on the TABLEM1 referenced by a MATT4 entry is poorly defined
such that the extrapolated viscosity versus temperature results in a
negative viscosity (which is physically impossible) then the run fails
without message in the EMG module.
Avoidance: Make sure that the viscosity versus temperature curve always
remains in the (+,+) quadrant.
@
7480; ; DMAP - Delivery Database Build - SFE 4276 in XCSA
When building a delivery database in more than one run, if any subDMAP
fails compilation in the first run, which also compiles the NDDL, then a
subsequent run may fail with BIOMSG error 3001 in GREAD and SYSTEM FATAL
MESSAGE 4276 in XCSA.
 Avoidance: If this message is encountered, the delivery database is
corrupted and must be rebuilt. To avoid corrupting the database again, do
not compile any subDMAPs in the first run.
@
7481; ; DMAP - Delivery Database Build - Poor Message, SFE 4276 in LINKE
R
When building a delivery database in more than one run, if a subDMAP which
is called by a main subDMAP fails during compilation in one run and then a
link of the main subDMAP is attempted in the next run, then the run will
fail with BIOMSG error 3001 in GREAD and SYSTEM FATAL MESSAGE 4276 in
LINKER. The run should have failed with USER FATAL MESSAGE 443 or 444.
Avoidance: Recompile the offending subDMAP and relink the main subDMAPs
which call the offending subDMAP.
@
7495; 69; P-elements - VIEWP, TETRA, FEEDGE - Excessive CPU
For large models of TETRA p-elements with curved edges (i.e. FEEDGE bulk
data entries), the VIEWP module takes excessive CPU time. The time spent
in the VIEWP for TETRA p-elements with curved edges is approximately ten
times a model that does not have curved edges. The overall run time may be
increased approximately 30 percent.
Avoidance: There is no avoidance.
@
7498; 69; Transient Heat Transfer - DPD - UFM 2066, 2079, SWM 2072
If the grid points are given TEMP for the initial conditions without the
use of a TEMPD bulk data entry, then you the following error message is
encountered:
USER FATAL MESSAGE 2079, GRID OR SCALAR POINT ******** HAS AN UNDEFINED
COORDINATE REFERENCED ON A NOLINI CARD.
USER FATAL MESSAGE 2066, UNDEFINED GRID POINT 0 ON DAREA-, RLOAD- OR
TLOAD- CARD.
SYSTEM WARNING MESSAGE 2072, CARD TYPE 7107 NOT FOUND ON DATA BLOCK.
Avoidance: Add a TEMPD bulk data entry.
@
7511; 69; Eigenvalue Analysis - REIGL, EIGRL - Wrong Answers
In rare cases Lanczos may produce wrong answers and no warning message is
issued. MSC/NASTRAN Version 69 corrects this situation by issuing
appropriate messages.
Avoidance: Prior to Version 69, it is the user's responsibility to verify
that the actual eigenvalues found agree with all sturm checks. If the
results do not agree then they can not be correct. The eigrl entry should
be modified by increasing maxset to 15 from the default of seven and the
job should be re-run. Changing the frequency range of interest such that
f1 is blank, not 0.0, often corrects the problem.
@
7513; 70.5; Optimization/Design Sensitivity - DOM6, DRESP2, PARAM,OPTEXIT -
SFM 3001
When only DRESP2 responses are defined in an optimization job, the
requested OPTEXIT=4 or -4 results in two different errors:
 1. In Version 68.x, SYSTEM FATAL MESSAGE 3001 occurs in DOM6.
2. In Versions 69 and 70, the DSCM2 matrix is not generated in an f06
file.
Avoidance:
There is no avoidance for v68.x. A dmap avoidance is available for v69
and v70:
compile exitopt
alter 'if ( ((','if ( (('
if ( ((dsprint or dsunform or dsexport) and mdtrkflg=0) and
(NORSP2>0 or nodscmg>0) ) THEN $
endalter
@
7515; 69.0.4; Direct Transient Analysis - CBAR,PLOAD1,LOADSET/LSEQ - Wrong Ans
wers
In a direct transient dynamic analysis, if a CBAR element is loaded by a
PLOAD1 entry through the LOADSET/LSEQ combination, then the output titled
"force distribution in bar elements" will be wrong.
Avoidance: There are two possible avoidances:

1. Replace the CBAR element with a CBEAM element.

2. Replace the PLOAD1/LOADSET/LSEQ combination with equivalent point
loads using the DLOAD/TLOADi/DAREA
combination.
@
7518; 70.5; Post-processing - PARAM,POST,-5, PARAM,OUGCORD,BASIC - Missing O
utput
If PARAM,POST,-5 is specified with PARAM,OUGCORD,BASIC then PARAM,POST,-5
is effectively ignored and the expected output will be missing.
Avoidance: Insert following DMAP alter:
compile sedrcvr
alter 'ELSE IF ( NOT(post',''
ELSE IF ( post>-3 and PATVER<3. and OUGCORDX='BASIC' ) THEN $

@
7539; 69; Optimization/Design Sensitivity - PARTN - UFM 2166
If a statics and a modes subcase are attempted in the same geometric shape
optimization run, then the following fatal message is issued, "USER FATAL
MESSAGE 2166, MATRIX TO BE PARTITIONED IS OF SIZE 22806 ROWS BY 7 COLUMNS.
ROW PARTITION SIZE IS 22806 COLUMN PARTITION SIZE IS 5 (INCOMPATIBLE)."
Avoidance: Include the following DMAP avoidance:
compile bndshp nolist noref
alter 'shpcas'
type parm,,i,n,nt $
MDCASE CASEXXu/
CASESTAT,,,,,,,,,,,,,/
S,N,STATCC $
tabpt casestat/ $
paraml casestat//'trailer'/1/s,n,nstat $
 paraml ygbndr//'trailer'/1/s,n,nyg $
nt=nstat+nyg $
delete /cvec,,,, $
matgen ,/cvec/6/nt/nstat/nyg $
@
7542; 69; Optimization/Design Sensitivity - DRESP1, DOM9 - Unexpected EOF
in DOM9
If a data deck has multiple weight or multiple volume responses that are
referenced by the objective or are
constrained, unexpected results occur and the answers are wrong. System
Fatal Messages such as "Unexpected end of file in DOM9" may also occur.
Avoidance: Remove the multiple DRESP entries.
@
7546; 69; Optimization/Design Sensitivity - DRESP1 - SFM 7127
In an optimization run, if a series of DRESP1 entries makes alternating
references between different types (such as PSHELL and ELEM) in the PTYPE
field, then the message "SYSTEM FATAL MESSAGE 7127 (DOPR3B) AN ELEMENT
WHICH REFERENCES PTYPE **** ** PID *** IS NOT DEFINED IN THE ANALYSIS
MODEL" is issued.
Avoidance: Put all DRESP1's that reference a particular type in
consecutive order, where the order is given by the DRESP1 ID.
@
7547; ; Documentation - Slip Ratio - Reference Manual
Documentation that defines the meaning of "Slip Ratio" for gap and
slideline elements does not exist. The following paragraph should be
added to the Section 15.3.12 of the Reference Manual.
The slip ratio is defined as the (shear force)/(coefficient of friction *
normal force). Thus, a ratio of 0.0 indicates no shear (or no contact), a
slip ratio of 1.0 indicates slipping, and a value in between 0.0 and 1.0
is a normalized value which indicates how close to impending slipping the
shear force is."

@
7550; 69; Optimization/Design Sensitivity - PARAM,MODEPT - Parameter not s
et Correctly
The MODEPT parameter in SOL 200 is intended to indicate whether the design
model EPT (element property table) differs from the analysis model. There
are a number of implementations errors associated with this parameter:
The DOPR1 module has an error which results in the MODEPT parameter being
set to true only when the FINAL
designed property differs from the analysis value.
The DESOPT driver does not correctly pick up the MODEPT value set in
subDMAP PREDOM because of an error
in the call to PREDOM.
The design intent should be that the EPTN datablock is only equivalenced
to EPT when MODEPT is true. Currently it is always equivalenced with the
results that the finite element analyses on the
database gets destroyed on a restart when they should not.
@
7552; 69; Installation - MSC/NASTRAN will not make the icons on Beta NT sy
stems
For Windows NT Version 4.0, which is the Windows 95 interface on top of
Windows NT, the installation of MSC/NASTRAN will not create the icons or
group for the Beta versions of this Operating System.
Avoidance: Do not upgrade your operating system to 4.0 until the non
Beta software is available. Or manually create the group (MSC) and create
the icons in the group.
@
7553; 70; Restarts - - PBCOMP - SWM 2072 CARD TYPE 3002
If a model contains a pbcomp entry, and a restart is performed going from
SOL 103 to SOl 111 on a DEC
AlPHA OSF machine, then the job may terminate with the following messages
in the f06 file.
SWM 2072 - CARD type 3002
On some DEC ALPHA machines, no messages may appear in the f06 file. In
all cases, the following messages appear in the log file:
time: command terminated abnormally
Avoidance: There are two possible avoidances:
1. Perform the analysis in one pass using solution 111 without using
solution 103 for restart.
2. Replace the PBCOMP with an equivalent PBEAM entry.
3. Replace the PBCOMP record in the NDDL with the following:
RECORD=PBCOMP(5403,55,349),
PID,I, MID,I, A,RS, I1,RS, I2,RS, I12,RS, J,RS, NSM,RS, K1,RS,
K2,RS, M1,RS, M2,RS, N1,RS, N2,RS, nsect(c),i,
either,back(nsect)=0,
undef,20,
or,1,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,2,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,3,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,4,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,5,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,6,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,7,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
 endentry,count=nsect,
or,8,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,9,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,10,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,11,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,12,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,13,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,14,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,15,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,16,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,17,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,18,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,19,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
or,20,
entry=section, Y,RS, Z,RS, C,RS, MID,I, undef,1,
endentry,count=nsect,
endeith,
EOR,
@
7555; 68.2.9; Miscellaneous - PUNCH - SFM 4276, Invalid Floating Operation
A request for PUNCH output may result in the following traceback on
Fujitsu computers running MSC/NASTRAN Version 68.2:
jwe0292i-e Invalid floating operation exception was detected.
execution mode = advanced
error occurs at ofppun_ line 156 loc 00857c10 offset 00001800
ofppun_ at loc 00856410 called from loc 0083be68 in ofp_
line 1340
ofp_ at loc 008366f0 called from loc 00006eb8 in xsem14_
line 124
xsem14_ at loc 00006850 called from loc 00000928 in MAIN__
line 108
MAIN__ at loc 00000500 called from o.s.
taken to (standard) corrective action, execution continuing.
 Avoidance: The error can be avoided by setting system(210)=1. It is
recommended that this system cell be set in the system rc file of all
Fujitsu systems. Please be aware that other machines have experienced
problems when system(210) is set. These problems are very rare however,
and are not expected to effect the Fujitsu system.

@
7556; 70; Documentation - Frequency Response - Response of Rigid Body Mode
s
When an excitation is made at zero frequency for a model with rigid body
modes, the contribution from the rigid body modes (which is theoretically
infinite) is discarded. Only the finite modal displacements from modal
DOFs with non-zero natural frequencies are retained. An xy plot of
element stresses from zero to small excitation frequencies will be
continuous. An xy plot of displacements will grow rapidly as zero
excitation frequency is approached, then drop to values of the same
magnitude as response at frequencies away from natural frequencies. The
solution at zero frequency is analogous to an inertia relief solution,
while the other solutions are made in an inertial basis. The following
remark should be added under the FREQi entries:
Remark 4. In modal analysis, solutions for modal DOFs from rigid body
modes at zero excitation frequencies may be discarded. Solutions for non-
zero modes are retained
@
7557; 70; Documentation - PBCOMP, I12 - Incorrect Default
The default value for the I12 field of the PBCOMP entry is 0.0 and not
I1*I2-(I12)**2 as described in the MSC/NASTRAN Quick Reference Guide. The
correct description for I12 is as follows:
I12 Area product of inertia. See Remark 6. (Real, default=0.0,
but I1*I2-I12**2 > 0.0).
@
7561; 70; Documentation - BUFFPOOL - UFM 1286
Remark 2d in the section "How Can I Reduce the Size of the SCRATCH DBset?"
of the Common Questions and Answers Guide, Second Edition, should be
changed from

d. Increase the available memory to the functional modules by setting the

"smem" and BUFFPOOL parameters
to zero......
to
d. Increase the available memory to the functional modules by setting the
"smem" parameter to zero......

Starting in Version 68, the default solver is the sparse solver. BUFFPOOL
cannot be turned off with the
sparse solver.
@
7564; 69; Frequency Response/Random Analysis - RANDOM, K2PP - SFM 4276 FRE
AD
When K2PP matrices or FREQ looping are present in random analysis, the
RANDOM module fails.
The problem has been identified as the wrong case control data block being
passed to the RANDOM module. The RANDOM module should get CASEXX instead
of CASECC. The following is a DMAP avoidance.
compile sol30 or SOL26
alter 'random',''
RANDOM XYCDB,DIT,PSDL,OUGV2,OPG2,OQG2,OES2,OEF2,CASExx/PSDF,AUTO/
S,N,NORAND $
@
7583; 68.2.8; Miscellaneous - INPUTT4 - Various Fatal Messages
MSC/NASTRAN may not be able to read INPUTT4 files on certain machines.
Known problem machines are Fujitu and DEC ALPHA. The Fujitsu systems gave
the following traceback in the INPUTT4 module:
error occurs at pack_ line 113 loc 00e577b0 offset 00000360
pack_ at loc 00e57450 called from loc 00712bb0 in inptt4_
line 359
inptt4_ at loc 00711580 called from loc 000024c0 in xsem02_
line 125
xsem02_ at loc 00001e60 called from loc 00000868 in MAIN__
line 84
MAIN__ at loc 00000500 called from o.s.
error summary (fortran77 ex system)
While the DEC ALPHA OSF system completed but issued the following
messages:
Unaligned access pid=15155 <analysis> va=14011d924 pc=1201febe0
ra=1201feb4c type=ldt
Unaligned access pid=15155 <analysis> va=14011d924 pc=1201febe0
ra=1201feb4c type=ldt
@
7591; 70; Databases - MATREDU - Wasted Space in DBALL
The module MATREDU is using the DBALL as its scratch space instead of
SCR300 or the regular SCRATCH space.
Avoidance: There is no avoidance.
@
7599; 69; Documentation - PARAM,TSTATIC - Can Be Used in Case Control
The MSC/NASTRAN Version 68 Quick Reference Guide states that param,tstatic
can only be used in the Bulk Data section. It can be used in the Bulk
Data and Case Control.
@
7600; 70.5; Complex Eigenvalue Analysis - EIGC, CLAN - Missing Roots
A complex lanczos eigenvalue analysis of a model with zero damping will
likely not find all the roots that are found by a real lanczos eigenvalue
analysis with the same shifts. Likewise in models with damping, complex
 lanczos is likely to have missing roots.
Avoidance: The only avoidance is to add additional shift points. In
Version 69.1 the epsilon has been adjusted so that correct behavior should
be achieved for real world models. Thus, a Version 69.1 run is likely to
find more roots then the same run in 69.0.
@
7615; 70.5; DMAP - ALTER - UFM 3063
If a DMAP contains a long alter followed by short alters followed by a
long alter the following message may be encountered:
SYSTEM FATAL MESSAGE 3063 MODULE = PREFACE
MEMORY ALLOCATION FAILURE IN SUBROUTINE XALTER ERROR CODE: -1
MEMORY CORRUPTION FOUND. PREVIOUSLY ALLOCATED MEMORY AREAS HAVE BEEN
INCORRECTLY MODIFIED.
USER ACTION: CONTACT MSC CLIENT SUPPORT.
Avoidance: There are two avoidances. The user can either re-arrange the
alters so that the long and short alters are placed together or split
alters that are more than 112 lines into smaller segments.

@
7616; 69; Aeroelastic Analysis - APD, PFAERO - No Message, Wrong Answers
Specifying a theta-1 array and no theta-2 array for an aerodynamic
interference body may yield results that do not include all the
interference body effects. For example, if a PAERO2 specifies that an
interference body has 3 elements, of which elements 1 and 2 specify a
theta-1 array, and element 3 does not, the program will run and yield
wrong results.
Avoidance: The wrong results can be avoid by properly specifying all
interference elements with a theta array.
@
7617; ; Loads - PLOAD2, PLOAD4 - SFM 2105, SFM 4577, Missing Elements
If PLOAD2 or PLOAD4 entries specify a range of element ids with the "thru"
option and an element does not exist for each id in the range, the job
will fail with SYSTEM FATAL MESSAGE 2105 and SYSTEM FATAL MESSAGE 4577.
This limitation is not clearly documented and limits the usefulness of the
thru option and should be corrected.

@
7619; ; Documentation - Nonlinear Analysis Data Recovery Restarts
The current documentation does not make it clear how one should prepare
input files to perform data recovery restarts with a
nonlinear analysis. If the user expects to obtain any data recovery from
solution 106, 129, 153 or 129, the input file should be prepared to
perform the complete solution and not rely on the database to contain the
necessary datablocks. For example, flux output must be explicitly
requested in the cold start if one wants to obtain it on the restart. The
user may specify flux(plot) if he chooses not to see the printed output in
the cold start.
@
7627; 69; Heat Transfer Analysis - VIEW - Poor Performance, SFM 4276
Heat Transfer analysis that involves third body shadowing calculations for
 a large number of surface element in the VIEW module may experience poor
performance or fail with the message SYSTEM FATAL MESSAGE 4276.
Avoidance: There is no guaranteed avoidance; however, if parallel
processing is being selected, deselecting parallel may avoid the problem.
Furthermore, even though the error is possible on all machine types a
specific problem may run on one machine but not another. Thus, running on
another machine may avoid the problem.
@
7632; 70.7; Design Sensitivity - SENSITY, SOL 103 - UFM 4287
Material sensitivity analysis in SOL 103 fails on Cray computers with the
following message, "USER FATAL MESSAGE 4287 , ILLEGAL THICKNESS FOR PLATE
ELEMENT ID = xxx".
Avoidance: Use SOL 200.
@
7633; 69; Optimization/Design Sensitivity - Premature Convergence for Opti
mization
As explained in Section 3.4 of the MSC/NASTRAN User's Guide for Design
Sensitivity and Optimization, a design task is considered converged if the
objective is not changing and the design is feasible. This has been found
to be sometimes unconservative in the case where the previous design was
unfeasible. That is, if the design changes from being infeasible to
feasible without a weight change, it may be that further iterations will
produce a further improvement in the objective.
Avoidance: If the next to last design was infeasible, as indicated by its
maximum constraint value being
greater that GMAX, it is worthwhile to start a new design task starting
from the final design of the
run using the DESVAR values from the punch file.
@
7637; 69; Case Control - ECHO=NOSORT - Missing Output
If ECHO=nosort is used in the case control section, bulk data echo is not
printed to the output.
Avoidance: use ECHO=unsort.

@
7644; GL; Transient Heat Transfer - CELAS2, TEMPBC,TRAN - UFM 315
In transient thermal analysis, if the user has CELAS2 entries in the input
test deck and there is also TEMPBC,TRAN bulk data entries, then the
following fatal message will be encountered:

0*** USER FATAL MESSAGE 315 (IFPDRV)

FORMAT ERROR ON BULK DATA ENTRY TEMPBC 102
SORTED CARD COUNT = 761

This error occurs because the TEMPB,TRAN bulk data entry internally
produce CELAS2 entries for the large conductance method. These entries
then conflict with explicitly defined CELAS2 entries.
 Avoidance: Instead of CELAS2 bulk data entry uses CELAS1,and PELAS to
define the thermal conductance value.
@
7655; 70.6; Nonlinear Eigenvalue Analysis - PARAM,NMLOOP - Missing Eigenvect
ors
If a nonlinear normal modes analysis (param,nmloop) is requested in a SOL
106 run, and there are more than one subcase and the displacement output
is not requested for all the subcases, then the nth eigenvector will be
missing for the nth subcase that does not have a displacement output
request. In the following example, eigenvectors 1,3,4,5,6,7,8,9, and 10
will be printed. Eigenvector number 2 will be missing.
sol 106
.
subcase 1
load 1
nlparm = 10
disp = all
subcase 20
load = 2
nlparm = 20
subcase 35
load = 3
nlparm = 30
disp = all
Begin bulk
.
param,nmloop,20
eigrl,10,,,10
.

Avoidance: Request displacement output for all subcases.

@
7659; 69.1; DMAP - COMPILER - Various Failures
If an "ELSE" statement appears in a DMAP without a matching set of IF and
ENDIF statements, various random failures may occur. In one instance, a
"FILPOS" error exit was taken upon entering the subdmap where the error
was made. In another instance, a USER FATAL MESSAGE was issued indicating
that there was "NO MASS MATRIX AVAILABLE FOR EIGENVALUE ANALYSIS"."
Avoidance: There is no avoidance only the recommendation that the
statements inside if-endif clauses be indented to insure that all else
statements are bracketed by if-endif statements. DIAG 44 is also useful
for pinpointing the location of the error when this type of error is
suspected.

@
7660; ; P-Elements/Superelements - GP4 - UFM 2101A
The following error message will be issued if a grid point appearing on a
p-element edge is constrained and
not in an upstream superelement.
*** USER FATAL MESSAGE 2101A MODULE = GP4
GRID POINT xxxx COMPONENT y ILLEGALLY DEFINED IN SETS UM UB
 For superelements, there is a general restriction in p-elements for
boundary degrees of freedom ( exterior degrees of freedom in
superelements). That is there is an internal MPC created for SPCs on edges
and faces (GMSPC, GMBC). Consequently, the constrained degrees of freedom
are subject to the rules for MPCs. The dependent degrees of freedom
(m-set) have to be in the upstream superelement, they can not be exterior.
@
7668; 70; Documentation - PCOMP, TREF - Missing Default
The PCOMP bulk data entry description in the MSC/NASTRAN Quick Reference
Guide does not give the default for TREF field. The default is "0.0".
In addition, remark 3 would be clearer if instead of:
3. TREF given on the PCOMP entry will be used for all plies of the
element; it will override values supplied on material entries for
individual plies. If the PCOMP references temperature dependent material
properties, then TREF given on the PCOMP will be used as the temperature
to determine material properties. TEMPERATURE Case Control commands are
ignored for deriving the equivalent PSHELL and MATi entries used to
describe the composite element.

it said the following:

3. The TREF specified on the material entries referenced by plies are not
used. Instead TREF on the PCOMP entry is used for all plies of the
element. If not specified, it defaults to "0.0".
If the PCOMP references temperature dependent material properties, then
the TREF given on the PCOMP will be used as the temperature to determine
material properties. Note that MAT8 data entries do not support
temperature dependent material properties.
TEMPERATURE Case Control commands are ignored for deriving the equivalent
PSHELL and MATi entries used to describe the composite element.
@
7678; 69; Matrix Methods - Sparse DECOMP - UFM 6137 RANK DEFICIENT
Sparse decomposition in Version 69 has been changed to all rank deficient,
singular matrices, to go through. The action taken is now the same as the
old active column decomposition, i.e., it puts a very small term on the
diagonal and proceeds with the solution. The user is then warned about
the degrees of freedom causing the high maxratio.

@
7680; 69.1; Rigid Elements and Multipoint Constraints - RSSCON - Wrong Answe
rs
The RSSCON element produces wrong answers if the shell grid is not on the
straight edge connecting the two solid grids. This is evident because the
element in this condition will not pass a rigid body check.
The magnitude of the error is related to the magnitude of the offset. In
Version 69 an offset up to 5% of the height of the solid is allowed. The
mpc relations generated by the element are adjusted as if the grid point
was moved onto the straight line but the actual geometry is not adjusted.
 Avoidance: Other than manually moving the grid points so that they line
along the straight line there is no avoidance.
In V69.1, the shell grid point is moved on to the straight edge connecting
the two solid grids if the shell grid lies close to the edge within a
tolerance. The tolerance is user modifiable with PARAM, TOLRSC, eps where
eps is 0.05 by default. The absolute value of the tolerance is then eps
times the length of the solid edge. Shell grid points outside the
tolerance will cause a fatal error.
@
7682; 69.0.2; Elements - RSSCON - UWM 6695, References Incorrect GRID IDs
If a shell grid point lies outside the straight line between the two
connecting solid grid points when using the RSSCON element, then the
following warning messages will be issued.
*** USER WARNING MESSAGE 6695 (CONSSD )
RSSCON NUMBER aaa WILL MPC ONLY GRID bbb OF QUAD ELEMENT ccc TO
HEXA ELEMENT ddd
IT WILL NOT MPC GRID ddd; THE CORNER SHELL GRIDS MUST LIE ON OPPOSITE
EDGES OF A FACE
OF THE SOLID ELEMENT.
This is the proper message to issue; however, the error message may
reference an incorrect offending shell grid point "bbb".
Avoidance: Modify the shell grid point to lie on a straight line between
the two connecting solid grid points.
@
7689; 69.0.2; Optimization - PBARL - Various Fatal Messages
The new capability in MSC/NASTRAN Version 69 which allows users to define
their own beam cross sections has two undocumented limitations. There is
a limit of 250 user defined beam cross section types, and there is a limit
of 7 dimensions, including the Non-Structural Mass (NSM) on user defined
sections. If either of these limits are exceeded various error types may
be generated. In optimization, the fatal message may be issued from the
module DOPR1H, or it may be issued from the API with the phrase buffer
overflow.
Avoidance: Do not exceed the limits. The limits have been raised to 2000
beam cross section types and unlimited dimensions on a beam section for
the next release.

@
7690; 70.5; Documentation - DMAP Module Dictionary - MATMOD Option 5
Example 3 of the MATMOD option 5 documentation is incorrect. This example
shows how to transform KGG from a global to a basic coordinate system.
Following the documentation results in the wrong matrix. In the SMPYAD
shown, the last and not the first TRANSG matrix should be transposed.
However, SMPYAD will not let you transpose the last matrix of a triple
multiply, so you must transpose TRANSG before SMPYAD. The incorrect and
corrected documentation follow:
Incorrect:
 matmod cstm,sil,bgpdt,,,/transg,/5//-1 $
smpyad transg,kgg,transgt,,,/kggbasic/3////1 $
Correct:
matmod cstm,sil,bgpdt,,,/transg,/5//-1 $
trnsp transg/transgt $
smpyad transg,kgg,transgt,,,/kggbasic/3//// $ No transpose flags set
@
7692; 70; Superelements - MATGPR - Overflow (******)
When using the enhanced superelement features (parts), the boundary points
of the upstream superelements may be sent downstream. Internally, these
boundary points are assigned very large ID numbers, numbers that are much
larger than the user can specify on the grid entries. These large numbers
do not fit in the format for the MATGPR module. If MATGPR is used with
the enhanced superelement, the IDs for these grids will be printed
as (*******). The matrix values are, however, correct.
Avoidance: Use MATPRN instead of MATGPR and use PARAM,USETPRT and the
boundary sequence table to identical the external degree of freedom.
@
7697; 70; Heat Transfer - SPC - Poor message, UFM MATMOD-4
Currently the transient heat transfer analysis capability does not support
multiple SPC conditions. An analysis can only have one SPC condition
apply through all subcases. If multiple boundary conditions are attempted
the message "USER FATAL MESSAGE MATMOD-4" will be issued.
Avoidance: Multiple boundary conditions must be executed in separate runs.
@
7699; 69.0.2; P-Elements - GP0, GMLOAD, FEFACE - UFM 6380
If the GMLOAD entry references an FEFACE entry that is undefined the
following fatal message is correctly issued but the FEFACE ID is not
properly identified.
USER FATAL MESSAGE 6380, (GP0F),
THE FEFACE ID = 0 REFERENCED BY GMLOAD IS UNDEFINED.
Avoidance: There is no avoidance to make the code identify the incorrect
entries. The user must find and correct the input error.
@
7705; 70; Documentation - EIGRL - Performance/Crashes
The MSC/NASTRAN Version 69 Quick Reference Guide description of the EIGRL
entry documents a new parallel Lanczos capability. This new capability
was not documented in the Version 69 Release Guide because it did not meet
MSC's strict standards for quality. The capability and documentation has
several errors:
1. The Quick Reference Guide fails to mention that the new capability is
only available on the CRAY and SGI
systems.
2. The following FATAL messages have occurred while using this capability:

*** SYSTEM FATAL ERROR 4276.SFUPFI

*** USER FATAL MESSAGE 815 (RDINIT)
 (Note that bad answers have never been encountered as a result of this
capability.)
3. The purpose of this capability was to reduce the elapsed time. It has
been found that parallel lanczos helps reduce elapsed time on a very
small set of problems, and that elapsed time actually INCREASES for
many runs.
Avoidance: Do not set the NUMSEG keyword (a.k.a. SYSTEM(197). Do not set
NUMSEQ on the EIGRL entry.
@
7706; 69.1; SSSALTER - evds200a.v69 - UFM 149
If the evds200a.v69 sssalter is used, the job will fail with the following
messages:

*** USER FATAL MESSAGE 149 (LINKAR )

CALL STATEMENT IFPS IN SUBDMAP IFPL HAS AN ARGUMENT LIST OF DIFFERENT
LENGTH THAN THE CALLED SUBDMAP

*** SYSTEM INFORMATION MESSAGE 480 (SEMINT)

JOB TERMINATED DUE TO ABOVE ERRORS IN THE DMAP/NDDL LINKAGE EDITOR
(LINKER)
Avoidance: Replace the following statements:
type parm,,logical,n,noedom2,ambs9,nogoifp9,ngoxsrt9
type parm,,char8,n,dball9
$
call IFPS ,,bulk2/BULK9,
$ GEOM1Q,EPT,MPT,EDT,DIT,DYNAMIC,GEOM2,GEOM3,GEOM4,
GEOM1Q9,,MPT9,EDT9,DIT9,DYNAMIC9,GEOM29,GEOM39,GEOM49,
$ MATPOOL,AXIC,DEFUSET,DMI,DMINDX,DTI,DTINDX,PVTX,
MATPOOL9,AXIC9,DEFUSET9,DMI9,DMINDX9,DTI9,DTINDX9, ,
edom2,DEQATN9,DEQIND9,CONTACT9,OINT9,EPTA9/
DBALL9/AMBS9/SEID9/AUXMID9/NOGOIFP9/NGOXSRT9/noedom2 $
$

with

type parm,,logical,n,noedom2,ambs9,nogoifp9,ngoxsrt9,selist9
type parm,,char8,n,dball9
$
call IFPS ,,bulk2/BULK9,
$ GEOM1Q,EPT,MPT,EDT,DIT,DYNAMIC,GEOM2,GEOM3,GEOM4,
GEOM1Q9,,MPT9,EDT9,DIT9,DYNAMIC9,GEOM29,GEOM39,GEOM49,
$ MATPOOL,AXIC,DEFUSET,DMI,DMINDX,DTI,DTINDX,PVTX,
MATPOOL9,AXIC9,DEFUSET9,DMI9,DMINDX9,DTI9,DTINDX9, ,
edom2,DEQATN9,DEQIND9,CONTACT9,OINT9,EPTA9/
 DBALL9/AMBS9/SEID9/AUXMID9/NOGOIFP9/NGOXSRT9/noedom2/
selist9 $
$

@
7707; 69.1; Nonlinear Analysis - CBUSH - Incorrect SPCFORCE Output
Nonlinear static analysis does not include the effect of the CBUSH element
forces in the forces of single point constraint output.

Avoidance: There is no avoidance.

@
7711; 70; Frequency Response - SEUPPLOT - SFM 4276, Time Command Terminate
d Abnormally
IF an SEUPPLOT request is made in solution 108, the job may fail with the
message time command terminated abnormally or, SYSTEM FATAL MESSAGE 4276.
Avoidance: Use SEPLOT instead of SEUPPLOT.
@
7713; 69; P-elements - FEFACE, CIDBC, MODGM4 - Stops With No Message
In MSC/NASTRAN Version 68.2 on the DEC/ALPHA, if a secondary coordinate
system is referenced on a FEFACE (CIDBC field), then the job may terminate
in the MODGM4 module without any error messages. Another symptom of this
error is the generation of a core file in the submittal directory.
Avoidance: If feasible, do not reference a secondary coordinate system on
the CIDBC field.
@
7715; 69.0.2; P-Elements - GMINTC, GMBNDC - SFM 6425 (EMA)
On machines that enforce boundary alignment such as the HP 9000 the
following fatal message may be issued when the GMINTC entry is used:
*** SYSTEM FATAL MESSAGE 6425 (EMA).
THE KDICTP TABLE AND KDICTP TRAILER ARE INCONSISTENT.
USER ACTION: CONTACT MSC CLIENT SUPPORT.
This error does not occur on the IBM/RS6000 or the CRAY machines.
Avoidance: There is no avoidance.
@
7716; 69.0.2; Miscellaneous - SYSTEM(210)=1 - Excessive Output
SYSTEM(210) can be used to correct punch errors on many systems; however
when SYSTEM(210)=1 is used on the Sun/Solaris machines, excessive output
is generated causing the .f06 and/or .pch file to be very large.
Avoidance: There are no known problems that require Sun/Solaris systems to
set system(210)=1. Thus, on this system do not specify SYSTEM(210)=1.
@
 7717; 69.1; Database - Neutral File Transfer with p-elements, GP2 - SFM 3001
The following message is issued when an attemt is made to load the neutral
database (created by dbunload) via sol DBTRANS. This error occurs only if
the database datablocks are related to p-element analysis.
SYSTEM FATAL MESSAGE 3001 MODULE = GP2
THE INPUT DATA BLOCK NAMED EQEXINS AT POSITION 2 DOES NOT EXIST.
Avoidance: There is a rather complex avoidance which requires an NDDL
recompilation; however, it is not deemed practical.
@
7718; 69.1; Virtual Mass - MPRES - Wrong Answers, SFM 4276 EC 7777
If MPRES output is requested and PARAM,DDRMM,-1 is not specified wrong
answers or a SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777 may appear.
Avoidance: In Version 68 simply specify PARAM,DDRMM,-1. For Version 69
the following alter from the sssalter library will improve efficiency and
correct the error.
$ FILE VMEA.V69
$ LAST REVISED 17 DEC 96 M. A. GOCKEL
$ VIRTUAL MASS EFFICIENCY IMPROVEMENTS, FOR SOLS 103,111,112.
$ FOR NAST69 DATED 10/18/96, WITH DEL. DB TIME STAMP OF 961018160119
ECHOOFF $
COMPILE SEMRM nolist NOREF $
ALTER 'CALL MGEN1','' $
$ MODERNIZE NAMES
EQUIVX MCHI/CEE/ALWAYS $
EQUIVX MCHI2/BEC/ALWAYS $
EQUIVX MLAM/MEE/ALWAYS $
EQUIVX MLAM2/MEC/ALWAYS $
EQUIVX GEG/GGE/ALWAYS $
CALL DBSTORE CEE,MEE,BEC,MEC,MAR//SEID/0/' '/0 $ FOR MGEN2
CALL FDRMGEN1 USET,GM,GOA/RGA $
IF (NOA>-1) THEN $
MPYAD GGE,RGA,/GEA/1 $
ELSE $
TRNSP GGE/GEA $
ENDIF $
VECPLOT ,,BGPDTS,EQEXINS,CSTMS,,,,/RBGRT/GRDPNT//4 $ RB VECTORS FROM
$ GEOMETRY. For weight summary. Always made, but summary output
$ must be requested by parm, grdpnt
TRNSP RBGRT/RBGR $ RBGR HAS G ROWS, 6 COLUMNS
MPYAD GGE,RBGR,/GER/1 $
CALL DBSTORE GEA,GER,,,//SEID/0/' '/0 $ FOR MODERS
IF (SEID>0 AND NOOSET>-1) THEN $
CALL MGEN2 CASES,/MVMAA/SEID $
ADD MLAA1,MVMAA/MLAA $
ENDIF $
COMPILE MODERS nolist NOREF $
ALTER 'REIGL KXX'(,1) $ AFTER ENDIF FOR READ OR REIGL
CALL MGENRS PHIx,USET,LAMA,CASES// $ PHIAZ AND LAMA ARE UPDATED
COMPILE SEDISP NOLIST NOREF $ REMOVE FOR V69.1
ALTER 293,299 $ OLD MGEN BRANCH MOVED TO SEDRCVR
COMPILE SEDRCVR noLIST NOREF $ REVISE FOR V69.1
ALTER 43 $ MOVE MDATA FROM SEDISP, AFTER SDR2 FOR NON-DDRMM
$ CHECK FOR VIRTUAL MASS PRESSURE OUTPUT REQUESTS. LIFTED FROM NAST69T2
$ REMOVE MOST OF THIS WHEN NAST69T1 SUPPORTED
CALL DBFETCH /MEA,MAR,,,/SEID/0/0/0/0 $ AVOID CALL CHANGES
DBSTATUS Mea//S,N,NOMGEN $
IF ( NOMGEN>0 AND (APP='FREQRESP' OR APP='TRANRESP' OR
APP='REIG ' OR APP='CEIGEN ') ) THEN $
PARAML USET//'USET'//////'A'/S,N,NOASET $
IF ( LUSETS=NOASET ) THEN $
EQUIVX UG/UL1/-1 $
ELSE $
UPARTN USET,UG/UL1,,,/'G'/'A'/'S'/1 $
ENDIF $
MDATA CASEDR,,MAR,MEA,UL1,OL2/OEP1/APP $
ENDIF $ NOMGEN>0
ALTER 166,166 $ ADD OEP1
SDR3 OUG1F,OQMG1 ,OEP1,,,/OUG2F,OQMG2,OEP2,,, $
ALTER 226 $ INSERTIONS
OFP OEP1//S,N,CARDNO $
ALTER 184 $ BEFORE OTHER SORT2 OFPS
OFP OEP2//S,N,CARDNO $
ALTER 196 $
IF ( NOMGEN>0 AND (APP='FREQRESP' OR APP='TRANRESP' OR
APP='REIG ' OR APP='CEIGEN ') AND
APP1<>'MMREIG' ) THEN $
XYTRAN XYCDBDR,,,,OEP2,/XYPPLOT/
APP/'DSET'/S,N,PFILE/S,N,CARDNO/S,N,NOPPLOT $
IF ( NOPPLOT>-1 ) XYPLOT XYPPLOT// $
ENDIF $ NOMGEN>0
COMPILE FDRMGEN1 NOLIST NOREF $
SUBDMAP FDRMGEN1 USET,GM,GO/RGA $
TYPE PARM,,I,N,NOOSET,NOMSET,NOA,NOSSET $
$
$ PRODUCES REDUCTION FROM G-SET TO F-SET
$
PARAML USET//'USET'//////'G'/S,N,NOGSET/'A'/S,N,NOASET/
'F'/S,N,NOFSET/'N'/S,N,NONSET/'O'/S,N,NOOSET $
IF (NOGSET=NOASET) NOA=-1 $
IF ( NOA=-1 ) THEN $
MATGEN, /RGA/1/NOGSET $
RETURN $
ENDIF $
IF ( NOMSET>-1 ) THEN $
MATGEN ,/INN/1/NONSET $
UMERGE1 USET,GM,INN,,/RGN/'G'/'M'/'N'/1 $
ENDIF $ NOMSET>-1
$
IF ( NOSSET>-1 ) THEN $
MATGEN ,/IFF/1/NOFSET $
UMERGE1 USET,,IFF,,/RNF/'N'/'S'/'F'/1 $
ENDIF $ NOSSET>-1
$
IF ( NOMSET>-1 AND NOSSET>-1 ) THEN $
MPYAD RGN,RNF,/RGF $
ELSE IF ( NOMSET>-1 ) THEN $
EQUIVX RGN/RGF/-1 $
ELSE IF ( NOSSET>-1 ) THEN $
EQUIVX RNF/RGF/-1 $
ENDIF $
IF ( NOOSET>-1 ) THEN $
MATGEN ,/IAA/1/NOASET $
UMERGE1 USET,GO,IAA,,/RFA/'F'/'O'/'A'/1 $
 IF ( NOMSET>-1 OR NOSSET>-1 ) THEN $
MPYAD RGF,RFA,/RGA $
ELSE $
EQUIVX RFA/RGA/-1 $
ENDIF $
ELSE $
EQUIVX RGF/RGA/-1 $
ENDIF $ NOOSET>-1
RETURN $
END $ FDRMGEN1
COMPILE MGEN2 nolist NOREF $
SUBDMAP MGEN2 CASES,PHIAZ/MavZZ/SEID $
$ COMPUTES VIRTUAL MASS IN REDUCED BASIS, ALSO RIGID BODY MASS OUTPUT
$ OLD NOTATION NEW NOTATION
$ MCHI CEE
$ MCHI2 BEC
$ MLAM MEE
$ MLAM2 MEC
TYPE PARM,,I,N,NF,N1,N2,NX,NZP6,ny,nw $
TYPE PARM,,I,Y,GRDPNT=-1 $
TYPE PARM,,RS,Y,WTMASS=1.0 $
TYPE PARM,,CS,N,MASSWT $
TYPE PARM,NDDL,I,N,SEID $
type parm,,char8,y,solvit='no' $
CALL DBFETCH /cee,,,,/seid/0/0/0/0 $
paraml cee//'presence'////s,n,nocee $
if (nocee=-1) return $
CALL DBFETCH /ger,MEE,BEC,MEC,GEA/seid/0/0/0/0 $
PARAML PHIAZ//'PRESENCE'////S,N,NX $
PARAML CASES//'DTI'/-1/194//S,N,MPRESID $ MAKE MEA ONLY IF MPRES
REQUESTED
$ this logic somewhat flawed for superelements. if mpres anywhere,
$ it goes to all se subcases, including se0. It shouldn't.
IF (NX>-1) THEN $
MPYAD GEA,PHIAZ,/PRHEZ $
ELSE $
EQUIVX GEA/PRHEZ/ALWAYS $
ENDIF $
PARAML CEE//'TRAILER'/1/S,N,NE $ NUMBER OF FLUID ELEMENTS
PARAML BEC//'PRESENCE'////S,N,NC $ NUMBER OF CAVITIES
PARAML PRHEZ//'TRAILER'/1/S,N,NZ $ NUMBER OF MODES or a-set
PARAML GEA//'TRAILER'/1/S,N,NA $ SIZE OF A-SET
NF = NE $
equivx prhez/pez/always $ temp
EQUIVX PRHEZ/PRHFZ/NC $
EQUIVX GER/GFR/NC $
append pez,ger/pey $
Ny = NZ + 6 $
IF (MPRESID<>0) THEN $ APPEND GEA TO THE MODAL SHAPES
nw = Ny + NA $
MATGEN ,/vwya/6/Nw/Ny/NA $
APPEND Pey,GeA/Pew $
ELSE $
EQUIVX Pey/Pew/ALWAYS $
ENDIF $
IF (NC>-1) THEN $ CLOSED VOLUME OPTION
PARAML BEC//'TRAILER'/1/S,N,NC $ NUMBER OF CAVITIES
NF = NE + NC $
N1 = NE - NC $
N2 = 2*NC $
 MATGEN, /VFEC/6/NF/NE/NC $
CALL MAKVE1C BEC/VE1C/NE/N1/NC $
MATGEN, /VC/6/NC/0/NC $ UNIT VECTOR
MATGEN, /V2CC/6/N2/NC/NC $
MERGE VE1C,VC,,,,VFEC/VF12/0 $
partn pew,,ve1c/p1w,pcw,,/0 $
nx = 2 * nc $
matgen,/icc/1/nc $
ENDIF $
IF (NC=-1) THEN $ DO THE FREE SURFACE METHOD
if (solvit><'yes') then $
putsys (0,209) $ use old solver
solve cee,pew,,,/uew $
putsys (1,209) $ restore new solver
else $
matmod pew,,,,,/va01,/12/s,n,nonull/1 $ 1-set is null columns
if (nonull>-1) then $
partn pew,va01,/pe0,,pe1,/0 $ remove null columns
else $
equivx pew/pe0/always $
endif $
type parm,,i,n, unit6 $
unit6 = getsys(unit6,2) $ see where .f06 going now. v69 only
putsys(50,2) $ send .f06 to unit 50. makes large, useless file
solvit cee,pe0,,,,,,,,,/ue0,,//1//////-1 $ jacobi preconditioner
putsys(unit6,2) $ reset .f06 to original file
if (nonull>-1) then $
merge ue0,,,,va01,/uew/0 $ insert null columns
else $
equivx ue0/uew/always $
endif $
endif $
mpyad mee,uew,/mew $
ELSE $ DO THE ENCLOSED FLUID METHOD
merge pcw,,,,,v2cc/p2w/0 $
PARTN BEC,,VE1C/B1C,BCC,,/1 $
TRNSP B1C/B1CT $
TRNSP BCC/BCCT $
PARTN CEE,VE1C,/C11,CC1,C1C,CCC/-1 $
type parm,,i,n,old209 $
old209 = getsys(old209,209) $
putsys (0,209) $ use old solver
DECOMP C11/L11,U11, $
putsys (old209,209) $ restore new solver
FBS L11,U11,C1C/X1C $
trnsp x1c/x1ct $
trnsp c11/ct11 $
decomp ct11/lt11,ut11, $
trnsp ccc/ccct $
MPYAD X1C,X1C,ICC/IXCC/1 $
SOLVE IXCC,x1ct,,,/xxcc $
trnsp cc1/cc1t $
MPYAD C11,Cc1t,/AX1C $
MPYAD C1C,CCCt,AX1C/A1C $
MPYAD CC1,Cc1t,/AXCC $
MPYAD CCC,CCCt,AXCC/ACC $
MERGE A1C,,B1C,,V2CC,/A12/0 $
TRNSP A12/A21 $
MERGE ACC,BCCT,BCC,,V2CC,/A22/-1 $
DECOMP A22/L22,U22, $
 CALL FBSX L11,U11,lt11,ut11,XxCC,x1c,A12,P1w/UX2,UA21w $
MPYAD A21,UA21w,P2w/PA2w//-1 $ MINUS PRODUCT
FBS L22,U22,PA2w/UB2w $
MPYAD A21,Ux2,/PB2 $
FBS L22,U22,PB2/UC2 $
MATGEN, /I22/1/N2 $
ADD I22,UC2/UD2//-1.0 $
DECOMP UD2/LD22,UD22, $
FBS LD22,UD22,UB2w/U2w $
MPYAD A12,U2w,P1w/PC1w//-1 $ MINUS PRODUCT
CALL FBSX L11,U11,lt11,ut11,xXCC,x1c,PC1w,/U1w, $
partn u2w,,v2cc/ucw,qcw,,/0 $
merge u1w,ucw,,,,ve1c/uew/0 $
mpyad cee,uew,/cuw/1 $
mpyad mee,cuw,/mcuew $
mpyad mec,qcw,mcuew/mew $
ENDIF $
IF (MPRESID<>0) THEN $
PARTN mew,VwyA,/mey,,meA,/0 $
ELSE $
EQUIVX mew/mey/ALWAYS $
ENDIF $
CALL DBSTORE MEA,,,,//SEID/0/' '/0 $
MATGEN ,/VyZ6/6/Ny/NZ/6 $
PARTN mey,VyZ6,/meZ,,me6,/0 $
MPYAD PeZ,mez,/MZZ/1////6 $
IF (GRDPNT>-1) THEN $ DO WEIGHT SUMMARY
mpyad ger,me6,/m66/1 $
TRNSP m66/m66t $
MASSWT = (0.5,0.0)/CMPLX(WTMASS) $
ADD m66,m66t/VMRR/MASSWT/MASSWT $
MESSAGE //'RIGID BODY VIRTUAL MASS MATRIX, IN WEIGHT UNITS' $
MATPRT VMRR // $
ENDIF $ WTMASS
TRNSP MZZ/MZZT $
ADD MZZ,MZZT/MAVZZ/0.5/0.5 $ CAVALIER AVERAGING
RETURN $
END $ MGEN2
COMPILE MGENRS noLIST NOREF $ 0
SUBDMAP MGENRS PHIAZ,USET,LAMA,CASES// $ PHIAZ AND LAMA ARE UPDATED
$ TO IMPROVE THE EIGENSOLUTION WITH VIRTUAL FLUID MASS EFFECTS

TYPE PARM,,CHAR4,Y,DEBUG='NO' $
PARAML CASES//'DTI'/-1/166//S,N,MFLID $
if (mflid=0) return $ this doesn't work for se.s
$ for some unknown reason mflid gets into se 0 case control
CALL DBFETCH /cee,,,,/0/0/0/0/0 $
paraml cee//'presence'////s,n,nocee $
if (nocee=-1) return $ belt and suspenders
paraml phiaz//'trailer'/2/s,n,nz $ number of modes
CALL MGEN2 CASES,PHIAZ/MVMZZ/0 $
LAMX, ,LAMA/LMAT/-1 $
MATMOD LMAT,,,,,/MZ,/1/4 $ 4TH COL IS GEN. MASS
MATMOD LMAT,,,,,/KZ,/1/5 $ 5TH COL IS GEN. STIFFNESS
MATMOD KZ,,,,,/KZZ,/28 $ DIAGONALIZE
MATMOD MZ,,,,,/MZZ1,/28 $
ADD5 MZZ1,MvmZZ,,,/MZZ $
IF (DEBUG='YES') THEN $
 MESSAGE //'ROOTS WHEN ONLY STRUCTURAL MASS'/
' PRESENT, DEBUG OUTPUT' $
OFP LAMA// $
ENDIF $
REIGL KZZ,MZZ,,,,,,,,,/LAMAZ,PHIZZ,MIZ,EIGVMATZ,/
'MODES'/S,N,NEIGV////-1///NZ/0 $ FIND THE SAME NUMBER OF MODES
MESSAGE //'IMPROVED ROOTS FOR VIRTUAL MASS' $ OFP IS AFTER ALTER
MPYAD PHIAZ,PHIZZ,/PHIAZ1 $
EQUIVX PHIAZ1/PHIAZ/ALWAYS $ OVERWRITE INPUTS
EQUIVX LAMAZ/LAMA/ALWAYS $
RETURN $
END $ MGENRS1
COMPILE MAKVE1C nolist NOREF $
SUBDMAP MAKVE1C BEC/VE1C/NE/N1/NC $
$ select a c-set which produces stable c11 matrices
TYPE PARM,,I,N,NE,N1,NC,NX,II=1,JJ=1,NEm1,CI,IIM1,NCOMP,
prec,non,nonm1,lasti $
FILE besum=save/ve1c=save $
IF (NC=1) THEN $
NEM1 = NE - 1 $
MATGEN ,/VE1C/6/NE/0/1/NEMRI $
RETURN $
ENDIF $
matgen ,/besum/6/ne/ne $ a null vector
copy besum/ve1c $
DO WHILE (JJ<NC+1) $
MATMOD BEC,,,,,/BEJ,/1/JJ $ EXTRACT A COLUMN
if (jj>1) then $
partn bej,,besum/bok,bng,,/0 $ bng used up on prior passes
paraml bok//'presence'////s,n,boksiz $
if (boksiz>-1) paraml bok//'trailer'/5/s,n,boksiz $
if (boksiz<1) then $
non=0 $
jump usedup $
endif $
endif $
ADD BEJ,Besum/BESUMJ $
equivx besumj/besum/always $
PARAML BEj//'TRAILER'/5/S,N,NZWDSB $
PREC = PRECISON() $
NZWDSB = NzWDSB/PREC $
IF (NZWDSB<Ne) THEN $
PARTN BEj,,BEsum/BENULL,BENON,,/0 $ separate the null, non-null terms
PARAML BENON//'TRAILER'/2/S,N,NON $ NUMBER OF NON-NULL TERMS
ELSE $
NON = 0 $
endif $
label usedup $
if (non>0) then $
paraml bej//'trailer'/5/s,n,nzbj $
nzbj = nzbj/prec $
nx = nzbj - 1 $
matgen, /vnic/6/nzbj/0/1/nx $ 1.0 in first row, null elsewhere
merge, ,vnic,,,,bej/ve1i/0 $
else $ do a dmap binary search until you find the first non-zero term
partn bej,,ve1c/beok,beused,,/0 $ don't use a c-set already used
paraml beok//'trailer'/2/s,n,nok $
ii = 1
do while (ii<nok+1) $
 paraml beok//'dmi'/1/ii/s,n,beji $
if (beji<>0.0) then $
iim1 = ii - 1 $
nx = nok-ii
matgen, /veoki/6/nok/iim1/1/nx $
merge veoki,,,,,ve1c/ve1i/0 $
jump loopiend $
endif $
ii = ii + 1 $
enddo $
label loopiend $
endif $
add ve1i,ve1c/ve1cj $
equivx ve1cj/ve1c/always $
JJ = JJ + 1 $
ENDDO $ JJ
return $
END $ MAKVE1C
COMPILE FBSX nolist NOREF $
SUBDMAP FBSX L11,U11,lt11,ut11,xXCC,x1c,P1,P2/U1,U2 $
$ SOLVE FOR TWO BLOCKS OF VECTORS ON ONE PASS
$ [C11*C11 + C1C*CC1]*U = P, WHERE L11 IS FACTOR OF C11
$ P1 MUST BE PRESENT. P2-P4 MAY BE PURGED
TYPE PARM,,I,N,N1,N2,nt,NTM2 $
PARAML P1//'TRAILER'/1/S,N,N1 $
PARAML P2//'TRAILER'/1/S,N,N2 $
NT = N1 + N2
IF (N2>0) THEN $
NTM2 = NT - N2 $
MATGEN ,/VN12/6/NT/Ntm2/N2 $
APPEND P1,P2/P12 $
ELSE $
EQUIVX P1/P12/ALWAYS $
ENDIF $
FBS L11,U11,P12/P $
MPYAD XXCc,P,/XYCZ $
MPYAD X1C,XYCZ,P/Y//-1 $
FBS Lt11,Ut11,Y/U $
IF (N2>0) THEN $
PARTN U,VN12,/U1,,U2,/0 $
ELSE $
EQUIVX U/U1/ALWAYS $
ENDIF $
RETURN $
END $ FBSX
$ END OF VME ALTER PACKAGE
ECHOON $

@
7719; 70.7; Material Nonlinear Analysis - HEXA, PENTA, TEMP - Wrong Answers
In non-linear analysis the HEXA and PENTA elements will give incorrect
results if the element references a MATS1 or MATT1 or both and has a
non-constant temperature distribution. The displacements are correct. The
center stresses are correct. The Gauss/Grid Point Stresses have an
incremental error which is proportional to the temperature gradient.
Avoidance: There is no known avoidance.
@
7720; ; Documentation - MSC/NASTRAN Reference Manual - Wrong Dimension
The dimensioning of the channel-section as shown on page 5.2.1-8 of the
Version 68 MSC/NASTRAN Reference Manual is incorrect. The b value should
only extend to the median line of the web instead of all the way to the
outside edge of the web.
@
7728; 69.1; Database - DBUNLOAD, DBLOAD - Inefficient Restart
When a neutral database created by DBUNLOAD is migrated to CRAY and is
dbloaded for the purpose of a subsequent restart, the restart job goes
through unnecessary steps and recalulates the stiffness matrix.
@
7734; 69.0.2; Elements - BEAM, PBAR1 - Incorrect K2 Formulation for Beam Libra
ry I-beam.
The PBAR1 and PBEAM1 area shear factor term, K2, for the I-beam cross
section is incorrectly calculated. It is formulated to be the area of
flanges divided by the total area, when it should be 5/6 of this amount.
Avoidance: The error can be avoided by specifying the correct properties
on a PBAR or PBEAM entry or the user could add their own section. The
error has been corrected in MSC/NASTRAN Version 69.0.2. If you have
already been delivered 69.0.0 or 69.0.1, then 69.0.2 must be explicitly
requested.
@
7735; 69.0.2; Heat Transfer - QVECT, CHBDYE or CHBDYG with TYPE REV - Wrong An
swers
When QVECT loads are applied to CHBDYE or CHBDYG elements with surface
type REV and midside nodes, the orientation vectors are not calculated.
This error results in zero directional heat loads for this type of
element.
Avoidance: There is no avoidance. The error has been corrected in
Version 69.0.2. Version 69.0.2 will not be automatically delivered to
clients who have received 69.0.0 or 69.0.1, it must be requested.
@
7737; 69.0.2; Optimization/Design Sensitivity - DOM9 - SFE 4276 EC 3001
In certain cases solution 200 jobs that contain DRESP2 entries may fail in
the DOM9 module with the following error:
*** SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0
Avoidance: There is no simple avoidance. The error is corrected in
MSC/NASTRAN Version 69.0.2. Version 69.0.2 will not be automatically
delivered to clients who have already received Version 69.0.0 or 69.0.1.
The client should request this corrected version.
@
7738; 70; Documentation - PBARL - Wrong Information
Figure 1 under the PBARL description on page 793 and PBEAML on page 810 of
the MSC/NASTRAN Version 69 Quick Reference Guide indicates TYPE=BOX where
it should be BAR and vice versa.
@
7740; 69.1; Superelements - SEP1X, RBAR, RBE2 - BIOMSG 3002
Superelements with RBAR and RBE2 elements may cause an error in the sep1x
module which results in the message "BIOMSG 3002" being output.
Avoidance: Replace the RBAR and RBE2 elements with other types of rigid
elements.
@
7742; ; Superelements - MIRROR, REPEAT - Possible Wrong Answers
When using the enhanced superelement features, if the mirror or repeat
option is used to generate a copy of a primary superelement, and any of
the GEOM4-type data (e.g., MPC, RBEi, SPCi, ASETi, OMITi, BSETi, CSETi,
etc.) for the mirrored or repeated superelement is modified as compared to
the primary superelement, then the answers may be wrong.
Avoidance: If any of the GEOM4-type data for the mirrored or repeated
superelement must be different than the primary superelement, create a new
primary superelement instead of using the secondary superelement.
@
7743; GL; Submittal - notify=yes May Not Appear to Work.
If a user has set the message state to off, e.g., with the 'mesg n'
command, the notification message cannot be sent. This is most noticable
when a job is run in the background, as the notification message is the
only positive indication the job has completed. The user may receive no
indication at all, or may receive the cryptic message "permission denied"
when the notification is actually attempted.
Avoidance: The user must enable messages for the notification message to
be received under batch=yes.
@
7748; 70; Miscellaneous - Authorization - UFM 3060
If the memory specified on the 'mem' parameter is too high, then
MSC/NASTRAN may abort with a UFM 3060. The MSCID printed will be 0:
0*** USER FATAL MESSAGE 3060, SUBROUTINE MODEL - OPTION NAST NOT IN
APPROVED LIST.
SYSTEM DATE (MM/DD/YY): 11/11/96
SYSTEM MSCID: 0 (DECIMAL) 00000000 (HEXADECIMAL) SYSTEM
MODEL NUMBER: -1 , SYSTEM OS CODE: 1
Avoidance: Reduce the memory request.
@
7757; 69.1; Documentation - 'symbol' Keyword Documentation Error
The documentation for the symbol keyword in Section 4.1.2 and in Appendix
B.1 of the MSC/NASTRAN Version 69 Configuration and Operations Guide
(UNIX) is incorrect. The correct syntax for the keyword is:
symbol=<symbolic_name>=<string>
where <symbolic_name> is the name of the symbol to be defined and <string>
is the value of the symbol.
The specific example in 4.1.2 should read:
 SYMBOL=DATADIR=/dbs/data
The specific example in Appendix B.1 should read:
SYMBOL=DBSDIR=/dbs
@
7773; 69.0.2; Miscellaneous - Job hangs on HP-UX 10.20 when submitted through
at(1).
On HP-UX 10.20, an MSC/NASTRAN V69 job may hang if the job was run using
the at(1) facility. There are two ways to run a job using at(1), the most
obvious is to use the after keyword. Another is to create a queue command
such as
submit=at -q %queue% -m -f %job% %after%
Once the job is submitted, it will hang in the limits special function
with the following error message
displayed on your window or sent in an email:
sh[29]: ulimit: The specified value exceeds the user's allowable limit.

Avoidance: This problem is corrected by inserting the line

unset MSC_ARGV
just before the line
%0% limits
in the file
<install_dir>/bin/nast69.lcl
@
7775; 69.0.2; Plotting - PlotPS - Contour lines are scrambled
Contour plots are incorrect on the Solaris Version 69 system. These plots
have the contour lines going to the origin instead of to each contour
point. The effect is a set of vectors going from the origin to the
contour points.
Avoidance: Use a post-processor such as MSC/PATRAN to view stress
contours. The error has been fixed in Version 69.0.2 and Version 69.1.
@
7776; ; Superelements - BSET - UFE 4156, Poor Message
When using the enhanced superelement capabilities introduced in Version
69, only the boundary points can be assigned to the BSET degree of
freedoms, and this is the design intent. However, if any interior point
on a part (superelement) is assigned to the BSET, then the job will fail
with the following unclear error messages.

0*** USER FATAL ERROR 4156--- DIMENSIONS OR TYPE OF DATA BLOCK KLAAV ,
SUPERELEMENT y TRAILER=( xx xx x
x x xxx) ARE INCONSISTENT WITH SUPERELEMENT MAP.

Avoidance: If an interior degree of freedom on a part is desired to be

placed in the BSET, then assign that grid point to the ASET. Note that
 when using parts (superelements), if a degree of freedom at a grid point
is assigned to the ASET, then all six degrees of freedom for that grid
point will also be assigned to the ASET.
@
7780; 70.7; Executive System - ACQUIRE, DBLOAD, ENDJOB - UFM 453
If you specify the ACQUIRE NDDL, DBLOAD, and ENDJOB statements and the
"del=" keyword on the submittal command then UFM 453 is issued.
Avoidance: Do not use the "del=" keyword. Add ASSIGN statements for
MASTERA, MSCOBJ, and MSCSOU.
@
7789; 70; Superelements - PARAM,METHCMRS - UFM 3032
If superelements are present and PARAM,METHCMRS is specified for the
residual structure, then UFM 3032 will be issued stating that the program
is unable to find the selected set (EIGRL entry).
Avoidance:
Verify that the referenced EIGRL entry exists in the input file. If it
does, then include the following DMAP alter in the input file:
$
$ for V69
$
compile ifpl
alter 'call ifps'
TYPE PARM,,I,N,DOIT=0 $
CALL DBSTORE DYNAMIC,,,,//0/0/'MASTER'/S,DOIT $
$
compile semr3
alter 'REIGL.*CMKXX,CMMXX','REIGL.*CMKXX,CMMXX'
TYPE PARM,,I,N,OK $
CALL DBFETCH /DYNAMIC,,,,/0/0/0/0/S,OK $
REIGL CMKXX,CMMXX,DYNAMIC,CASES,,,,USET,EQEXINS,SILS,VYXW/
CMLAMA,CMPHIXZ,CMMI,EIGVMATS,/
'MODE'/S,N,NOZSET////METHRS////////CMSET $
$
$ DMAP to fix error 7789 - V70.5
$
compile ifpl
alter 'call ifps'
TYPE PARM,,I,N,DOIT=0 $
CALL DBSTORE DYNAMIC,,,,//0/0/'MASTER'/S,DOIT $
$
compile cmpmode
alter 'MATMOD.*CASES,DYNAMICS.*METHTYP.*METHC','' $
$
TYPE PARM,,I,N,OK $
CALL DBFETCH /DYNAMIC,,,,/0/0/0/0/S,OK $
MATMOD CASES,DYNAMIC,,,,/,/23/S,N,METHTYP//METHC $
alter 'READ.*CMKXX,CMMXX.*DYNAMICS',''
READ CMKXX,CMMXX,,,DYNAMIC,USET,CASES,VYXW,SILS,USET,,,EQEXINS/
CMLAMA,CMPHIXZ,CMMI,CMOEIGS,EIGVMATS,/
'MODE'/S,N,NOZSET//METHF/CMSET/METHCMRS $
@
7791; 70.5; Databases - DBLOAD, GP1 - SFM 307 in IFPSIZ
 If a neutral database is DBLOAD'ed and GP1 is executed without first
running IFP, then the job will fail in GP1 with:
0*** SYSTEM FATAL MESSAGE 307 (IFPSIZ)
THE NAME OF BULK DATA ENTRY CORD1C IS INVALID.
PROGRAMMER ACTION:IF THE ABOVE NAME IS A VALID BULK ENTRY,
THEN CHECK THE HASHING FUNCTION TABLE(S),
AND MAKE SURE THERE ARE NO "LOST" ITEMS.
0*** SYSTEM FATAL MESSAGE 3007 MODULE = GP1
0ILLEGAL INPUT TO SUBROUTINE IFPSIZ
0FATAL ERROR
Avoidance: Always run DBTRANS when DBLOADing a database.
@
7792; ; Documentation - Reference Manual, PBEND, FSI=3 - UWM 4260
The documentation in Volume 3 of the MSC/NASTRAN Reference Manual for USER
WARNING MESSAGE 4260, BENDING CHARACTERISTIC PARAMETER EXCEEDS RANGE OF
APPLICABILITY FOR ELEMENT ** indicates an acceptable range of .5 to 1.0.
In fact the acceptable range is .05 to 1.0.

@
7796; GL; Superelements - PLOTS - Missing Features, Missing Plots
The "LABEL GSPC" and "LABEL EPID" are not supported with the new
superelement features in Version 69.

@
7798; ; Documentation - Parallel Processing EMA, SYSTEM(107) - Disabled
by Default
On most of the systems supporting parallel operation with MSC/NASTRAN, the
Parallel EMA function shows little or no performance gain over the
non-parallel EMA. Because of this, by default, Parallel EMA will be
disabled on all platforms for MSC/NASTRAN Version 69.
If a you wish to enable PEMA, you may do so by setting the MSC_PEMA
environment variable to a value of YES (the test for "YES" is
case-insensitive). In this mode, the standard SYSTEM(107) enable/disable
tests for PEMA are active.
@
7799; 69; Performance - Infinite Cpu Loop of Cray - Missing Insufficient S
pace Message
On the Cray, if there is insufficient disk space in the directory that you
submit your job, then the job may be trapped in an infinite loop when
attempting to report the error message and shut down. Messages similar to
what is shown below, including "not enough space" or "no space left on
device", should appear in your log file. In version 69, the job will
terminate with similar messages as shown below, except the "not enough
space" or "no space left on device" message may or may not appear.
Process Id = <pid>
Attempting to bypass eie layer : Not enough space for cache pages
apparent state: unit 6 named <path>.f06
last format: (13X ,I8, 1H-, 8X, 10A 08 )
lately writing sequential formatted external
 SIGNAL - Signal SIGABRT(6) caught, job aborting.
See signal(2), CRI manual SR-2012 for further information.
TB001 - BEGINNING OF TRACEBACK
- $TRBK WAS CALLED BY RECOVR AT 561544a (LINE NUMBER 33)
- RECOVR WAS CALLED BY sysrcv AT 561764a (LINE NUMBER 112)
- sysrcv WAS CALLED BY __handlr AT 1074600a
- __handlr WAS CALLED BY killm AT 1070722a
- killm WAS CALLED BY abort AT 775403d (LINE NUMBER 22)
- abort WAS CALLED BY _ferror AT 1170241d (LINE NUMBER 46)
- _ferror WAS CALLED BY _errhndl AT 1167630b (LINE NUMBER 22)
- _errhndl WAS CALLED BY $ioerp AT 1174051b (LINE NUMBER 31)
- $ioerp WAS CALLED BY $WFF AT 1203202c
- $WFF WAS CALLED BY XSOWRT AT 3705646a (LINE NUMBER 322086)
- XSOWRT WAS CALLED BY XSOPIV AT 3700566c (LINE NUMBER 32290)
- XSOPIV WAS CALLED BY XSORT AT 3674767d (LINE NUMBER 321937)
- XSORT WAS CALLED BY $SEGCALL AT 12323d (LINE NUMBER 316417)
- $SEGCALL WAS CALLED BY XSEM01 AT 3667237b (LINE NUMBER 316417)
- XSEM01 WAS CALLED BY $SEGCALL AT 12323d (LINE NUMBER 11778)
- $SEGCALL WAS CALLED BY MAINAL AT 4636a (LINE NUMBER 11778)
- MAINAL WAS CALLED BY main AT 32c (LINE NUMBER 15)
- main WAS CALLED BY $SEGRES AT 12270d (LINE NUMBER 369)
- $SEGRES WAS CALLED BY $START$ AT 330525a (LINE NUMBER 369)
TB002 - END OF TRACEBACK
lib-1208 /nast/msc682/exe/unicos/analysis: UNRECOVERABLE library error
An I/O statement was invoked while another was active.
apparent state: unit 101 named
last format: 5001(' BIOMSG: ERROR',I6,' HAS OCCURRED IN ROUTINE ',2A4)
lately writing sequential formatted external
lib-1208 /nast/msc682/exe/unicos/analysis: UNRECOVERABLE library error
An I/O statement was invoked while another was active.
apparent state: unit 101 named
last format: 5001(' BIOMStime: command terminated abnormally.
Avoidances: There are two possible avoidances.
1. Increase the size of the directory that you are submitting your job
from.
2. Send the output files to a different directory that has plenty of
space (e.g., the scratch directory)
using the "out=..." keyword on the submittal command, and then copy
the files back to your submittal
directory afterward.
@
7808; 69.1; Optimization/Design Sensitivity - PARAM,DESPCH - Missing Output
If PARAM,DESPCH is set to a value greater than one, then there should be
DESVAR and GRID output in the punch file for every design cycle. However,
in Version 69 the GRIDS are not punched.
Avoidance: Insert the following DMAP alter:
compile desopt
alter 'dom12,',''
DOM12,
XINIT,XO,CVALO,PROPI,PROPO,OPTPRMG,HIS,DESTAB,
 geom1,coordo,edom,mtrak/
HISADD,NEWPRM,,NEWDES/
DESCYCLE/OBJIN/OBJOUT/S,N,CNVFLG/1/OPTEXIT//
MDTRKFLG/DESPCH/MODETRAK/eignfreq $

Grid output for the last design cycle can be obtained by setting
PARAM,DESPCH,0.
@
7814; ; P-Elements - GMBCs - Wrong Answers
When boundary conditions in shell p-elements are specified on an edge
using a global coordinate system, the answers become wrong for p=9 and the
program crashes for p=10. The answers are correct up to the order of p=8.
@
7818; ; Optimization/Design Sensitivity - TRD2, NOLINi - Quits With No M
essage
If a modal transient analysis is performed in Sol 200 (ANALYSIS=MTRAN) and
the job contains NOLINi entries, the job may abort in module TRD2 with no
message except a machine dependent message in the log file.
Avoidance: NOLINi entries are not supported in solution 200. The modal
transient analysis may be run in solution 112.
@
7829; 70.7; Restart - DBLOCATE read-only restart - UFM 1127
If the LOGICAL keyword is specified on the RESTART statement and the
corresponding database on the ASSIGN statement specifies a path for the
MASTER DBset which is different from the path in which the job is run then
the DBALL and any other associated DBsets will not be auto-assigned. For
example, if an input file is located at /home/zzz/test.dat, and the
desired databases are /tmp/fender.MASTER and /tmp/fender.DBALL, and
test.dat contains:
ASSIGN abc='/tmp/fender.MASTER'
RESTART LOGICAL=abc
then only the MASTER DBset is attached and all other DBsets will have the
status of "OFFLINE" echoed in the
.f04 file. Also, UFM 1127 will probably be issued.
Avoidance: Move the DBsets to the same path that the job is run in or
specify ASSIGN statements for all
DBsets.
@
7832; ; Elements - PBAR - Missing Torsional Mass Moment
The BAR element in MSC/NASTRAN does not include a term for the torsional
mass moment of inertia in its formulation (even in the coupled mass
matrix). This is documented in the Basic Dynamics User's Guide, but not in
the Quick Reference Guide.
Avoidance: If torsional mass moments of inertia are needed, use BEAM
elements or provide the values using CONM elements.
@
7839; 69.1; P-Elements - DBUNLOAD - SFM 4276, EC 7777, Segmentation Fault
A DBUNLOAD of a database with p-elements may produce SYSTEM FATAL MESSAGE
 4276 or a segmentation fault followed by a traceback.

@
7841; 69.0.2; DMAP - REIGL - UFM 6413, Insufficient Memory
The REIGL module was enhanced in Version 69 to allow a dmap writer to call
it without the input of a dynamics or casecc input data blocks. The data
they contain is input on some new parameters added to the end of the REIGL
call. If REIGL is run by itself, it quits with an "insufficient memory"
error:
USER FATAL MESSAGE 6413 (DFMSA)
INSUFFICIENT MEMORY FOR THE SYMBOLIC PHASE OF SPARSE DECOMPOSITION.
USER ACTION : INCREASE MEMORY BY AN ESTIMATED 9401 WORDS.
Avoidance: An avoidance is to run reigl without case control inputs after
a call with the case control input. The first call will initalize a
variable properly for the second call. The first call could contain
a small set of matrices from matgen, for instance.
@
7844; 69.0.2; Acoustic Analysis - ACMG - SFM 4276, Segmentation Fault, Access
Violation
Acoustic jobs may fail unexpectedly in ACMG. The error was discovered by
looking at the code, to our knowledge the error has never been
encountered.
Avoidance: There is no avoidance. The error has been corrected in
MSC/NASTRAN Version 69.0.2. This release is not being shipped to clients
who have already received Version 69.0.0, it must be specifically
requested.
@
7851; 70; Buckling Analysis - SUPORTi - SFM 3011
When a SUPORT entry is present in sol 105 the following error message
appears:
*** SYSTEM FATAL MESSAGE 3011 MODULE = RBMG4
ATTEMPT TO WRITE A TRAILER ON FILE 201WHEN IT HAS BEEN PURGED
Avoidance: Replace the SUPORT entry with a selected SPCi entry. In
Version 70 the suport entry is treated as an SPC which is consistent with
solution 101 when param,inrel is not selected. This error occurs in sol
103 also.
@
7852; 70; Inertia Relief - DMIG, UACCEL - Wrong Answers
When using inertia relief without using the rigid body acceleration
(DMIG,UACCEL), and any one of the following entries
BNDFL, DMIAX, ELIST, MFLUID, RADCAV, RADSET, RADLIST, RADMTX, DMIG
is present in the input file, then a conventional static analysis is
performed instead of inertia relief. A symptom of this error is the
printout of the URA intermediate matrix will be missing. A review of the
spcforce output will also reflect that a conventional static analysis is
performed instead of inertia relief.
Avoidances: There are two possible avoidances.
 1. Remove or comment out the above offending entries since they are not
used in a static analysis anyway.
or
2. Add the following dmap avoidances:
compile statrs $
alter 'paraml.*matpool' $
nompool = -1 $
endalter $

@
7854; ; DMAP - TYPE - Possible Wrong Values for Parameters
If a parameter is defined in the NDDL, and a DMAP "TYPE" statement
declares it as changeable in CASE CONTROL or BULK DATA ("Y") but not as an
"NDDL" parameter, then the program will incorrectly use the NDDL default
value and TYPE for the parameter. If the parameter is listed as a
qualifier (QUAL) in the NDDL, the default
value works correctly in the DMAP. For example, the following subDMAP in
SOL 100
SUBDMAP TEST input/output
type parm,,i,y,maxit=10 $ parameter is in the NDDL, NDDL default = 4
type parm,,i,n,modacc=10 $parameter is in the NDDL, but "N" is used on the
TYPE
type parm,,i,n,maxratio=1 $ parameter is in the NDDL as RS, but "N" is
used on the TYPE
type parm,,i,y,mesh=1 $ different type than NDDL
type parm,,i,y,seid=345 $ test a qualifier
$
RETURN
END
will result in the following values for the parameters:
maxit = 4 - NDDL default value in place of user-requested default
modacc = 10 - user requested default
maxratio = 1 - integer rather than RS as defined in NDDL
mesh = 4, but UFM 1053 indicating wrong parameter type
SEID = 345 - qualifiers work ok

Avoidance: If a parameter is defined in the NDDL, either define it as

TYPE NDDL or do not allow it to be changed by case control or bulk data.
NOTE: this only affects the default value, changes in DMAP or values in
BULK DATA or CASE CONTROL will work.
i
@
7855; 69.1; Buckling - TAFF, PARAM,FOLLOWK - Traceback
If a PARAM,FOLLOWK,YES is requested to include follower force affects to
the differential stiffness and loads that do not exist are referenced the
job will stop in the TAFF module and the following traceback will be
issued in the log file.
Traceback:
 Offset 0x000000ac in procedure tafest_
Offset 0x000000e8 in procedure tafldt_
Offset 0x00000684 in procedure taff_
Offset 0x000002d0 in procedure xsem10_
Offset 0x00000950 in procedure mainal
Avoidance: Remove the reference to the loads that do not exist.
@
7856; 69.1; ESTIMATE Utility - Missing GRID Entry
The ESTIMATE utility may flag GRIDs as missing when in fact the GRID
entries are present. The error occurs when a line that exceeds 80
characters precedes an otherwise valid line. Under certain circumstances,
the valid line may be skipped.
Avoidance: As a work around, delete all trailing blanks from the data
file, expand all tabs to spaces, and
truncate any lines exceeding 80 characters.
@
7857; 69.0.4; Superelements - SELA - SFM 4276, EC 1200
Superelement jobs may fail in module SELA with the message SYSTEM FATAL
MESSAGE 4276 ERROR CODE 1200. Encountering the error is dependent upon
the size of the load vector and the amount of memory requested.
Avoidance: Increase memory and re-run the job.
@
7860; 68.2.8; Optimization/Design Sensitivity - FRRD2 - SFM 4276, EC 7777
Solution 200 with analysis=dfreq may fail in the FRRD2 module with the
message:
SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777
The log file may contain a traceback from subroutine frrd2.
Avoidance: The error has been corrected in Version 69. It has also been
corrected in Version 68.2.8, a re-release for Fujitsu systems. The
following DMAP avoidance may also be applied that replaces FRRD2 with
FRRD1:
compile fea list noref $
$ 285 FRRD2 CKDD,CBDD,CMDD,,PDF,FOL1/DETA/0.0/0.0/0.0 $
$ FRRD1 CASES,DIT,CKDD,CBDD,CMDD,K4DD,PDF,FRL,FOL/
$ UDVF/SOLTYP/0 $ RF8 - FRRD /UDVF
alter 'FRRD2 CKDD','' $
frrd1 casers,dit,ckdd,cbdd,cmdd,,pdf,frl,fol1/deta/'direct'/0 $
@
7871; 70; Heat Transfer Analysis - GPFORCE - Missing Output
GPFORCE output is missing for steady state heat transfer analysis, SOL
153, in Version 69.0.
Avoidance: Place the following alter in Version 69.0.
compile sedrcvr list$
alter 299,299 $ V69.0.2
IF ( GPFDR ) THEN $ GRID POINT FORCE
@
7874; 2001; Miscellaneous - Extra plot request - Punch File
If an output request has two plot requests, e.g.,
disp(plot,plot) = all
then in addition to the plot file, a punch file containing the referenced
quantities will
also be generated.
Avoidance: Remove the extra plot request.
@
7875; 69; P-Elements - GP1, GMCORD, GMLOAD - SFM 3001
If a GMLOAD references a GMCORD in Version 68.2, then the job will fail
with the following error messages.
*** SYSTEM FATAL MESSAGE 3001 MODULE = GP1
THE INPUT DATA BLOCK IN POSITION 5 DOES NOT EXIST.
Avoidance: Use a CORDij instead of GMCORD when using GMLOAD.
@
7878; ; Optimization/Design Sensitivity - TEMP - Wrong Answers
An optimization job may yield wrong sensitivity results when temperature
loads are applied but certain designed elements (either DVGRIDs or
DVPRELi) are not temperature loaded.
Avoidance: A possible avoidance is to add a small temperature load on
unheated elements in Version 68.2.
@
7881; 69.1; Postprocessing - PARAM,POST,0 in SOL 200 - Missing tables in .xd
b file
Design sensitivity coefficient tables ( R1SEN and R2SEN datablocks) are
missing from the XDB.
Avoidance:
1. Copy the nddl.dat file from the MSC/NASTRAN DEL directory.
2. Replace the DATABLK r1MAPR and R2MAPR statements with the following.
DATABLK R1mapR TYPE=TABLE PATH=geomi, LOCATION=DSOr,
EOF
DATABLK R2mapR TYPE=TABLE PATH=geomi, LOCATION=DSOr,
EOF
3. Insert the following in your executive control section above all
compile statements:
compile nddl noref nolist
incl nddl.dat
@
7892; 69.1; Matrix Methods - SOLVIT - Iterative Solver Gives Wrong Answers
If the solvit module is given a diagonal matrix it returns a null output
matrix.
Avoidance: Add at least one small off-diagonal term.

@
7893; 69.1; Plotting - SEPLOT - UFM, Problems Parsing the SEPLOT Command
When plotting superelements, the SEPLOT module may have problems parsing
when the command extends multiple lines. This may lead to a USER FATAL
MESSAGE pointing at the SEPLOT command.
Avoidance: If possible put the plot command all on one line. If not
possible, rearranging the command may help.
@
7903; 70.5; Documentation - QRG, SYSTEM(148), DBVERCHK - Misleading Document
ation
The MSC/NASTRAN Quick Reference Guide documentation for SYSTEM(148),
a.k.a. DBVERCHK, states that this system cell permits restart from a
pre-Version 69 database into Version 69. This system cell simply bypasses
the check in MSC/NASTRAN that ensures the Version of the database is the
same as the Version being executed and to that extent it is applicable to
all Versions of MSC/NASTRAN. The system cell does nothing to make the
databases compatible.
Avoidance: In general databases are not compatible between Version 68, 69
or 70. The only way to restart is to perform a DBUNLOAD and a DBLOAD:
Step 1: Unload database in V68
assi dbunload='______.u68'
dbunload
endjob
Step 2: Load V68 database into V69
assi dbload='______.u68'
dbload
$
sol dbtrans
compile dbtrans
alter 'PARAML BULK'(,-1)
dbequiv datablk=bulk where (seid=0 and peid=0) ,
convert(seid=-1;peid=-1) $
alter 'EQUIVX XGEOM1.4/XGEOM1Q',''
type db geom1s,geom1q $
EQUIVX XGEOM1.4/GEOM1/-1 $
EQUIVX XGEOM1.4/GEOM1q/-1 $
alter 'if ( rsonly'(2),'call setq'(,-1)
IF ( RSONLY ) THEN
equivx geom1/geom1s/-1 $
ELSE IF ( NOT(IMAGE OR EXTER) ) then $
SEP2 GEOM1,GEOM2,,,,,SLIST,EMAP,,,/
GEOM1s,t1,,,,,t6,t7,t8,,/SEID $
 ENDIF $ (RSONLY)
GP1 GEOM1s,GEOM2s,,,,/
g1,g2,GPDTs,CSTMs,BGPDTs,g3,/
S,N,LUSETs/0/0 $ OUTPUTS ARE for SEP1
cend
enddata
@
7904; ; Documentation - Reference Manual - sensitivity of materials, sol
103

@
18854; ; Postprocessing - PARAM,POST,-1 CBAR - Incorrect Margins of Safet
y
PATRAN P3 will display margin of safety results for bar elements even
though no allowables were specified in the analysis.
Avoidance: Either specify allowables or ignore the results.
@
25363; 69.1; Miscellaneous - NOTIFY=YES - Extraneous Output Printed to the Sc
reen
Extraneous output will be written on Cray (and others?) when submitting a
job. The text "is y" or "is n" may appear on the screen just after the
system ID line is written as shown below:
mwali[136]: nast69t2 d10101d bat=n app=y scr=y
MSC/NASTRAN V69.0t2 (CRAY UNICOS 8.0.4.2) Tue Dec 17 09:12:20 1996
is y
MSC/NASTRAN beginning job d10101d.
This is the output from the mesg(1) command. This output should have been
captured by the nastran driver, and will result in the nastran driver not
being able to warn to user if they set notify=yes and their current
message state is "n".
@
25365; 70; Statics - DCMP, SGI R4400 - Segmentation fault
In Version 69, when running a large problem on the SGI R4400, if there is
insufficient memory allocated for SPARSE DECOMP, then the job may fail
with no error messages in the f06 file. A symptom of this error is the
printout of the following unclear message in the log file.
Receiving signal number 11: Segmentation fault.
Note that in Version 69 sparse decomp is designed with an enhanced spill
logic so that it should be able to execute in significantly less memory.
Avoidance: Increase the memory substantially to avoid spill.
@
25375; 69.0.2; DMAP - Complex Double Precision Functions - SFM 4276, EC 7777
Double precision complex functions in DMAP cause the following error on
 the HP 9000/800 Version 69.0.0 release of MSC/NASTRAN.
SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777 PID= 0
Avoidance: Set nastran system(100)=255 or request Version 69.0.2 where
this error has been corrected.
@
25377; ; Documentation - MATMOD option 25
matmod option 25 dmap dictionary article incomplete
This option is designed to produce two output vectors for only unsymmetric
matrices (form 1). For symmetric matrices (form 6) only the O1 vector is
output. This condition can be detected by the sym output parameter, which
is set to -1 for symmetric matrices, and 0 for unsymmetric matrices.

@
25378; 69.1; ACCESS - Cannot Build ACCESS Sample Programs
Any attempt to build the ACCESS sample programs will fail with the error
message:

make: Cannot find a rule to create target sol301.o from dependencies.

This error can be corrected by taking the revised makefile from Scopus tan
2523, or editing the
file <install_dir>/msc69/access/makefile:
Change the line:
build: nastran emas ddl
to
build: nastran ddl
Remove the trailing backslash in SECOND line of the following two:
install: $(EXE)/demo1 $(EXE)/demo2 $(EXE)/smplr $(EXE)/d
dlprt
$(EXE)/ddlqry
And delete the immediately following line:
$(EXE)/sol301 $(EXE)/sol302

@
25379; 70; Matrix Methods - DECOMP - UWM ZERO DIAGONAL TERMS ENCOUNTERED
In MSC/NASTRAN Version 69, if the decomp module runs out of memory and the
DECOMP parameter is set to 4 or 32, incorrect warning messages may be
issued that indicate "ZERO DIAGONAL TERMS ENCOUNTERED". The message that
should be output is:
USER FATAL (or WARNING) MESSAGE 6413 (DFMSA)
INSUFFICIENT MEMORY FOR THE SYMBOLIC PHASE OF SPARSE DECOMPOSITION.
USER ACTION : INCREASE MEMORY BY AN ESTIMATED XXXXX WORDS.
Avoidance: Increase the memory and re-run.
@
25422; 69.1; Static Nonlinear Analysis - PARAM,KDIAG - UFM 4551, Failed Conve
rgence
In nonlinear analysis (SOL 106), if the model has zero stiffness in the
undeformed configuration, then the solution may fail to converge or it may
converge to a wrong answer. This error first appears in Version 69, and
is associated with an error in the "KDIAG" processing.
Avoidance: Put PARAM,KDIAG,-1.-30 in the input file to avoid the error or
apply the following alter:
in NLSTATIC replace statement
ADD5 KAAL,KLLTX,KLLRH,,/KAALX///-1. $
with the statement
ADD5 KAAL0,KLLTX,KLLRH,,/KAALX///-1. $
@
25423; 69.1; Installation - SGI IP19 is Incorrectly Identified
The Silicon Graphics IP19 system is an R44 system, but is able to run
IRIX64. SGI had categorically claimed that IRIX64 would *only* be found
on R8/R10 systems. As a result of the presence of IRIX64, M/N V69 will
attempt to run the incompatible R8/R10 executable.
An IP19 system can only run the "irix" executable.
The work-around for V69.0.2 is to:
1) Modify the shell script "<install_dir/bin/msc69" and change the case
label "IRIX" to "IRIX*".
Note, this will leave the succeeding "IRIX64" case active, but
ineffective.
2) Modify the file "<install_dir>/bin/nast69.ini" to include the line:
s.arch=irix
This error should not occur in V69.0.0 or V69.0.1. If it does, the
solution is identical.
@
25425; 69.0.4; Optimization/Design Sensitivity - Beam Library Optimization is I
nefficient
Beam Library optimization is inefficient. For a model with 1000 designed
properties, a speedup of a factor of 70 has been demonstrated. The poorly
performing modules are DOM6 and DOM9. MSC/NASTRAN Version 69.1 corrects
this inefficiency.
Avoidance: There is no avoidance.
@
25439; 69.1; Optimization/Design Sensitivity - Beam Library - Incorrect Sensi
tivities
If the design optimization feature of the beam library is utilized with
multiple designed stress recovery points, the sensitivities may be in
error. The error is most likely to show up when the stress recovery
coordinate (e.g., C1 or C2) is not a function of all the designed
dimensions. This error can occur for both the MSC provided sections and
for user provided sections.
 Avoidance: There is no avoidance.
@
25451; 69; Submittal - Windows NT - Input File Not Found
Version 68 of MSC/NASTRAN on Windows NT must have .dat extensions
specified for input files. This error has been fixed in Version 69.

@
25456; 69.1; Optimization/Design Sensitivity - DRESP2 - Zero Gradients
It is possible to generate the message "Zero Gradients of Violated
Constraints" in error if both the lower and upper bounds of a constraint
are active for the DOT optimization algorithm and if the underlying
response was generated using a DRESP2 entry. This condition will produce
a fatal message and the job will terminate. In other cases where both
constraints are active, the optimizer may become confused, but it will not
terminate.
Avoidance: Be careful when specifying tight limits on a DRESP2 entry and
be aware that this may lead to difficulty. The error has only been
diagnosed when using the SQP algorithm (METHOD=3 on the DOPTPRM entry).
@
25470; 69.1; Optimization/Design Sensitivity - PARAM,INREL - Poor Performance
Calculation of design sensitivities for static subcases with inertia
relief can be CPU intensive.
Avoidance: A DMAP avoidance is available.

@
25472; 69.1; Inertia Relief - SOL 200, INREL - Various fatal messages or wro
ng answers
If the value of PARAM,INREL is specified differently in one or more
subcases, then various fatal messages
will be issued or wrong answers may result.
Avoidance: Do not specify PARAM,INREL in the Case Control. Always
specify in the Bulk Data.
@
25501; 70; Superelements - SEPLOT - Missing Plots
If there are no SEMOVES, REPEAT, or MIRROR superelements in the deck and
if there is a CSTM in a substructure there is a possible memory error
causing the substructure not to get plotted or if plotted not in its
correct orientation.
Avoidance: If the user is missing plot output of a substructure or
suspects that it is in the wrong location put in a dummy semove
substructure and the problem should go away.
@
25531; 70; Submittal - Remote Execution via 'node=' May Fail
The ability to run an MSC/NASTRAN job on a remote system using the 'node'
command line keyword will fail in Version 69.0 if the command is not named
'nastran' on the remote system. The nastran command is supposed to use
 the actual command name used to run the job on the local system when
running the job on the remote system. The V69.0.X release will only use
the generic 'nastran' command name on the remote system.
Avoidance: To correct this problem, the generic nastran name must be
available on the remote system.
@
25532; 69; Miscellaneous - Large field format for TABLED1 - Possible Wrong
Answers
The correct specification of a long-field format entry requires a blank
entry to be included for blank fields. For example, fields 4 thru 9 of
the TABLED1 entry are not used, thus the correct specification of a
long-field formatted TABLED1 entry requires the second entry to be blank.
In Version 68, if this rule is not followed for the TABLED1 and x-y pair
data is specified on the second entry the data is ignored. The program
performs a linear interpolation from the data on subsequent entries for
the ignored region and this is not likely to be the intent of the user.
Version 69 corrects this error by issuing "USER FATAL MESSAGE 315
(IFPDRV)".
Avoidance: An example of an incorrect entry and the correct specification
follows:
TABLED1*1
*AO11000
*AO11000 .0000 -1397.00 .0007
-1398.00*AO11001
*AO11001ENDT
A blank line must be included after the first entry as shown below.
TABLED1*1
*A1
*A1
*AO11000
*AO11000 .0000 -1397.00 .0007
-1398.00*AO11001
*AO11001ENDT
@
25545; 69.1; Eigenvalue Analysis - REIGL, EIGRL - SFM 4276, EC 7777
In rare circumstances, Lanczos eigenvalue analysis may fail in the REIGL
module and the message "SYSTEM FATAL MESSAGE 4276, ERROR CODE 7777" will
be issued.
Avoidance: The error is more likely when a large buffsize is set. The
error may be easily avoided by either increasing or decreasing the
buffsize.
@
25555; 69.1; Miscellaneous - SSSALTER - Old Released Alters
The SSSLATER library released for V69.0.2 is not the current library
intended for release.
Avoidance: Contact your MSC representative for the correct versions of
the alters.
@
25595; 69.1; Submittal - Command Line Option "cpu" Causes a Memory Fault
For Alpha/Unix Version 69.0.2 the use of the command line option cpu
causes a memory fault.
Avoidance: Set ppcdelta to .1 in the system rc file.

@
25673; 70; Modal Transient Analysis - TRD1 failure with SYSTEM(146)=2 and u
nsym. DMIG input
If in a modal transient analysis uses Lanczos with SYSTEM(146)=2 and an
un-symmetric direct matrix input with B2PP is activated, then the TRD1
module fails on an SGI R4400 machine with the log-file message:
Receiving signal number 10: Bus Error
Avoidance: Do not set SYSTEM(146)=2.
@
25675; 70; Documentation - CTRIA3, NASTRAN system cell T3SKEW - Missing Doc
umentation
The MSC/NASTRAN Version 68.2 Release Notes document the T3SKEW system cell
which allows the user to control the the minimum vertex angle for TRIA3
elements at which USER WARNING MESSAGE 5491 is issued. The documentation
should have been but was not transferred to the Version 69 Quick Reference
Guide.

@
25681; 70; Documentation - SETREE - Unexpected answers
If SETREE entries are used to define a multi-level processing tree, then
according to Remark 6 under the Quick Reference Guide description of the
SETREE entry superelements that are not specified on the SETREE entry will
not be included in the analysis. However, if the non-listed superelements
are attached to grid points which belong to the residual structure, they
will be placed in the tree immediately above the residual structure and
included in the analysis.
@
25688; 69.1; SSSALTER - rflagb.v69 - Wrong Answers for Modal Analysis
The rflagb.v69 alter gives correct answers for motion at the enforced
motion DOFs, but wrong answers for all other DOFs. The cause of the
error is in the assembly of the modal matrices. This error applies to the
alter dated November 96. The version delivered with Version 69.1 dated 10
Jan 97 has this error corrected.
Avoidance: If you do not have access to Version 69.1, you may request a
copy of the updated alter through client support.
@
25691; 69.1; Nonlinear Analysis - OFP - SFM 4276, EC 7777
A nonlinear analysis that requests stresses for plane strain elements in
MSC/NASTRAN Version 68 may fail in the OFP module with the message SYSTEM
FATAL MESSAGE 4276 ERROR CODE 7777 on CONVEX computers.
@
25700; 69.1; Restart - Mini-database Restarts with Superelements - Missing Ou
tput, UWM 4410
If superelements are present in a mini-database restart using a dynamic
solution sequence such as SOL 103, 107--112, 129, 145, and 146, or dynamic
analysis in 200, then data recovery output for the residual structure will
be missing and UWM 4410 will be issued.
Avoidance: Insert following DMAP alter:
compile phase0
alter 'IF ( (DYRD',''
IF ( (DYRD<>0 OR (SEID<>0 AND METH<>0)) AND
(NOT(STATICS) OR AERO) and not(saverstl) ) THEN $
@
25713; 69.0.2; Rigid Body Elements - RSSCON - Wrong Answers
In Version 69.1, the shell-solid connector RSSCON does not allow the
connection of midside noded elements with non midside noded elements when
the ELEM option is used. For example, a CQUAD8 shell element can not be
connected to a CHEXA with 8 nodes. A fatal message is issued.
In Version 69, non matching elements were connected with the ELEM option.
Spurious zero energy modes appeared
as a result of the incomplete connection. In Version 69.0.2 the fatal
message:
USER FATAL MESSAGE 6689 (GP4RS2)
BULK DATA RSSCON 201 ELEMENT 1 CONNECTED TO ELEMENT 5
IS AN INCOMPLETE ATTACHMENT.
ONLY A SUBSET OF THE SHELL GRID AND EDGE DEGREES OF FREEDOM ARE
CONSTRAINED
is issued, wrong answers are avoided and thus the error is considered
corrected.

Avoidance: If non matching elements need to be connected, the GRID option

must be used.

@
25715; 70; Rigid Body Element and Multipoint Constraints - RSSCON - UFM 669
2 (SECNSD)
If more than one shell element is connected to the same solid element with
the RSSCON element using the ELEM option, conflicts in the multi-point
constraints may occur. A message of the form:
USER FATAL MESSAGE 6692 (SECNSD )
RSSCON NUMBER *** CANNOT CONNECT AN EDGE OF TRIA ELEMENT TO TETRA
ELEMENT **
will be output for those shell grids which can not be connected. The
Version 70 Quick Reference Guide documents this limitation closing this
report.
Avoidance: In general, it is NOT recommended to connect more than one
shell element to the same solid element.
@
25724; 70; Documentation - CELAS1, CELAS2 - UFM 316
The MSC/NASTRAN Version 69 Quick Reference Guide description of the CELAS1
and CELAS2 indicates that multiple degrees of freedom may be attached on a
single entry. This is not the case and if attempted will result in the
message:
*** USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA ENTRY CELASi *** SORTED ENTRY COUNT =
*
@
25730; 70; Postprocessing - DBC - Excessive CPU Time
DBC can take an inordinate amount of time to write to the .xdb file in
some sort1 jobs. For instance, a 130000 dof dynamics job that is writing
100 modes spent 50 minutes cpu time to calculate the modes, but 16 hours
writing to the .xdb file.

Avoidance:
Use the following alter which forces sort2 only for writing to xdb.
compile sedrcvr souin=mscsou list, noref
$ force sort2 output in a sol 103, normals modes run
alter 329,329 $
SDR3 OUGV1,OPG1,OQG1,OEF1X,OES1X,OSTR1/
OUGV2,OPG2,OQG2,OEF2,OES2,OSTR2 $ SORT2 OUTPUT
DBC OPG2,OUGV2,OEF2,OES2,OQG2,GPLS,EGPSTR,EGPSF,GPDCT,ELDCT,
ONRGY1,OGPFB1,ERROR1,,,,,,,//
'OPG'/'OUG'/'OEF'/'OES'/'OQG'/'GPL'/'GPS'/'SVF'/
'GPDCT'/'ELDCT'/'OESE'/'OGPF'/'PERROR'////////
-1/DBCPATH/S,N,CP/APP1/ICYCLIC/GEOMU/LOADU/POSTU/
DBCDIAG/DBCPROG/DBCOVWRT/DESITER/////ADPTINDX/LUSETS $
@
25746; ; Matrix Methods - MPYAD - Segmentation Violation
A client deck fails with a segmentation violation while performing a
sparse MPYAD in the SDR1 module.

Avoidance: The cause of this is under investigation, but the following

alter avoided the problem by deselecting sparse MPYAD before the module
where the error occurred.
compile sedisp, souin=mscsou, nolist, noref
alter 217
type parm,,i,n,sys126=0 $
sys126 = getsys(sys126,126) $
putsys(24,126) $
alter 218
putsys(sys126,126) $ put original value back
@
25747; 70.5; Documentation - PARAM,POST,0 - DBCMSG Warning Message 3 from Sub
routine CHEXA
The DBC module which creates the graphics database may issue unclear
 messages such as those that follow.
CHEX ELEMENT 19 HAS BEEN UPDATED.
This message is simply an indication that an xdb existed prior to the run
and the element cited has been updated or changed.

@
25773; ; Rigid Elements - RSSCON - UFM 6689
RSSCON elements do not correctly connect warped solid elements to shell
elements. In the case of a CHEXA with midside nodes out of alignment with
the corner grids by more than TOLRSC*(element edge length), the shell
attachment to the midside nodes results in the following improperly issued
FATAL message:
USER FATAL MESSAGE 6689 (CONSSD)
RSSCON NUMBER 200 CANNOT CONNECT HEXA ELEMENT 1 TO QUAD ELEMENT
3
USER ACTION: CHECK GEOMETRY
Avoidance: Other than modifying the geometry to eliminate the warped
elements there is no avoidance.
@
25814; 70.5; DMAP - AUTOSPC Incorrectly Constrains USET.
If a user places a degree of freedom into the user sets (U1-U6), and the
degree of freedom is singular, then when GPSP moves it to the s-set it
also removes it from the user sets. The user sets are mainly used by DMAP
writers
Avoidance: Use the USET table produced by the GP4 module rather than the
updated version produced by the GPSP module.

@
25817; 70; Postprocessing - Stress sort PARAM SRTOPT, 1 and 3 Do Not Work.
Module STRSORT which is activated by PARAM,SRTOPT does not work as
documented for values 1 and 3. The module does not sort in ascending
order.
Avoidance: There is no avoidance.

@
25819; 69.1; Superelements - Mini-Database (scr=mini) - UFM 4252
The mini-database option only supports data recovery restarts, i.e., the
coldstart run must be executed for all superelements plus the residual
structure; i.e., SEALL=ALL, before the data base is useful. In other
words, superelements cannot be generated in separate runs with SEALL=n
referring to one or more superelements under the mini-database option. If
this is attempted then UFM 4252 will be issued in module SEMA.
Avoidance: None.
@
25821; 70; Online Documentation - Installation of MSC/NASTRAN Encyclopedia
 When performing a custom install of the MSC/NASTRAN Version 69
Encyclopedia for WorldView 2.2 Architectures, the following choices for
systems are presented which contain duplicate menu choice labels for HP
and SGI systems:
----------------------------------------------------------------
MSC Installation and Setup - Interleaf WorldView V2.2 Architectures

Sel Architecture Mode Mbyte

1. Digital Alpha UNIX . . . . . . . . . . . . . . . . . . . . . No 9
2. Hewlett Packard 9000 No 9
3. Hewlett Packard 9000 . . . . . . . . . . . . . . . . . . . . No 9
4. IBM RISC System/6000 No 8
5. Silicon Graphics . . . . . . . . . . . . . . . . . . . . . . No 12
6. Silicon Graphics No 12
7. Sun SPARC Solaris . . . . . . . . . . . . . . . . . . . . . No 8
8. Sun SPARC SunOS Yes 7
Total space required 7
C. Continue with installation.
X. Completely exit this installation now.

Enter the number of the architecture to select/deselect,

C to go to the next menu, or X to exit this installation.
Enter 1-8, C or X (no default)?
----------------------------------------------------------------
There is no explanation as to the difference between the two HP 9000
systems (items 2&3) nor the two SGIs (items 5&6).
Avoidance: Choose either one.
@
25824; 70; SSSALTER - sm03a.v69 - SIM 6470
If the lanczos module is given a mass matrix which contains terms further
from the diagonal than occur for the stiffness matrix, various errors may
occur. One symptom is a fatal message mentioning "spurious modes".
Another is a message of the form
*** SYSTEM INFORMATION MESSAGE 6470 (SSMR0D)
MORE EFFICIENT PROCESSING IN LANCZOS EIGENSOLUTION MAY BE POSSIBLE
BY INCREASING THE BUFFER SIZE OR
THE VALUE OF SYSTEM CELL 199 SUCH THAT THE PRODUCT OF THESE TWO
VALUES IS AT LEAST 2266 WORDS.
This above message is actually a symptom of a fatal error which will occur
later, of the form
0*** SYSTEM FATAL MESSAGE 3001 MODULE = LANCZOS
THE EXECUTIVE FILE 0 DOES NOT EXIST.
This problem most commonly occurs with component mode synthesis, a
capability supported only by dmap alters
in sol 3, the solution sequence in which the lanczos module is delivered.
The error was encountered when
using an alter from the sssalter library, which replaces the reigl module
with the lanczos module for sol 103.
 An avoidance is to replace the stiffness matrix with a slightly shifted
matrix, as shown below. This will perturb the solutions very slightly
while avoiding the error.
Knew = Kold + a*M, where a is a small number
If SIM 6570 appears, it is also necessary to increase buffsize on the
nastran command.
This feature will be added to alter package sm03a.v69 on the MSC web site.
The factor a will be used only when the user sets the "a" parameter
described in the readme file for the alter.
@
25836; 70; Documentation - V69 Online Encyclopedia Installation Problems
The V69 MSC/NASTRAN ENCYCLOPEDIA (MNE) Installation does not work if the
user chooses to keep both collections on the CD-ROM. This is the minimum
installation option.
Avoidance: If you want to install only the WorldView browser and access
both V69 and V68.2 documentation collections from the CD-ROM (minimum
installation), you must do the following before you start the
installation.
1. Decide on the <install_dir> directory in which to install the MNE.
2. Before installation, create the directory <install_dir>/mne69
After installation, edit the <install-dir>/iview22/Iview file and change
the line
Iview*welcomeDoc:<install-dir>/mne69/v69docs/69doc/start69.pl
to
Iview*welcomeDoc:<install-dir>/mne69/v69docs-cd/69doc/start69.pl

Also, on some IBM RS6000 systems, when iview (MNE) is started, the
following error message may be issued:
Error: Can't open display: :0.0
In order for the iview browser to display properly on IBM RISC System/6000
Computers systems, the DISPLAY environment variable must contain the node
name. If the DISPLAY variable set by default on your system does not
contain the node name, include the following line in your .cshrc file (for
csh users):
setenv DISPLAY <nodename>:0.0
or the following line in your .profile file(for Bourne or Korn shell
users):
DISPLAY=<nodename>:0.0; export DISPLAY
@
25845; 70; Eigenvalue Analysis - EIGRL, REIGL - UFM 6578
A new feature in V69 allows the use of parameters on the REIGL DMAP
statement to be used to control Lanczos eigenvalue solutions. An error
exists that if this approach is used, then the a value must be provided
for parameter 9 (NDES) or the module will fail with UFM 6578 indicating
that the value for parameter 9 is out of bounds.
 Avoidance: If parameters on REIGL are used to control Lanczos, always
include a value for parameter 9 (NDES).
@
25846; 70; Eigenvalue Analysis - EIGRL, REIGL - SFM 6578 parameter 9 out of
bounds
In MSC/NASTRAN Version 69, it is possible to provide the control
parameters for the eigenvalue solution on the REIGL statement. If this is
done and no value is provided for ND (number of roots desired, the run
will fail with SFM 6578
stating that parameter 9 is out of bounds on the REIGL module.
Avoidance: Provide a large integer in the field for ND on the REIGL module
statement.
@
25851; 70; Optimization/Design Sensitivity - SCR=NO, ANALYSIS=AERO - UFM 11
26
If aeroelastic optimization is attempted using scr=no, errors in the NDDL
dependencies cause the run to
fail on the second design cycle with the message:
USER FATAL MESSAGE 1126 (GNFIST)
DMAP MODULE AMP IS ATTEMPTING TO OUTPUT DATABLOCK 202 NAME =
QKH , WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE QKH =OVRWRT
If a flutter analysis is included, the QKH datablock causes the initial
error. If this datablock problem is corrected and/or there is a static
aeroelastic subcase, the failure is in the DBSTATUS module when it is
checking for the presence of the AUG datablock.
Avoidance: Make the following NDDL changes:
depen qkh /qhh $
DEPEN AUG /IFPA(VP),lsall $
@
25852; 70; Optimization/Design Sensitivity - ANALYSIS=FLUTTER, DSAD14 - SFM
3007
If flutter optimization is used in conjunction with the PKNL flutter
method, SYSTEM FATAL MESSAGE 3007 is issued which is of limited usefulness
to clients.
Avoidance: The PKNL method is not currently supported in SOL 200. Use
the PK method.
@
25861; 70; Superelements - SCR=MINI - SFM 4276, EC 1044
In version 69 a new "minimum database" capability has been introduced for
MSC/NASTRAN. This capability is implemented by specifying "SCR=mini" on
the submittal statement. If this feature is used with the split database
capability of superelements, the following error may occur:
BIOMSG: ERROR 1044 HAS OCCURRED IN ROUTINE OPEN , FILE INDEX (FILEX) =
0
Avoidance: If this occurs, either use SCR=NO or SCR=YES on the submittal
line.
@
25865; 70; Superelements - BUSH - Wrong Answers
 If the BUSH element is used in a model with superelements, the
frequency-dependent capability will not work.
Avoidance: Either include the following DMAP alter in the input file:
compile phase1a
alter 'TA1.*,,ECTS,EPTS,BGPDTS'(,-1)
if(seid=0)then $
type db,ept $
dbview epttemp=ept (where wildcard=true) $
TA1 ,,ECTS,EPTTEMP,BGPDTS,SILS,ETT,CSTMS,DIT,ECTAS,EHT/
EST,,GEI,GPECT,,VGFD,DITID/
LUSETS/-1/S,N,NOSIMP/1/S,N,NOGENL/0///S,N,FREQDEPS $
else $
alter 'TA1.*,,ECTS,EPTS,BGPDTS'
endif $
or:
place all superelements in partitioned bulk data sections and place the
residual structure in the main bulk data. The frequency dependent
capability will work if the main bulk data does not have any
superelement-related entries.
@
25866; 70; Optimization/Design Sensitivity - DOM6 - SFM 4276 EC 7777 Sig 8
The overflow message, "SYSTEM FATAL MESSAGE 4276 ERROR CODE 7777" has been
observed on the CRAY in the DOM6 module and has been identified as being
caused by a combination of requesting prints of sensitivity data (OPTEXIT
= 4 or 7 the DSAPRT case control command) when both DLINK and DVPREL2
entries are present.
Avoidance: The only avoidance is to eliminate one of the three
contributing factors. The Beam Library enhancement in Version 69
alleviates the need for DVPREL2 entries and may be a desirable avoidance
in many cases.
@
25873; 70; Superelements - SCR=MINI - No Data Recovery for External Superel
ements
In MSC/NASTRAN Version 69 a minimum database capability was introduced. If
external superelements are used and SCR=MINI is used for the residual
structure, then data recovery is not possible for the upstream external
superelements.

Avoidance: Add the following line to the FMS of the external superelement
data recovery run:
dblocate param=(saverstl), where (peid=0), logical = rsdb
(where RSDB is the logical name used for the database with the residual
structure solution)
@
25884; 70; Superelements - SESEB - Missing UFM 6794, Possible Wrong Answers
Version 69 on CRAY computers may not correctly detect duplicate boundary
points and thus may fail to issue:
USER FATAL MESSAGE 6794 (SEPSEB)
DUPLICATE BOUNDARY POINTS DETECTED CONNECTING TO SUPERELEMENT ** AT GRID
 ** FROM DOWNSTREAM **.
The failure to issue this message may result in an upstream superelement
being connected to an unintended degree of freedom resulting in wrong
answers.
Avoidance: When using the automatic superelement connection logic, be
careful in maintaining unique grid locations for boundary points.
@
25888; 69.1; Eigenvalue Analysis - GIV - Wrong Answers, UIM 4158
Using the GIVENS method to calculate modes of a structure which has a non
positive definite mass matrix, should result in the following error
message:
"USER FATAL MESSAGE 4646---- THE MASS MATRIX IS NOT POSITIVE
DEFINITE,USING THE GIVENS METHOD. USE MGIV INSTEAD"
However, the run may proceed and incorrect answers may be produced. Since
MSC/NASTRAN automatically omits massless degrees of freedom, this error
only occurs with models involving nearly massless mechanism. The user is
warned of the poor answers by a very large value for the factor diagonal
ratio in the following message:
USER INFORMATION MESSAGE 4158 ---STATISTICS FOR CHOLESKY DECOMPOSITION OF
DATA BLOCK MXX FOLLOW
MAXIMUM RATION OF MATRIX DIAGONAL TO FACTOR DIAGONAL =
The user may also be warned of the poor results by the issuance of a
machine dependent message indicating a floating point overflow.
Avoidance: Use MGIV or LANCZOS method of extracting eigenvalues for the
structures of these nature.
@
25891; 70.5; Rigid Elements - RBE3 - Poor Answers
If an RBE3 element is input with insufficient independent degrees of
freedom to define all 6 possible rigid body modes of the element, a set of
constraint equations may be produced that do not reflect the weighting
factors input by the user, and can also cause internal constraints
associated with stiffness and mass terms associated with degrees of
freedom attached to the RBE3 element. This error was observed on a model
which has an RBE3 element of the form
$ refc wti ci
RBE3, [eid], , [refgrid],123456, 1.0, 123, gida, gidb
As the two attached grid points have only translations coupled, (123 input
for ci), the element is not capable of reacting the component of moment
applied at refgrid that is parallel to a line drawn between gida and gidb.
The element was processed without warning or fatal messages, and produces
different answers on different machine types.
A method to detect this type of error is to use param,checkout,yes in sols
101 or 103, and look for non-zero terms in EMH.
A fix for this particular problem is to add a third grid point,
non-collinear with gida and gidb.
In Version 70.5, an element of this type will cause a fatal error, with
the first offending RBE3 element identified. The test performed to
 identify a poor solution is find the largest number in [C*CINV-I], where C
is the weighting matrix used to normalize the constraint equations for the
element, CINV is its computed inverse, and I is the identity matrix. All
numbers in the matrix of the brackets should be computational zeros. If
any is greater than 1.0D-10 a fatal error exit is taken.
@
25893; 70; Superelements - External Superelement Data Recovery in SOL 103 -
UFM 1126
In external superelement data recovery restart runs in which there are
multiple normal modes solutions, the run may fail with UFM 1126 on data
block PHA.
Avoidance: Insert the following alter:
compile extdrin
alter 1
file ul=ovrwrt/ol=ovrwrt $

@
25909; 70; RESTARTS - PBCOMP - ERROR 7777 HAS OCCURRED IN ROUTINE tbk,FILE
INDEX (FILEX)=0
Restarts of decks with PBCOMP entries may fail the message:
ERROR 7777 HAS OCCURRED IN ROUTINE tbk,FILE INDEX (FILEX)=0
Avoidance: The simplest avoidance is to execute the analysis as a cold
start.
@
25916; 2001; p-Elements - CBEAM - UFM 6367 Undefined Grid Point
A client p-element model using the adaptive beam failed with the following
message;
USER FATAL MESSAGE 6367, (POLYD), GRID OR POINT ID =******** FOR ELEMENT
ID = *** IS UNDEFINED
even though the grid points were correctly defined for all elements in the
model.
Avoidance: The cause of this error is still under investigation; however,
modifying the sixth field of the PVAL entry from explicitly specifying "0"
to blank as follows caused the job to run successfully.
PVAL 1 1 1 1 0 SET 2
to
PVAL 1 1 1 1 SET 2

@
25939; 70; Optimization/Design Sensitivity - OFREQ - SFM 4276 EC 3002
If an OFREQ entry is specified in the case control of a solution 200,
modal frequency response analysis the following error message is issued:
SYSTEM FATAL ERROR 4276.FREAD ERROR CODE 3002 PID= 0
 Avoidance: Remove the OFREQ entry from the case control. A DMAP avoidance
is under investigation.
@
25948; 70; Miscellaneous - CORD1i - UFM 2006 (GP1C), Undefined Coordinate S
ystem
In MSC/NASTRAN Version 69, If two coordinate systems are defined on a
CORD1i entry, then the second coordinate system definition will be
ignored. This error will likely lead to the message:
USER FATAL MESSAGE 2006 (GP1C)
GRID POINT *** REFERENCES UNDEFINED COORDINATE SYSTEM ***
Avoidance: Other than editing the input bulk data to place the CORD1i
definitions on separate entries, there is no avoidance.
@
25950; 70.5; Eigenvalue Analysis - SUPORT, MGIV, GIV, MHOU, HOU - Wrong Answe
rs
For a two degrees freedom system, if one of the tridiagonalization
eigenvalue methods (e.g., MGIVENS) is used and the model contains a SUPORT
entry, then the answers will be wrong. A symptom of this error is the
existence of an extra rigid body mode and zero generalized mass for the
flexible body mode.
Avoidance: There are three possible avoidances.
1. Use the Lanczos eigenvalue solver.
2. Add an extra degree of freedom.
3. Remove the SUPORT entry if feasible.
@
25970; 70; Frequency Response Analysis - CBUSH, PBUSHT - Wrong Answers
If the KN field is used on the PBUSHT entry in a frequency response
solution, incorrect answers will result.
Avoidance: If BUSH elements are used in a frequency response solution, do
not use the KN entries.
@
26022; 67.5; DMAP - INPUTT2/OUTPUT2 - FORTRAN I/O ERROR 933, Unclear Message
When using INPUTT2 and ALL of the following conditions are met:
1) Multiple rewind/reads off same INPUTT2 file (as would happen in a
looping situation)
2) Using a neutral (formatted) INPUTT2 file
3) An EOF marker (OUTPUT2, ,,,,//-9/iunit $) was NOT used when writing out
the output2 file
then the job will terminate in the f06 file without any messages. The log
file file may contain any
one of the following messages:
1)FORTRAN I/O ERROR 933 - Job aborts reading INPUTT2
2)1525-001 The READ statement on the file xxxxx cannot be completed
because the end of the file was reached. The program will stop.
 Avoidance: Add an EOF marker by adding an output2 statement similar to
the statement shown below:
output2, ,,,,//-9/junit $
@
26060; 70; Submittal - Memory Allocation Error on HP-UX - UFM 3060
MSC/NASTRAN V69.X on HP-UX systems may be unable to acquire necessary
shared memory or queue resources due to lack of available resources.
These errors will be manifested as the inability to allocate memory or UFM
3060 errors citing NUSR.
The problem is that MSC/NASTRAN's shared memory and queues may not get
properly cleaned up if MSC/NASTRAN does not exit normally, e.g., is
terminated via kill -9. As a result, system resource remain committed to
jobs that have terminated, and are unavailable later on when needed for
another job.
Rebooting the system will clear up these resource, but we don't recommend
this as the fix as it's more than a little disruptive. The following
changes to the local template will give MSC/NASTRAN an additional
opportunity to clean up after itself.
MSC/NASTRAN V69.0: In <install-dir>/bin/nast69.lcl, just after the line:
%s.arch=hpux:%%debug=n:% $tcmd %a.solver% %a.rel=68:% $asgf %mem%
add the lines:
%s.arch=hpux:%pid=`grep 'Process Id' %out%.log | awk '{print $4}'`
%s.arch=hpux:%test -n "$pid" && {
%s.arch=hpux:% ipcs -mqp | grep 0x | while read type id key mode
owner group pid1 pid2
%s.arch=hpux:% do
%s.arch=hpux:% test "$pid1" = "$pid" && {
%s.arch=hpux:% case "$type" in
%s.arch=hpux:% m ) ipcrm -m "$id" ;;
%s.arch=hpux:% q ) ipcrm -q "$id" ;;
%s.arch=hpux:% esac
%s.arch=hpux:% }
%s.arch=hpux:% done
%s.arch=hpux:%}

MSC/NASTRAN V69.1, the variable "s.arch" was changed to "MSC_ARCH", and

the change becomes:
In <install-dir>/bin/nast691.lcl, just after line (247):
%MSC_ARCH=hpux:%%debug=n:% $tcmd %a.solver% %a.rel=68:% $asgf %mem%
add the lines:
%MSC_ARCH=hpux:%pid=`grep 'Process Id' %out%.log | awk '{print $4}'`
%MSC_ARCH=hpux:%test -n "$pid" && {
%MSC_ARCH=hpux:% ipcs -mqp | grep 0x | while read type id key mode
owner group pid1 pid2
%MSC_ARCH=hpux:% do
%MSC_ARCH=hpux:% test "$pid1" = "$pid" && {
%MSC_ARCH=hpux:% case "$type" in
 %MSC_ARCH=hpux:% m ) ipcrm -m "$id" ;;
%MSC_ARCH=hpux:% q ) ipcrm -q "$id" ;;
%MSC_ARCH=hpux:% esac
%MSC_ARCH=hpux:% }
%MSC_ARCH=hpux:% done
%MSC_ARCH=hpux:%}
These changes are completely safe as they only clean up system resources
that were in use by the MSC/NASTRAN executable that just finished
execution.
If the generated script is terminated, the above code still won't get a
chance to do its work. In that case, the following script can be used:
-------------------- ipc_clean -------------------
#!/bin/sh
if [ $# -eq 1 ] ; then
USER=${1}
else
USER=`whoami`
fi

if [ "$USER" = "root" ] ; then

echo "Error, will not remove ipcs for user: root"
exit
elif [ "$USER" = "daemon" ] ; then
echo "Error, will not remove ipcs for user: daemon"
exit
fi
#echo "setting user = <${USER}>"
ipcs | grep ${USER} | while read type id addr perms owner group
do
# echo "$type $id $owner"
if [ "$USER" = "$owner" ] ; then
if [ "$type" = "m" ] ; then
echo "removing mem segment: $id"
ipcrm -m $id
elif [ "$type" = "s" ] ; then
echo "removing semaphore: $id"
ipcrm -s $id
elif [ "$type" = "q" ] ; then
echo "removing message queue: $id"
ipcrm -q $id
fi
fi
done
-------------------- ipc_clean -------------------
This script can be invoked as:
ipc_clean
to remove all shared memory segments, queues, and semaphores associated
with the current username, or
ipc_clean _user_
 to do the same for the specified username. This script will remove *all*
shared memory, semaphores, and queues associated with a specific user.
Make sure the user isn't actually using these resources, e.g., is
currently running MSC/PATRAN or MSC/NASTRAN.

@
26069; 70; Restarts - RANDOM - Inefficient Processing
In MSC/NASTRAN Version 69, if a random analysis is executed as a restart
to a frequency response analysis the FRRD1 module will be unnecessarily
re-executed.
Avoidance: Apply the following DMAP avoidance.
compile sedfreq
alter 'if ( not noph2'(2),''
if ( false ) then $
@
26076; 70; Superelements - DBLOCATE - Multi-Master Data Recovery Restarts m
ay Fail
If the database for a superelement has changed from the original run, then
a multi-master data recovery restart may fail in DBLOCATE.

@
26078; 70; Static Analysis - PARAM,CURV1, PARAM,OG - Traceback in CURV1
If PARAM,CURV,1 is specified to request that the CTRIA3 and CQUAD4 element
stress and strain output be computed in the material coordinate system and
PARAM,OG,-1 is specified to avoid calculation of the stresses at the grid
points, the run will fail in the CURV module with the following traceback
printed in the log file or other machine dependent symptoms such as
excessive times or stopping without any messages.
Traceback:
Offset 0x00000a78 in procedure curv1_
Offset 0x000000d4 in procedure curv_
Offset 0x000000e4 in procedure xsem14_
Offset 0x0000097c in procedure mainal

Avoidance:
The CURV module grid output in the user defined coordinate system is
obtained by using PARAM,S1G,1 for stresses and PARAM,S1AG,1 for strains.
This grid output can either be turned off by removing these parameters,
setting them to "0", or using PARAM,OG,-1. Since, PARAM,OG,-1 is not
working, use one of the other two methods to turn off the CURV module grid
output.
@
26111; 70.5; Optimization/Design Sensitivity - MODTRK - UFM 6656 (MODTRK) PHG
REF Not Found
If mode tracking is attempted with multiple boundary conditions and if two
or more DRESP1 entries identify the same mode, mode tracking will fail and
 the message following message will appear.
USER FATAL MESSAGE 6656 (MODTRK) PHGREF CANNOT BE FOUND
Avoidance: Do not list the same mode number on multiple DRESP1 entries.
If the same mode number is
being tracked in multiple boundary conditions, a single DRESP1 entry can
be invoked by separate DCONSTR entries for each of the boundary
conditions.
@
26127; 70; Miscellaneous - EMG - SIGNAL 10: Bus Error, Segmentation Violati
on
MSC/NASTRAN Version 69 stops in EMG with no messages in the F06 file when
materials with duplicate IDs exist. The log file indicates either:
SIGNAL 10: Bus Error
or:
SIGSEGV - Segmentation violation.

In Version 68.2, USER FATAL MESSAGE 2017 MATERIAL PROPERTY IDENTIFICATION

xxx WAS DUPLICATED
was correctly issued.
@
26128; 70; Differential Stiffness - STATSUB, BAILOUT, DECOMP - Unexpected F
atal Message
PARAM,BAILOUT has the incorrect default in solutions 103, 105 and 107 when
the STATSUB command is specified; i.e., differential stiffness is
requested. This may cause jobs to stop and issue fatal messages in
DECOMP. Also, PARAM,BAILOUT,-1 is ignored in SOLs 103, 105, and 107 when
the STATSUB command is specified.
Avoidance: Insert the following DMAP avoidance:
compile statics2
alter 'call phase1c'(,-1)
bailout=-1 $
endalter
@
26132; GL; Optimization/Design Sensitivity - STATIC, SAERO - UFM 4574
Analysis types STATIC and SAERO, do not work together in solution 200. If
they are attempted in the same run the following message is issued, "USER
FATAL MESSAGE 4574, DATA RECOVERY STARTED BUT NO DISPLACEMENT VECTORS
AVAILABLE"
Avoidance: There is no avoidance.
@
26173; 2001; Documentation - GPFORCE(PLOT)=ALL - Missing Documentation
Although it is not documented in the QRG, the plot option is supported for
the GPFORCE entry.

@
26180; 70; P-Elements - TEMPF, ADAPT - Traceback
TETRA and PENTA p-elements with TEMPF thermal loading fail in the ADAPT
module with the following traceback.
Traceback:
Trap encountered:
SIGTRAP - Trace trap
Traceback:
Offset 0x00000414 in procedure gxstrn_
Offset 0x00000b80 in procedure adpegd_
Offset 0x00000900 in procedure adpeld_
Offset 0x00000c70 in procedure adp_
Offset 0x0000134c in procedure adapt_
Offset 0x000003ac in procedure xsem14_
Offset 0x0000094c in procedure mainal
--- End of call chain ---
Avoidance: Use another method to define the thermal loads.
@
26187; ; Documentation - DMAP Module Dictionary, MAKEOLD - Missing Docume
ntation
In the Release Notes for Version 69 it states that the MAKEOLD module may
be needed to convert datablocks to the old format. This module is not
documented anywhere and there is no list of the datablocks it works on.
Avoidance: A brief description follows:
Description: Used to convert datablock GEOM1, GPDT, BGPDT, CSTM, AXIC and
EST to old format for OUTPUT2 module.

Format:
MAKEOLD
IDB1,IDB2,IDB3,IDB4,IDB5/ODB1,ODB2,ODB3,ODB4,ODB5/P1/P2/P3/P4/P5/P6/P7/P8
P9/P10 $
IDBi = Input Data Block, targeted datablocks are GEOM1, GPDT, BGPDT, CSTM,
AXIC and EST.
ODBi = Output Data Block.
P1:P5 Character input, default is blank. Then generic name of the
corresponding input data block; e.g. P3 corresponds to IDB3, etc.
P6:P10 Character input, default is blank. The generic name of the
corresponding output data block; e.g. P8 corresponds to ODB3, etc.
Examples:
MAKEOLD
BGPDTS/BGPDT,SIL,GPL,EQESIN/'BGPDT'////'BGPDT'/'SIL'/'GPL'/'EQEXIN' $
MAKEOLD ESTS,BGPDTS/EST/'EST'/'BGPDT'////'EST' $
@
26190; ; Heat Transfer Analysis - RADBC, RADM - Wrong Answers, Missing Wa
rning Message
If the user intends to use RADBC, radiation to space, the absorptivity
value on the RADM bulk data entry must be specified. Otherwise, the
ambient space temperature is ignored and unexpected answers will result.
Avoidance: Make sure that when you have RADBC in the problem, you specify
the absorptivity value.
@
26204; 70; Heat Transfer Analysis - CTRIAX6 - No Warning Message, NaN in Ou
tput
No warning message is issued if convection or radiation is specified on
the axis of rotation for the axisymmetric CTRIAX6 element. This is an
impossible physical boundary condition that may lead to "NaN" being output
for the "HEAT FLOW INTO HBDY ELEMENTS".
Avoidance: The user should check whether he is applying this ill-posed
boundary condition.
@
26221; 70; MSC/ACCESS - DBCTA1 Datablock Description
The MSC/ACCESS sample programs need to be updated to the same version of
MSC/NASTRAN that they are delivered with, including having the most recent
versions of MSC/NASTRAN source in them and updating the program so they
work.
They currently contain v67.5 element definitions and, thus, cannot be used
as a guide to MSC/ACCESS users on how to access information on modern
elements.
Avoidance: Send the latest version of the MSC/NASTRAN source file
"dbctbd.f" to the client, and have the client update the "sampler.f"
sample program as practice and then update their own program in a similar
way.
@
26230; 70.5; Buckling and Differential Stiffness - GPFDR, TEMP(LOAD) - Wrong
Answers
Grid point force data recovery for static analysis in SOL 105 is incorrect
when thermal loads are applied.
In version 69, the error exists also in SOLs 103 and 107. In version 70,
the error exists also in SOLs 103 and
107 through 112.
Avoidance: A DMAP avoidance is given below for Versions 68.2, 69, and 70.
compile phase1a
alter 'call selg'
equivx ptelem/ptel/-1 $
call dbstore ptel,,,,//seid/0/' '/0 $
compile sedrcvr
alter 'GPFDR CASEDR',''
call dbfetch /ptel,,,,/seid/0/0/0/0 $
GPFDR CASEDR,UG,KELMx,KDICTx,ECTS,EQEXINS,GPECT,PG1,QG,
BGPDTS,SILS,CSTMS,VELEM,PTEL/
ONRGY1,OGPFB1/APP2/TINY//CYCFLG $
@
26231; 70; Documentation - RESTART - UFM 146 (LINKTP)
Version 69.1 cannot RESTART from a Version 69.0 database. If attempted
the the message:
USER FATAL MESSAGE 146 (LINKTP )
A 'TYPE' STATEMENT DEFINES A DATABLOCK PVTBC IN SUBDMAP SEMFREQ WHICH
IS NOT DEFINED IN THE 'NDDL'
will be encountered.
Avoidance: It is possible to perform a DBLOCATE restart.
@
26244; 70.7; Optimization/Design Sensitivity - PARAM,POST - Optimized Soluton
not on OP2 File
Solution 200 has been enhanced in MSC/NASTRAN Verison 70.7 to output the
optimized solution to the op2 file. Prior to this enhancement it was
required to update the bulk data file with the punch file solution results
and rerun a static solution to obtain an updated op2 file.
@
26250; 70.5; Documentation - DBDELETE Module Not Documented
The new DMAP module DBDELETE in Version 69 is not documented.
This module is useful for those writing looping dmaps. It was introduced
late in Version 69 development, and is not in current documentation.
Examples of its use may be seen in the fea subdmap.
Format:
DBDELETE [DATABLK=(datablk-list)] WHERE (where_expr)
The where_expr rules are similar to those for the dbdict module. A
wildcard option is available.
Example:
DBDELETE KJJ WHERE (WILDCARD)
Function: All copies of KJJ will be erased from the database,
regardless of the values of their qualifiers.
@
26267; 70; Heat Transfer Analysis - RESTART, TEMP - UFM 2166
If enforced temperatures are applied in a transient heat transfer analysis
cold start, then a subsequent restart may fail with the message:
USER FATAL MESSAGE 2166, MATRIX TO BE PARTITIONED IS OF SIZE ** ROWS
BY *** COLUMNS.
ROW PARTITION SIZE IS ** COLUMN PARTITION SIZE IS ***
(INCOMPATIBLE).
Avoidance: Insert the following alter into the Executive Control section:
compile nltran
alter 'copy ysnt/ysnr',''
endalter
@
26274; 70; Documentation - CBARAO - Plotting of Intermediate CBAR Results
In order to get a CBAR element with a plot code of 100 which provides for
data recovery at intermediate locations on the bar, a PLOAD1 must be
specified for the element in question. Remark 1 under the Quick Reference
Guide description on the CBARAO incorrectly indicates that this a
requirement for the SCALE=LE option only.

@
26278; 70; Heat Transfer Analysis - CHBDYG,AREA8, VIEW - Wrong Answers
Radiation heat transfer using the backside of the CHBDYG,AREA8 or
CHBDYG,AREA6 for view factor calculation yields incorrect results.
Avoidance : Always use the front side for the CHBDYG,AREA8 and AREA6
elements for view factor calculations.
The normal directions can be reversed on the CHBDYG,AREA8 by changing the
grid connectivities.

@
26283; 70; Documentation - Transient Heat Transfer - Incorrect Punched Outp
ut
The PARAM,CURVPLOT,1 is used to force SORT1 output for transient thermal
analysis in MSC/NASTRAN Version 68. However, the DMAP has been changed in
Version 69 so that PARAM,CURVPLOT,1 can no longer force SORT1 output in
transient thermal analysis. To request SORT1 output in Version 69 specify
THERMAL(SORT1) or FLUX(SORT1) and remove the OUTPUT(XYPLOT) section from
the case control section.
Avoidance: If you remove the OUTPUT(XYPLOT) section, then version 69 can
output the SORT1 in transient thermal analysis
@
26285; 70; Matrix Methods - Parallel Decomposition - Job Hangs
Eigenvalue analysis using the Housholder method may hang on the VAX/VMS
Version 69 system if parallel decomposition is selected.
No sparse parallel decomp has been blocked on all systems in Version 70.
Avoidance: Do not select parallel processing.
@
26294; ; Utilities - tabtst.f - illegal structure for unformatted file
In the MSC/NASTRAN Version 68.2 Installation and Operating Instructions
Guide for Windows NT for INTEL, Chapter 2 How to install MSC/NASTRAN,
Section 2.1, Installing MSC/NASTRAN, section 2.1.1, Before You Begin, it
says:
"If you are going to build the Utility Programs using the supplied source,
you must also have a suitable compiler. The supplied source has been
compiled and tested using:
WATCOM FORTRAN 77/32, Version 9.5c
WATCOM C/C++ 9.5c
.
.
 Note: Later versions of the WATCOM compilers should work but have not been
tested. The supplied source should work with other compilers available
for Windows NT but has not been tested."

This should be changed to indicate that if the customer's program reads

binary files, it probably won't work due to binary file format differences
between most Window's NT Fortran compilers.
The following is an example of an error message received when running a
binary file reading program that has an incompatible binary file format
with the WATCOM compiler used on MSC/NASTRAN:
"illegal structure for unformatted file"

@
26298; 70; Response Spectra - DDRMM, PARAM,OPTION,NRL - Incorrect von Mises
Stresses
In a response spectrum analysis, if the NRL summing option is used, then
the Von Mises stresses are calculated differently than what is described
in the NAVSEA 0908-LP-000-3010 specification. In the NAVSEA spec, the von
Mises stresses are calculated for each mode based on the sigma-x, sigma-y,
and tau-xy for that mode. The von Mises stress for each element is then
summed using the NRL criteria based on the von Mises stresses for the
modes. In MSC/NASTRAN, the component stress (e.g., sigma-x) for each
element is calculated using the NRL criteria based on the component
stresses for the modes. The von Mises stress for the element is then
calculated using the component stresses.
Avoidance: There is no avoidance.
@
26304; 70; Aeroelastic Analysis - ZONA51 - Wrong Answers
The supersonic aerodynamics code (ZONA51) used in the AERO II option in
MSC/NASTRAN has been observed to produce incorrect results when the z
distance between boxes on interfering non-coplanar surfaces is less than
half the width of a spanwise strip of boxes in the aerodynamic mesh.
Avoidance: A possible avoidance is to decrease the width of the maximum
strip. An alternative avoidance is to artificially modify the geometry of
the aerodynamic model by either making the interfering surfaces coplanar
or by increasing the z separation of the two surfaces. Another option is
to make the surfaces non-interfering by putting them in interference
groups (IGID on the CAERO1 entry).
@
26315; 70; Optimization/Design Sensitivity - CORD2i - SFM 4276, EC 7777
A shape optimization job may produce a numerical overflow (division by
zero in tranms/d) or produce the wrong updated mesh when it uses a chain
of CORD2i entries. The error may occur when the following conditions are
true:
1) At least two CORD2i entries define a parent and child coordinate
systems where the child system is a non-basic coordinate system and,
2) The non-basic child system has a different type of coordinate system
from its parent.
 For example, the following two entries will trigger the error:
CORD2R 1 2 ...
CORD2S 2 1 ...
where the first entry is the parent and the second entry is the child.
Note that the second condition above is met because the parent is of type
rectangular while the child is spherical.
Avoidance: Use the same type of coordinate system for the parent and all
the child coordinate systems when a chain of CORD2i's are used. Else use
CORD1 type definitions.

@
26329; 70.5; Optimization/Design Sensitivity - AEROS, REFC - Wrong Answers
Solution 200 supports the ability to simultaneously analyze and design
responses associated with static
aeroelasticity and flutter. If the user inputs a reference chord on the
AEROS entry that differs from
the reference chord input on the AERO entry (both fields are designated
REFC), an incorrect flutter analyses
will be performed with unpredictable results. That is, the run may
terminate with an unrelated error message or it may run to completion,
with wrong answers.
Avoidance: Make sure that the two REFC values are the same.
@
26331; 70; Nonlinear Analysis - RFORCE, NLCOMB - Segmentation Violation, Tr
aceback
Nonlinear Analysis that combines RFORCEs with other loads will fail with
the following traceback in the NLCOMB module:
Traceback:
Offset 0x0000099c in procedure nlcslb_
Offset 0x00000210 in procedure nlcslt_
Offset 0x00000234 in procedure nlcomb_
Offset 0x0000017c in procedure xsem15_
Offset 0x0000098c in procedure mainal

Avoidance: There is no practical avoidance.

@
26340; 70; Heat Transfer - RADBC, SPCFORCE - Missing Output
SPCFORCE output is missing in heat transfer. In solution 159, any
SPCFORCE output associated with an element with nonlinear capability is
overwritten by noise at all points. In both solutions 153 and 159 the
RADBC
element gives no heat to ambient points.
Avoidance: There is no avoidance.
@
26342; 70; Miscellaneous - CORDi - SFM 2005, UFM 6754
In Version 69.0, if there are multiple nested coordinate systems (e.g.,
 CORD a references CORD b and CORD b references CORD c, etc.), then the job
may fail with the following error messages on some machines. This error
has so far been observed on the SUN and HP machines when the nested
coordinate system is cylindrical and the z-axis lines up with one of the
axes of the basic coordinate system.
*** SYSTEM FATAL MESSAGE 2005 (GP1C)
INCONSISTENT COORDINATE SYSTEM DEFINITIONS DETECTED IN SUPERELEMENT
x
*** USER FATAL MESSAGE 6754 (GP1C)
NO GRID POINTS DEFINED FOR CORD1I ENTRIES IN SUPERELEMENT x
Avoidance: Changing the location of the z-axis slightly in the nested
cylindrical coordinate system may avoid the problem.
@
26349; 70; Optimization/Design Sensitivity - DVGRID - Poor Message, SFM 306
2
When a shape optimization job has DVGRID entries that refer to
non-existing grids, Version 69 produces the following message:

0*** SYSTEM FATAL MESSAGE 3062 MODULE = DOPR2

FAILURE OCCURRED IN UTILITY MAPXTI CALLED IN SUBROUTINE DSAJ1 .
REQUESTED EXTERNAL ID(S) NOT FOUND.
Version 68 produced the following message:
0*** USER FATAL MESSAGE 5492 (DSAJ1
AN UNDEFINED GRID ID = 9999 HAS BEEN REFERENCED BY A DVGRID BULK
DATA CARD WITH A DVID = 1
USER ACTION: CHECK THE DVGRID AND GRID BULK DATA CARDS
REQUESTED EXTERNAL ID(S) NOT FOUND.
Avoidance: Correct the offending DVGRID entry.
@
26351; 70; Optimization/Design Sensitivity - DLINK - UWM 7001 (DOPR1B), Wr
ong Answers
If the DLINK ID and the DDVID across multiple DLINK bulk data entries are
not in the same ascending order, then the message
USER WARNING MESSAGE 7001 (DOPR1B)

may be issued in error and the subsequent results will be incorrect.

Avoidance: Make sure the lowest DLINK ID corresponds to the lowest DDVID
and that subsequent ID, DDVID
pairs are in monotonically ascending order.
@
26352; 70; Optimization/Design Sensitivity - DRESP2, Superelements - UFM 71
39
If an optimization run with ANALYSIS=SAERO or ANALYSIS=FLUTTER has
superelements, then one of the following fatal messages are issued:
"USER FATAL MESSAGE 7139 (DOPR3C) ON DRESP2 entry ID = xxxx, RESP1 ID =
yyyy DOES NOT REFERENCE AN EXISTING DRESP1 ENTRY."
"SYSTEM FATAL MESSAGE 3007 MODULE = SEDR ILLEGAL INPUT TO SUBROUTINE
SEDR"
 "SYSTEM FATAL ERROR 4276.GETSTR ERROR CODE 3021 PID= 0"
Avoidance: The run must be made without superelements.
@
26354; 70; Superelements - CSUPEXT, SEELT - THRU Option is ignored
The CSUPEXT entry with the "THRU" option does not function correctly. Only
the first grid point is made exterior to the superelement and all other
points are ignored. The SEELT and the RELEASE entries also have problems
processing the "THRU" range. It has been experienced that the last
element of the range may be dropped. The "THRU" problem may apply to
other entries also.
Avoidance: Do not rely on the "THRU" range for the CSUPEXT, RELEASE or
SEELT entries. Explicitly specify the list of grid points or element ids
for these entries.
@
26356; 70; Utilities - BUILDSSS -
The buildsss.sh file in the <ins-dir>/msc691/util uses 'which', which is
not part of the standard UNICOS bsh environment.
This may result in the following error:
buildsss.sh[122]: test: argument expected
Avoidance:
Comment out the lines:
test -x $progexe || progexe=`which $PROG 2>/dev/null`
test -x $progexe || {
echo "$tool: prog=$PROG: executable not found."
exit 1
}
@
26376; 70; Optimization/Design Sensitivity - PARAM,CDIF,YES - UFM 1126 (GNF
IST)
If the central difference method is selected in a solution 200 run by
setting PARAM,CDIF,YES with Transient response the following error
message will be issued.
0*** USER FATAL MESSAGE 1126 (GNFIST)
DMAP MODULE DSVG1 IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME
= EGKGB , WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE EGKGB =OVRWRT
Avoidance: Insert the following DMAP alter:
compile pslgdv
alter 'egkgb'(,-1)
delete /egkgb,egmgb,egbgb,, $
delete /egkgc,egmgc,egbgc,, $
@
26377; 70; Optimization/Design Sensitivity - ANALYSIS=FLUTTER - Wrong Answe
rs
 If a design sensitivity run is made in SOL 200 that includes multiple
flutter subcases that use the same FMETHOD case control entry to invoke
the same FLUTTER bulk data entry, results will be repeated from the first
subcase rather than extracted from each subcase in turn.
Avoidance: Specify unique FMETHOD entries in each ANALYSIS=FLUTTER subcase
in SOL 200.
@
26404; 70; Virtual Mass - MFLUID - UFM 4258
In Version 69, when you select PLANE1 symmetry on an MFLUID bulk data
entry, MSC/NASTRAN incorrectly checks for PLANE2 symmetry instead. This
may incorrectly cause the following fatal message to be output if your
model violates the PLANE2 symmetry:
USER FATAL MESSAGE 4258. ELEMENT xxxx WHICH IS ON FLUID STRUCTURE
BOUNDARY, CROSSES A PLANE OF SYMMETRY
Avoidance: Put a coordinate system on the MFLUID bulk data entry such
that PLANE2 symmetry is not violated (even though it doesn't matter for
the problem). Note that PLANE1 symmetry is not checked.
@
26410; 70; Optimization/Design Sensitivity - DOM9 - Unexpected Zero Gradien
ts, UFM 6499
The optimization algorithm will stop if a violated constraint is found to
have a zero gradient. It has been
found that these zero gradients can be produced in error whenever the
amount of space required to hold the active solution vectors (that is
those that are associated with active constraints) exceeds the amount
available in the run. For example, 100 double precision solution vectors,
each having 250,000 degrees of freedom, would require 50 megabytes of
memory to hold all this information in core. If less than this is
available, all the sensitivity results will be in error and many will have
zero gradients.
Avoidance: Estimate the amount of memory required and request this amount
when running the job, or reduce the amount required by reducing the number
of solution vectors or the number of degrees of freedom or both.
@
26424; 69.1.1; Elements - PSOLID, MAT9, CORD*, TEMP, TEMPD - Wrong Answers
If thermal loads are applied to a structure and all of the following
conditions exist in the model:
1. A PSOLID property entry that references a MAT9 material property
2. The CORDM field on the PSOLID entry references a non-basic coordinate
system
then the answers will be incorrect. This error occurs on virtually all
Version 69.0 and Version 69.1; however, the error has been corrected in
Version 69 system that were built and released after April 1997. The
corrected Version 69.0 systems are the DEC VAX and DEC Alpha/NT Version
69.0.4 systems and the HP Exemplar SPP S CLASS (PA2) and NEC SX-4 Version
69.1 systems. This error applies to all CHEXA, CPENTA, and CTETRA
elements.
 Avoidance: Transform the Gij matrix on the MAT9 entry into the basic
coordinate system and references the basic coordinate system on the CORDM
field of the PSOLID entry.
@
26425; 69.1; Optimization/Design Sensitivity - TEMP(LOAD) - Wrong Answers
In MSC/NASTRAN Version 69, when an optimization job has TEMP load entries
and different types of elements in the model are subjected to the thermal
loads, wrong sensitivities may result.

Avoidance: There is no practical avoidance.

@
26427; 70; Geometric Nonlinear Analysis - ELAS3, ELAS4 - Wrong Answers
CELAS3 and CELAS4 elements may cause geometric nonlinear analysis, sol 4
in Version 69, sol 64 in Version 68, to stop in the GNFM module on DEC
Alpha systems.
Avoidance: Use CELAS1 or CELAS2 elements.
@
26429; 70; Optimization/Design Sensitivity - PARAM,OPTEXIT - Incorrect Term
ination Message
When param,optexit,4 is specified for an optimization job, the the correct
job termination message is:
RUN TERMINATED DUE TO PARAMETER OPTEXIT = 4.
However, when DESMAX on the DOPTPRM entry is set to 1 in the same deck,
the following job termination message will be produced:
RUN TERMINATED DUE TO MAXIMUM NUMBER OF DESIGN CYCLES = 1.
Avoidance: Remove DESMAX = 1 from the DOPTPRM entry, or simply ignore the
message.
@
26436; ; Superelements - PARAM,USETPRT PARAM,OPGEOM - Overflow of Print F
ormats
PARAM,USETPRT and PARAM,OPGEOM request a print of the user set table for
the g-set and the full set. When SENQSET entries are present in the main
bulk data deck, or when CAEROi entries exist, leading to k-set variable,
the degrees of freedom automatically generated have large ids which may
overflow the print formats or these parameters.
An avoidance is to use param, usetprt, 1 instead of ,0.
@
26438; 70; Installation - Version 69.1 SGI IRIX 64 level 6.2, R4400, R4000,
R5000
MSC/NASTRAN V69.1 installation will fail on SGI R4XX and R5000 machines
running the IRIX64 operating system. A clue to this problem will show up
when the mscsetup installation script incorrectly identifies the machine
as type R8000/R10000. A "uname -a" command will also echo IRIX64.
Avoidance:
Prior to installation set the environment variable MSC_ARCH to irix. In
 c-shell for example:
setenv MSC_ARCH irix
After this, the mscsetup script should correctly indicate the machine type
as R4XX/R5000 and the installation should complete normally.
@
26440; 70; Parallel Processing - OLOAD - Erroneous Output
If parallel processing is selected on SGI 4D computers, then erroneous
output that contains "nan" may appear.
Avoidance: Turn off parallel processing by removing the "parallel=x" from
the nastran statement.
@
26448; ; Superelements - LOADSET, LSEQ - UFM 4390, Poor Design
When the LOADSET command is used in partitioned bulk data superelements,
all superelements for which the LOADSET applies must have LSEQ entries. If
not, UFM 4390 will result.
Avoidance: Either have LSEQ entries for all partitioned superelements
using the LOADSET command or use the LOADSET command only in SUBCASEs
applying to the superelements which have LSEQ entries.
@
26450; 70; Plotting - OUTPUT(PLOT) with superelements - Bad Message, SWM 30
01
If there are superelements present along with an OUTPUT(PLOT) section but
there is no SEPLOT or SEUPPLOT
command in the OUTPUT(PLOT) section then SWM 3001 will be issued.
Avoidance: Specify an SEUPPLOT or SEPLOT command.
@
26455; ; Online Documentation - IVIEW - Can Not Print from Version 69 MNE
Printing using the print command defaults will not work in the Version 69
MSC/NASTRAN online encycopedia for the UNIX BSD4.2 operating system.
The UNIX System V default command lp -s ... gives the following error
messages:

Avoidance:

edit as root the file <NASTRAN-install-dir>/iview22/Iview as follows:

! example of printer definition for UNIX BSD4.2 default:
Iview*printerConfig: grz13p2 PS-doublesided, LaserWriter+, lpr -P%s %s;
no-default, LaserWriter+;
@
26464; ; Online Documentation - IVIEW - Printing to European Paper Size (
DIN A4)
Printing pages from the NASTRAN V69 Encyclopedia to DIN A4 paper (european
paper size) there is no reduced scaling. This means the contents of the
page is cut on the upper and right border.
This error exists since the first release of the on-line documentation.
 Avoidance:
Due to the non-function of the <NASTRAN-install-dir>/iview22/Iview entry
Iview*europeanPaper: True
use the following
Workaround to print NASTRAN V69 Encyclopedia pages to european paper size:
--------------------------------------------------------------------------
$ su
$ cd /msc/iview22/ps
$ cp -p psprolog.600 psprolog.600.org
Edit the standard postscript prolog file:
$ vi psprolog.600

Edit original line 49:

% modified by A. Hopf MSC/Germany 20-Mar-1997

----------------------------------
%/ia4{statusdict/a4tray known{rcignore statusdict begin a4tray end
rcrestore}if}bdf
/ia4{statusdict/a4tray known{rcignore statusdict begin a4tray 0.88 0.88
scale 50 -7 translate end rcrestore}if}bdf
%/ia4{statusdict/a4tray known{rcignore statusdict begin a4tray 0.88 1.04
scale 50 -80 translate end rcrestore}if}bdf
%------------------------------------------------------------------------
------
This creates a print with the default size of the xscale equal to the
yscale. If you want full paper size, increase the y-scaling. For
instance in the example above, comment the current /ia4 statement with %
and uncomment the third /ia4 statement.

If an extra page appears with the print message

Interleaf::pl2ps 6.0.0
you can prevent this extra printing by editing the file psprolog.600 as
follows:
original lines 71 and 72:
% modified by A. Hopf MSC/Germany 20-Mar-1997
----------------------------------
% (\tInterleaf::pl2ps )print
% SUB MINOR MAJOR vp
%------------------------------------------------------------------------
------

!!! Important: Print Encyclopedia pages with "Paper Type" A4 !!!

@
26475; 70; Submittal - MSC/NASTRAN Always Executes on first subcomplex
 The MSC/NASTRAN submittal command provides the 'subcomplex' keyword for HP
CONVEX Exemplar system to select a specific subcomplex to run the job on.
If this keyword is not set, the job is supposed to run on the current
subcomplex. An error in the driver always forces the job to run on the
*first* subcomplex, usually this is 'System'.
There are three choices:
1) Accept this restriction.
2) Force the submittal command to always choose the current subcomplex by
make the following changes to the local template file(s):
a) MSC/NASTRAN V69.0.X: In <ins-dir>/bin/nast69.lcl:
i) remove the line:
%s.arch=sppux:% echo "Subcomplex: %subcomplex%"
ii) Remove the two words "-sc %subcomplex%" from the line
%s.arch=sppux:% $tcmd /bin/mpa -sc %subcomplex% -DATA -STACK
...

b) MSC/NASTRAN V69.1: In <ins-dir>/bin/nast691.lcl:

i) remove the line:
%MSC_ARCH=sppux:% echo "Subcomplex: %subcomplex%"
ii) Remove the two words "-sc %subcomplex%" from the line
%MSC_ARCH=sppux:% $tcmd /bin/mpa -sc %subcomplex% -DATA
-STACK ...
3) Install the "S-Class" (a.k.a PA-RISC 2.0) version of MSC/NASTRAN V69.1,
this includes the correction to the submittal command and adds supports
for the PA-RISC 2.0 processor.
@
26500; GL; Optimization/Design Sensitivity - DBLOCATE - SFM 1033 (XIO)
When using DBLOCATE to continue any analysis into a subsequent
optimization analysis, the following error occurs:
0*** SYSTEM FATAL MESSAGE 1033 (XIO)
FILE 0 NOT FOUND IN FIST
0********* DUMP OF XIO PARAMETERS
0XIO NAME = 0
ATTEMPTED OPERATION = OPEN (UPDATE)
FILEXX = 0
.
.
.
0*** SYSTEM FATAL ERROR 4276.XIO ERROR CODE 4276 PID= 0
0*** USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH
DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION.
This is caused because the DBLOCATEd file is read only, but optimization
analysis requires the desiter qualifier to be set on the datablocks
Avoidance: Change the DBLOCATE statements to include "convert
(desiter=n)" where n is the design iteration number. Include a separate
DBLOCATE statement for every design iteration that will occur. For
instance, if three design iterations will occur the DBLOATE statements may
 look like the following:
ASSIGN BB='beambit.master'
DBLOCATE DATABLK=*, WHERE(SEID=9), LOGICAL=BB, CONVERT(DESITER=1)
DBLOCATE DATABLK=*, WHERE(SEID=9), LOGICAL=BB, CONVERT(DESITER=2)
DBLOCATE DATABLK=*, WHERE(SEID=9), LOGICAL=BB, CONVERT(DESITER=3)
where beambit.master is the database that the second analysis is starting
from.
@
26508; 70; Miscellaneous - PARAM,BAILOUT Ignored - UWM 6137, UIM 4258, UIM
4698
It has been observed in a model which has obvious singularities that
Version 69 may continue with the solution after flagging these
singularities using UWM 6137 (rank deficient matrix) and UIM 4258 and 4698
indicating that MAXRATIO has been exceeded. This is in spite of explicitly
setting PARAM,BAILOUT,1.
Avoidance: There is no known avoidance for this error. Search the output
file for message 6137. If it occurs look for UIM 4258 and 4698. The most
obvious evidence is large displacements (possibly as large as 1.E=12) with
low stresses.
@
26527; 70; PK-Method Flutter Solution - FLFACT - Unexpected Results when Ve
locity = 0.0
If a PK flutter analysis is attempted at zero velocity, unpredictable
results can occur, ranging from
bad answers to floating point divide checks. This error has been
corrected in Version 70 by issuing a user fatal message whenever a zero
velocity is specified on a FLFACT entry.
Avoidance: do not specify velocities of 0.0 of the flfact bulk data entry.
Very low velocities should be
avoided in general, since they may cause the aerodynamics to be
extrapolated to reduced frequencies
above those regarded as acceptable for the aerodynamic theory.
@
26530; 70; Multipoint Constraints - MPC, RBEi, MCE1 - Wrong Answers
Models that contain MPC equations or RBEs may produce wrong answers in
Version 69. The symptoms which can cause this error
usually occur when the dependent points for the equations are widely
spread over the columns of the matrix, resulting in a need to pivot when
solving the equations. In this condition the pivoting may discard terms,
leading to a plausible but incorrect solution.
Avoidance: Models with rigid elements should be run through the checka
alter delivered in the sssalter library. SYSTEM(247) may be set to 1 to
avoid this error.
@
26547; 70; Rigid Elements - RSSCON - UFM 6694 Doesn't Identify Grid Point
In Version 69.1 the following fatal messages about RSSCON are properly
issued, but instead of printing the grid ID which is causing the error,
the printout is *** . The grid IDs are properly printed in Version
69.0.2.
*** USER FATAL MESSAGE 6694 (GP4RS4)
 RSSCON CAUSES CONFLICTING ROTATIONAL CONSTRAINTS AT SHELL GRID POINT
***
@
26557; 70; Nonlinear Analysis - PARAM,KDIAG - Undocumented Differences
In MSC/NASTRAN Version 68 and prior releases, in non-linear statics if
the nonlinear stiffness matrix contained nulls on the diagonal then if
kdiag<0 (the default) the program would automatically add a 1.0 to the 0.
diagonal term(s).
If kdiag>0 the program would always add a user specified value to the all
diagonal terms.
If kdiag=0, nothing was done.
In Version 69 a change was made: First the addition of a term to the
diagonal of the stiffness matrix was based on the total stiffness
Knl+Kslide-line+Kdiff-stiff and second terms were added to the right hand
side. If KDIAG<0 these additional terms will automatically be discarded
(though in Version 69 there was a variable error-see alter below to fix)
If kdiag>0, then these terms need to be turned off after the first subcase
since they can lead to incorrect convergence. They are turned off by a
subcase with KDIAG=0 in it. The advantage of these terms is a great speed
up in convergence. However the structure is grounded until kdiag is turned
off.
The kdiag parameter is used in the case where at the start of a nonlinear
iteration there is no stiffness in the structure.
In the case of a 3-dimensional structure. Terms of zero stiffness are
often not computed zeroes but rather small
noise zeros. Thus the need for the KDIAG>0 option
Avoidance: In Version 69 the following alter will correct the dmap
problem:
COMPILE NLSTATIC LIST NOREF
alter 'CFACT=CMPLX'(1,-2) $
IF ( KDIAG=0. ) EQUIVX KAAL0/KAAL/-1 $ FOR (NEWP>-1 AND MSCHG>-1) FALSE
alter 'IF ( KDIAG>0.)','ENDIF $ + ONE' $
IF ( KDIAG>0. AND NOT(NLHEAT)) ADD ILL,KAAL0/KAAL/CFACT//KOPT $ + ONE ALL
DIA
alter 'ADD5 KAAL,KLLTX,KLLRH','' $
ADD5 KAAL0,KLLTX,KLLRH,,/KAALX///-1. $
ENDALTER
@
26560; 70; Rigid Elements - RBE3 - Wrong Answers
In MSC/NASTRAN Version 69, if the 'UM' option is used on the RBE3 entry
and the GMi terms are not in ascending order then wrong answers result.
Avoidance: Specify the GMi terms in ascending order. This error may be
avoided by setting system(247)=1. Please also see error number 26530.
@
26571; 70; DMAP - TYPE, RD - BIOMSG: ERROR 7777
If a double precision parameter is specified as follows:
 type parm,,rd,,(dumvar=0.0)
then either incorrect values or garbage (e.g., #.#####-315) will be
assigned to this variable on the Digital ALPHA OSF/1 and INTEL NT Versions
of MSC/NASTRAN. Depending on the subsequent operations, a core dump
similar to the following messages may be produced.
BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE tbk, FILE INDEX (FILEX) =
0
STATUS = 0
********* NASTRAN FILE TABLE *********

Avoidance: Modify the type statement as follows:

type parm,,rd,,(dumvar=0.d0)
@
26574; 70; External Superelement - CSUPER, DBLOCATE - Possible Wrong Answer
s
When using the external superelement feature in Version 69 with the
DBLOCATE method, if data recovery for a superelement is requested in a
separate run from the residual structure run (as shown in the superelement
seminar notes), then the answers may be incorrect. Note that this error
report does not apply to the enhanced superelement features introduced in
Version 69, as external superelement using the enhanced superelement
technology is not supported in Version 69.
Avoidance: Change the following:
dblocate db=eqexins, where (peid=0), logical = rsdb
to:
dblocate db=bgpdts, where (peid=0), logical = rsdb
@
26575; 70; Miscellaneous - RCFILE - Enhancement Request, UFM 1019
Blank characters are significant in the symbol definitions in an RC file
or on the command line.
For example, in the following line:
SYMBOL = MTF = /u/jfc/csr
The symbol name will be taken as
"MTF "
and the value as
" /u/jfc/csr"
Write the symbol definition as
symbol=mtf=/u/jfc/csr
@
26586; 70; Heat Transfer Analysis - VIEW, CHBDYG,REV - Missing View Factor
Output
 Although the MSC/NASTRAN Quick Reference Guide description of the CHBDYG
element states that View factor calculations are not supported with the
REV option, Version 68 and prior releases supported the calculations when
no third body shadowing existed. In Version 69, even for this limited
case, View factor output for the CHBDYG element using the REV option is
missing.
Avoidance: There is no avoidance.

@
26609; 70; Post-processing - PARAM,POST,-2 - Missing Output
PARAM,POST,-2 does not properly output the geometry data blocks if
superelements exist.
Avoidance: Insert the following alter:
compile phase0
alter 'POST=-2 OR ',''
if ( false ) then $
compile super3
alter 'RSPECTRX='
CALL OUT2GEOM CSTM,GPL,GPDT,EPT,MPT,GEOM2,GEOM3,GEOM4,
GEOM1,DIT,DYNAMICS,EQEXIN,BGPDT,CASECC//
OUNIT2/0 $
@
26620; 70; Transient Response Analysis - SOLs 109 and 112 - Inefficient Loa
d Calculation
In SOLs 109 and 112, transient response loads are computed inefficiently
and, depending on the number of time steps, may require significant CPU.
Avoidance: Insert the following DMAP alter in SOL 109 or 112:
compile sedtran
alter 'trlg ','trd1 '(,-1)
copy usetd/usett $
paraml usett//'trailer'/4/s,n,usetr $
PARAML USETT//'USET'//////'m'/S,N,NomSET/
's'/S,N,NosSET/
'o'/S,N,NooSET $
if ( nomset>-1 ) usetr=usetr-1 $ turn off m-set bit
if ( nosset>-1 ) usetr=usetr-2 $ turn off s-set bit
if ( nooset>-1 ) usetr=usetr-4 $ turn off o-set bit
modtrl usett// / / /usetr $
TRLG CASEt,USETt,DLT,,BGPDTS,SILS,CSTMS,TRL,DIT,
,,rpd,EST,MPTS,MGG/
ppt,,,,pdt,tol, $
compile semtran
alter 'trlg ','trd1 '(,-1)
copy usetd/usett $
paraml usett//'trailer'/4/s,n,usetr $
PARAML USETT//'USET'//////'m'/S,N,NomSET/
's'/S,N,NosSET/
'o'/S,N,NooSET $
if ( nomset>-1 ) usetr=usetr-1 $ turn off m-set bit
if ( nosset>-1 ) usetr=usetr-2 $ turn off s-set bit
if ( nooset>-1 ) usetr=usetr-4 $ turn off o-set bit
 modtrl usett// / / /usetr $
TRLG CASEt,USETt,DLT,,BGPDTS,SILS,CSTMS,TRL,DIT,
,,rph,EST,MPTS,MGG/
ppt,,,,pht,tol, $
@
26624; 70.7; DMAP - VECPLOT, option 1 - Output Matrix Wrong for Fluid Problem
s
In a fluid problem (CD=-1 on GRID entry), if option 1 is used in module
VECPLOT, then the output matrix is bad. A symptom of this error is
indicated by a different matrix trailer size in the f04 file as compared
to the matrix size printed in the f06 file. Sometimes "BIOMSG: ERROR
211 HAS OCCURRED IN ROUTINE BLDPK" fatal message occurs. This output
matrix is not used in any standard SOL sequence, so only user-written DMAP
may encounter this error.
@
26633; 2001; DMAP - BNDSPC, APPEND - BIOMSG: ERROR 1404, Wrong Answers
The BNDSPC module is a new module introduced for partitioned bulk data
superelements in Version 69. Its function is to take spcs and spcds on
upstream superelement boundary points and transfer them to the downstream
superelement with the same boundary point. While this module has passed
many tests, different bad behaviors have been detected on some input
files, including:
[The s-set variables appear out of sequence when printed by use of param,
usetprt, 0. If the input datablocks (YSB and YSD) to the BNDSPC module
are null, then corrupted output datablocks may be generated which may lead
to failure in a subsequent module with the following messages: this may
be wrong. delete this sentence from error report]
BIOMSG: ERROR 1404 HAS OCCURRED IN ROUTINE GETSTR ,
This error has been observed to occur in the APPEND module. However, it
could appear in other downstream modules, if the APPEND module is branched
over and the corrupted data block are used elsewhere. Also the YS1 vector
is described as a table rather than a matrix when printed with MATPRN, and
its values are typically very small numbers such as 1.e-330. At other
times these data blocks may cause MATPRN to crash. The alter listed below
avoids this problem for null YS vectors.
This error is most likely to occur when user-dmap is present. This error
is still under investigation, and its effects may not all be known at this
time. The bad Ys vector symptoms were observed when diag 20 was turned on
before the module.
An avoidance is to prevent this module from moving terms by doing so
yourself when you suspect it is malfunctioning. That is, put spcs on
boundary points at the lowest level possible. After running the model,
and confirming that you moved all
of the spcs that were on the boundary, you can equiv the inputs to the
outputs of this module, and branch over it. Another avoidance which
sometimes works is to remove diag 20 from the input file.
@
26634; GL; Postprocessing - PARAM,POST, ANALYSIS - Missing Plot Output
Optimization jobs which have multiple analysis types and requests to the
MSC/NASTRAN plotter can cause problems with the OUTPUT2 file for
 MSC/PATRAN. The OUTPUT2 file may be missing data and contain extra data.
Avoidance: Do not combine NASPLOT requests with PARAM,POST,-1 in solution
200.
@
26677; 70; Single Point Constraints - SPC, SPCD - UWM 2051A, UFM 2052, SFM
3001
If enforced displacements are specified on the SPC or SPCD entries on grid
points that are not in the structure, then the following messages:
*** USER WARNING MESSAGE 2051A (GP4)
UNDEFINED GRID POINT x IN SPC SET y IGNORED. (ANY
OTHERS ALSO IGNORED)
*** USER FATAL MESSAGE 2052 MODULE = GP4
UNDEFINED GRID POINT x IN SINGLE-POINT CONSTRAINT SET
z
are issued. These messages are the correct messages and the job should
have terminated in GP4. MSC/NASTRAN is designed to continue when spcs are
specified on non-existent grid points, but not on spcd values. However,
the job will continue without producing a uset table. The next module
that needs the uset table will then fatal with the following error
message:
*** SYSTEM FATAL MESSAGE 3001 MODULE = xxxxx
THE INPUT DATA BLOCK NAMED USET0 AT POSITION a DOES NOT
EXIST.
Avoidance: Look at the grid point that UFM 2052 is pointing to in GP4 and
remove all illegally specified SPCD values.
@
26728; 70; Rigid Elements - RSSCON - UFM 4296, UFM 6692
The shell-solid connector RSSCON fails for certain irregular geometries
when the ELEM option is used.
In the example, the following fatal messages are issued
*** USER FATAL MESSAGE 4296. (SECNSD) ILLEGAL GEOMETRY FOR QUAD4
ELEMENT WITH ID = 129131
*** USER FATAL MESSAGE 6692 (SECSR )
RSSCON NUMBER 129780 CANNOT CONNECT AN EDGE OF SHELL ELEMENT
129131 TO SOLID ELEMENT 129079
The failure occurs when the shell element shares an edge with the solid
element and the shell plane is tilted
to the solid face.
Avoidance: Use the GRID option to make the connection.
@
26735; 70; Eigenvalue Analysis - EIGRL - UFM 17 SYMBOL MEMSAVE IS UNDEFINED
If the System Cell MEMSAVE=1 is specified instead of SYSTEM(229)=1, the
run results in a UFM 17.
@
26739; 70; Documentation - EIGRL, NASTRAN SYSTEM(229)=1 - Misleading Docume
ntation
The documentation for SYSTEM(229), a.k.a. MEMSAVE, indicates that setting
this system cell to 1 reduces scratch space usage in Lanczos by 67%. This
is the maximum possible savings. Jobs that execute three shifts or less
will experience no savings.
@
26747; 70; Superelements - MATREDU - SFM 4276 EC 3012, Wrong Answers
In superelement analysis, if rigid elements or mpcs exist for a
superelement but no spcs exist, and damping elements exist that do not
intersect the m-set DOFs the damping effects are discarded. This may lead
to SYSTEM FATAL MESSAGE 4276, ERROR CODE 3012 or wrong answers
Avoidance: There are two avoidances:
1. Add a token grid point to the superelement. Autospc will cause the
dummy grid point to generate spcs.
or
2. Insert the following DMAP alter:
compile semrb
alter 'matredu',''
paraml uset//'uset'//////'f'/s,n,nofset/
'a'/s,n,noaset/
'm'/s,n,nomset/
'o'/s,n,nooset/
's'/s,n,nosset $
matgen ,/iff/1/nofset $
equivx iff/rfa/nooset $
if ( nooset>-1 ) then $
matgen ,/iaa/1/noaset $
umerge1 uset,iaa,goa,,/rfa/'f'/'a'/'o'/1 $
endif $
equivx rfa/rna/nosset $
if ( nosset>-1 ) umerge1 uset,rfa,,,/rna/'n'/'f'/'s'/1 $
equivx rna/rga/nomset $
EQUIVX xGG/xNN/NOMSET $
if ( nomset>-1 ) then $
equivx gm/gmf/nosset $
if ( nosset>-1 ) upartn uset,gm/gmf,,,/'n'/'f'/'s'/2 $
equivx gmf/rma/nooset $
if ( nooset>-1 ) then $
upartn uset,gmf/gma,,gmo,/'f'/'a'/'o'/2 $
mpyad gmo,goa,gma/rma $
endif $ nooset>-1
umerge1 uset,rma,rna,,/rga/'g'/'m'/'n'/1 $
MCE2 USET,GM,xGG,,,/xNN,,, $
endif $ nomset>-1
IF (NOSSET>=0) UPARTN USET,xNN/,xSF,,/'N'/'F'/'S' $
smpyad rga,xgg,rga,,,/xaa/3////1////6 $
@
26752; 70; Frequency Response - SOLs 11, 111 and 200; SDAMP - Inefficient P
rocessing
 For modal frequency response (SOLs 11,111, and 200), if only modal
damping (SDAMP) exists in the model, then the uncoupled solution algorithm
should be used. However in Version 69.1, the coupled solution algorithm
will be used even though only modal damping exists in the model.
Avoidance: Insert the following dmap alter for SOLs 111 and 200:

$
compile freqrs souin=mscsou list
$$$$
alter ' frrd1 ','' $
FRRD1 CASES,DIT,CKDDF,CBDDF,CMDD,K4DDF,PDFI,FRLI,FOLI,
EDT,SILD,USETD,/
UDFI,FOLI1/SOLTYP/-1/////S,N,FIRSTBD0/FRRDSET $
$$$$
alter ' frrd1 '(2),''
FRRD1 CASES,DIT,CKDD,CBDD,CMDD,K4DD,PDFx,FRL,FOL,
edt,sild,usetd,/
UDF,fol1/
SOLTYP/-1/////s,n,firstbad/frrdset $
$

For SOL 11 (or 30) insert the following alter:

compile sol11
alter 'frrd1',''
FRRD1 CASExx,,KHH,BHH,MHH,,PH,FRL,OLST,edt,sild,usetd,/
UHV,olst0/
FORM/-1/////S,N,FIRSTBAD $
@
26762; 70; Eigenvalue Analysis - MGIV, GIV, HOU, MHOU - Missing Eigenvector
, UIM 5218
On long word machines like CRAY, if the following user information message
is encountered in an eigenvalue analysis using a transformation method
such as Givens or Househoulder, then some of the eigenvectors requested
may not be output.
USER INFORMATION MESSAGE 5218
EIGENVALUE APPROACHING INFINITY AT ** MODE. EIGENVECTORS WILL NOT BE
COMPUTED BEYOND THIS POINT.
The error only occurs when ND on the EIGR entry is less than the number of
eigenvalues calculated. In this case there will be one missing
eigenvector for each mode that UIM 5218 is output for.
Avoidance: Increase ND by the number of missing eigenvectors.
@
26772; GL; Nonlinear Transient Heat Transfer - TEMPBC,TRAN - Wrong Answers
If there are unreferenced type TRAN, TEMPBC entries in the test deck, then
different answers occur. This is a design limitation. TEMPBC entries
internally generate CELAS2 elements and DAREA input. The CELAS2 elements
 are generated whether the TEMPBC is referenced or not and this may cause
unexpected results.
Avoidance: Remove any unreferenced type TRAN, TEMPC entries.

@
26775; 70; Superelements - DEFORM - Wrong Answers
If DEFORM bulk data entries are included in a part superelement file (that
is, after a BDGIN SUPER= entry) they will be ignored.
Avoidance: Place DEFORM entries in the main bulk data deck.
@
26778; 70; Nonlinear Transient Heat Transfer - EMG - Wrong Answers
When multiple subcases are specified in a nonlinear heat transfer analysis
in MSC/NASTRAN Version 69, module EMG does not generate the nonlinear heat
transfer capacitance matrix BGGNL at the first iteration. This leads to
wrong answers.

Avoidance: Do not specify subcases.

@
26818; 70; Optimization/Design Sensitivity - ANALYSIS=SAERO, UMERGE1 - SFM
3001
The design intent in solution 200 is that all static aeroelasticity
subcases should have the same boundary condition but that other analysis
types can have different boundary conditions specified. The sensitivity
analysis for static aeroelasticity may produce a fatal error in UMERGE1 of
the form below if different boundary conditions are applied to static
aeroelastic subcases:
SYSTEM FATAL MESSAGE 3001 MODULE = UMERGE1
THE INPUT DATA BLOCK NAMED USET AT POSITION 1 DOES NOT
EXIST.
USER INFORMATION: THIS ERROR IS CAUSED BY ONE OF THE FOLLOWING:
Avoidance: This following is a dmap avoidance for the problem:
compile pslgdv
alter 'desaerdr'(1,-1)
CALL SETQ CASES//SEID/PEID/S,MTEMP/S,K2GG/S,M2GG/S,B2GG/S,MPC/
S,SPC/S,LOAD/S,DEFORM/S,TEMPLD/S,P2G/S,DYRD/S,METH/
S,MFLUID $
PARAML CASES//'DTI'/1/255//S,N,SUPORT $
CALL PMLUSET USET//S,NOASET/S,NOBSET/S,NOCSET/S,NOGSET/S,NOLSET/
S,NOOSET/S,NOQSET/S,NORSET/S,NOSSET/S,NOTSET/
S,NOVSET/S,NOA/S,NOSET/S,NORC/S,NOMSET/S,NOUSET $
compile fea
alter 'saersens'(1,-1)
CALL PMLUSET USET//S,NOASET/S,NOBSET/S,NOCSET/S,NOGSET/S,NOLSET/
S,NOOSET/S,NOQSET/S,NORSET/S,NOSSET/S,NOTSET/
S,NOVSET/S,NOA/S,NOSET/S,NORC/S,NOMSET/S,NOUSET $
@
26825; ; Nonlinear Slide-Line Analysis - BLSEG, BCONP - Wrong Answers
The statement below is in error. Replace unsymmetric with symmteric and
 vice-versa. Unsymmetric is the default and use the deafult option.
Rakesh Allahabadi
Wrong answers may occur for nonlinear slide-line analysis when the
unsymmetric search option, PTYPE=2 on the BCONP bulk data entry is
selected.
Avoidance: Use the symmetric option which is the default.

@
26842; 70.5; Superelements - ASETi, BSETi, CSETi Entries Ignored
When part superelements are present, aseti entries in the main bulk data
section are ignored. That is all free degrees of freedom are placed in
the a-set, and none in the omitted set. In a non-superelement model all
free degrees of freedom not placed on aseti entries are moved to the
omitted set.
Avoidance: An avoidance is to use BSETi entries in place of ASETi
entries
@
26844; ; Aeroelastic Analysis - SPLINEi, OMITi, ASETi - Wrong Answers
In SOL 144 and for ANALYSIS=SAERO in solution 200, wrong answers will
result if the user splines the aerodynamics to omitted degrees of freedom.

Avoidance: Do not spline to omitted structural degrees of freedom

@
26846; 70; Submittal - XNASTRAN - Job will not run with batch=yes
If a job is submitted using the XNASTRAN GUI with "batch=yes" and the
"after" keyword blank, the job will not execute.
Avoidance: Set the "after" keyword to a non-blank value. Setting
"after=now" will cause the job to run as soon as it's submitted. See the
at(1) man page for additional information and acceptable keyword values.
@
26855; 70; Postprocessing - PARAM,POST - Wrong Results Written to OUTPUT2 F
ile
If PARAM,POST,-1 or -2 is specified in a modal response run (e.g. SOL 111
or 112), and XYPLOTS are requested, or SORT2 output is requested, then the
displacements written to the OUTPUT2 file are the modal eigenvectors
instead of the frequency or transient response solutions.
Avoidances: Two avoidances follow, however, the first avoidance is
inefficient and the second avoidance is recommended.
1) Insert PARAM,DDRMM,-1
or
2) Insert the following DMAP alter (V69.x)
compile sedrcvr
alter 'if ( app1=\'mmreig\' and not(sort2) )',''
if ( app1='mmreig' ) then $
 alter ' sdr3 ',''
SDR3 qUGV1,OPG1,qQG1,qEF1,qES1,qSTR1/
OUGV2,OPG2,OQG2,OEF2,OES2,OSTR2 $ SORT2 OUTPUT
@
26866; 70; Miscellaneous - CURV - SFM 6806
In MSC/NASTRAN Version 69, if the CURV module is used with elements other
than QUAD4s or TRIA3s the run may abort with the following message:
*** SYSTEM FATAL MESSAGE 6806 (CURV)
RETURNED ERROR CONDITION -3, LOC CODE = 645, IN SUBROUTINE
CURV1, FILE NUM = 104
(NOTE ADDITIONAL MESSAGES.)
No CURV module output will occur. The non-CURV module output will be
available in the f06 file and is unaffected.
Avoidance: There is no avoidance.
@
26867; 70; Optimization/Design Sensitivity Analysis - SELA, INREL - UIM 457
0, UFM 4575
If inertia relief is requested in solution 200 with superelements then the
run may fail with UIM 4570 and UFM 4575.
Avoidance: Insert the following DMAP alter:
compile pslgdv
alter 'null egx'
type parm,,i,n,nopgs $
NOPGS=NDVAR*6 $
matgen ,/egpg6/7/lusets/nopgs $
@
26869; 70.0.1; DMAP - OUTPUT4, Sparse - Quits No Message
If an OUTPUT4 request is made for a sparse, formatted matrix, then the job
quits in the OUTPUT4 module with no message, no "END OF JOB" and the
OUTPUT4 file is empty.
Avoidance: Do not request sparse output.
@
26879; 70; Eigenvalue Analysis - SOL 3, 10, 11, 12, EIGRL - Poor performanc
e
A lanczos eigenvalue analysis, EIGRL entry, may require more memory in
solutions 3, 10, 11, and 12 than in Version 68.2 or the corresponding
structured solution sequence.
Avoidance: Insert the following DMAP alter:
malter 'cmset=\'v\''
cmset='a' $
@
26899; ; Composites - Buckling, PCOMP - Spurious Modes
Spurious modes may be observed in linear buckling (Sol 105) of composite
shells that have extremely soft transverse shear material properties. The
 error is under investigation but it is possible that the spurious modes
are be due to a limitation in Mindlin shell theory.
@
26953; 70.5; Documentation - V69 Release Notes - Error in sample problem inpu
t deck
The input deck for the V69 Release Notes sample problem "blexam.dat"
contains an error on the continuation card for PBARL 1 (should be "A"
instead of "B"). This error is in the file itself as well as the book
(pp. 55-57). If run as is, the input gives a fatal error. The corrected
test deck has been put on the Version 70.5 delivery in the archive
directory, the documentation can not be updated.
@
27029; 70.5; Loads - DEFORM, GRAV - Wrong Answers
If DEFORM entries are combined with GRAV entries with the same SID, wrong
OLOAD resultants and wrong answers result. It is documented that DEFORM
entries require unique SID; however, the following fatal messages should
be issued not wrong answers:
USER FATAL MESSAGE 628 (IFP1D)
THERMAL, DEFORMATION, AND EXTERNAL LOADS CANNOT HAVE THE SAME SET
IDENTIFICATION NUMBER.
or:
USER FATAL MESSAGE 2134 MODULE = GP3
LOAD SET 1 DEFINED FOR BOTH GRAVITY AND NON-GRAVITY LOADS.

@
27036; 70; Loads - PLOAD1, FRPR - Erroneous Output
When applying a PLOAD1 with a SCALE of FRPR, and the load direction is
parallel (or near parallel) to the beam element, then a zero load will be
generated in versions prior to Version 69. In Version 69, an undefined
load is generated which leads to output that contains the following
character "NANQ".

Avoidance: Since the load is parallel to the axis, use the SCALE of FR
instead.
@
27048; 2001; Optimization/Design Sensitivity - FREQ, DRESP1 - SFM 4276, Trap
Encountered
If a FREQ response type is used on the DRESP1 entry and the associated
eigenvalue is negative, the code
attempts to compute the frequency by taking the square root of the
negative eigenvalue, subsequently a trap occurs in subroutine dsad11 and
the job terminates without a meaningful explanation.
Avoidance: A negative eigenvalues typically occurs with unsupported rigid
body modes. It is probably not the
intent of the user to place design conditions on these roots. Therefore,
review the design intent and force the eigenvalue analysis to only find
non-negative eigenvalues.
@
27055; 70; Static Analysis - SNORM - UFM 316 (IFPDRV)
 Even though GRID ids may be up to one hundred million, the GRID ids
referenced on a SNORM bulk data entry must be less than one million. If
larger GRID ids are referenced the message:
USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA ENTRY SNORM
will be issued.
Avoidance: Other than renumbering the model so that the grid id's are
less than one million, there is no avoidance.
@
27065; ; DMAP - Inconsistent lusets - Possible Infinite Loop
When using dmap, the lusets parameter of the TA1 module must be set
consistent with the values in the SIL table. If these values are not set
consistently, the results are unpredictable. In one instance, when lusets
was set to 1 rather than the value of 150 set in the SIL data block, the
TA1 module produces a corrupted gpect data block. When this data block is
subsequently input to the EMA module, the job is trapped into an infinite
loop and terminates only after it runs into a resource limit (e.g.,
insufficient disk space).

Avoidance: Set the luset value equal to the second word in the trailer of
sil
@
27139; 70; DMAP - DBDICT - Missing Output
The datablk(local) option of the dbdict DMAP command which should list the
non-nddl datablocks contained in a subdmap does not work.
Avoidance: DBDIR ////64 $ can be used to provide the same information.
@
27147; 70; Nonlinear Analysis - CBUSH, PARAM,LGDISP - SFM 2085
Nonlinear jobs with PARAM,LGDISP,1 and CBUSH elements may fail with the
message:
SYSTEM FATAL MESSAGE 2085 MODULE = NLITER
ELEMENT TYPE 102 NOT SUPPORTED FOR NONLINEAR ANALYSIS.
SYSTEM FATAL MESSAGE 3007 MODULE = NLITER
ILLEGAL INPUT TO SUBROUTINE NLEMG

Avoidance: Other than not selecting PARAM,LGDISP there is no avoidance.

@
27247; 70.5; Static Analysis - SOL 1, LSEQ - Missing Output
Solution 1 accepts the LOADSET case control command and the LSEQ entries
it references. It calculates loads and solutions
for all of the load conditions listed on LSEQ entries, but outputs loads
only for the first loading condition.
Avoidance: Use subcases with LOAD commands instead.
@
27255; ; DMAP - TABPT, CASECC - SFM or Nan in Output
When a user dmap uses the TABPT module to print the CASECC datablock, a
system fatal message may occur on certain platforms if a parameter is
 specified in the case control. Other platforms may not give a system
fatal message, but instead print "nan" or "Nan" in the output.
Avoidance: Put the parameter in the bulk data if possible.
@
27257; 70.7; Parallel Processing - SYSTEM(107) - SWM 6299 (PARCPU)
In NASTRAN Version 69 through Version 70 on HP computers with pa2.0 chips
and on all NASTRAN versions on SUN Solaris Version 7, the following
message in the execution summary will indicate that no parallel processing
is occurring, when in fact it is:
*** SYSTEM WARNING MESSAGE 6299 (PARCPU)
NO PARALLEL PROCESS CREATED
Avoidance: Ignore the message.

@
27286; 2001; Transient Response Analysis - STRAIN(PUNCH) - Mislabeled Punch F
ile
When strain output is requested for transient analysis, the punch file
labels it "POINT ID" instead of "ELEMENT ID". The element ID numbers and
the "ELEMENT TYPE" labels are correct.
@
27291; 70; Eigenvalue Analysis - REIGL - Incorrect Hiwater Memory Output, P
oor Performance
The REIGL module does not keep track of the maximum amount of memory it
uses so the HIWATER value reported in the .f04 file may be underestimated
for eigenvalue jobs that go through REIGL. An eigenvalue run using mem=X,
where X is the HIWATER value from a previous run of the same deck, may run
slower, use a smaller MAXSET (Lanczos block vector size), require many
more solves, or may not have sufficient memory to complete.
@
27368; ; Executive - ASSIGN - Cryptic Message, SFM 1019
If a file exists and the keyword NEW is indicated on an ASSIGN statement
SYSTEM FATAL MESSAGE 1019 will be printed. Included in this message is
text returned by the operating system. It should be noted that under some
operating systems this text does not always clearly indicate the nature of
the file open problem and may include the somewhat cryptic message "Not a
typewriter (errno=25)".
@
27503; 70; Superelements - Split Database - UFM 1127
When using a split database with superelement in Version 69, if the
upstream database for a superelement is offline during a residual
structure run, then the job will fail with the following error messages:
USER FATAL MESSAGE 1127 (GNFIST)
DMAP MODULE XPURGE IS ATTEMPTING TO ACCESS DATA BLOCK 201 NAME = BTOPO
WHICH IS ON AN UNAVAILABLE DBSET
Avoidances: 1. Leave the upstream database online.
2. Modify the NDDL. Replace "PATH NLOOP ..." with:
" PATH NLOOP PEID,APRCH,HIGHQUAL,NLOOP $"
@
27515; ; SSSALTERS - rflagb.v69 - Restarts
Restarts using the rflagb.v69 alter in the sssalter directory have not
been tested. While some restarts may work correctly, this can not be
determined for any particular run without testing, and perhaps
modification of the alter.
Restarts are not supported for production work, and should be regarded as
an experimental technique when using the
alter.
Avoidance: Do not use the alter for restarts on production runs.
@
27516; 70; SSSALTER - effmassa.v69 - UFM 2025
The effmassa.v69 dmap alter included in the sssalter directory for V69 may
fail with the message:
USER FATAL MESSAGE 2025, UNDEFINED COORDINATE SYSTEM
Avoidance: The cause of this is under investigation. If you encounter this
error, contact MSC client support for an updated version of the alter.
@
27524; 2001; Superelements - DMIG - UFM 327
If a DMIG matrix is input after two or more BEGIN SUPER or BEGIN AUXMODEL
bulk file delimiters, and the DMIG matrix in each partitioned file has the
same name, an error message of the form
*** USER FATAL MESSAGE 327, BAD DATA OR FORMAT OR NON-UNIQUE NAME. DMIG
AXGGT SORTED CARD COUNT = 1
is produced.
Avoidance: An avoidance is to place a second dmig matrix in each file,
with a different name. Only a header entry is needed. The code keeps
only one name from the dmig entries in the bulk data deck stored between
repeated calls to ifp.
@
27533; 70; Matrix Methods - SPARSE - Wrong Answers on SGI R4400 Computers -
Very Serious
In Version 69 on SGI R4400 computers, if the memory specified with the
default memory or a mem= on the command line causes the MEMORY AVAILABLE
to be between the MINIMUM MEMORY REQUIREMENT and the SPARSE DECOMP
SUGGESTED MEMEORY as listed in USER INFORMATION MESSAGE 4157 of the f04
file, bad results may be calculated. One symptom of the bad results is a
very large epsilon in the f06 file. Larger problems are more likely to be
affected by this. Sometimes under these conditions the problem will
terminate with a Segmentation fault.
Avoidance: Specify more memory or set SYSTEM(205)=1 and SYSTEM(198)=1.
The latter avoidance can cause up to a 30% increase in decomposition
times. MSC RECOMMENDS adding the two system cell settings to
install_dir/conf/irix/nast69rc for V69 and install_dir/conf/irix/nast691rc
for V69.1 BEFORE continuing production runs. These are the architecture
specific rc files (see section 3.6 of the UNIX Configuration and
Operations Guide).
@
27561; 70; Post Processing - CHEXA, CPENTA, CTETRA - Wrong Answers in XDB
 The nonlinear stress values for the center point of any solid element are
wrong in the .XDB graphical database. The stress in the x direction is
swapped with the stress for the y direction. All other data is correct
(including nonlinear strains for the center point). The data blocks
affected are NHEXR, NTETR, and NPENR.
Avoidance: There is no avoidance.
@
27572; 69.1; Documentation - Logical Symbol, RC file - UFM (decode_keyword)
In Version 69, you can define a logical symbol in the RC file. However,
if the logical symbol is defined in the rc file as shown on page 43 of the
MSC/NASTRAN Configuration and Operations Guide, then the job will fail
with the following error messages:
*** USER FATAL MESSAGE (decode_keyword)
SYMBOL XXX=yyy
Set by ./.nast691rc[1].
This keyword is not recognized.
Avoidance: Replace the following statement

SYMBOL DATADIR=/dbs/data
with
SYMBOL=DATADIR=/dbs/data
@
27580; 70.0.1; PLOT - SEUPPLOT - Incorrect Plots
If a superelement run is sent to the plotter and the residual contains no
plot information, the resulting plots may be bad for SEUPPLOT 0.
Avoidance: Include at least one grid/scalar which has a finite
displacement.
@
27597; 70.5; Postprocessing - PARAM,POST,-1 or -2 - Mislabeled Datablocks
Solutions 1-16 are new in Version 69. If PARAM,POST,-1 or -2 is used in
these solution sequences, the output datablocks are mis-labeled. For
example, GEOM1 will be mis-labeled as IGEOM1 on the output file and I-DEAS
or MSC/PATRAN will not interpret the data properly.
Avoidance: Other than using solution 101 thru 200 there is no avoidance.
@
27598; 70.5; Heat Transfer Analysis - Restarts with APP COUPLED - UFM 9086
A restart into a structural analysis with APP COUPLED from a previous heat
transfer run fails with the message:
USER FATAL MESSAGE 9086 (GETLOOP)
THE DATA BASE IS INSUFFICIENT FOR PARAM,LOOPID RESTARTS. CHECK FOR BULK
DATA CHANGES RELATED TO
GEOMETRY, CONNECTIVITY, AND PROPERTIES, WHICH CAUSE THE RECALCULATION OF
NONLINEAR DATA.
Avoidance: Insert the following DMAP alter:
compile super1
alter 'PARAML CASECC',''
 type parm,nddl,i,y,subid $
PVT PVT,CASECC/PVTX $
PARAML CASECC//'DTI'/SUBID/251////S,N,PATH $
compile getloop
alter 1
type parm,nddl,char8,n,aprch $
if ( GETSYS(1,21)=5 and APRCH=' ' ) APRCH='HEAT ' $
@
27611; 70; Optimization/Design Sensitivity - SPC, SPCD, BC - Wrong Answers
In Design Optimization (Sol 200) with multiple boundary conditions, an
enforced displacement of 0.0 is not preserved after the initial design
cycle when one subcase puts a non-zero enforced displacement (using SPC or
SPCD) on the same degree of freedom that a different subcase puts an
enforced displacement of 0.0. When these conditions apply, the initial
design cycle will be correct, but the subsequent design cycles will put
the nonzero enforced displacements on all subcases.
Avoidance: Use the following DMAP avoidance:
compile update nolist, noref $
alter 129 $ V69.0
$alter 132 $ V69.1
paraml ysb//'null'////s,n,nullysb $
if (nullysb < 0) then $
matgen ,/yg1/7/lusets/1 $
else $
alter 130 $ V69.0
$alter 133 $ V69.1
endif $
@
27617; 70; SSSALTERS - alter9ga.v69 - Infinite cpu Loop
If alter9ga.v69 is used in a frequency response solution, an infinite cpu
loop will result. This problem did not occur in previous versions of the
alter.
Avoidance: Call MSC and request an updated version of the alter.
@
27623; 70; Superelements - SELA - Wrong Answers
If there are more than 5 superelements at a single level, then the static
load matrices in the downstream superelement will be wrong.
Avoidance: Increase BUFFPOOL on the NASTRAN statement or insert the
following DMAP alter:
compile fea noref nolist
alter 'return'(,-1)
dbdelete datablk=(zuzr11) where(zname='pav' and zuzr1=*) $
$
compile sela1 souin=mscsou noref nolist
ALTER 'TEMP=SEID','END $ SELA1'(1,-2) $ V68.2.3 and up
call dbfetch /pav,,,,/s,seid/0/0/0/0 $
add5 pav,pjc,,,/pg $
$
compile phase0 noref nolist
$
$ The alter and GP1 call are for V68.2.3 and up
$
ALTER 'IF ( IMAGE OR EXTER ) THEN'(1,+1) $ V68.2.3 and up
GP1 GEOM1I,GEOM2Scr,,GDNTAB,MEDGE/
xGPLS,xeQEXIN,xGPDT,xCSTM,xBGPDT,xSIL,xVGFS/
S,N,xLUSET/0/NOGPDT $
call dbstore xeqexin,xsil,,,//s,seid/0/' '/0 $
$
$
compile phase1dr noref nolist
alter 1 $ All versions from V68.1 and up
type parm,,i,n,iseid=0,isetyp $
$ generate 6x6 mirror transformations
matgen ,/i1/6/6/0/1 $
matgen ,/i2/6/6/1/1 $
matgen ,/i3/6/6/2/1 $
matgen ,/i4/6/6/3/1 $
matgen ,/i5/6/6/4/1 $
matgen ,/i6/6/6/5/1 $
add5 i1,i2,i3, ,/t001x/-1./ 1./ 1. $
add5 i4,i5,i6,t001x,/t001y/ 1./-1./-1. $
add5 i1,i2,i3, ,/t002x/ 1./-1./ 1. $
add5 i4,i5,i6,t002x,/t002y/-1./ 1./-1. $
add5 i1,i2,i3, ,/t003x/ 1./ 1./-1. $
add5 i4,i5,i6,t003x,/t003y/-1./-1./ 1. $
matmod t001y,,,,,/t001,/28 $
matmod t002y,,,,,/t002,/28 $
matmod t003y,,,,,/t003,/28 $
mpyad t001,t002,/t012 $
mpyad t001,t003,/t013 $
mpyad t002,t003,/t023 $
smpyad t001,t002,t003,,,/t123/3 $
ALTER 'SEP2DR SLIST,EMAP' $ V68.2.3
iseid=iseid+1 $ superlement counter
$
ALTER 'RETURN'(1,-2) $ V68.2.3
$
if ( seid>0 ) then $
if ( image or exter ) then $
call dbfetch /xeqexin,xsil,,,/s,seid/0/0/0/0 $
isetyp=(iseid-1)*7+7 $ word location of type
paraml emap//'dti'/1/isetyp//s,n,setyp $ type
setyp=setyp/64 $ right shift by six bits
if ( exter ) setyp=0 $
if ( setyp=1 ) then $ x
matmod t001,xsil,,,,/tcd,/5 $
else if ( setyp=2 ) then $ y
matmod t002,xsil,,,,/tcd,/5 $
else if ( setyp=4 ) then $ z
matmod t003,xsil,,,,/tcd,/5 $
else if ( setyp=3 ) then $ xy
matmod t012,xsil,,,,/tcd,/5 $
else if ( setyp=5 ) then $ xz
matmod t013,xsil,,,,/tcd,/5 $
else if ( setyp=6 ) then $ yz
matmod t023,xsil,,,,/tcd,/5 $
else if ( setyp=7 ) then $ xyx
matmod t123,xsil,,,,/tcd,/5 $
else $ external or identical w/no reflections
 paraml xsil//'trailer'/2/s,n,xluset $
matgen ,/tcd/1/xluset $
endif $
dbstatus maps//s,n,nomaps $
if ( nomaps>0 ) then $
mpyad maps,tcd,/tcd1 $
equivx tcd1/tcd/-1 $
endif $ nomaps>0
else $
equivx eqexins/xeqexin/-1 $ for call sela0
PARAML USET0//'USET'//////'A'/S,N,NOASET $
matgen eqexins/intext/9/0/lusets $
upartn uset0,intext/tcd,,,/'g'/'a'/'o'/1 $ a(int) by g(ext)
endif $
dbstatus pav//s,n,nopav $
if ( nopav>0 ) call sela0 tcd,eqexins,sils,xeqexin/pav /
sedwn $
endif $
$
compile sela0 noref nolist
subdmap sela0 tcd,eqexins,sils,xeqexin/
pav/
sedwn $
type parm,,i,n,sedwn,temp $
mpyad tcd,pav,/paup/1 $
matmod paup,xeqexin,,,,/dmigpa,/16 $
delete /pav,paup,,, $
temp=sedwn $
dbview eqexind=eqexins where (peid=temp) $
dbview sild =sils where (peid=temp) $
paraml sild//'trailer'/2/s,n,lusetd $
mtrxin ,,dmigpa,eqexind,sild,/paup,,/lusetd $
call dbfetch /pav,,,,/s,sedwn/0/0/0/0 $
add5 paup,pav,,,/paupd1 $
equivx paupd1/pav/-1 $
call dbstore pav,,,,//s,sedwn/0/' '/0 $
delete /pav,,,, $
return $
end $ sela0
$
compile dbmgr noref nolist
alter 1
putsys(1,109) $
alter 'return '(,-1)
putsys(0,109) $
compile dbfetch noref nolist
alter 1
putsys(1,109) $
alter 'return '(,-1)
putsys(0,109) $
compile dbstore noref nolist
alter 1
putsys(1,109) $
alter 'return '(2,-1)
putsys(0,109) $
@
27629; 70.5; Buckling - EIGB,SINV - UWM 6278
A buckling analysis in Version 69.1 that specifies the SINV method will
 issue the warning message:
*** USER WARNING MESSAGE 6278 (MTMD23)
THE BULK DATA ENTRY SELECTED BY THE CASE CONTROL COMMAND METHOD = xx WAS
NOT FOUND
If there is no fatal message printed in the f06 file, this warning message
can be safely ignored.
Avoidance: There are two possible avoidances:
1. Ignore the extraneous warning message.
2. Use Lanczos method.
@
27647; 70; Post Processing - CBEAM - Orientation Vector Information not Wri
tten to XDB
When the GRID option is used to define the orientation of the CBEAM
element, the orientation information is not written to the XDB file.
Avoidance: Other than using the vector option, there is no avoidance.

@
27659; 70; Optimization/Design Sensitivity - BC, STATSUB - Wrong Answers
Eigenvalue sensitivity analyses (normal modes and buckling) require the
associated eigenvectors. In Version 69, the
appropriate set of normal modes eigenvectors is selected based on the BC
case control command. For buckling, the
STATSUB command is used. If BC or STATSUB does not result in a unique
eigenvector set, wrong answers can result.
Avoidance: Use BC and STATSUB requests in a way that ensures uniqueness.
@
27674; 69; Security - UFM 3060, SUBROUTINE MODEL - OPTION NUSR NOT IN APPRO
VED LIST
The message "USER FATAL MESSAGE 3060, SUBROUTINE MODEL, OPTION NUSR NOT IN
APPROVED LIST" will appear when running on Windows NT systems that allow
multiple users to run concurrently, e.g., Windows NT 4.0 with "remote
shell" support. The problem has been fixed in MSC/NASTRAN V69.0.4 on the
updated Windows NT CD released with the Digital Alpha NT system.
@
27677; 70; Documentation - SEMPLN,NORMAL - Unsupported Features
When defining the mirror plane for a partitioned, image superelement, only
the MTYPE=PLANE option on the SEMPLN entry is supported. The Version 69.1
Release Guide incorrectly states that the MTYPE=NORMAL option is also
supported.
Avoidance: Use the MTYPE=PLANE option on the SEMPLN entry.
@
27685; 70; Post Processing - CTRIAX6 - Missing Information in XDB
The "negative" key is not written to the STX6R Object to distinguish
between Von Mises and Octahedral Shear stress results for the CTRIAX6
element.
@
27704; 2001; Nonlinear Transient Heat Transfer - CP is Temperature Dependant
- Wrong Answers
If CP, specific heat is a function of termperature, in a nonlinear heat
transfer transient analysis then wrong answers will be produced.
Avoidance: There is no avoidance
@
27717; 70.5; Postprocessing - BAR, BEAM, ROD, CONROD - Wrong Margins of Safet
y in XDB
The Bar, Beam, Rod and Conrod elements all get wrong answers for the
margins of safety in the .XDB file.
Avoidance: There is no avoidance.
@
27722; 70; Documentation - QRG, PARAM,PRETEQXIN - UFM 5199
The MSC/NASTRAN Version 69 Quick Reference Guide incorrectly documents
PARAM,PRETEQXIN, the parameter should be PARAM,PRTEQXIN. If PRETEQXIN is
specified USER FATAL MESSAGE 5199 may result.
@
27734; 70; Documentation - QRG - ESE With Thermal Loads
Prior to Version 70 element strain energy output was incorrect in the
presence of thermal loads. This limitation was documented in the Quick
Reference Guide under Remark 4 of the ESE Case Control Command. This
limitation has been removed in Version 70 but the Version 70 QRG had
already been sent to press, so Remark 4 was not removed.
@
27739; 70; Nonlinear Heat Transfer - SDR3, XDB - UFM 4276
Nonlinear Heat Transfer jobs that request an XDB (i.e. param,post,0) may
fail with USER FATAL MESSAGE 4276 on HP 9000/700 systems. The error
appears to be dependent on problem size, machine and operating system.
The error has been corrected in Version 70.
Avoidance: Other than running on a different system there is no
avoidance.
@
27747; 70.5.1; Heat Transfer Analysis - CHBDYi, SSG1 - SFM 3007, Wrong Answers
If the model has nonunique element IDs among with heterogeneous type of
CHBDYP,CHBDYG, or CHBDYE , MSC/NASTRAN does not know which CHBDYx element
ID should be used in the convection or radiation. This confusion in the
code may lead to wrong answers or SYSTEM FATAL MESSAGE 3007 in the SSG1
module. The fix to the code allows the program to flag the non-unique
CHBDYP and CHBDYG in the GP2 module.
Avoidance: Change the Element IDs of CHBDYP and CHBDYG so that the IDs are
unique.
Duplicate CHBDYi element ids may result in the message:
@
27755; ; Eigenvalue Analysis - EIGRL, REIGL - SFM 5299
Small models with many repeated modes may require smaller values for
MAXSET else the following message may be encountered:
*** SYSTEM FATAL MESSAGE 5299.UNRECOVERABLE TERMINATION FROM LANCZOS
ITERATION
INTERNAL ERROR CODE = -3 USER ERROR RETURN = -33
Avoidance: Reduce the value of MAXSET on the EIGRL entry.

@
27757; 70; P-Elements - FEFACE,SURFID, EMA, GPSP - UWM 4398, BIOMSG ERROR
1305
If p-elements have FEFACE records that use SURFID to specify the geometry
of the edges of the face, bad answers or one of
the following messages may result.
Module EMA may give the message:
BIOMSG: ERROR 1305 HAS OCCURRED IN ROUTINE ENDPUT,
FILE INDEX (FILEX) = 0
Module GPSP may give the message:
USER WARNING MESSAGE 4398 (GPSP)
GRID POINT nnnnnnnnn FOR TRANSLATION DEGREES OF FREEDOM DOES NOT
PASS SYMMETRY CHECK. NO SINGULARITY TESTS MADE ON THESE TERMS.
Avoidance: This error occurs when several elements are connected to the
edge with conflicting geometry. The only known avoidance is to identify
the edges which are failing, and use consistent geometry for the FEFACEs
which are associated with the edge.
@
27761; 70; Submittal - Can't Modify the Input File Suffix
In Version 69, there is no simple method to specify a user or
site-dependent default input file suffix for the nastran command.
The V70 nastran command allows the keyword 'j.dat' in the '.ini' file to
define the default input file suffix. For example, 'j.dat=.bdf' sets the
default suffix to '.bdf'.
Avoidance: There is no simple avoidance. The nastran script could be
modified by the user to add a suffix whenever one is not provided. Since
the order of input to the nastran command does not matter, the script
would have to search for the input with no equal sign and then append the
desired suffix to that input.

@
27762; 2001; P-Elements - EMG - SFM 3008
A p-element run may terminate with a fatal message if a shell p-element
 and a solid p-element share a face and the p-order of the edges of the
face is higher than 2. There is no avoidance to this errror. The following
fatal message is issued
0*** SYSTEM FATAL MESSAGE 3008 MODULE = EMG
0INSUFFICIENT MEMORY AVAILABLE FOR SUBROUTINE PSHELD
0FATAL ERROR
@
27774; 70; Documentation - EXTRN - Incorrect Format
The format for the EXTRN entry as shown in the Version 69.1 Release Guide
is incorrect. The correct format is as follows:
EXTRN GID1 C1 GID2 C2 GID3 C3 GID4 C4
GID5 C5 -etc-
Avoidance: Correct the entry as shown above.
@
27775; 70; Documentation - SEBULK - Incorrect TYPE
When defining the superelement boundary search options for the external
superelement using the partitioned bulk data method, the TYPE field on the
SEBULK entry should be specified as EXTERNAL. The V69.1 Release Guide
incorrectly states that the TYPE field is either EXTRNLG or EXTRNLA. The
documentation will be corrected in the Version 70 Quick Reference Guide.
Avoidance: Use TYPE is equal to EXTERNAL.
@
27789; 69.0.4; Matrix Methods - PARALLEL, SYSTEM(107) - Wrong answers
If parallel processing is selected by setting SYSTEM(107) or the PARALLEL
keyword on MSC/NASTRAN Version 69.0.2 for Windows NT wrong answers may
result.
Avoidance: Other than not selecting parallel there is no avoidance. A
new version of MSC/NASTRAN Version 69 (V69.0.4) is in the process of being
released and the problem will be corrected in this version. Also, the new
executable (analysis.exe) is available on the MSC FTP site, in file
mscnv69.exe in the /pub/outgoing directory. This is a self-extracting ZIP
file that contains a single file, analysis.exe. To update an
installation, copy this mscnv69.exe file to the install_dir\msc69\i386
directory and run it by entering "mscnv69". Reply "y" to the request to
overwrite the existing executable (or rename the existing executable
before running mscnv69). This updated version also corrects some UFM 3060
problems that exist with multiple concurrent users (i.e., NUSR failures).
@
27804; ; Installation - Authorization File not Updated in .ini File - UFM
3060
If during an installation of MSC.Nastran on an NT system an existing
authorization file is selected and that file's name is different from the
default, the .ini file will not be updated with the location of the
authorization file. This will lead to USER FATAL MESSAGE 3060.
@
27811; 70.5; Documentation - PBARL, QRG - Missing Information
In the MSC/NASTRAN Quick Reference Guide it states that elements which use
the PBARL entry to define their properties have the origin of the element
coordinate system at the shear center of the cross-section. The manual
 does not state that the PBARL does not account for offsets between the
neutral axis and shear center. Therefore, the CHAN, CHAN1, and CHAN2
cross-sections may produce incorrect results.
Avoidance: If a cross-section does not have coincident shear center and
neutral axis, use a PBEAML.
@
27819; 2001; Miscellaneous - GPSTRESS, Von Mises, Plane Strain - Wrong Answer
s
In a plane strain analysis, the von Mises stress output using the grid
point stress option is incorrect. The component
stresses using the grid point stress option are, however, correct.
Avoidance: If feasible, use the von Mises stress output from the element
stress output.
@
27820; ; Nonlinear Analysis - PELAST - Possible Wrong Answers
When the nonlinear property PELAST is used for the scalar spring, CELAS1
or CELAS3, the linear stiffness, Ki, must be greater than zero. If Ki is
non-zero the actual value is replaced by the PELAST function. However, if
0.0 is entered then the value of 0.0 is used.
@
27821; 70.5; ESTIMATE - Estimate Utility does not search standard MSC/NASTRAN
RC files
The ESTIMATE utility does not examine the standard nast*rc files to
determine default settings for any of its keywords.
Avoidance: In order to set default values, you must set them in the
$MSC_BASE/$MSC_VERSD/$MSC_ARCH/estimate.ini, $HOME/.estimaterc, or the
./.estimaterc files.
ESTIMATE V70.5 will use the nastran command to search the standard
MSC/NASTRAN RC files. In order to do this properly, (i) the nastran
command executable must present in the same directory containing the
ESTIMATE executable, and (ii) ESTIMATE must be passed the same values for
<version> and <rcf> (if any) that will be passed to the nastran command.
Items set in the ESTIMATE RC files and on the ESTIMATE command line will
override the MSC/NASTRAN RC files.
@
27822; 70.5; Optimization/Design Sensitivity - MODTRAK - UFM 316
The Quick Reference Guide appears to allow the user to leave the HIGHRNG
field blank with the algorithm defaulting to
the highest extracted mode. However, if the user has set LOWRNG and
leaves HIGHRNG blank, the following fatal message is issued:
USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA ENTRY MODTRAK
Avoidance: Either remove the LOWRNG value or specify the HIGHRNG value.
@
27824; 70; Miscellaneous - SDR3 - SFE 4276 EC 7777, SIGNAL 10 ABORTING
On the HP that uses the PA2.0 processor, if the SDR3 module is being
exercised (whether it is through user-dmap or executed automatically by
the SOL sequence, e.g., SOL 111), the job may fail with the following
error messages:
 ***** RECEIVED SIGNAL 10, ABORTING ...
BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX)
= 0
0*** SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777 PID= 0
The module SDR3 is used to convert sort1 output to sort2 output.
Avoidance: If feasible, use sort1 output instead of sort2 output to avoid
execution of the SDR3 module.
@
27826; 70; Optimization/Design Sensitivity - Mode Tracking and Acoustics -
UFM 6677
If fluid elements are present in SOL 200 and mode tracking is requested
the run may fail with:
USER FATAL MESSAGE 6677 (DOM12E)
RUN TERMINATED DUE TO MODE TRACKING FAILURE. SEE USER FATAL MESSAGE 6677
(MTFTRD) PRINTED ABOVE FOR INFORMATION.
Avoidance: Insert the following DMAP alter:

COMPILE FEA
ALTER 'MODTRK','MODTRK'
MODTRK CASEM,EDOM,LAMAS,MGG,MAA,PHG,PHSA,
PHGREF0,PHAREF0/
MTRAK,NEWLAMA,NEWPHG,NEWPHA,PHGREF,PHAREF/
DESCYCLE/S,N,NOTRACK $
@
27828; 70; Heat Transfer Analysis - RADMT - UFM 4658
Large models with RADMT bulk data entries may fail in NLITER with the
message:
USER FATAL MESSAGE 4658 . DIFFICULT GEOMETRY PREVENTS FURTHER
COMPUTATIONS FOR ELEMENT WITH ID = ***

Avoidance: The only avoidance is to increase memory.

@
27833; ; Documentation - Getting Started with MSC/NASTRAN
The "Getting Started with MSC/NASTRAN User's Guide" states that "In any
coordinate system, all angular input is in degrees, but output ... is in
radians". This statement needs clarification. The "input" refers only to
geometric locations. If an angular enforced displacement is specified the
input is in radians.
@
27867; 70; Superelements - SEMA - SFM 3007, ILLEGAL INPUT TO SUBROUTINE MPG
TKG
Superelement jobs with SETs with large thru ranges may encounter the
following fatal message:
SYSTEM FATAL MESSAGE 3007 MODULE = SEMA
0ILLEGAL INPUT TO SUBROUTINE MPGTKG
 0FATAL ERROR
Avoidance: There is no practical avoidance.
@
27872; 70; Dynamics - PARAM,APHAi - Raleigh Damping Ignored
Raleigh damping is ignored when using param,alpha1 (proportional to mass)
and alpha2 (proportional to stiffness).
Avoidance: Add the following dmap avoidance.
compile gma $
alter 'call.*xadd5.*mhh.*khh.*bhh','' $
add5 mhh,khh,bhh,,/bhhx/alpha1/alpha2/cu $
alter 'call.*xadd5.*mdd.*kdd.*bdd','' $
add5 mdd,kdd,bdd,,/bddx/alpha1/alpha2/cu $
@
27879; 70; Post Processing - Optimization, Superelements - Missing Data fro
m XDB
In solution 200 with superelements, requested data may be missing from the
XDB for any design cycle after the initial analysis.
Avoidance: Insert the following DMAP alter:
compile update
alter 'do while'(,-1)
descycle=descycle-1 $
alter 'descycle=descycle-1',''
@
27880; 70; Superelements - Postprocessing, PARAM,POST,0 - SWM 6065, 6066
Superelement jobs that set PARAM,POST,0 to produce an .XDB file may
encounter the following messages when the new partitioned bulk data
superelements are used:
SYSTEM WARNING MESSAGE 6065 (DBC)
*** DBCMSG ERROR MESSAGE 11 FROM SUBROUTINE DBCEQG
PREMATURE EOR FOR MSC/NASTRAN LOGICAL FILE GEOM1S FOR SEID
0, DBSET 0.
*** SYSTEM WARNING MESSAGE 6066 (DBC)
*** DBCMSG ERROR MESSAGE 34 FROM SUBROUTINE DBC
THE DBC MODULE HAS BEEN TERMINATED BECAUSE OF THE ABOVE FATAL DBC
ERROR.
USER INFORMATION: SUBSEQUENT DBC CALLS WILL NOT BE EXECUTED, AND THE
DATABASE CONVERSION IS INCOMPLETE.
@
27881; 70; Miscellaneous - IFPDRV - UFM 316
Large input files (<1 million entries) can cause overflow in the sorted
entry count of an error message. For instance:
*** USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA ENTRY MPC 1 SORTED ENTRY COUNT = ******
^^^ USER FATAL MESSAGE 9002 (IFPS)
^^^ ERROR(S) ENCOUNTERED IN THE MAIN BULK DATA SECTION
 ^^^ SEE MESSAGES ABOVE. ERROR ENCOUNTERED IN MODULE IFP,

@
27917; 70.5; Superelements - External Superelement, SEP1X - SFM 3001
In Version 69.1, when using the parameter EXTOUT=dmigop2 option for
external superelements, and more than one external superelement is brought
in, then the job will fail with the following error messages:
*** SYSTEM FATAL MESSAGE 3001 (SEP1X )
THE EXECUTIVE FILE 0 DOES NOT EXIST.
PROGRAMMER INFORMATION: SUBROUTINE SEPMAB
0FATAL ERROR
Avoidance: If more than one external superelement is to be brought in,
use the dmigpch option.
Or insert the following DMAP avoidance:
compile phase0
alter 'dbview geom1f','dbview geom1f'(,3)
dbview geom1f=geom1qs where ( peid<>0 ) $
dbview geom2f=geom2s where ( peid<>0 ) $
dbview geom4f=geom4qs where ( peid<>0 ) $
dbview sgpdtf=sgpdts where ( peid<>0 ) $
@
27918; 70; Miscellaneous - DEQATN - SFM 6048
A DEQATN entry is validated for syntax by evaluating the equation with all
variable set to zero. If this evaluation results in an invalid
operation, like the square root of a negative number, then the following
message may be unnecessarily issued:
SYSTEM FATAL MESSAGE 6048 (XQFNC7)
Avoidance: The error should be able to be avoided by adding appropriate
absolute value operators to the equation.
@
27927; 70; Documentation - QRG, PARAM,ACOUT - Poor Description
The description of the ACOUT parameter in the Version 69 Quick Reference
Guide states;
To obtain sound pressure level in units of dB and dBA given by the FORCE,
DISPLACEMENT, and PRESSURE commands, a peak reference pressure must
be specified with PARAM, PREFDB.
The sentence should read:
To obtain sound pressure level in units of dB and dBA given by the FORCE
command, a peak reference pressure must be specified with PARAM, PREFDB.
@
27930; ; Acoustic Analysis - Modelling Error - SFM 6169, Poor Message
When a plate is immersed in a fluid, then the structural grid points will
be attached to up to eight fluid HEXA elements (more if PENTAs and/or
TETRAs are use) and two fluid/structure interfaces, one for each side of
 the plate structure. The fluid grids on each side of the plate are on
different interfaces and thus must have unique coincident grids. Thus, at
a point in space there should be two fluid grids and one structural grid.
If only one fluid grid is erroneously used, that is, the fluid elements on
both sides of the interface use the same grid, then the following error
will occur:
*** USER FATAL MESSAGE 6169, FOUND TWO OR MORE STRUCTURAL GRID POINTS WITH
IDENTICAL COORDINATES
WHICH LIE ON THE FLUID/STRUCTURE BOUNDARY.
USER ACTION: CHECK GRIDS AND/OR SET1 CARDS WHICH BELONG TO PANEL OR
ACMODL CARDS.
GRID ID = nnnn
GRID ID = xxx

@
27936; 70; Documentation - External Superelement - Missing FMS statements
Item 1 of Step 3 of the data recovery procedure as described on page 8 of
the MSC/NASTRAN Version 69.1 Release Guide is incomplete.

Avoidance: Add the following FMS statements.

Assign sexx='step1.MASTER'
Restart logical=sexx
where step1.MASTER is the database from the first step.
@
27937; 70; Miscellaneous - NASTRAN - NASTRAN Command Ignored.
The NASTRAN command must be the first entry in an input file, disregarding
comment lines. If the NASTRAN command is mixed in with FMS commands such
as ASSIGN statements, the command may be ignored.

@
27938; 70; Cyclic Symmetry - EMG - UFM 2016, Various UFM
APP HEAT will be ignored in cyclic symmetry jobs. The job may fail in
various ways, one likely way is:

USER FATAL MESSAGE 2016 MODULE = EMG

NO MATERIAL PROPERTIES EXIST.
FATAL ERROR
Avoidance: Apply the following dmap avoidance.
COMPILE SUPER1 NOLIST NOREF
ALTER 'IF (GETSYS(1,56)=1)'(,-1) $
PUTSYS(1,56) $ HEAT TRANSFER
@
27944; 70; Rigid Elements - RSSCON - UFM 6692
In nonlinear statics (Sol 106), the following user fatal messages are
issued when a bulk data entry RSSCON with option
 "ELEM" is in the residual structure.
*** USER FATAL MESSAGE 6692 (GP4RS1)
BULK DATA ENTRY RSSCON CAN NOT FIND ELEMENT IDS, USER ACTION: CHECK
SUPERELEMENT STRUCTURE.
*** SYSTEM FATAL MESSAGE 3001 (GP4 )
THE INPUT DATA BLOCK IN POSITION 11 DOES NOT EXIST.
Avoidance: Use the following DMAP alter which provides the element
connection GEOM2S to
the GP4 module.
compile nlstatic souin=mscsou, nolist, noref
alter 'GP4 CASESX2,GEOM4S','GP4 CASESX2,GEOM4S' $
GP4 CASESX2,GEOM4S,EQEXINS,SILS,GPDTS,BGPDTH,CSTMS,,,,GEOM2S,/
RGNL,YS1H,USET1,/
LUSETS/S,N,MPCF1/S,N,MPCF2/S,N,NOSSETX/S,N,NOOSETX/
S,N,NORSET/0/0/S,N,NOSETX/S,N,NOLSET/S,N,NOMSO/0 $
endalter $

@
27964; ; Old Design Sensitivity - DSVG1, PARAM,EIGD - SFM 4276 EC 1409 IN
ROUTINE GETSTR
When param,eigd,yes is used in solution 103 to calculate the eigenvector
derivatives, SYSTEM FATAL MESSAGE 4276 ERROR CODE 1409 is issued from
module DSVG1 in subdmap DSAMODES.
Avoidance: Use the alter evds103a.v691 provided in the SSSALTER library.
@
27968; ; Submittal - PSL: NUSR: unable to initialize (0 expired keys)
The following error is issued in the log file when an abnormality occurs
in the NUSR queues in the system. (i.e. if a nastran
job gets killed some of the NUSR queues will not be removed from the
system and following message will be issued when a subsequent job is
started.
PSL: NUSR: unable to initialize (0 expired keys).

Avoidance: Flush the queues by rebooting the machine.

or,
Remove nastran related queues by the command
ipcrm -q PROCESSID of the queue.
@
27996; 70; Superelements - Plotting, PLTSET - SFM 3062, Subroutine CNSTRC
Use of a plot set defining a list of elements and grids to be used by the
structure plotter cause a SFM 3062 whenever the exclude option is used.
For example:
 SET 1001 =110106 THRU 110452 110553 220101 THRU 220120 120274 THRU,
120292 120134 THRU 120178 230274 THRU 230286 220208 220218,
.
.
5033 5080 241 5034 5078 240 209 218 520 545 208 219,
exclude grid points 1 thru 999999
results in the SFM 3062 from the PLTSET module.
Avoidance: Commenting out the exclude set from the set definition allows
the run to proceed but the plots are not consistent with the request of
the set. That is, more information is plotted than requested.
@
28008; 2001; DMAP - MATMOD Option 2 - SFM 4276, Traceback
This option of matmod filters small numbers from the input matrix. When
given a complex matrix, a floating point overflow occurs at 1.E19 on
rs/6000 computers, and other relatively small numbers on other computer
types. Since the range of this computer type is greater than 1.E30, this
is not a reasonable limitation.
The diagnostics from such a crash are machine dependent. On the rs/6000,
the following appears in the .log file:
SIGTRAP - Trace trap FP overflow
Traceback:
Offset 0x00000028 in procedure pack_
Offset 0x0000042c in procedure mtmd02_
Offset 0x000028e4 in procedure matmod_
Avoidance: Use the NORM module to first find the largest number in the
matrix, scale it with the inverse of this number before using matmod(2),
then re-scale the results. The NORM module can handle numbers larger than
matmod(2) can.
@
28009; 70; Miscellaneous - TABLED1, TABLEM1 - SFM 4276, EC 7777
Extrapolation of data from either TABLED1 or TABLEM1 entries beyond the
end points of the table may result in numerical overlfow. Version 70 has
been enhanced to trap machine overflow due to extrapolation and will issue
USER FATAL MESSAGE 6917 to address this error.
Avoidance: Add additional data points at each end of the curve to make the
curve flat so that extrapolation to any x value can not be unbounded.
@
28011; 70.5; Documentation - CHEXA, CPENTA, CTETRA, Plot Codes - Missing Defi
nitions
The stress item codes 8, 16 and 22 in the Quick Reference Guide are
incorrectly labeled for the CHEXA (page A-7), CPENTA (page A-9 and A-10),
and CTETRA (page A-17). Modern textbooks arrange the principal stresses
in this order:
First Principal> Second Principal> Third Principal. MSC/NASTRAN arranges
the principals as: First Principal> Third Principal> Second Principal.
MSC/NASTRAN arranges the principals this way because the principal
stresses are the roots of a cubic polynomial and the this is the order
that the method that extracts the roots puts out. The stress item codes
8, 16, and 22 for the solid elements should have the labels changed to the
 following,
8 - First principal
16 - Third principal
22 - Second principal
A note should also be added to the element descriptions that indicate that
the principal stresses are output such that A>B>C.

@
28027; 70; Elements - TETRA, TEMP(LOAD), MAT9 - Wrong Answers
In MSC/NASTRAN Version 69 on the IBM RS6000 power2 computer, the 10 noded
TETRA element in conjunction with MAT9 and thermal loads produces wrong
answers whenever there is also HEXA elements in the model.
Avoidance: The non-power2 executable can be selected with the exe
submittal line command but this executable will run significantly slower.
@
28047; 70.5; Material Nonlinear Analysis - TABLES1 - Wrong Answers
If stress-strain data is entered on the TABLES1 entry in descending order
wrong answers may results.
Avoidance: Enter data in ascending order.

@
28052; 70; Dynamics - OFREQ and OTIME, SOLs 108 and 109 - Wrong answers
The OFREQ and OTIME case control commands will produce wrong answers in
SOLs 108 and 109 in MSC/NASTRAN Version 69.1.
Avoidance: Insert following DMAP alter to avoid this error:
compile disprs
alter 'upf1'
equivx upf1/upf/-1 $
@
28063; 70.5; Acoustic Analysis - SDR2,OUG1F - Poor Performance
Acoustic analysis jobs that request data recovery sets with the thru
option may experience very poor performance in the SDR2 module.
Avoidance: There is no avoidance.
@
28070; 70; Postprocessing - PARAM,POST,0 - XDB File Left on Remote Node
When using the <node> keyword, the XDB file is left on the remote node.
This file should have been deleted after it was copied from the remote
node to the local node.
Avoidance: Add the following line:
%trans=[ya]:% %s.rsh% %node% -l %username% /bin/rm -f %r.out%.xdb
just before the following line in the <install-dir>/bin/nast*.rmt file:
 %trans=[ya]:%}
@
28073; 70; Optimization/Design Sensitivity - QUAD4, CTRIA3, SNORM - SFM 306
2
In Version 69.1 of MSC/NASTRAN, optimization and design sensitivity runs
with shell normals may fail with SYSTEM FATAL MESSAGE 3062.

@
28087; 70; Miscellaneous - ESTIMATE - SMEM and BPOOL Parameter Not Passed t
o Utility
If the ESTIMATE utility is being used by the nastran command to estimate
the memory requirements of a job, the nastran command fails to pass in
either "smemory" or "bpool". If set to non-default values, they could
have a significant impact on ESTIMATE's accuracy.
Avoidance: Either use the default "smemory" and "bpool", or set the values
in the following files:
UNIX: $HOME/.estimaterc or <jid-dir>/.estimaterc
NT: %HOME%\estimate.ini or <jid-dir>\estimate.ini
VMS: SYS$LOGIN:estimate.ini or <jid-dir>:estimate.ini
@
28088; 70; Miscellaneous - ESTIMATE, SYMBOL - SYMBOLS Defined in RC File No
t Used
The nastran command will not use SYMBOLs defined in the RC files on the
command line.
For example, if the following line exists in one of the system RC files or
the $HOME RC file,
symbol=MYDIR=/my/main/project
The following will fail
msc69 nastran MYDIR:myjob ...
Avoidance: Set the SYMBOL values that will be used in the command line as
environmental variables, e.g.,
export MYDIR=/my/big/project # Korn shell
MYDIR=/my/big/project; export MYDIR # Bourne shell
setenv MYDIR /my/big/project # C-shell
and use the environment variable on the command line.
msc69 nastran $MYDIR/myjob ...
@
28102; 70; Submittal - System RC Files on Remote Node Cannot Contain 'nastr
an' Statements
The system RC files on a remote system cannot contain a 'nastran'
statement. For example
$ msc691 nastran myjob node=remotenode ...
 The system RC files on remotenode cannot contain a 'nastran' statement.
Note: This restriction *only* occurs if the "node" keyword is used.
Avoidance: Remove the 'nastran' keyword and set system cells by number.
For example, change
nastran system(196)=1, massbuf=10
to
system(196)=1
system(199)=10
@
28104; ; DMAP - TABPT of CSTM and BGPDT - Wrong Answers on Double Precisi
on Machines
If the TABPT module is used to print the CSTM or BGPDT tables on double
precision, short-word machines, then incorrect data for the tables is
printed.

Avoidance: Use TABPRT (Formatted print of selected table data blocks) to

print the CSTM and BGPDT table.
TABPRT CSTM/////1$
TABPRT CSTM//'CSTM'$
TABPRT BGPDT/////1$
TABPRT BGPDT//'BGPDT'$
$TABPT CSTM,EQEXIN,BGPDT/ $
@
28123; 70; Installation - Invalid Media for V69.1 (R3) CD for HP
The MSC/NASTRAN V69.1 (r3) CD has an error when attempting to install on
an HP system with a local CD. The installation script, mscsetup, will
incorrectly report the following error:
MSC Installation and Setup - Invalid Media
The specified CD pathname
/CDROM
does not contain any of the installation filesets.

Please ensure that the pathname is correct, and the correct

installation
media is mounted.

Press ENTER or RETURN to continue...

Avoidance: There are two avoidances:
1) Copy the installation script to a local directory
cp /CDROM/MSCSETUP.\;1 mscsetup
Change line 3475 from
 iv_file="$iv_dev/${iv_file};1"
to
iv_file="${iv_file};1"
Install the CD using the modified script
mscsetup
2) Use the installation script in TAN #2828.

@
28124; 70; Nonlinear Analysis - PELAST - UFM 4551, No Convergence
Nonlinear analysis with CELAS elements that reference PELAST may encounter
difficult converging. The problem is caused by an error in the element's
coordinate transformation.
Avoidance: Replace the nonlinear spring by a ROD element with nonlinear
elastic material.
@
28128; 70.5; Documentation - CMASS1, CMASS2 - Incorrect Description in QRG
The description of the CMASSi bulk data entries in the Version 69 and 70
Quick Reference Guide indicates that multiple degrees of freedom, "up to
six unique integers", may be specified for the component degree of
freedom. This is incorrect, only a single degree of freedom may be
specified.

@
28158; 70; Submittal - Remote Execution will fail if MSC_BASE is not Proper
ly Defined.
The remote execution capability of the nastran command will fail if the
MSC_BASE on the remote machine is not properly defined. The command will
fail even if the nastran command on the remote machine was able to
determine the proper MSC_BASE value.
Avoidance: Make sure the MSC_BASE value is properly defined in the
<install_dir>/bin/nast70 script.
@
28169; 70; Miscellaneous - Cray T90 Hardware Error May Cause Wrong Answers
Cray Research has reported a T90 hardware error that has the potential to
return random results for certain vector operations. The specific
MSC/NASTRAN routines that may be affected by this error have been
identified.
It is very important to note that:
1) In our testing, and in specific testing of MSC/NASTRAN by Cray
Research, the effects of this error have NOT been seen.
2) This error is ONLY present on T90 systems.
 3) No client has ever reported a problem related to this error.
The specific MSC/NASTRAN executables that are affected by this error are:
MSC/NASTRAN Version Executables
--------------------- ---------------------------------------
V68.0 unicos/analysis.C90
V68.1 unicos/analysis.C90
V68.2 unicos/analysis.c90
V68.2.3R2 unicosc90/analysis, unicosts/analysis
V69.0 unicosts/analysis
V69.1 unicosts/analysis
V70 BETA unicosts/analysis
In summary, this affects all C90 and IEEE T90 executables for V68.X, and
only the IEEE T90 executables for V69.X and V70.

We have recompiled the IEEE T90 V70 executable with a new compiler that
prevents this error from occurring.

Avoidance: There are two avoidances:

1) Do not use one of the above executables.
2) Disable bi-directional memory using the cpu(8) utility. This may
result in up to a 15% performance penalty.
See the man page for details as the command may vary at each
installation.
@
28184; 70.5; Documentation - PARAM,COUPMASS - Incorrect Description
The definition of PARAM,COUPMASS in the Version 69 and 70 MSC/NASTRAN
Quick Reference Guide is incorrect. It says:
COUPMASS Default=-1
COUPMASS>=0 requests the generation of coupled rather than
lumped mass matrices for elements with coupled mass capability, ...

It should say:
COUPMASS Default=-1
COUPMASS>0 requests the generation of coupled rather than
lumped mass matrices for elements with coupled mass capability, ...
Note that the action of the code has never changed. To obtain coupled
mass a setting of 1 or greater has always been required.
@
28186; 70.5; Submittal - Job will fail if <sdir> is set to non-local filesyst
em.
Jobs on an SGI R8K, R10K system will fail if the <sdirectory> keyword or
TMPDIR environment variable is set to a non-local, i.e., NFS-mounted,
filesystem.
 Avoidance: There are two avoidances:
1) Do not set <sdirectory> or $TMPDIR to an NFS-mounted filesystem.
2) In <install-dir>/bin/nast*.lcl, change the line
"TMPDIR=%sdir%; export TMPDIR"
to
"TMPDIR=%TMPDIR%; export TMPDIR"
@
28194; 70; Documentation - QRG, CQUADR, CTRIAR, PCOMP - Missing Output
The Quick Reference Guide indicates that a PCOMP entry may be referenced
by the CQUADR and CTRIAR element. It should be noted that composites are
not supported for these elements and it is not possible to obtain lamina
stresses and strains.

@
28199; ; External Superelement - SEBULK, EXTRN - Possible Wrong Answers
When using the enhanced external superelement capability in Version 69.1,
if more than one external superelement is dblocated in a single assembly
run using the matrixdb option, then the answers may be wrong. Also see
error 27917.
Avoidance: Attach one external superelement at a time in the assembly
run. Attach additional external superelements by making successive
restarts--one additional restart for each additional external
superelement.
@
28205; 69; Post Processing - Special Version of HP-UX dbiolib.a for Catia
Catia Versions 4.1.6 and 4.1.7 require a shared dbiolib.a to link with
MSC/NASTRAN Version 68.0 and Version 68.1 on HP 9000 systems. The
standard dbiolib.a delivered with V68.0/V68.1 was not built this way (all
subsequent releases were). This standard library is available from the
anonymous ftp site under the URL
ftp://ftp.macsch.com/pub/msc-products/NASTRAN/V68/v68-hp9000-dbiolib.a+z
This file is NOT required for MSC/NASTRAN V69 and later versions.
@
28210; 70.5; Documentation - RFALTER and RFINCLUDE - Commands No Longer Exist
Starting with Version 69, the RFALTER executive control statement and the
RFINCLUDE FMS statement no longer work. This is due to the removal of
RFALTERs from the program. The Quick Reference Guide incorrectly still
lists these as valid commands.
Avoidance: Use the INCLUDE command.
@
28218; 70; External Superelement - SOL 101, NO SPC AND SCR=NO - UFM 9050
If the reduced external se is unconstrained and scr=no is used, then the
job will fail with UFM 9050 (run failed due to
excessive pivot ratio). The subsequent residual restart run will fail
with the following messages:
 SYSTEM FATAL MESSAGE 6551 (INPTX2)
AN END-OF-FILE WAS DETECTED ON UNIT 30 DURING BINARY INPUTT2
PROCESSING
PROGRAMMER INFORMATION:
1. A KEY VALUE WAS BEING READ
2. THE LAST KEY VALUE READ: 0
FATAL ERROR.
0 NO. FORTRAN RECORDS READ = 0
NO. FORTRAN WORDS READ = 0
0FATAL ERROR
Avoidance: There is no known avoidance.
@
28219; 70.5; Documentation - QUADR, PCOMP - UWM 4385
The Quick Reference Guide indicates that the QUADR element may reference a
PSHELL or PCOMP property entry. Although referencing a PCOMP entry may
not cause a fatal error, technically the QUADR element does not support
composites and referencing a PCOMP will lead to the following message:
USER WARNING MESSAGE 4385, STRESS PROCESSING FOR NON-COMP ELEMENTS IN
SUBCASE XX
NOT COMPLETE BECAUSE OF INSUFFICIENT DATA IN DATA BLOCK OES1A
@
28226; GL; Elements - QUAD4 - Incorrect Corner Stresses and Forces with Var
iable Thickness
The corner stresses and forces are incorrect for the QUAD4 element if
variable thickness is specified.
Avoidance: Do not request corner stresses, request stresses at the
centroid (the default).
@
28227; 70.0.1; Optimization/Design Sensitivity - DSAE - UFM 3003, Wrong Answers
Beginning in Version 69 a shape optimization deck with superelements whose
grids are designed may either fail with the message "USER FATAL MESSAGE
3003 MODULE = DSAE, READ PAST DATABLOCK TABEVS" or result in wrong
answers.
Avoidance: Either do no use superelements or use Version 68.
@
28234; 70.5; Acoustic Analysis - SOL 111 with CAABSF elements - Inefficient C
PU
Use of frequency dependent material property elements such as the CAABSF
element may make frequency response analyses take an order of magnitude
more cpu time.
Avoidance: Insert the following DMAP alter, in Version 69 or V69.1
solutions 108 or 111, to improve performance.
compile sedfreq
alter 'call freqrs'(,-1)
call dbstore vgfs,,,,//0/0/' '/0 $
compile semfreq
alter 'call freqrs'(,-1)
call dbstore vgfs,,,,//0/0/' '/0 $
 compile freqrs
alter 'do while'(,-1)
call dbfetch /vgfs,,,,/0/0/0/0/0 $
trnsp vgfs/vgfst $
matmod vgfst,,,,,/vgsf,/12//1 $
alter 'ckddf1',''
call fd3mpy kggf1,vphfs,vgsf,vgfs,rgdfd/ckddf1 $
alter 'cbddf1',''
call fd3mpy bggf,vphfs,vgsf,vgfs,rgdfd/cbddf1 $
alter 'k4ddf',''
call fd3mpy k4ggf,vphfs,vgsf,vgfs,rgdfd/k4ddf $
compile fd3mpy
subdmap fd3mpy kgg,vphfs,vgsf,vgfs,rgx/kdd $
matmod kgg,,,,,/nzk,/12//1 $
add5 vgsf,nzk,,,/vgnzf $
partn kgg,vgnzf,/kfgg,,, $
add5 vgfs,nzk,,,/vgnzs $
partn kgg,vgnzs,/ksgg,,, $
paraml ksgg//'null'////s,n,noksgg $
if ( noksgg>-1 ) then $
partn rgx,VPHFS,Vgnzs/rsgx,,,/1 $
smpyad rsgx,ksgg,rsgx,,,/ksdd/3////1////6 $
endif $
paraml kfgg//'null'////s,n,nokfgg $
if ( nokfgg>-1 ) then $
partn rgx,VPHFS,VgnzF/,,rfgx,/1 $
smpyad rfgx,kfgg,rfgx,,,/kfdd/3////1////6 $
endif $
MERGE ksdd,,,kfdd,VPHFS,/kdd///6 $
return $
end $ fd3mpy

@
28275; 70; Documentation - Superelements - UFM 6794
The example jet101a.dat in the MSC/NASTRAN Version 69 Release Notes
section describing superelement enhancements contains an error that will
lead to:
UFM 6794 (SEPSEB)
The error is caused by duplicate boundary points connecting to
superelement 4 at grid 26.
The input requires the following changes to work:
$GRID,4001,,-1.975,-3.4125,0.
$GRID,4002,,-1.925,-3.9875,0.
GRID,4003,,-1.925,-3.9875,1.
$SELOC,5,1,3,6,4001,4002,4003
SELOC,5,1,3,6,4101,4102,4003
@
28281; 70.5; Documentation - QRG, PARAM,PRTRESLT PARAM,PRTMAXIM - Missing Out
put
The MSC/NASTRAN Quick Reference Guide indicates that PARAM,PRTRESLT and
PARAM,PRTMAXIM are supported for solution 106. They are not supported and
their specification will yield no output. These parameters on supported
 in solution 106 starting in Version 70.5.
@
28304; ; Documentation/Online Encyclopedia - Missing Footnotes
The element summary table (Table 1. Element Summary - Small Strain
Elements) in the Encyclopedia Reference Manual chapter 5.1 pg 158-164 is
missing important footnotes. The symbols appear in the table but the
corresponding footnotes are missing. The critical footnote that is
missing is that that midside nodes are not allowed for elements in
geometric nonlinear analysis.
Avoidance: See the printed version of the Reference manual for the proper
footnotes.
@
28338; 70.5; Optimization/Design Sensitivity - CBUSH - UFM 2042
When a design optimization analysis has a DRESP1 entry with RTYPE=WEIGHT
and CBUSH elements, it may result in the following error message:
*** USER FATAL MESSAGE 2042 MODULE = WEIGHT
MISSING MATERIAL TABLE ******** FOR ELEMENT xxx.
Avoidance: Specify VOLUME instead of WEIGHT as the RTYPE.
@
28348; 70.5; Postprocessing/Buckling Analysis - PARAM,POST - Missing Output
In buckling analysis, solution 5, the static results are not written to
the xdb file. The buckling results
are written properly.
Avoidance: Insert the follwing DMAP alter:
compile sol5
alter 'EQUIVX CASECC\/CASEXX',''
if ( post=0 ) soltype='statics' $
CASE CASECC,/CASEXx/'TRAN'/1 $
alter 'DBC OSTR1'
soltype='bkl0' $

@
28373; 70.5; Installation - mscsetup hangs when starting FLEXlm
As documented in the INSTALL.html and INSTALL.txt files, and in the
mscsetup output, installing FLEXlm requires root privileges.
Specifically, root is required to modify the /etc/inittab file.
If non-root attempts to install FLEXlm AND modify /etc/inittab AND start
lmgrd before the next system boot, mscsetup may appear to hang after it
writes the following message to the screen:
Starting /etc/inittab command now...
The system is actually waiting for the password of the account that owns
the lmgrd process.
Avoidances: There are two avoidances-
 1) Install FLEXlm as root.
2) Enter the password for the user that owns the lmgrd process if mscsetup
appears to hang as described above.
@
28383; 70.5; Limitation on Problem Size - Version 70 on HP-UX Doesn't Support
Large Files
MSC/NASTRAN V70 on HP-UX does not support large files. Both the PA-RISC
1.1 and PA-RISC 2.0 executables should have supported large files, i.e.,
files larger that 2GB. This support is missing.
Avoidance:
Use FMS commands to partition the database into multiple files, each of
which is smaller than 2GB.
@
28395; 69; Nonlinear Normal Modes - PARAM,NMLOOP - SFM 3002
When performing a nonlinear normal modes analysis (param,nmloop,x), if the
load increment is anything other than one, then
the job will fail on the NT system in Version 68.2.5 with the following
error messages:
*** SYSTEM FATAL MESSAGE 3002 MODULE = CASE
EOF ENCOUNTERED WHILE READING DATA BLOCK CASEI (FILE 201) IN SUBROUTINE
CASE
FATAL ERROR
Avoidance: If the number of load increments is more than one, break the
subcase into multiple subcases--one load increment for each subcase.
@
28421; 70.0.1; Nonlinear Analysis - CBUSH, COR1S, CORD2S, CORD1C, CORD2C - Wron
g Answers
If the CBUSH entry references a non-cartesian coordinate system, wrong
answers result.
Avoidance: Use a cartesian system.

@
28483; 70.5; P-Elements - GMSURF, GMCURV - UFM 4659
If an element edge being mapped to a GMSURF or GMCURV has tangents which
are anti-parallel at the endpoints, the mapping is incorrect, possibly
causing the following fatal message in EMG.
*** USER FATAL MESSAGE 4659 (EMGPSD)
DURING INTEGRATION FOR ELEMENT WITH ID = 1
BAD GEOMETRY HAS BEEN DETECTED.
An example of an element that will cause this error is an element edge
spanning a semicircle.
Avoidance: Move one of the endpoints or split the element into multiple
elements, such that the tangents are not anti-parallel.
@
28508; 70; SSSALTERs - alter1ga.v691 - UFM 5423 (DUMERG)
 Use of alter1ga.v691 from the sssalter library may result in USER FATAL
MESSAGE 5423 point to the trailers for matrices GOAQ and AIDENT matrices.
Avoidance: If this error occurs, call your local MSC office for an updated
version of the alter. If you wish to make the correction yourself, change
the line:
UMERGE USET,QIDENT,/AIDENT/'A'/'Q'/'T' $
to:
umerge1 uset,qident,,,/aident/'A'/'Q'/'T' $
@
28509; 70; Parallel Processing - SYSTEM(107) - SFM 427, EC 7777
It has been reported on a Cray Version 69 system that when more than one
processors is requested, then the job may fail with the following error
messages:
0*** SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777 PID= 0
Avoidance: We have been unable to duplicate this error at MSC. To avoid
the error simply do not select parallel processing.

@
28525; 70; Documentation - V70 COG: HP Remote CD Install, V70 No such file
The "MSC/NASTRAN V70 Configuration and Operations Guide UNIX Edition",
Section 2.2.2, for remote CD installation of V70 on HP systems specifies
file_set = "/CDROM/MSCSETUP.TAR;1"
It should specify
file_set = "'/CDROM/MSCSETUP.TAR;1'"
@
28533; ; Acoustic Analysis - VELO(PHASE) - Wrong Answers
If the user requests VELO(PHASE) output and the resultant velocity is near
zero, then NASTRAN will not convert the result to the phase format and
will print the real/imaginary result in an F rather than an E format. An
example of the case control request and the incorrect output is:

VELO(PHASE,SORT2) = 100
X-VELOCITY
0.0
-0.0159
Avoidance: Correct E format results may be obtained by selecting
VELO(REAL,SORT2), the phase option is not applicable.
X-VELOCITY
0.0
1.591795E-02
@
28554; ; Matrix Methods - Iterative Solver, ITER - SFM 6723
The iterative solver may issue the message "SYSTEM FATAL MESSAGE 6723" NON
POSITIVE DIAGONAL ELEMENT IN SCALAR ROW INCREASE IPART VALUE ON BULK DATA
CARD".
Avoidance: IPAD may be increased by 1, the current value is given in
SYSTEM INFORMATION MESSAGE 4157 in the f04. However, SFM 6723 is often an
indication of an ill-conditioned model. Thus, the model should be
reviewed by plotting and/or by using the checka alter in the sssalter
library.
@
28557; 70.0.1; Frequency Response - MPYAD - SFM 4276, EC 1305, EC 32
In Version 70, if a large number of frequencies is requested in a
frequency response run, then the job may fail with one of
the following error messages:
BIOMSG: ERROR 1305 HAS OCCURRED IN ROUTINE ENDPUT , FILE INDEX (FILEX)
= 0
STATUS = 0
0*** SYSTEM FATAL ERROR 4276.ENDPUT ERROR CODE 1305 PID= 0
BIOMSG: ERROR 32 HAS OCCURRED IN ROUTINE GETSTR , FILE INDEX (FILEX)
= 0
0*** SYSTEM FATAL ERROR 4276.GETSTR ERROR CODE 32 PID= 0
Avoidance: Reduce the number of frequencies. In modal frequency
response, the new freq3, freq4 and freq5 entries may be used to guarantee
that a solution is requested at each of the modes without specify a large
number of frequencies.
@
28558; 70.5; Installation - rld: Fatal Error: cannot successfully map soname
'c_code.so'
The "Minimum" installation of MSC/NASTRAN Version 70 on the SGI
R8000/R10000 will fail to load a file required by the MSC/NASTRAN solver
executable, as a result, MSC/NASTRAN will fail with the following error
message:
$ msc70 nastran myjob
MSC/NASTRAN V70.0 (Silicon Graphics IRIX64 6.1) Tue Jul 29 08:52:34 1997
MSC/NASTRAN beginning job myjob
21692:/scratch1/dnl/msc/msc70/irix64/analysis: rld: Fatal Error: cannot
successfully map soname 'c_code.so' under any of the filenames
/scratch1/dnl/msc/msc70/irix64/c_code.so:/usr/lib64/c_code.so:/lib64/c_co
e.so:/usr/lib/64bit/c_code.so:/lib/64bit/c_code.so:
$
This file is loaded with the "standard", "full", and "custom"
installations.
Avoidances: There are two avoidances:
1) Use the "standard" or "full" installation.
2) If a minimum installation is required, or has already been performed,
the following commands will unload the
 missing file directly from the CD-ROM:
$ cd _install_dir_
$ tar xvf /CDROM/msc70/irix64.tar msc70/irix64/c_code.so.gz
x msc70/irix64/c_code.so.gz, 522805 bytes, 1022 blocks
$ msc70/irix64/gunzip -v msc70/irix64/c_code.so.gz
msc70/irix64/c_code.so.gz: 61.4% -- replaced with
msc70/irix64/c_code.so
$
@
28565; 70.0.1; Design Optimization/Aeroelastic Analysis - FLFACT - UFM 7087
The message "USER FATAL MESSAGE 7087" is issued when a velocity requested
in a flutter design task does not correspond to one used in the analysis.
This could be a user input error, but it has been found that it can also
be encountered when the user uses the thru option on the flfact entry. An
equality check is being used to make the comparison and the numerical
values that appear to be the same can differ slightly when they are
generated from the thru data.
Avoidance: Do not use the thru option on the flfact entry to specify
analysis and/or design flutter velocities in design optimization (SOL
200).
@
28566; 70.0.1; Optimization/Design Sensitivity - DSDJ2D, DSDJ2S - SFM 3007
The message "SYSTEM FATAL MESSAGE 3007" can appear when performing adjoint
sensitivity analysis if the adjoint solution vector is null. The adjoint
solution vector will only be null if the resultant sensitivity is null.
Thus, the message indicates a poorly posed design task in that a critical
response is unaffected by the design variables. An example is asking that
an acceleration response be greater than zero at zero hertz. Since this
response can only be zero, the error is triggered.
Avoidance: Remove the responses that trigger the message since they cannot
affect the design.
@
28573; ; Normal Modes - REIGL with auto-omit - SFM 3007
When computing normal modes using Lanczos, if the number of generalized
coordinates assigned is larger than the actual number of component modes
available or auto-omit is performed, then the job will fail with the
following messages:

*** SYSTEM FATAL MESSAGE 3007 (REIGL )

ILLEGAL INPUT TO SUBROUTINE MERGE
Avoidance: Remove the extra generalized coordinates or insert the
following DMAP alter avoidance:
compile moders
alter 'EQUIVX MKAA\/KXX\/NOARED'(,-1)
noared=-1 $
@
28581; 70.5; Security - FLEXlm - Unable to serve licenses on SGI R8K, R10K Sy
stems
If the FLEXlm license server is installed on an SGI R8K, R10K system and
 mscsetup installs the license file, the DAEMON line will incorrectly
reference the irix64 architecture (which isn't used on FLEXlm). As a
result, the license server will not function.
Avoidances: Either of the following two avoidances can be used:
1) Modify the <install_dir>/flexlm/licenses/license.dat file and change
the "irix64"
reference on the "DAEMON" line to "irix"
2) Create a link in the <install-dir>/flexlm directory from irix64 to irix
with the command
ln -s irix irix64
@
28582; 69; DMAP - SUPER3 - OPG1, OQG1, etc do not Exist Upon Return from SU
PER3
In SOL 101, the tables containing containing the results in "OFP" format
are not returned from subDMAP SUPER3. This means if a client DMAP alter
tries to use these tables (OPG1, OQG1, etc) in SESTATIC, they will not
exist.
There is no effect on the solution or output.
Avoidance:
Insert the following DMAP alter in runs using DMAP in SOL 101:
COMPILE SEDRCVR, SOUIN=MSCSOU, NOLIST, NOREF
ALTER 2,2
@
28604; 70.5; Miscellaneous - PUNCH files Cannot Exceed 2gB on HP
Punch files can not exceed the 2 giga-byte limit on HP and possibly other
machines. The 2 giga-byte limit still applies to all FORTRAN I/O files.
Avoidance: The only avoidance is to break-up the output requests so that
the punch files remain under the 2 giga-byte limit.
@
28605; ; Matrix Method - HP-Examplar PA-RISC 1.1 - SFE 4276, SFM 1013
If a large job is run on a HP-Exemplar PA-RISC 1.1 system in Version 70,
then the job may fail in DCMP with one of the following error messages:

*** SYSTEM FATAL MESSAGE 1013 (GBKMGR)

CANNOT FIND CORRECT CLUSTER POINTER RECORD
BIOMSG: ERROR 4276 HAS OCCURRED IN ROUTINE GBKMGR , FILE INDEX
(FILEX) = 0

*** SYSTEM FATAL ERROR 4276.GBKMGR ERROR CODE 4276 PID= 0

***** RECEIVED SIGNAL 2, ABORTING ...

BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX
(FILEX) = 0
STATUS = 0
Avoidance: Use the iterative solver instead of the sparse solver. The
 iterative solver can be activated by adding
the following nastran statement to the top of the input file.
nastran iter=yes
@
28615; 70; SSSALTER - alter1ga.v691 - Wrong Answers, Error in Transformatio
n Matrix
An error in the transformation matrix used in alter1ga.v691 may result in
wrong answers for displacements during data recovery or in applied loads.
Avoidance: Modify the following line in the alter:
UMERGE USET,GOAQ,/GOGT/'G'/'O'/'A' $
to:
UMERGE USET,GOAT,/GOGT/'G'/'O'/'A' $
or request an updated copy of the alter from MSC.
@
28648; ; Nonlinear Transient Heat Transfer - CHEXA - Wrong Answers
Transient heat transfer - Convection to a negative temperature - wrong
answers
If a model has a center node ,for example 8 elements within a cubic. The
center node temperature actually will increases the temperature despite
the fact the whole solid should be cool off by convection to an ambient
temperature odf -20 degree.
Avoidance: use a single CHEXA element to represent the model.
If I increase the timestep from 100 seconds to 1000 seconds,then my center
node temperature will eventually comes down for my 8 hexa problem.
However,I have no problem with a single CHEXA using a timestep of 0.1
seconds.
When I double the mesh by changing it to 64 elements, then the center node
temperature does go down.
See the test deck called /net/kodiak/app/etl/s28648_64hexa.dat
@
28650; 70.7; Aeroelastic Analysis - USPARSE=0 - Wrong Answers
If unsymmetric sparse is turned off using NASTRAN system cell 209
(USPARSE=0) and a complex left-handed fbs is required,
wrong answers are produced and there is no warning message. This error
affects all SOL 145 and 146 runs that use
doublet lattice aerodynamics. Also any solution 200 run that uses doublet
lattice aero in a flutter analysis.
Avoidance: Do not select USPARSE=0.
@
28655; 70.5; Nonlinear Slide-Line Analysis - BGP - MN4W Poor Performance
The performance suffers on the NT running slide-line contact problem using
SOL 106 as memory is increased.
Avoidance: Limiting the memory to 4 million words will improve
performance.
@
28662; 70.5; Frequency Response/Random Analysis - TABRND1, IFP - UFM 315
If the TABRND1 entry uses the log-log option with very small data the
following fatal message may result:
USER FATAL MESSAGE 315 (IFPDRV)
FORMAT ERROR ON BULK DATA ENTRY TABRND1
Avoidance: This message has been encountered for data on the order of
e-12. The only avoidance is to either use the linear option or to use
different units to scale the data up.
@
28669; 2001; Eigenvalue Analysis - PLOAD1, FORCE - Missing Output
If element force output is requested in SOL 103 and one or more PLOAD1
entries exist in the bulk data, there will be no element force output
provided.
Avoidance: Remove the PLOAD1 entries from the input.
@
28673; 70.6; Eigenvalue Analysis - READ, EIGR - SFM 4276, EC 1401 Using GIVEN
S
If GIVENS or MODIFIED GIVENS method is used for an eigenvalue solution and
spill is required, it is possible that the run may terminate with:
SYSTEM FATAL ERROR 4276.GETS68 ERROR CODE 1401
Avoidance: Either provide additional memory so that spill is not required
or use the Lanczos method.
@
28675; 70.5; Documentation - Plotting, QRG - MAXIMUM DEFORM
The documentation for the MAXIMUM DEFORM plot entry is deficient. Remark
2 is not true, and should be replaced with the remark below. Remarks 3,
and 4 should be added:
2. If you wish the FIND command to utilize the d value, a MAXIMUM
DEFORMATION command should precede the FIND command.
3. If you wish the plot deformation scaling to be different then the FIND
scaling a different MAXIMUM DEFORMATION command can appear first before
the PLOT command.
4. For nonlinear plotting MAXIMUM DEFORMATION d and the MAXIMUM
DEFORMATION field on the PLOT command should have the same value.
@
28679; 70.5; Dynamic Analysis - Multiple FREQ, TSTEP, DLOAD Subcases - Ineffi
cient Processing
In SOLs 111 and 112, if multiple FREQ or TSTEP and/or DLOAD subcases are
specified and default data recovery (PARAM, DDRMM, 0) is performed then
SDR1 and SDR2 are re-executed every subcase even though their outputs do
not change.
Avoidance: If no superelements are present, then use SOLs 11 or 12.
@
28680; 70.5; Acoustic Analysis - PARAM,RESVEC, MERGE, MPYAD - SFM 3007, UFM 5
423
When a fluid/structure model with rigid body modes (requiring a SUPORT
card) is run with PARAM,RESVEC,YES the following FATAL errors are issued:
*** SYSTEM FATAL MESSAGE 3007 (REIGL )
ILLEGAL INPUT TO SUBROUTINE MERGE
0FATAL ERROR
or
*** USER FATAL MESSAGE 5423---ATTEMPT TO MULTIPLY OR ADD INCOMPATIBLE
MATRICES, (SMPYAD)

Avoidance: Apply the following DMAP avoidance:

compile modefsrs
alter 'call resvec'(,-1)
call dbstore vafs,,,,//0/0/' '/0 $
compile resvec
alter 'px1',''
call dbfetch /vafs,,,,/0/0/0/0/0 $
partn pab,,vafs/px1,,,/1 $
alter 'px3,,,'(,-2)
merge pxf,,,,,vafs/pxfs/1 $
equivx pxfs/pxf/-1 $
alter 'phux1'(3,-1)
partn ux2,,vafs/ux2s,,,/1 $
equivx ux2s/ux2/-1 $
endalter
@
28684; 70.5; Dynamic Analysis - STRAIN and SPCF output in SOLs 11 - Missing o
utput
In SOLs 11 with default dynamic data recovery (PARAM, DDRMM,0) and
multiple FREQ subcases,
element strain and SPCforce (if PARAM,DYNSPCF=NEW) output will be missing
in the 2nd and subsequent subcases.
Avoidance: Insert the following alter:
compile sol11
alter 1
file mus=append/kus=append/bus=append/istr1=append $
file mum=append/kum=append/bum=append/istr2=append $
@
28685; 70.5; Superelements - CASE, Multiple Dynamic Subcases - SFM 3002
In SOLs 108, 109, 111 and 112 with multiple FREQ and TSTEP (or DLOAD)
subcases, if a superelement has fewer subcases than the residual structure
then the run will fail with the message SYSTEM FATAL MESSAGE 3002 in the
CASE module. For example,
disp=all
subcase 10
super=10
subcase 100
 freq=1
subcase 200
freq=2
Avoidance: Specify SUPER=ALL or the same number of subcases for the
superelements and the residual structure or insert the
following DMAP alter:
compile super3
alter 'casedrm',''
paraml casedr//'trailer'/1/s,n,norec $
if ( dloop>norec ) then $
CASE CASEDR,/CASEDRM/'tran'/norec $
else $
CASE CASEDR,/CASEDRM/APP/DLOOP $
endif $
@
28686; 70.5; Acoustic Analysis - Modal Participation Factors - Missing Outpu
t, Wrong Answers
In solution 111, modal frequency response, modal particicipation factor
output is wrong or missing when PARAM,ACSYM,YES, which is the default.
Avoidance: Insert PARAM,ACSYM,NO or the following DMAP alter:
compile mpf
alter ',ah,'
type parm,,char8,y,acsym='yes' $
if ( acsym='yes' ) then
delete /ah,,,, $
PARTN bHH,VPHFS,/,mAH,, $
add5 mah,,,,/ah/-1. $
endif $
alter 'z2x'
if ( acsym='yes' ) then $
add5 z2x,,,,/z2xx/-1. $
equivx z2xx/z2x/-1 $
endif $
@
28687; 70.5; Eigenvalue Analysis - ASETi, SPOINT, EIGRL, EIGR, SUPORT - SFM 3
007, UFM 4405
In an eigenvalue solution, if null dofs (e.g., SPOINTS) are placed in the
aset, and a SUPORT entry exists, then depending on the eigenvalue solver
requested, the job will fail with one of the following error messages:
*** SYSTEM FATAL MESSAGE 3007 (REIGL )
ILLEGAL INPUT TO SUBROUTINE MERGE
FATAL ERROR
or
USER FATAL MESSAGE 4405,0
NO EIGENVECTORS COMPUTED FOR COMPONENT MODE SYNTHESIS OR SYSTEM
SOLUTION.

Avoidance:
 1. Do not place the unreferenced SPOINTs in the ASET.
2. Remove the SUPORT entries if feasible. This avoidance only works for
the Lanczos method.
@
28688; 70.5; Acoustic Analysis - EMG, CAABSF - SFM 4276, EC 7777
When a solution at zero frequency is requested and frequency dependent
acoustic absorber elements, CAABSF, are specified, then EMG fails with a
divide by zero. At zero frequency, the frequency dependent acoustic
absorber element should put out zero impedance.
Avoidance: Do not request a solution at zero frequency.
@
28697; 2001.0.1; Acoustic Analysis - ACSRC, FORCE, DPD6 - SFM 3007
If an ACSRCE Bulk Data entry has the same DAREA ID (field 3) for loads
applied to structural grid points and the loading applies to more than one
degree of freedom at a given structural grid point then the run will fail
in the DPD module with SFM 3007. The run should fail with UFM 6825 (see
CSR 4618).
Avoidance: See CSR 4618.
@
28710; 70.5; Heat Transfer - MAT4, CONVM - UFM 6715, Wrong Answers
If UFM 6715 shown below is issued in SOL 153 (or 106) then the run will
not terminate but will continue and produce wrong answers.
*** USER FATAL MESSAGE 6715 (ECNVMD)
THERE IS NO DYNAMIC VISCOSITY SPECIFIED ON THE MAT4 ENTRY FOR CONVM
ELEMENT 11.
USER INFORMATION: FORCED CONVECTION ELEMENTS REQUIRE SPECIFICATION OF
DYNAMIC VISCOSITY
ON A MAT4 ENTRY IF ON THE PCONVM ENTRY THE REYNOLDS
NUMBER CONVECTION EXPONENT
IS GREATER THAN THE PRANDTL NUMBER CONVECTION
EXPONENT.

Avoidance: Insert the following DMAP alter:

compile nlstatic
alter 'call super1'
putsys(0,82) $
@
28712; 70.5; Eigenvalue Analysis - REIGL, DECOMP - Insufficient Disk Space
MSC/NASTRAN Version 70 introduces a new re-ordering method in sparse
decomposition for CRAY and SGI computers. While this method has been
found to have performance advantages in nearly all cases, there may be
cases where the improved performance is at the expense of requiring more
memory. If the memory is not available, which is more likely the case on
very large jobs, additional spill may be required and the job may fail due
to insufficient space.
Avoidance: In the single problem case encountered to date the problem was
resolved by increasing NQS memory. An error was found in the error
handling when NQS memory was insufficient. Although we don't think it
 should be necessary for the user to ever have to deselect the new ordering
method, SYSTEM(206) controls the re-ordering method as follows:
0: method selected automatically in symbolic factoring phase
1: minimum degree ordering
2: modified minimum degree ordering for indefinite matrices
3: no ordering (uses given sequence)
4: Extreme ordering (Cray and SGI only; version 70)

@
28713; 70; Security - NT Only - UFM 3060
A problem with MSC/NASTRAN Version 69 NT system causes authorization codes
that would normally produce expiration warning messages to issue USER
FATAL MESSAGE 3060.
Authorization expiration warning messages should be produced under the
circumstances listed below. The only avoidance is to request a new auth
code for a longer duration.
1. If the duration of an auth code is four months or longer, an expiration
warning message is displayed when the code is within 3 months of
expiring.
2. If the duration of an auth code is three months, an expiration warning
message is displayed when the code is within 2 months of expiring.
3. If the duration of an auth code is two months, an expiration warning
message is displayed when the code is within 1 month of expiring.
If the duration of an auth code is one month, no expiration warning
messages are displayed. The expiration threshold is calculated based
on the end of the month of expiration, not on the actual end date.
For Version 69, the following options can be used to circumvent this
problem:
1. Generate auth codes with at least a four month term, and add 3
additional
months to the code. For example, if you want a code to be valid from
August, 1997 thru December, 1997, use the start date of 8/97, and an
end
date of 3/98. If this method is used, I suggest editing the license
file and license report delivered to the customer to use the original
expiration date.
2. Generate a series of one-month codes for each of the months in the
contract period. The customer would add each of these codes to a
single license file, to bypass the error.
@
28716; 70.5; SSSALTERS - rflagb - Wrong Answers for Fluids with Enforced Moti
on
Wrong answers have been observed when the rflagb.v70 alter is used on
fluid-structure problems, and enforced pressures are input via the
Lagrange Multiplier Technique. The problem has been observed only on
short word machines, and does not occur on long word computers such as
Cray. The alter package has been modified to test for wrong answers, and
 to switch to the avoidance procedure automatically when wrong answers are
found. An investigation is underway to determine if this type of wrong
answer could be achieved in any other manner, in particular when no dmap
alters are present. The only cases found to date were for the
circumstances given above.
Avoidance: An avoidance is to cause the unsymmetric matrix solver to run
instead of the symmetric solver. There is some economic penalty for this
method, but it has been shown to work on all problems encountered to date.
One method to cause the unsymmetric solver to be selected is to put in a
k2pp matrix with one very small term on the diagonal, and with a FORM of 1
on the dmig header card. This says the matrix may be unsymmetric (even if
it isn't as in this example), and forces the unsymmetric solver to be
selected.

@
28730; 70.5; Miscellaneous - GPFORCE, BUFFSIZE - Possible Longer Runtime
A significant efficiency improvement has been added to the GPFDR module in
Version 70 which can cause GPFORCE output to run much faster than previous
versions. This improvement is most significant for large models with
multiple subcases and open sets (using the thru option). A 20 to 1 cpu
improvement has been observed on a client model. However, for smaller
models, the GPFDR time may take longer than the previous versions,
especially when using a large buffsize is specified.
Avoidance: Use a smaller buffsize and the run time should then be
comparable to previous versions. As the model gets larger, the runtime
will become faster as compared to previous versions.
@
28757; 70.5; Frequency Response - SOLs 108 and 111, TPL, K2PP, M2PP, B2PP - W
rong answers
If the Case Control commands TF, K2PP, M2PP, or B2PP, change between
subcases in SOLs 108 and 111, then the frequency response solution is
repeated but the answers will not be correct.
Avoidance: There is no avoidance.
@
28762; 70.0.5; Buckling - CBAR, MATi, ALPHA, TREF - Possible Wrong Answers
In a buckling analysis, if the thermal coefficient of expansion on the
MAT1 entry is non-zero, and no thermal loads are applied, then the
differential stiffness for the BAR element will be incorrect.
Avoidance: If there are no thermal loads then remove the thermal
coefficient of expansion since this quantity is not needed for a
mechanical load.
@
28765; ; Documentation - Aeroelastic User's Guide
Equation 8-43 in Section 8.10 of the Aeroelastic User's Guide indicates
that the Imaginary part of Q31 is used to calculate the yawing moment
coefficient due to sideslip rate. This is a typographical error. It
should be using the Real part of Q31. See the Version 65 Aeroelastic
Handbook equation 18 on page 6.2-190.
@
28778; 70.7; SSSALTERs - rflagba - Segmentation Fault
 A frequency response run on an SGI R8000 using the sssalter rflagba.v691
may fail in FRRD1 with "signal 11 : segmentation fault" in the log file.
Avoidance: The problem is under investigation at this time. Try to use
the large mass approach in place of the alter.
@
28782; 70; User Modifiable NASTRAN - XGPIDG - SFM 52
When adding new user modules to the MSC/N user modifiable in version 69.1,
the job may fail with the following error messages:
** SYSTEM FATAL MESSAGE 52 (XGPIDG )
NAMED COMMON /XLINK/ IS TOO SMALL
This error is most likely to occur when adding more than one new module.
Avoidance: Contact your local support for the corrected source to
semdbdf.f.
@
28798; 2001; Nonlinear Analysis - SPCD - Wrong Answers
Applying an enforced displacement with SPC or SPCD while using the CONV
field set to UPW causes wrong answers.

@
28813; ; DMAP - PARAM - CHAR8 Variables on PARAM Entries Must Be BCD, UFM
6011
The Quick Reference Guide states that character parameters are allowed as
inputs on the PARAM entry. However, the
full range of character values are not allowed. The Old Exec construct of
"BCD" parameters is enforced by the preface. This construct states that
only alphanmumeric values are used, that is, the first character must be a
letter of the alphabet. An input of param, char, -1, where char is typed
as a char8 variable in dmap, results in the fatal diagnostic:
USER FATAL MESSAGE 6011 (IPVDFT)
THE MAIN SUBDMAP VPS TYPE OF PARAMETER CHAR DIFFERS FROM THE TYPE
DETERMINED FROM PARAM CARD DEFINITION.
USER INFORMATION: VPS TYPE IS CHARACTER. PARAM TYPE IS INTEGER.
@
28817; 70.5; Documentation - PARAM,BAILOUT - Incorrect Default Value for BAIL
OUT in QRG
The V70 Quick Reference Guide states the default for BAILOUT is -1. In
Version 70 of MSC/NASTRAN, the default is 0.

@
28829; 2001; Superelements - SEP2D, SUPER - UFM (SEP2D), Poor Message
When the SUPER=i,j feature is used to input different load ids in subcases
for different superelements, messages of the form
*** USER FATAL ERROR ---- (SEP2D)
 CASE CONTROL RECORDS FOR SUPERELEMENT 2 CONTAIN A MIXTURE OF
ZERO AND NON-ZERO LOAD SEQUENCE NUMBERS.
may occur. The error message does not cause the program to exit, however.
In all cases studied to date it continues to completion, and produces
correct answers.
Avoidance: The error message may be ignored. Using a fully expanded case
control instead of using SUPER=ALL above the subcase level and SUPER=i,j
within the subcase will also avoid the problem.
@
28840; ; File Management Section - ASSIGN - DELETE Option does not work,
UFM 476
If an ASSIGN statement is used to define the database and the ASSIGN
statement specifies the DELETE option, a pre-existing database may not be
deleted and the run may fail with the message:
USER FATAL MESSAGE 476 (GPFILX)
THE LOGICAL NAME DBALL_1 ASSIGNED TO A PHYSICAL FILE NAME OF THE
PRIMARY DATA BASE, DOES NOT EXIST.
USER INFORMATION: PHYSICAL FILE NAME =

Avoidance: Delete unwanted databases before executing the run

@
28846; 70.5; Matrix Methods - FBS - Wrong Answers for Complex FBS with 2 by 2
Pivots
FBS of complex symmetric, non-positive semi-definite matrices may produce
wrong answers. Standard structural elements do not produce non-positive
semi-definite matrices so this error is unlikely to occur in most runs but
it is possible in poorly conditioned models and users dmaps that solve
Ax=b, where A is complex symmetric and not positive semi-definite. The
error has been observed on runs using the rflagb.v70 alter, which uses
Lagrange Multiplier Method techniques which cause indefinite systems of
equations.
Avoidance: Use the unsymmetric solution method. In the current version
of the rflagb.v70 alter, the model is run first with the symmetric solver,
usually the least expensive option. A check of the solution is made
immediately after the frrd1 module, and if the accuracy is poor, a second
solution is generated with the unsymmetric solver.
If it is suspected that this error is being encountered in a standard
solution, an avoidance to add a K2pp matrix with a form of 1 (square, but
not symmetric). This data is input in the IFO (field 4) of the dmig entry
used for K2pp effects. A very small number may be placed on a diagonal
term, so that this change does not modify the model significantly. This
will cause the unsymmetric option to be selected.
@
28848; 70.5; Inertia Relief - PARAM,INREL, MERGE - UFM 6143, Segmentation Fau
lt
On Digital UNIX Version 69.1 only, if inertia relief is requested in some
subcases, but not all, then the run will fail with a "segmentation fault"
in the MATGEN module. On all other machines in Version 69.1 and 70, if
inertia relief is requested in some subcases, but not all, in the presence
of enforced displacements then the run may fail with UFM 6143 in the MERGE
module.
 Avoidance: Insert the following DMAP alter:
compile selr
alter 'IF ( INREL0',''
IF ( INREL=-1 and not(delta) ) THEN $
@
28853; 70.5; Installation - mscsetup - Not all the files could be decompresse
d
On Version 70 HP systems, mscsetup may incorrectly report the total or
available disk space on a file system during the query or installation
phases.
If the error occurs during the installation phase, the installation may
terminate with the following error message:
Not all the files could be decompressed.
<req> MB free space are required.
<avail_insdir> MB were available in the installation directory
<insdir>.
<avail_tmpdir> MB were available in the work directory <tmpdir>.
Please make more space available and
redo this installation using the command
./mscsetup -t <work_directory> <insdir>/mscsetup.pbk
where <work_directory>
is an alternate directory with at least <req> MB free
space that can be used for scratch purposes
during the installation process.
Avoidance:
The mscsetup script can be corrected by copying the script to a temporary
directory and changing line 3269 from
$awk 'NR==1{avail=$3} NR==2{printf("%d %d\n",avail/MULT,$1/MULT)}'
MULT=$gs_mult -`

to
$awk 'NR==1{avail=$(NF-3)} NR==2{printf("%d
%d\n",avail/MULT,$1/MULT)}' MULT=$gs_mult -`
This will select the fourth-to-last item in the line instead of the third
item in the line. In a "well-formed" line, the third item *is* the
fourth-to-last item, but if the filesystem's actual device name is short,
it won't be.
You can redo the failed installation with the command
mscsetup <insdir>/mscsetup.pbk
where <insdir> is the installation directory specified in the failed
attempt.
@
28857; GL; Optimization/Design Sensitivity - Multi Master - SFM 4276 EC 300
1
The multi-master technique, allows the user to develop a design task in
pieces. That is, an automotive design task might have a chassis in one
model, the body in another and doors and windows in other models. As part
of the adjoint sensitivity development, an efficiency was introduced,
which skips the generation of pseudo-loads if there are no active
responses for a particular subcase. Unfortunately, this breaks the random
order capability of the multi-master technique which would allow the user
to have pseudo-load information available on a standby basis in case a
later job submittal requires it. If a later run needs the pseudo-load
information from a previous run, which did not calculate it, then the
message "SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0" is issued.
Avoidance: The avoidance is to calculate the responses prior to (or in
the same run as) the pseudo-load calculation.
@
28891; 70.5; Documentation - MATHP - Equations in QRG Incorrect
Remark 1 under the Version 70 Quick Reference Guide description of the
MATHP entry shows the equation 2D1=K2D1=K, it should read 2D1=K.

@
28892; 70.5; Optimization/Design Sensitivity - AERO - Wrong Answers
If, in design optimization, a static aeroelastic subcase is present and
their are no static aeroelastic specific responses (TRIM or STABDER on a
DRESP1), the mass effects in the sensitivity analysis are ignored and
wrong answers can result.
Avoidance: Include a token static aeroelastic response. It could be
constrained with very generous limits so that it
is never needed for the design task, but at the same time the DSCREEN
parameters should be set to force retention of the
unimportant constraint.
@
28893; 70.5; Documentation - MODTRL - Various UFM, Missing Documentation
In Version 70, the capability to modify the type and/or precision of a
matrix with the MODTRL module was removed. DMAPs that use MODTRL in this
way will fail with a message of the form:
USER FATAL MESSAGE (MODTRL)
AN ATTEMPT TO MODIFY MATRIX 101 TYPE OF TERMS HAS BEEN MADE.
USER INFORMATION: USE MATMOD OPTION 34 TO CREATE TWO OUTPUT REAL
MATRICES FROM A COMPLEX INPUT MATRIX.
USE ADD OR ADD5 TO CREATE A COMPLEX OUTPUT MATRIX
FROM ONE OR TWO REAL INPUT MATRICES.
Unfortunately there is no documentation in the Version 70 Release Guide
and the diagnostic messages from the MODTRL module
above may be misleading.
Avoidance:
 1. In order to change the precision of a matrix, use ADD5. If the new
precision does not match the machine precision specify PUTSYS(new
precision,55)
before ADD5. For example, on a double-word machine:
a. single to double
ADD5 SINGLE,,,,/DOUBLE $
b. double to single
PUTSYS(1,55) $
ADD5 DOUBLE,,,,/SINGLE $
PUTSYS(2,55) $
or
MPYAD DOUBLE,,DOUBLE/SINGLE///1 $

2. In order to change the type (complex or real) of a matrix, use ADD

to convert real to complex and MATMOD(34) for complex to real. For
example,
a. real to complex
ADD REAL,/CMPLX//(0.,1.) $
b. complex to real
MATMOD CMPLX,,,,,/REAL,/34 $

@
28894; 70.5; Eigenvalue Analysis - Residual Vectors (PARAM,RESVEC) - Wrong An
swers
If there is a c-set and/or r-set and component modes are augmented with
residual vectors (PARAM,RESVEC,YES or PARAM,INRLM,-1) then the answers
will be wrong.
Avoidance: Insert the following DMAP alter:
compile semr3
alter 'call resvec',''
call resvec pg,uset,gm,kfs,ys,bgpdts,scstm,cstms,medge,mff,
cmpho0,loo,koo,cmlama0,moo,got,koo,cmphrc/
cmpho,cmlama/
nozset/norc/seid/'o ' $
compile resvec
alter 'if ( nobset',''
if ( nobset>-1 ) upartn uset,got/,,gorc,/'t'/'b'/'c'/2 $
@
28917; 70.5; Nonlinear Analysis - MATHP - Poor Answers Using MATHP with Isotr
opic Materials
Hyperelastic elements have been designed to allow incompressible
solutions. The elements yield poor results when used with isotropic
material properties.
 1. In the Version 70 QRG entry for MATHP, Remark 1. Wrong: "2D1 = K2D1
= K" Correct: "2*D1 = K"
2. The "MEAN PRESSURE" is wrong for the hyperelastic element.
@
28919; 70; Matrix Methods - DCMP, Spill- SFM 4276.SFPCKI EC 62
The message, "SYSTEM FATAL ERROR 4276.SFPCKI ERROR CODE 62 PID= 0", is
supposed to be issued when the matrix to be decomposed exceeds the maximum
allowable size. However, prior to MSC/NASTRAN Version 70, it may be
issued erroneously, when the DCMP module starts to spill. This is because
the spill logic may overwrite the variable that contains the matrix size.
Avoidance: If possible, increase memory to avoid spill. Otherwise,
change the available memory, so the spill logic will work differently, and
perhaps not overwrite the variable with a number larger than allowed.
@
28936; 70.5; Elements - CQUAD4, OLDQ4K - Unexpected Answers
The NASTRAN system cell OLDQ4K is ignored in Version 69.1 and 70 systems.
Avoidance: The avoidance is to set SYSTEM(173)=1.

@
28939; 70.5; Composites - PCOMP, PSHELL - SFM 3002 MODULE = SDRCOMP
If a model requests STRESS output for some plate elements referencing
PCOMPs and other plate elements referencing an equivalent PSHELL/MAT2
combination generated in a prior MSC/NASTRAN run, then no composite ply
stress output or homogeneous stress output occurs and the following error
message is issued:
0*** SYSTEM WARNING MESSAGE 3002 MODULE = SDRCOMP
0EOF ENCOUNTERED WHILE READING DATA BLOCK EPTS (FILE 103) IN
SUBROUTINE SDRCMA
Avoidance: MID1s, MID2s, MID3s, MID4s that are created on the equivalent
PSHELL/MAT2s are equal to 100000000 plus the PCOMP PID. This is a flag to
NASTRAN that they are from PCOMPs. If the MIDis in the punched output are
change to smaller numbers, then composite ply stresses for the PCOMP
elements and homogeneous stresses for the PSHELL elements may be obtained.
@
28959; 70.5; Superelements - EXTRN - UFM 6894
The Quick Reference Guide documentation for the EXTRN bulk data entry
states that a thru option is available for listing connected points. When
this option is used for scalar points the following fatal error messages
are output:
FOR EXTERNAL SUPERELEMENT ID= 3, SCALAR POINT ID= 712 HAS AN
ILLEGAL COMPONENT DESIGNATED ON THE
EXTRN BULK DATA ENTRY.
Avoidance: Remove the thru range and list each point explicitly.
@
28965; ; Nonlinear Analysis/Hyperelastic Elements - Lumped Mass, MATHP -
Wrong Answers
The 20-noded hyperelastic hexa element produces wrong answers when used
 with lumped mass. The mean pressure output is wrong for all hyperelastic
elements.
Avoidance: Use consistent mass, ignore the mean pressure output.
@
28972; 70.5; Optimization/Design Sensitivity - AERO, FLFACT, DSAD - SFM 3007
If the velocity specification for a flutter design response is a positive
number on the associated FLFACT entry while the
FLFACT velocity specification for the analysis is negative (to indicate
that eigenvectors are required at that velocity), the above fatal message
will occur.
Avoidance: The simple avoidance is to not specify negative velocities on
either the analysis or the design FLFACT entries.
A slightly more complex avoidance is to not design at velocities at which
flutter eigenvectors are desired.
@
28987; 70.5; Submittal - V68.X and earlier used nice(1) with bat=yes, V69 and
later doesn't.
The nastran command, introduced with Version 69, removed the nice(1)
command from the "batch=yes" submittal. This can be returned by setting
the "-bg" queue command to "nice %job%" in any RC file. For example, the
line
submit=-bg=nice %job%
can be placed in any RC file.
@
29002; 70.5; Frequency Response, Flutter, Complex Eigen - DLOAD, CMETH - UWM
6498, Signal 11
In Version 70, if the DLOAD Case Control command references a non-existing
DLOAD or RLOADi entry, or a CMETHOD Case COntrol command references a
non-existing EIGC entry, othen depending on the machine, the job may fail
with one of the following unclear messages:
*** USER WARNING MESSAGE 6498 (MsgGetStr)
Message number is not found
PROGRAMMER INFORMATION:
Check call to MSGPOP from FRLG with the message number 3031
Severity is USER WARNING
Parameter 1 is 2
Parameter 2 is DLT
Parameter 3 is FRRD
USER ACTION: Please contact MSC Client support
Signal 11 (in the log file)
Avoidance: Correct the offending DLOAD/RLOADi entries or CMETHOD/EIGC
entries.
@
29005; 70.5; Miscellaneous - nastxxrc, symbol - Overwrite Ignored
Once a symbol is defined in an rc file, it cannot be overwritten by a
later RC file that you would expect to take precedence. This error
prevents a site from providing a default value in a system RC file and the
 user from overriding that value in their local RC file. Similarly, a user
cannot provide a default in the HOME RC file and then override it in the
local RC file.
Avoidance: There are two possible avoidances:
1. Use a different symbol
or
2. Override the RC file value with an environment variable as follows
(assuming the symbol name is MYDIR):
export MYDIR=/my/big/project # Korn shell
MYDIR=/my/big/project; export MYDIR # Bourne shell
setenv MYDIR /my/big/project # C-shell
@
29011; 70.5; Optimization/Design Sensitivity - ANALYSIS=MODES - Missing Outp
ut
A solution 200 shape optimization analysis that contains multiple subcases
and ANALYSIS=MODES will not output the eigenvectors for the final design
cycle.
Avoidance: Apply the following DMAP avoidance:
compile presens
alter 'deseig=false',''
@
29013; 70.5; Documentation - DRESP2 - Questions about Optimization Input
Remark 2 under the DRESP2 entry should be changed to read:
DRESP1 entries can not span analysis types or superelements.

@
29016; 70.5; Documentation - QRG, EXTRN - Blank Field Feature not Documented.
The EXTRN bulk data entry, first documented in the Version 70 Quick
Reference Guide, allows pairs of blank fields. The pair is read as if it
weren't there. This makes it easier to repair or modify EXTRN entries.

For example, suppose that an EXTRN entry were of the form:

$EXTRN GID CID GID CID GID CID GID CID
$ GID CID (ETC)
EXTRN, 12 123456 14 123456 22 123456 24 123456, $
etc.
and the model is changed so that GID 13 should be added to this list.
Since the sequence of points on the list is of critical importance, it is
necessary to move the input entry for every point after GID 13 to meet the
sequencing requirements. This change can be accomplished as shown:
$EXTRN GID CID GID CID GID CID GID CID
$ GID CID (ETC)
EXTRN, 12 123456
, 13, 123456, 14 123456 22 123456 24 123456, $ etc.
@
29043; 70; Installation - Folded license.dat Files may cause Auth Code Erro
r, UFM 3060
Various mail tools may fold text that exceeds a pre-determined line
length. Generally, the tool will break the line between words, i.e., at
whitespace. In normal text this is, at worst, annoying. In a license.dat
file, this can result in the failure of lmgrd.
Avoidance: Use the mscsetup script in TAN #2778 to install a license.dat
file. If the file has already been installed, but has been folded, copy
the file to another location and reinstall it using the revised mscsetup.
@
29044; 2001; Documentation - PELAS, PELAST - Requires Initial Stiffnes on PEL
AS
When a CELAS1 element is used with a PELAST property entry to represent a
nonlinear spring in nonlinear analysis, an initial, non-zero stiffness
must be specified on the PELAS entry. Other capabilities in MSC/NASTRAN
require the flexibility to allow the stiffness field on the PELAS entry to
be blank so the code will not be changed to trap this condition. A remark
will be added to the PELAST documentation stating the requirement.
@
29047; 70.5; Optimization/Design Sensitivity - AERO - SWM 6911 (QCBFIN)
The warning message:
*** SYSTEM WARNING MESSAGE 6911 (QCBFIN) MORE THAN ONE KHH
appears when a solution 200 run is restarted from a solution 145 run. The
message is, most likely, benign, but does have the potential for producing
wrong answers.
Avoidance: Apply the following DMAP avoidance:
compile respsen
alter 'khhx','qhhx'
type parm,,i,n,ncycl $
dbview khhx=khh where (wildcard and desiter=ncycl) $
dbview mhhx=mhh where (wildcard and desiter=ncycl) $
dbview bhhx=bhh where (wildcard and desiter=ncycl) $
dbview qhhx=qhhl where (wildcard and desiter=ncycl) $
alter 'dscmg'(2)
ncycl = desiter $

@
29050; 70; SSSALTERs - resflexa, RESTART - SFM 3003
It is possible that the resflexa.vxxx alter will not automatically remove
static vectors which are a linear combination of the elastic modes. If
this error occurs, the program may encounter SYSTEM FATAL MESSAGE 3003,
indicating an attempt to read past the end of a record.
Avoidance: The alter has been updated to correct the error. Either
request an updated version of the alter or include the following in the
alter:
mpyad phiajc,pfac,/stamod/0 $ <== existing line
 add stamod,uarsflx/resvec//-1. $ <== existing line
$ filter null columns if they exist - Sept, 1997
matmod resvec,,,,,/nullcol,/12/s,n,nonull/1 $
if(nonull>0)then $
message //'at least one residual vector can be represented' $
message //'by the elastic modes. the following matrix has 1.0' $
message //'in the column associated with those vectors.' $
message //'these columns will be removed automatically' $
message //' ' $
matprn nullcol// $
partn resvec,nullcol,/restemp1,,,/1 $
equivx restemp1/resvec/always $
nloads=nloads-nonull $
endif $
@
29052; 70.5; P-Elements - ADAPT, PSHELL - Negative Eigenvalues, Wrong Answers
, UFM 9050
In eigenvalue analysis, negative eigenvalues may occur in shell p-elements
if bubble functions are present and the distribution of p-values is non
uniform. In static analysis wrong answers or USER FATAL MESSAGE 9050 may
occurBubble functions may be present whenever the p-level is less than 3.
Avoidance: To avoid this error, bubble functions must be turned off by
setting system cell 225 to 1.
@
29053; 70.5; Rigid Elements - RSSCON - Wrong Answers on SGI
The RSSCON (rigid shell to solid connector element) has been found to
produce wrong answers for one problem on SGI computers.
Avoidance: The problem is not yet fully understood. In the case where
the answers were wrong the oload and spc resultants did not agree. Thus,
the user is advised to closely check these results. If they do not agree,
then if possible run the job on another computer or use another method
such as rbe2s to connect the shell elements to the solids.
@
29064; 70.5; Nonlinear - Modes, Preload - Wrong Stresses for Prestressed Eige
nvectors
If a Sol 106 run calculates a preloaded condition and then solves for the
modes, the stress output for the modal portion of the analysis will be
wrong whenever the stiffness matrix has been updated because the stresses
are calculated using the initial geometry. The nonlinear stresses are
still correct.
Avoidance: A DMAP avoidance for Version 70 is:
$
$ s29064.v70 11-13-97
$
$ ALTER FOR SOL 106 (VER 70) FOR NMLOOP DATA RECOVERY
$ AVOIDS LARGE SOURIOUS STRESSES (CSR 29064)
ECHOOFF
COMPILE NLSTATIC $
ALTER 'APPEND *UGNI' $
IF ( NLOOP = NMLOOP ) CALL DBSTORE UGNI,,,,//0/0/' '/0 $
$
COMPILE SEDRCVR $
ALTER 'OPG1,OQG1,OUGV1','' $
 TYPE PARM, ,I,Y,NMLOOP=0 $
TYPE PARM,NDDL,I,Y,LGDISP $
IF ( APP1 = 'REIG' AND NMLOOP > 0 AND LGDISP > 0 ) THEN $
CALL DBFETCH /UGNI,,,,/0/0/0/0/0 $
VECPLOT UGNI,BGPDTS,,CSTMS,CASEDR,,,/UGVBAS/0/0/3 $
MATMOD UGVBAS,BGPDTS,,,,/BGPDTU,/11 $
ELSE $
EQUIVX BGPDTS/BGPDTU/-1 $
ENDIF $
SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,,ETT,OL2,BGPDTU,PJ1,QG,UG,EST,
XYCDBDR,OINT,PELSETS,VIEWTB,GPSNTS,DEQATN,DEQIND,DITID/
OPG1,OQG1,OUGV1,OES1,OEF1,PUGV/APP1/S,N,NOSORT2/NOCOMPS/////
ACOUT/PREFDB/-1./-1./ADPTINDX/ADPTEXIT/FALSE $
ENDALTER $

@
29068; 70; Matrix Methods - UDSFA, FRRD2 - UFM 6136, Insufficient Memory in
FRRD2
It has been observed that for some frequency response problems that an
incorrect estimate is made for the amount of core required for the incore
decomposition. This error results in the message:
*** USER FATAL MESSAGE 6136 (UDSFA) insufficient memory in FRRD2
regardless of how much memory the job is given.
Avoidance: The avoidance is to set system(166)=1 so that if this error is
encountered the job will revert to standard decomp and avoid this error.
@
29077; ; Documentation - CORD3G - Incomplete Feature
The Quick Reference Guide for Version 70 contains information on the
CORD3G Bulk Data entry. This feature is still being developed and is not
ready for use.
@
29082; 70.5; Installation - mscsetup - noclobber Causing mscsetup to Crash
If the noclobber flag is set, mscsetup will fail when it starts.
Avoidance:

Insert the line

set +C
as the first executable line of mscsetup and rerun.
@
29087; 70.5; Aeroelastic Analysis - MKAERO - Lack of precision in Dynamic Aer
oelasticity
The aerodynamic matrices required in SOL 146 are interpolated from
aerodynamics calculated at reduced frequencies input on the MKAEROi bulk
data entries. The calculations are performed in single precision and it
has been observed that this can produce bad answers when a poor choice of
MKAERO data is entered. The problem has not shown up on long word
 machines.
Avoidance: Restrict the number of k's input on the MKAERO to no more than
eight. Avoid k's that are close to one
another (certainly no closer than 0.005). Check answers on a long word
machine if possible.
@
29096; 70.5; Superelements - MPCFORCE - Answers do not Include Inertial Force
s
If MPCFORCEs are requested for MPCs or R-type elements (RBAR, RBE2, etc)
in a transient response or frequency response solution, the resulting
output will not include the inertial forces due to attached masses.
Avoidance: Insert the following alter into the input file:
compile cforce, souin=mscsou, list, noref
alter 'IF ( APP=',''
IF ( APP='reig' or seid>0)then $
$
@
29105; 70.5; Optimization/Design Sensitivity - AERO - Wrong Answers for rtype
=trim or stabder
If a static aeroelastic subcase retains a STABDER or TRIM response types,
but no element or grid type responses, the required sensitivity analyses
will not be performed and wrong answers will be produced for the
sensitivities.
Avoidance: Include element and/or grid response in the subcase and force
its retention through appropriate DSCREEN values.
@
29126; ; Installation - Uninstall Problems on Windows NT, MSCEXE file sti
ll running
If MSC/NASTRAN or MSC/PATRAN on Windows NT are unistalled and the system
may not reboot and may issue the message "MSCEXE file still running".
Avoidance: When performing an uninstall of MSC/NASTRAN or MSC/PATRAN, the
user must kill the FLEXlm service daemon. This may be done through the
following steps; type ctrl+alt+del, select task manager; find the mscexe
process, and then kill it. The system may then be rebooted.

@
29134; ; External Superelements - SEBULK, EXTRN - SFM 3007 (SEP1X)
If the external superelement capability is used with the enhanced
superelement method (begin super=seid), and the EXTRN entry is missing in
the assembly run, then the job will fail with the following unclear
messages:
*** SYSTEM FATAL MESSAGE 3007 (SEP1X )
ILLEGAL INPUT TO SUBROUTINE SEPEXT
FATAL ERROR
Avoidance: Provide the proper EXTRN entry.
@
29139; 70.5; Documentation - QRG, CONVM - Incorrect Documentation
 The description of the CONVM bulk data entry in Quick Reference Guide
incorrectly states that the EID may be that of a CHBDYE or CHBDYP element.
This entry may only reference the CHBDYP entry and only when the FTUBE
option is specified.
@
29140; 70.5; Optimization/Design Sensitivity - SIGSEGV Segmentation Violation
- Problem Size
The message "SIGSEGV - Segmentation violation" can appear in the log file
of short word machines for large optimization design tasks. The cause is
an integer overflow calculation. A general limit on problem size is that
the 2*ndv + ncon SQRT< 2**31-1 = 46,341 (where ndv is the number of design
variables and ncon is the number of constraints,
including property limits but not side constraints.
Avoidance: Reduce the problem size to fit within the above limit or run on
a long word machine where the comparable limit is 3 billion and therefore
not meaningful.
@
29144; 70.5; Optimization/Design Sensitivity - DSAD, PCOMP, CTRIA3 - UFM 5626
(DSAD13)
The above message can occur in error when a DRESP1 entry asks for a
CFAILURE response and the deck contains only CTRIA3's or CQUAD8's and no
CQUAD4's.
Avoidance: Add a dummy, disjoint QUAD4 that references the same PCOMP
entry and constrain its response.
@
29150; 70.5; Cyclic Symmetry, Buckling - SOL 16 (or 77), EIGRL - UFM 4418
If a buckling analysis is performed using a cyclic symmetry solution
sequence (SOL 16), and the Lanczos eigenvalue solver is requested, then
the job will fail with the following error messages:
0*** USER FATAL MESSAGE 4418,
0 NO EIGENVALUE EXTRACTION DATA IS AVAILABLE.
Avoidance: Insert the following DMAP alter:
compile sol16
alter 'matmod',''
MATMOD CASEds,DYNAMICS,,,,/,/23/S,N,NP/S,N,LANCZOS $
endalter

@
29165; 70.5.1; P-Elements - Shell Elements - Wrong Answers
The bending energy is calculated incorrectly for the error estimator for
p-shell elements. Thus, p-shell element models subjected to bending may
produce poor or wrong answers.
Avoidance: There is no avoidance using p-elements.
@
29178; 70.5; Superelements - ASETi - ASETi Entries Ignored, Reduction Not Per
formed
 The aseti entries are ignored if they are placed in the main bulk data
section using the enhanced superelement capabilities (begin super=seid).
Avoidance: There are two possible avoidances:
1. Perform a component mode reduction instead of a Guyan reduction on the
residual structure.
or
2. Place only aset grids in the residual structure and put the rest of
the grid points in a separate superelement.
@
29199; 70.5; Cyclic Symmetry - LOADCYT, CYCLIC1 - SFE 4276, EC 7777
Cyclic symmetry problems on the Cray T90/IEEE Version 70 release of
MSC/NASTRAN that contain LOADCYT entries may fail with the following
traceback in the CYCLIC1 Module.
Beginning of Traceback:
Called from line 42 (address 15262370c) in routine 'TRBK'.
Called from line 39 (address 202316c) in routine 'RECOVR'.
Called from line 112 (address 1101772a) in routine 'sysrcv'.
Called from address 14765131c in routine '__handlr'.
Called from address 1717073a in routine 'CY1CYT'.
Called from line 119 (address 1714533c) in routine 'CY1CYD'.
Called from line 36 (address 1711300c) in routine 'CY1CCL'.
Called from line 267 (address 1747550d) in routine 'CYCL1'.
Called from line 116 (address 10753001b) in routine 'XSEM12'.
Called from line 342 (address 2117a) in routine 'MAINAL'.
Called from line 306 (address 4110d) in routine '$START$'.
Avoidance: Set the "METHOD" field to 1 on the LOADCYT to avoid the error.

@
29200; ; Superelements - PARAM,POST,-1 (-2) - Erroneous Output
In a superelement analysis, if param,post,-1 or param,post,-2 is used, and
is placed inside a subcase, then some of the output for
the output2 file will be sent to the log file. A symptom of this error is
the erroneous output printed in the log file similar to the following:
Process Id = 20562
X 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ$/+-()'=.*"#%&,;:<>?@![]^_\
020080
61^GGPL F^4^66^810/81BK33333^71^60^3^61^GGPL
F^72^60^3^81BK^813.813*813"813#813
Avoidance: There are two possible avoidances:
1. Place the param,post,-1 or param,post,-2 in the bulk data section or
above the first subcase.
or
2. Add the following dmap alter and place "param,ounit1,12" inside the
same subcase as "param,post,-x".
 compile out2geom $
alter 'type.*parm.*ounit1','' $
type parm,,i,y,otape2,ounit1 $

@
29202; 70.5; Optimization/Design Sensitivity - ADJOINT - Wrong Answers for Mu
ltiple B.C.s
The release guide for V70 states that the adjoint sensitivity method is
not available when there are multiple static boundary conditions. In
fact, the adjoint method may still be selected by the code but will yield
wrong answers.
Avoidance: The following alter avoids the problem:
compile descon
alter 'IF ( DOSTATX'(,-1)
mbcflg=false $
alter 'bcdr '(,-1)
mbcflg=true $
@
29203; 70.5; Optimization/Design Sensitivity - SUPORT1, FBS - SFM 2198
The message "SYSTEM FATAL MESSAGE 2198, INPUT DATA BLOCK 101 HAS BEEN
PURGED" may be produced when a design sensitivity task contains static and
normal modes subcases and the normal modes subcase contains a SUPORT1
entry that the static subcase does not.
Avoidance: Use the following DMAP alter
compile desopt
alter 'fbs '(,-1)
CALL SETQ CASESx//
SEID/PEID/S,MTEMP/S,K2GG/S,M2GG/S,B2GG/S,MPC/
S,SPC/S,LOAD/S,DEFORM/S,TEMPLD/S,P2G/S,DYRD/S,METH/
S,MFLUID $
PARAML CASESx//'DTI'/1/255//S,N,SUPORT $
@
29210; 70.5; Aeroelastic Analysis - AEROS - Incorrect NDDL Description for AE
RO Datablock
The first record of the AERO datablock was changed for Version 70, but the
NDDL description was not updated. Three additional words were added to the
datablock to support running static aero and flutter in the same run while
using different reference chord lengths. The words are the reference
chord length from the third field of the AEROS entry, plus the aerodynamic
and reference coordinate system ID's from field 7 and 8.
Avoidance: The correct NDDL description for Version 70 is:
Desc={Aerodynamic Matrix Generation Data }
$ Desc={Contains information for control of }
$ Desc={aerodynamic matrix generation and flutter }
$ Desc={analysis }
$ Desc={ }
 RECORD=HEADER,
NAME(2),CHAR4,{Data block name }
EOR,
RECORD,
ND,I,{Y sym flag SYMXZ }
NE,I,{Z sym flag SYMXY }
REFCS,RS,{reference chord from AEROS }
REFB,RS,{reference span or 0 }
REFS,RS,{reference wing area or 0 }
REFC,RS,{reference chord from AERO }
ACSID,I,{ }
RCSID,I,{ }
EOR,
RECORD,
M,RS,{value of m }
K,RS,{value of K }
EOR,
RECORD=TRAILER,
WORD1,i,{ = 1 }
WORD2,i,{ = 0 }
WORD3,i,{ = 0 }
WORD4,i,{ = 0 }
WORD5,i,{ = 0 }
WORD6,i,{ = 0 }
EOR,
$ Note={ Record 2 is terminated with an integer '-1' }
EOF
@
29213; 70.5; Documentation - STRAIN - Clarify Remark in QRG
Remark 1 of the Quick Reference Guide description of the STRAIN case
control command claims that 'Strains for nonlinear elements are requested
by the STRESS command and appear in the nonlinear stress output'. This
remark should state that 'Nonlinear strains for nonlinear elements are
requested by the STRESS command and appear in the nonlinear stress
output'.
@
29310; 70.5; Optimization/Design Sensitivity - DRESP2, DVPREL2 - UFM 316
The message "USER FATAL MESSAGE 316 ILLEGAL DATA ON BULK DATA ENTRY
DVPREL2" will be produced in error if the input deck contains DRESP2
entries and the highest numbered DRESP2 entry references DVPREL1 data.

Avoidance: Remove the DVPREL1 references from the highest numbered DRESP2
entry. In almost all cases, this is a simple replacement to a DESVAR
reference.
@
29311; ; Iterative Solver - SITDR2, SOLVIT, ITER - UFM 6724 (SB1RM)
The iterative solver may fail with the message:
USER INFORMATION MESSAGE 6447 (SITDR2)
ITERATIVE SOLVER DIAGNOSTIC OUTPUT
EPS : 1.00000E-06
BICWELL PRECONDITIONING
USER FATAL MESSAGE 6724 (SB1RM)
TOO MANY RESTARTS.
 USER ACTION: INCREASE IPAD VALUE ON BULK DATA CARD ITER.
0FATAL ERROR
Avoidance: This message may be caused by a negative term on the factor
diagonal.. A potential cause for getting a negative diag. term is that
the matrix is ill-conditioned. The error may be avoided by increasing
the IPAD parameter on the ITER entry. Another possibility is- to use
BICILL instead of BICWELL.
@
29315; 70.5; Accounting - MSCACT, acct, acid, acdata - Missing Documentation
& Examples
To add accounting of the account ID (acid) and site-dependent account data
(acdata) to the acct and MSCACT utilities, the following steps are
required:
1) Modify the nastXX.lcl, where XX is "69", "691", "70", to pass the acid
and acdata values to acct. For example, in the nast70.lcl file, change
%acct=y:% %a.archdir%/acct $acctf $startdate %user% %out%
MSC/%appl% %version%
to
%acct=y:% %a.archdir%/acct $acctf $startdate %user% %out%
MSC/%appl% %version% "%acid%" "%acdata%"
This change adds the two values, in quotes, to the acct command line.
2) Compile the modified mscact.c source as described in the COG, chapter
6.
# cd <ins-dir>/msc70/util
# msc70 util mscact
This modified program adds "summary=acid" and "summary="acdata".
3) Install the new mscact executable as mscact *and* acct in the
architecture directory, for example,
# cp mscact <ins-dir>/msc70/<arch>/mscact
# cp mscact <ins-dir>/msc70/<arch>/acct
4) Use the "acct=y", "acid", and "acdata" keywords as required.
5) Generate reports using mscact, for example,

msc70 mscact summary=acid <ins-dir>/acct/msc97*.acc

@
29316; 70.5; Optimization/Design Sensitivity - DVPREL1 referenced on a DRESP2
- SIGTRAP FP
The message SIGTRAP TRACE TRAP FP OVERFLOW may appear in the log file when
a DVPREL1 entry is referenced by a DRESP2 entry and sensitivity results
are requested.
Avoidance: Replace the DVPREL1 reference with a reference to the
associated DESVAR's or remove the request for sensitivity output.
@
29324; 70; SSSALTERs - rflagba.v691, TRD1 -Receiving signal number 10 : Bus
error
 When running a large problem in Version 69.1 on the SGI, if insufficient
memory is available, then the job may fail in the TRD1 module with no
error messages in the f06 file. A symptom of this error is the presence
of the following error message in the log file:
Receiving signal number 10 : Bus error
This error has been observed in a large transient dynamic run when using
the rflagba.v691 sssalter.
Avoidance: Increase the memory.
@
29329; ; Virtual Mass - MPRES - Missing Element Lables from Output, Missi
ng Output
Element labels for complex "PRESSURES FOR ELEMENTS ON FLUID STRUCTURE
BOUNDARY" may be missing if SORT2 output is requested.
Avoidance: Request SORT1 output.
@
29338; ; Postprocessing/Nonlinear Analysis - PCOMP, PARAM,POST,-2 - Missi
ng Output
In solution 106 with lgdisp,1 and composite elements, ply stresses are
printed in the f06 file but not in the param,post,-2 output2 file. Also
homogeneous stresses are put in the output2 file although they are not
requested.

Avoidance: Use the following alters:

V68:
$
$ the following alter is used to work-around a problem when
$ using PCOMP in SOL106 with INTOUT=YES on NLPARM (V68.2.3)
$ originally the NLSTATIC DMAP writes OESCP multiple times
$ with PARAM,POST,-1 which causes problems for PATRAN...
$ the fix is to APPEND OESCP within the loop and write it
$ afterwards.
$ modified 10-9 to add I-DEAS output (POST=-2)
$
$ output2 blocks: OESCP laminar stresses
$ OESTRCP laminar strains
$
COMPILE NLSTATIC SOUIN=MSCSOU NOLIST NOREF $ v68.2
$
ALTER 30 $ allow blocks to be overwritten
FILE OESCP=APPEND/OESTRCP=APPEND/OEFIT=APPEND
$
ALTER 433,434$ change intermediate output block names
$ and then append to original block names
$
SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT/
OESCP1,OEFIT1,/0 $
OFP OESCP1,OEFIT1//S,N,CARDNO $
APPEND OESCP1,/OESCP/2 $
APPEND OEFIT1,/OEFIT/2 $
$
ALTER 444,445$ change intermediate output block names
$ and then append to original block names
$
SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT/
OESTRCP1,OEFITX,/1 $
OFP OESTRCP1//S,N,CARDNO
APPEND OESTRCP1,/OESTRCP/2 $
$
ALTER 447,447 $ move output2 outside the loop!
ALTER 464 $
IF ( (POST=-1 AND PATVER>=3.) OR POST=-2 ) OUTPUT2,
OESCP,OESTRCP//0/OUNIT2//OMAXR/'OESCP'/'OESCP' $
$

$FILE**** cpalter
$
$ updated: replaced line numbers with string based alters
$ for V69.0 21-mar-97 L.Proctor/MSC. Should help in subsequent
$ versions.
$
$ the following alter is used to work-around a problem when
$ using PCOMP in SOL106 with INTOUT=YES on NLPARM (V68.2.3)
$ originally the NLSTATIC DMAP writes OESCP multiple times
$ with PARAM,POST,-1 which causes problems for PATRAN...
$ the fix is to APPEND OESCP within the loop and write it
$ afterwards.
$
$ modified 10-9 to add I-DEAS output (POST=-2) $
$ output2 blocks: OES1C laminar stresses
$ OESTR1C laminar strains
$ OEFIT failure indices
$
$
COMPILE NLSTATIC NOLIST NOREF $
$
alter 'FILE.*SAVE' $ allow blocks to be overwritten
FILE OES1C=APPEND/OSTR1C=APPEND/OEFIT=APPEND
$
alter 'SDRCOMP'(1,0),''(1,1) $ change intermediate output block names
$ and then append to original block names
$
SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/
OESCP1,OEFIT1,/0//loadstep $
OFP OESCP1,OEFIT1//S,N,CARDNO $
APPEND OESCP1,/OES1C/2 $
APPEND OEFIT1,/OEFIT/2 $
$
alter 'SDRCOMP'(2,0),''(2,1) $ change intermediate output block names
$ and then append to original block names
$
SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/
OESTRCP1,OEFITX,/1//loadstep $
OFP OESTRCP1//S,N,CARDNO
APPEND OESTRCP1,/OSTR1C/2 $
$
alter 'OUTPUT2.*OESCP','' $ move output2 outside the loop!
alter 'PARAML.*UGNR' $ this is just outside the loop
IF ( POST=-2 ) THEN $
OUTPUT2, OES1C,OSTR1C//0/OUNIT2//OMAXR/'OES1C'/'OSTR1C' $
 OUTPUT2 OEFIT//OTAPE2/OUNIT2//OMAXR $
ENDIF $

V69:
$
$ updated: replaced line numbers with string based alters
$ for V69.0 21-mar-97 L.Proctor/MSC. Should help in subsequent
$ versions.
$
$ the following alter is used to work-around a problem when
$ using PCOMP in SOL106 with INTOUT=YES on NLPARM (V68.2.3)
$ originally the NLSTATIC DMAP writes OESCP multiple times
$ with PARAM,POST,-1 which causes problems for PATRAN...
$ the fix is to APPEND OESCP within the loop and write it
$ afterwards.
$
$ modified 10-9 to add I-DEAS output (POST=-2) $
$ output2 blocks: OES1C laminar stresses
$ OESTR1C laminar strains
$ OEFIT failure indices
$
$
COMPILE NLSTATIC NOLIST NOREF $
$
alter 'FILE.*SAVE' $ allow blocks to be overwritten
FILE OES1C=APPEND/OSTR1C=APPEND/OEFIT=APPEND
$
alter 'SDRCOMP'(1,0),''(1,1) $ change intermediate output block names
$ and then append to original block names
$
SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/
OESCP1,OEFIT1,/0//loadstep $
OFP OESCP1,OEFIT1//S,N,CARDNO $
APPEND OESCP1,/OES1C/2 $
APPEND OEFIT1,/OEFIT/2 $
$
alter 'SDRCOMP'(2,0),''(2,1) $ change intermediate output block names
$ and then append to original block names
$
SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/
OESTRCP1,OEFITX,/1//loadstep $
OFP OESTRCP1//S,N,CARDNO
APPEND OESTRCP1,/OSTR1C/2 $
$
alter 'OUTPUT2.*OESCP','' $ move output2 outside the loop!
alter 'PARAML.*UGNR' $ this is just outside the loop
IF ( (POST=-1 AND PATVER>=3.) OR POST=-2 ) THEN $
OUTPUT2, OES1C,OSTR1C//0/OUNIT2//OMAXR/'OES1C'/'OSTR1C' $
OUTPUT2 OEFIT//OTAPE2/OUNIT2//OMAXR $
ENDIF $
$

V70:
$
$ V70.
 $
$ the following alter is used to work-around a problem when
$ using PCOMP in SOL106 with INTOUT=YES on NLPARM (V68.2.3)
$ originally the NLSTATIC DMAP writes OESCP multiple times
$ with PARAM,POST,-1 which causes problems for PATRAN...
$ the fix is to APPEND OESCP within the loop and write it
$ afterwards.
$
$ modified 10-9 to add I-DEAS output (POST=-2) $
$ output2 blocks: OES1C laminar stresses
$ OESTR1C laminar strains
$ OEFIT failure indices
$
$
COMPILE NLSTATIC NOLIST NOREF $
$
alter 'FILE.*SAVE' $ allow blocks to be overwritten
FILE OES1C=APPEND/OSTR1C=APPEND/OEFIT=APPEND
$
alter 'SDRCOMP'(1,0),''(1,1) $ change intermediate output block names
$ and then append to original block names
$
SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/
OESCP1,OEFIT1,/0//loadstep $
OFP OESCP1,OEFIT1//S,N,CARDNO $
APPEND OESCP1,/OES1C/2 $
APPEND OEFIT1,/OEFIT/2 $
$
alter 'SDRCOMP'(2,0),''(2,1) $ change intermediate output block names
$ and then append to original block names
$
SDRCOMP CASESX2,MPTS,EPTS,ETT,EST,OESXC,OEFXC,DIT,bgpdth/
OESTRCP1,OEFITX,/1//loadstep $
OFP OESTRCP1//S,N,CARDNO
APPEND OESTRCP1,/OSTR1C/2 $
$
alter 'OUTPUT2.*OESCP','' $ move output2 outside the loop!
alter 'PARAML.*UGNR' $ this is just outside the loop
IF ( (POST=-1 AND PATVER>=3.) OR POST=-2 ) THEN $
OUTPUT2, OES1C,OSTR1C//0/OUNIT2//OMAXR/'OES1C'/'OSTR1C' $
OUTPUT2 OEFIT//OTAPE2/OUNIT2//OMAXR $
ENDIF $
$

@
29339; ; Elements - Fluid Points, CELASi, CDAMPi, CMASSi - UFM 2012
If a scalar element is attatched to a fluid point, the following error
message may occur:
** USER FATAL MESSAGE 2012 (GP1GSM)
IDENTIFICATION 1 SAME BETWEEN GRID, SCALAR OR POINT IN
SUPERELEMENT -1
Avoidance: Use DMIG entries for these coefficients or attach the scalar
elements to the 1 dof instead of the 0 dof of the fluid point.
@
29340; 70.5; Heat Transfer Analysis - MATT4 - Wrong Answers
Transient Heat transfer - CBAR element - Version 69 and version 70 only
 If MATT4 references TABLEM1 or TABLEM2 entries that define constant
values, then wrong answers result.
Avoidance: Do not use the MATT4 if the material properties are constant.
Use the MAT4 to specify these properties.
Or if you are using the CBAR element,changed to crod element
@
29341; 70.5; Static Analysis - GPFORCE, GPFDR - SFE 4276 FILPS2 EC 1044
If grid point force output is requested for only some of the subcases, and
the set does not contain all the grid points, then the job may fail with
one of the following error messages below. Similarly in superelements, if
there is a gpforce request above the subcase level that references a set
and the set does not contain a grid point in each of the superelements
then the job will fail.
SYSTEM FATAL ERROR 4276.FILPS2 ERROR CODE 1044 PID= 0 GPFDR
*** USER WARNING MESSAGE 2354 (GPFDR)
MODULE IS UNABLE TO CONTINUE AND HAS BEEN TERMINATED DUE TO ERROR
MESSAGE PRINTED ABOVE OR BELOW THIS MESSAGE.
THIS ERROR OCCURRED IN CODE WHERE THE VARIABLE -NERROR- WAS SET =
43
*** SYSTEM WARNING MESSAGE 3002 (GPFDR )
EOF ENCOUNTERED WHILE READING DATA BLOCK SCRATCH (FILE 301) IN
SUBROUTINE GPFDFR
0FATAL ERROR
Avoidances: There are 3 possible avoidances:
1. Put the gpforce output request above the first subcase.
2. Change the gpforce output to all.
3. Put in a gpforce(plot)=x in the subcase that gpforce printout
is not desired. The value of x must reference the same set
as the other subcases.
@
29342; GL; Superelements - External SE with CMR, CSET - CMR DOF AUTOSPC'd i
n SE0
If a component mode DOF from an external SE is not being attached into the
residual structure as anticipated it is an indication that the mode is
either a linear combination of the static transformation vectors or the
first mode calculated. MSC/NASTRAN will eliminate this mode via the
autospc and INREL DMAP mechanism by design. Two avoidances exist.
1) remove the cset entries from the external SE run.
or
2) ask for more modes to be calculated and add a second spoint putting it
into the qset.

@
29344; 70.5; Elements - CTRIAR, MCID - Error in Material Orientation, Wrong A
nswers
 Specifying a MCID (integer, material coordinate system) in field 7 of the
CTRIAR entry does not work. The resulting THETA (real, orientation angle)
will be a large number ( 512.*(MCID+1) ), yielding incorrect stiffness
for anisotropic materials. It may also lead to wrong element mass matrix.
Avoidance: Do not use an integer in field seven of the CTRIAR entry, use
the CTRIA3 element instead. DIAG 38 may be used to print the material
angles.

@
29348; 70.5; Acoustic Analysis - PARAM,STRUCTMP, PARAM,FLUIDMP - Missing Outp
ut
Requests for structural modal participations and fluid modal participation
factors do not work.
Avoidances: Structural modal participations can be obtained for Version
69 and later systems with the following dmap alter:

compile mpf
alter ',ah,'
type parm,,char8,y,acsym='yes' $
if ( acsym='yes' ) then
delete /ah,,,, $
PARTN bHH,VPHFS,/,mAH,, $
add5 mah,,,,/ah/-1. $
endif $
alter 'z2x'
if ( acsym='yes' ) then $
add5 z2x,,,,/z2xx/-1. $
equivx z2xx/z2x/-1 $
endif $
and fluid modal participation factors can be obtained with the following
alter:
alter 'ENDIF \$ FLUIDMP\<NOFHSET'(,-1)
else $
equivx PHDFH1/PHDFH2/-1 $
@
29356; 70.5; Optimization/Design Sensitivity - DESVAR, mem - SFM 3007
If an optimization job does not have enough memory allocated to it, it may
fail with the message " *** SYSTEM FATAL MESSAGE 3007 (DOM9 ) ILLEGAL
INPUT TO SUBROUTINE DOM9P4".
Avoidance: Allocate more memory (using mem keyword on NASTRAN command
line). If more memory is not available, and the model contains only
DVPREL1 entries with one associated DESVAR, then you can reduce the memory
requirement by setting DELP > DELX on the DOPTPRM entry.
@
29367; 70.7; Nonlinear Transient Analysis - SADD5 - UFM 4551, UFM 5423, Miss
ing Output
When a nonlinear transient model has difficulty converging, NASTRAN issues
 some variation of the message "USER FATAL MESSAGE 4551". The job stops at
that point and results which have been obtained up to that point are
printed out. However, if the variation "USER FATAL MESSAGE 4551, STOPPED
PROBLEM DUE TO TWO DIVERGENCES IN THE SAME TIME STEP." is issued, then no
results are printed, and it is followed by the fatal message:
0*** USER FATAL MESSAGE 5423---ATTEMPT TO ADD INCOMPATIBLE MATRICES,
(SADD5)
COLS ROWS FORM TYPE
NZWD DENS
TRAILER FOR PST1 IS 27 22 2 1
2 875
TRAILER FOR IFS IS 28 22 2 1
19 9740
Avoidance: Correct the model so that it will not diverge.
@
29398; 70.7; Nonlinear Analysis - CGAP - Incorrect SPC Force Output
When the CGAP element orientation is defined by the orientation vector in
fields 6 through 8, the moment due to frictional shear force is not
calculated correctly. The resulting moment ignores the length of the
element and is incorrectly calculated as the (shear force)/2. As a
consequence, the moment component in SPC forces is incorrect.
Avoidance: Use CID on CGAP to define the orientation of the Gap element.
@
29411; 70.5; Optimization/Design Sensitivity - ANALYSIS=BUCK - Missing Eigenv
alue Table
If buckling analysis is used in an optimization run, then only the initial
eigenvalue table will be printed.
Avoidance: Insert the following dmap alter:
compile moders $
alter 1 $
type parm,,logical,n,eigprt $
alter 'if.*nasout.*ofp.*lama'(,-1) $
eigprt=nasout $
nasout=true
alter 'if.*nasout.*ofp.*lama' $
nasout = eigprt $
endalter $
$

@
29412; 70.01; Security - FLEXlM - Missing License Server on Windows NT
The initial delivery of the MSC/NASTRAN Version 70 CD-ROM does not contain
the FLEXlm network licensing servers for Windows NT.
If you want to use your Intel Windows NT system as the license server and
do not already have this software, then you can get the network software
from MSC by calling our toll-free Order Processing hotline at:
1-800-400-4672. A revised Version 70 CD is available, the Version 70.5 CD
will also have the Intel Windows NT license server.
 For Digital Alpha Windows NT systems, we do not have the network license
server available.
We will update the Version 70 Windows NT CD-ROM to contain both licensing
servers in the near future.
@
29421; ; Miscellaneous - after keyword - Maximum File Size 2 Giga Bytes
On an SGI IRIX 6.4, use of the NASTRAN keyword "after" will lower the
maximum filesize from unlimited to 2Gb. This is error is caused by a bug
in the SGI operating system.
Avoidance: There is no avoidance except not to use the "after" keyword.
@
29425; 70.5; Elements - CTRIA6, CQUAD8, ELFORCE - SFM 4276, EC 7777, Trap Enc
ountered
If element force data recovery is requested for the CTRIA6 elements, then
the job may fail on the IBM RS6000, HP, and possibly other machines.
Depending on the machine, the job may fail with the following error
message in the f06 file,
***** RECEIVED SIGNAL 8, ABORTING ...
BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX)
= 0
STATUS = 0
or with the following error messages in the log file without any messages
in the f06 file.
Trap encountered:
SIGTRAP - Trace trap FP overflow
Traceback:
Offset 0x00000024 in procedure gmmatd_
Offset 0x00005044 in procedure str61d_
Offset 0x0000075c in procedure s2bbrd_
Offset 0x00000c54 in procedure sdr2b_
Offset 0x00000260 in procedure sdr2_
Offset 0x00000194 in procedure xsem13_
Offset 0x00000970 in procedure mainal
Avoidance: If strains and mpcforces are not needed then the following
dmap alter will avoid the error:
COMPILE SEDRCVR $ AVOID CSR 29425
ALTER 'OQMG1,,OSTR1'(,-1) $
JUMP NOSTR1 $
ALTER 'OQMG1,,OSTR1' $
LABEL NOSTR1 $
@
29442; 70.5; Post Processing - GPFORCE - Can Not Post Process GPFORCE Results
An error in the access key for grid point force results in the xdb
 prevents these results from being post processed in transient and modes
runs. MSC/PATRAN Version 7.5 supports post processing of these results
but MSC/NASTRAN Version 70.5 will be required to produce the corrected
xdb.

@
29478; 69.1; Documentation - Release Notes/jet101.dat - UFM 6794 (SEPSEB)
The file jet101.dat on the V69 delivery CD, in the
msc69/nast/misc/doc/relnotes directory, contains an error which will
result in the following message:
USER FATAL MESSAGE 6794 (SEPSEB)
DUPLICATE BOUNDARY POINTS DETECTED CONNECTING TO SUPERELEMENT 4 AT
GRID 26 FROM DOWNSTREAM 0.
LIST FOLLOWS:
4101 4001
This error is caused by grid points 4001 and 4002 which are not needed.
Avoidance: Remove the definition for grid 4001 and 4002. The correct
file is on the delivery CD for Version 69.1.
@
29480; 2001; Nonlinear Transient Analysis - PBUSHT, KN, TABLED1 - BIOMSG ERRO
R 1044
Nonlinear transient analysis with CBUSH elements that refer to a
PBUSHTentry that defines stress dependent properties may fail in the
NLTRD2 module with SYSTEM FATAL MESSAGE 4276 BIOMSG ERROR 1044.
Avoidance: There is no avoidance.
@
29489; 70.5; P-elements - GMSURF with zero derivatives - Poor Answers
If a GMSURF has a point with zero derivatives, such as the center of a
cylinder or sphere, linear element edges with grids at the singularity may
have skewed parameterizations. This may result in poor answers.
Avoidance: An avoidance is to specify an FEEDGE for any such element
edge. No GMCURV or POINTs need to be specified for a linear edge.
@
29508; ; Nonlinear Transient Heat Transfer - THERMAL, DBX - Segmentation
Fault
If a nonlinear transient heat transfer analysis, sol 159, specifies
multiple THERMAL case control requests the job may fail with the
following message in the log file.
Process Id = 4393
Receiving signal number 11: Segmentation fault
Job aborting...
dbx: Not found
Unable to call dbx for trace: Error 0
Avoidance: The data deck that encountered this error was attempting to
request printed output for a specific set and punched output for all. The
correct way to make this request is to use the OTIME entry to limit the
printed request.
@
29526; 70.5; Random Response Analysis - TABRND1 - UFM 316 (IFPDRV)
The SKIP feature on the TABRND1 entry does not work when LOG-LOG option is
used.
The following messages are issued.
*** USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA ENTRY TABRND1 *****
SORTED ENTRY COUNT = **
Avoidance: Do not use the SKIP feature when LOG-LOG is used.

@
29531; 70.5; Superelements - SEBULK, SECONCT, SEP1X - SIGSEG, Segmentation v
iolation
When using the enhanced superelement capabilities (begin super=x), only
one grid point may occupy the same location for a particular superelement
at a particular boundary location. This is a documented limitation.
Logically a user may believe that one way to get around this limitation is
to use the SECONCT command; however, the use of the command for this
purpose will cause the job to terminate with no message in the f06 file
and the traceback shown below in the log file. The proper way to get
around this limitation is to use the SEEXCLD command.

Trap encountered:
SIGSEGV - Segmentation violation
Traceback:
Offset 0x00000a90 in procedure sepmc1_
Offset 0x000003f8 in procedure sepman_
Offset 0x00000180 in procedure sepsrh_
Offset 0x00000754 in procedure sep1x_
Offset 0x00000458 in procedure xsem08_
Offset 0x00000970 in procedure mainal
Avoidance: Use the SEEXCLD command to exclude all extra boundary points.
Note that the SEEXCLD command only works with partitioned bulk
superelement.
@
29532; 70.5; Superelements - SEEXCLD, SEMA - SFM 4192, 3062
If the enhanced superelement capabilities (begin super=x) are used, and
there is more than one boundary point that occupies the same location, the
SEEXCLD can be used to exclude the extra boundary points. However, if the
model also contains the old superelement type (seset), then the job may
fail with the following misleading error messages:
0*** SYSTEM FATAL message 4192 (sema)
GRID POINT 10 IN SUPERELEMENT 0 IS NOT DEFINED IN BGPDT
FOR DOWNSTREAM SUPERELEMENT
0*** SYSTEM FATAL MESSAGE 3062 MODULE = SEMA
FAILURE OCCURRED IN UTILITY MAPBTI CALLED IN SUBROUTINE SEMA .
REQUESTED EXTERNAL ID(S) NOT FOUND.
0FATAL ERROR
 Avoidance: Reordering the grid sequencing for the superelements may avoid
this error. Reorder the grids such that the lowest superelement number
has the lowest ids, the next superelement has the next group of higher
ids, and so on. Note that the SEXCLD entry only works with the
partitioned bulk superelement.
@
29547; 70.5; Aeroelastic Analysis - MKAERO - UFM 1126
The message:
*** USER FATAL MESSAGE 1126 (GNFIST)
DMAP MODULE ADD IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME
= AJJ0 , WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE AJJ0 =OVRWRT
is produce when the input deck has m,k (Mach number and reduced frequency)
pairs that are identical.
Avoidance: Remove duplicate mk pairs.
@
29548; 70; SSSALTER - seaspa.v70 - Possible Wrong Answers
An error in the seaspa.v70 alter (wrong value of app to some apd calls)
results in un-needed data being checked improperly. It is unlikely that
the error in the alter will cause errors in answers; however, the alter
has been fixed to remove this error and it also gives better diagnostics
if the auxmodel input file is given a wrong value. The revised alter may
be down loaded from the web and has a revised date of Oct 23, 1997. The
appropriate sample input files have also been updated and posted on the
web at http://www.macsch.com.

@
29577; 70.5.2; P-Elements - Singularity in Stresses When Interior Angle of Elem
ent is Small
P-Elements may encounter a data recovery error when the interior angle of
the element is small.
Avoidance: Other than modifying the mesh, there is no avoidance. In
Version 71 an new system cell 269 is introduced to allow the user to
specify a threshold ratio for the Jacobian determinant at a VU-grid with
respect to it p-element center. If the ratio is smaller than the threshold
specified, then the Jacobian determinant at the VU-grid is replaced with
its element-center value when computing the stress and strain.
@
29580; 70; Documentation - V69/70 COG - Sec 4.8.3 does not include HP Exemp
lar
In the COG, Section 4.8.3 describes a method to flush the F04 and F06
files to disk during a long running job. The method is stated to work on
Convex, Cray, and SGI systems. The method also works on HP Exemplar
systems. The documentation should also explicitly state that after the
buffers are flushed the job is terminated.
Edition 2 of the Version 70 Configuration and Operations Guide has been
clarified.
@
29581; 70.5; DMAP - MATMOD OPTION 13, COPY - SFM 1269 (SLCDRV)
 Copying the lower triangular factor from DECOMP or DCMP module may lead to
the fatal message:
*** SYSTEM FATAL MESSAGE 1269 (SLCDRV)
THE ASSOCIATED FILE FNUMERIC OF DATABLOCK 101 IS ALREADY OPEN.
USER ACTION: CONTACT MSC CLIENT SUPPORT.
0FATAL ERROR
Avoidance: The error occurs only in Version 70, Version 70.5 corrects the
error and prior versions also work.
@
29613; 70.5; Documentation/P-Elements - PVAL - Recommended Minimum p-order.
In general, an analysis started at P=1 will not produce the same results
as starting at other p-levels. Although it is not suggested in the PVAL
documentation where the p-order is set by either default or
recommendation, a minimum starting p-level of 3 is recommended for
accuracy and efficiency.
@
29615; 70.5; Documentation - DMIG, UACCEL - Incorrect Example in QRG
The format and example for the UACCEL entry in the Quick Reference Guide
implies that enforced accelerations are input as one column. The correct
input format is one column for each enforced acceleration case. A better
example with more typical values follows:

$ 2 3 4 5 6 7 8 9
dmig, uaccel, 0, 9, 1, , , , 4
dmig, uaccel, 2, , , 2, 3, 386.4
dmig, uaccel, 3, , , 2, 4, 3.0
dmig, uaccel, 4, , , 2, 6, 1.0
Remark 2 implies that the SUPER case control command must be present.
This is not true. A better remark would be:
"2. The load sequence number interpretation depends on the value of NCOL
field on dmig, uaccel, 0 entry. The recommended method is to set it equal
to the number of loading conditions, 4 in this example. The load sequence
number L is then the sequence number of the subcase to which the applied
acceleration will be applied. For this example, the following subcase
structure might be used:
subcase 11
label = Conventional loads, no enforced accelerations
load = 11
subcase 12
label = L value of 2 used here
subcase 13
label = L value of 3 used here
subcase 14
label = L value of 4 used here
If NCOL is left blank, the number of loading conditions is the number of
columns input on the dmig, uaccel entry. In this example only 3 subcases
would be allowed. Subcase 11 would have to be deleted and the column for
L=2 (the first column input) would apply to subcase 12 ( the first subcase
 remaining).
@
29619; 70.5; Elements - PBARL, PBEAML - Wrong Answers for K1, K2, Cw, J
The BEAM library capability in MSC/NASTRAN Version 70 calculates the K1,
K2 (the shear stiffness factor and CW (the warping coefficient) for the
the BOX section incorrectly.
Avoidance: Calculate the correct cross sectional properties and enter use
the pbar and pbeam entries or create a beam server library with the
correct cross sectional properties.
@
29622; 70.5; Cyclic Symmetry - SOL 15, EIGRL - SFM 3007
If the EIGRL Bulk Data entry is selected and the CYSUP entry is also
specified in SOL 15 then the run may fail with SFM 3007 in REIGL.
Avoidance: Run SOL 115, use EIGR entry or insert the following DMAP
alter:
compile sol15
alter 'reigl ',''
REIGL Kxx,Mxx,DYNAMICS,CASEcc,,MR,DMx,vxcompr,,,vaxw/
LAMA,phixc,MI,EIGVMAT,OUTVEC/READAPP/S,N,NEIGV/RECNO/
///////////nreigl $
@
29625; 70.5; Nonlinear Analysis - PARAM,POST - DBC Segmentation Violation
Nonlinear jobs that fail due to insufficient time or failure to converge
that have param,post,0 may continue and issue a traceback from the DBC
module:
Traceback:
Offset 0x000006c4 in procedure dbcaol_
Offset 0x00000db8 in procedure convrt_
Offset 0x00001180 in procedure dbc_
Offset 0x000002d4 in procedure xsem14_
Offset 0x00000984 in procedure mainal
--- End of call chain ---

@
29639; 70.5; Documentation - V70 Release Notes - Missing DMAP Non-upward Comp
atibilities
The MSC/NASTRAN Version 70 Release Guide did not contain the description
of the DMAP module non-upward compatibilities. Below is the information
that should have been included. The information is also soon to be
included on the MSC external web site. The location will be included in
this report when it becomes available.

This section summarizes DMAP module changes from Version 69 and 69.1 to
Version
70. This information is intended to help you convert your Version 69 and
69.1
DMAP alters or sequences to run in Version 70.
A structured solution sequence excerpt shows how the module is
actually invoked in the structured solution sequences and the additions or
deletions are shown in lowercase in the Version 69 or 70 excerpt,
respectively.

The following modules have changed in Version 70. (SEP1X changed in

V69.1).

DSAD DSAH DSAL DSTAP2 DSVG1P GPSP MDCASE SEP1X

_______________
DSAD Module
Add 1 data block, BGPDTS, to the end of the input list. Add 4 data blocks;
CASDSN, DRDUG, DRDUTB, AND CASADJ, to the end of the output data block
list.
These data blocks are needed for adjoint sensitivity analysis.

Version 69 in SubDMAP DESCON

DSAD RSP1CT, R1TAB, RESP12, OBJTAB, CONTAB, BLAMAX,LAMASX,
DIVTAB,AUXTAB,STBTAB,FLUTAB,OUGVS1,OESS1,OSTRS1,
OEF1AA, EFITS,ES1CS,EA1CS,OQGS1,
DSCREN, XINIT, COORDN, OLX, FRQRSP,
CASEDS, CASERS, UGX, OPTPRM, PROPI /
R1VAL,R2VAL,RSP2R,R2VALR,CVAL,CVALR,OBJTBR,CNTABR,
R1TABR,R1VALR,DRSTBL,FRQRPR,UG1,R1MAPR,R2MAPR/
WGTS/VOLS/S,N,OBJVAL/S,N,NR1OFFST/S,N,NR2OFFST/
S,N,NCNOFFST/APP/DMRESD/SEID/DESITER/EIGNFREQ $

Version 70 in SubDMAP DESCON

DSAD RSP1CT, R1TAB, RESP12, OBJTAB, CONTAB, BLAMAX,LAMASX,
DIVTAB,AUXTAB,STBTAB,FLUTAB,OUGVS1,OESS1,OSTRS1,
OEF1AA, EFITS,ES1CS,EA1CS,OQGS1,
DSCREN, XINIT, COORDN, OLX, FRQRSP,
CASEDS, CASERS, UGX, OPTPRM, PROPI,bgpdts/
R1VAL,R2VAL,RSP2R,R2VALR,CVAL,CVALR,OBJTBR,CNTABR,
R1TABR,R1VALR,DRSTBL,FRQRPR,UG1,R1MAPR,R2MAPR,
casdsn,drdug,drdutb,casadj/
WGTS/VOLS/S,N,OBJVAL/S,N,NR1OFFST/S,N,NR2OFFST/
S,N,NCNOFFST/APP/DMRESD/SEID/DESITER/EIGNFREQ/
s,n,adjflg/pexist/mbcflg $

DSAH Module
Remove data blocks UGSN amd QGSN, and their subsequent commas, from the
input
list. Remove data blocks UGDSN amd QGDSN, and their subsequent commas,
from
the output list. Add input parameters ADJFLG and SEID to end of the
parameter
list.
Structured solution sequence excerpts:
Version 69 in SubDMAP RESPSEN
DSAH DRSTBL,R1TABR,CASEDS,TABECN,BLAMAX,LAMASX,
ugsn,qgsn,OLB,DIVTAB,FRQRPR,VIEWTBDS,CASERS,
BUGX,PHGX/
BUDG,PHDG,CASEDSF,LBTAB,BDIAG,LFTAB,
COGRID,COELEM,ugdsn,qgdsn,DSEDV,OINTDSF,PELSDSF/
APP/DMRESD/NDVTOT $

Version 70 SubDMAP RESPSEN

DSAH DRSTBL,R1TABR,CASDSN,TABECN,BLAMAX,LAMASX,
OLB,DIVTAB,FRQRPR,VIEWTBDS,CASERS,csnmb,
BUGX,PHGX/
BUDG,PHDG,CASEDSF,LBTAB,BDIAG,LFTAB,
COGRID,COELEM, DSEDV,OINTDSF,PELSDSF/
APP/DMRESD/NDVTOT/adjflg/seid $

DSAL Module
Add two data blocks, DRDUTB and ADELU, to the end of the input list. Add
input
parameter ADJFLG to the end of the parameter list. These data blocks and
parameter are needed for adjoint sensitivity analysis.
Structured solution sequence excerpts:
Version 69 in SubDMAP RESPSEN
DSAL DRSTBL,DELWS,DELVS,DELB1,DELF1,COGRID,COELEM,
OUGVDSN,OESDSN,OSTRDS1,OEFDSN,EFITDS,ES1CDS,EA1CDS,
R1VALR,TABDEQ,OLB,DSDIV,DELX,DELS,DELFL,FRQRPR,DELBSH/
DSCM/NDVTOT/DELTAB/EIGNFREQ $

Version 70 in SubDMAP RESPSEN

DSAL DRSTBL,DELWS,DELVS,DELB1,DELF1,COGRID,COELEM,
OUGVDSN,OESDSN,OSTRDS1,OEFDSN,EFITDS,ES1CDS,EA1CDS,
R1VALR,TABDEQ,OLB,DSDIV,DELX,DELS,DELFL,FRQRPR,DELBSH,
drdutb,adelu/
DSCM/NDVTOT/DELTAB/EIGNFREQ/adjflg $

DSTAP2 Module
Remove the last 4 data blocks, DRSTBLG, FOL1V, TOL2V, and FRQRPRG from the
end
of the input list.
Structured solution sequence excerpts:
Version 69 in SubDMAP EXITOPT
DSTAP2 R1TABRG,RSP2RG,drstblg,fol1v,tol2v,frqrprg/DSCMCOL $
Version 70 SubDMAP EXITOPT
DSTAP2 R1TABRG,RSP2RG /DSCMCOL $

DSVG1P Module
Add two data blocks, GPSNTS and ESTDV2B, to the end of the input list.
These
data blocks are required for optimization with p-elements.
Structured solution sequence excerpts:
Version 69 in SubDMAP PSLGDV
DSVG1P EST,ESTDV2F,BGPDVP,CSTMS,MPTS,DIT,DEQATN,DEQIND,UGX,,DSPT1/
EGKP,/
COUPMASS/K6ROT/ALTSHAPE/WTMASS/NOPSLG/1 $

Version 70 SubDMAP PSLGDV

DSVG1P EST,ESTDV2F,BGPDVP,CSTMS,MPTS,DIT,DEQATN,DEQIND,UGX,,DSPT1,
gpsnts,estdv2b/
EGKP,/
COUPMASS/K6ROT/ALTSHAPE/WTMASS/NOPSLG/1 $

GPSP Module
For correct processing of MPC's and rigid elements during PARAM AUTOSPC
operations specify KMM in the second position of the input data block list
and
the string 'KMM' in the tenth parameter position.
Structured solution sequence excerpts:
Version 69 in SubDMAP SEKR
GPSP KNN, ,USET0,SILS,GPLS,,GEOM4S,EQEXINS/USET,/S,N,NOSSET/
AUTOSPC/PRGPSTX/SPCGEN/EPZERO/ACON/
S,N,SING/EPPRT/S,N,NOSET/S,N,NGERR $

Version 70 SubDMAP SEKR0

IF ( NOMSET>=0 ) UPARTN USET0,KGG/KMM,,,/'G'/'M'/'N' $
GPSP KNN,kmm,USET0,SILS,GPLS,,GEOM4S,EQEXINS/USET,/S,N,NOSSET/
AUTOSPC/PRGPSTX/SPCGEN/EPZERO/ACON/
S,N,SING/EPPRT/S,N,NOSET/S,N,NGERR/'kmm' $

MDCASE Module
Add 1 data block, EDOM, to the end of the input data block list and 1
output
parameter, WVFLG to the end of the parameter list. This data block and
parameter are needed for adjoint sensitivity analysis.
Structured solution sequence excerpts:
Version 69 in SubDMAP DESOPT
MDCASE CASEXX/
CASESTAT,CASEMODE,CASEBUCK,CASEFREQ,CASEMTRN,CASESAER,
CASEDVRG,CASEFLUT,,,,,CASESADV,CASESNMB,CASEXN/
S,N,STATCC/S,N,MODECC/S,N,BUCKCC/S,N,DFRQCC/S,N,MFRQCC/
S,N,MTRNCC/S,N,SAERCC/S,N,DVRGCC/S,N,FLUTCC/
////S,N,DESOBJ/S,N,DESGLB/S,N,OBJSID/TRUE $

Version 70 SubDMAP DESOPT

MDCASE CASEXX,edom/
CASESTAT,CASEMODE,CASEBUCK,CASEFREQ,CASEMTRN,CASESAER,
CASEDVRG,CASEFLUT,,,,,CASESADV,CASESNMB,CASEXN/
S,N,STATCC/S,N,MODECC/S,N,BUCKCC/S,N,DFRQCC/S,N,MFRQCC/
S,N,MTRNCC/S,N,SAERCC/S,N,DVRGCC/S,N,FLUTCC/
////S,N,DESOBJ/S,N,DESGLB/S,N,OBJSID/TRUE/s,n,wvflg $

SEP1X Module
In Version 69, SEP1X was executed only if partitioned superelements were
present. In Versions 69.1 and 70, SEP1X is now executed for all models,
regardless of the presence of superelements. SEP1X was enhanced to refine
the
grid point connections of RSSCON elements. Three parameters, SEP1XOVR,
SEBULK,
and TOLRSC, are added to control the operation of the module. SEP1XOVR
and
TOLRSC are described in the MSC/NASTRAN Version 70 Quick Reference Guide.
SEBULK is a logical parameter which is TRUE if partitioned superelements
are
present.
Structured solution sequence excerpts:
Version 69 in SubDMAP PHASE0

IF ( SEBULK
) SEP1X SELIST,GEOM1F,GEOM2F,GEOM4F,SETREE,SGPDTF/
EMAP,SGPDT,SCSTM/
S,N,NOSE/CONFAC/'PEID'//S,N,PARTRS/NQSET/'XEID' $

Version 70 SubDMAP PHASE0

SEP1X SELIST,GEOM1F,GEOM2F,GEOM4F,SETREE,SGPDTF/
EMAP,SGPDT,SCSTM/
S,N,NOSE/CONFAC/'PEID'//S,N,PARTRS/NQSET/'XEID'/
sep1xovr//sebulk/tolrsc $
@
29643; 70.5; Documentation - COMPILER - Missing Documentation
The design intent of the COMPILER LIST statement is to list all subDMAPS
called by a solution sequence. If you wish to list only specific subdmaps
you must input a specific COMPILE LIST statement for each subdmap listing
desired. Remark 4 should be clarified to discuss this design intent.
@
29645; 70.5; Static Analysis - GPFORCE, SUBCOM, SET - Wrong Answers
If GPFORCE force is requested for a set of grid points and SUBCOM is used,
then the the gpforce output from the SUBCOM subcase will be incorrect.
Avoidance: There are two possible avoidances:
1. Request GPFORCE=all
or
2. Use SUBCASE instead of SUBCOM
@
29646; 70.5; Optimization/Design Sensitivity, Differential Stiffness - STATSU
B - UFM 1126
When performing a buckling analysis in Optimization (SOL 200) with no mass
on any unconstrained degree of freedom in the model and with scr=yes on
the nastran command line, then the message "USER FATAL MESSAGE 1126
(GNFIST) DMAP MODULE UPARTN IS ATTEMPTING TO OUTPUT DATABLOCK 202 NAME =
KSF, WHICH ALREADY EXISTS. USER ACTION: DELETE THE DATABLOCK, OR USE FILE
KSF=OVRWRT", may be issued.
The error may also occur in SOLs 103 and 107 through 112 if differential
stiffness is requested with the STATSUB command.
Avoidance: There are three avoidances:
1. Insert a dummy mass element on any unconstrained degree of freedom.
2. Insert the following DMAP avoidance in V70.
compile sekr nolist, noref $
$
alter 10 $
$
$ check for KFF
$ if KFF exists, then so does KFS, KSF, KSS
$ if it doesn't exist generate KFF, KFS, KSF, KSS
$
dbstatus kff//s,n,nokff $
if (nokff < 1) then $
$
$ if in this if statement, then either first time to
$ generate KFF, or KFF was scratch datablock
$ either way KFF needs to be generated
$
alter 13, 13 $
 if (nosset >= 0) then $
$
$ delete upartn output in case this is scratch
$ run and KFS, KSF, KSS are on database
$
delete /KSF,KFS,KSS,, $
UPARTN USET,KNN/KFF,KSF,KFS,KSS/'N'/'F'/'S' $
$
endif $
$
alter 15 $
endif $
3. Specify scr=no on the command line.

@
29648; 70.5; Eigenvalue Analysis - EIGRL - SFM 4276, EC 7777
MSC/NASTRAN Version 70 on Dec Alpha OSF may fail in the EIGRL module with
the message SYSTEM FATAL MESSAGE 4276, ERROR CODE 7777. The following
traceback will be issued in the log file:

handler(0x11fffdff8,0x11fffdf98,0x8,0x0,0x11fffdff8) L. 28
<<?? pc=0x12180e0b4>>(0x0,0x0,0x0,0x8,0xc00000007) L. 28
dot3rd_(0x140602658,0x1208397bc,0x1409cbddc,0x12164ccc8,0x1409eb378) L.
93
lncrdd_(0x1404c1678,0x165000b38,0x1404c1678,0x1409eb378,0x165001708) L.
266
lncr3d_(0x165001c30,0x1407ac4f0,0x165001c50,0x165001c70,0x165000b38) L.
357
lncr2d_(0x1650023c8,0x165002550,0x1650026d8,0x1650008f8,0x140645210) L.
429
lncr1d_(0x165002550,0x165001cc8,0x1650026d8,0x165000250,0x1650002f8) L.
457
lncs1d_(0x1405d2348,0x14064522c,0x140645230,0x140645228,0x1406450c8) L.
893
Avoidance: In the single case that this error has been encountered the
error was avoided by changing the EIGRL entry from specifying ND only to
specifying a specific frequency range.
@
29655; 70.6; SSSALTERs - propa.v70, propf.for - Wrong Answers
The sssalter propa.v70 and accompanying fortran program propf.for enables
the flutter analysis of propellers. Two errors have been discovered in
this package. The first is that the propeller blade aspect ratio is based
on the propeller diameter when it should be based on the radius. The
second is that the example case based on NASA TN 1807, uses a reference
semichord that is twice that specified in the NASA TN.
Avoidance: Either input the aspect ratio into the preprocessing routine
directly or modify the supplied FORTRAN code to use radius in the aspect
ratio calculation. For the test case, divide the input chord lengths by
two.
@
29663; 70.5; Elements - CRAC3D, TEMP(LOAD), GP3D - SFM 3007
If TEMP is applied to a CRAC3D element that has 32 grids or more the run
 will fail with the message:
*** SYSTEM FATAL MESSAGE 3007 (GP3 )
ILLEGAL INPUT TO SUBROUTINE GP3D
Avoidance: If possible use TEMPD to apply the temperature load.
@
29672; 70.5; PCOMP, STRAIN - Missing Output
In post Version 68 systems, the case control command STRAIN=n is supposed
to write ply strains to the .f06 file. This command only works if stress
is also requested. Thus, there is no way to get ply strains only.
Avoidance: To get only strain output, specify the STRESS=m command where
SET m specifies an undefined element.

@
29691; 70.5; Optimization/Design Sensitivity - ANALYSIS=FLUTTER, K2GG, M2GG B
2GG - SFM 6296
If flutter optimization is attempted with CASE CONTROL commands K2GG, M2GG
or B2GG then the run will fail and the message SYSTEM FATAL MESSAGE 6296
Unable to locate a member of datablock QHHL" will be issued.
Avoidance: Don't use the K2GG, M2GG or B2GG commands in an
ANALYSIS=FLUTTER subcase in SOL 200. A complex workaround involving
TABEDIT's is available if this avoidance is unacceptable.
@
29699; 70.5; Postprocessing - PARAM,OUNIT1 PARAM,OUNIT2 - Misplaced Output
The parameters ounit1 and ounit2 are ignored in Version 70 when used with
param,post,-2.
Avoidance: If the ounit1 and ounit2 are needed, the following DMAP alter
may used:
$
$ for v70.0
$
compile super3 nolist, noref $
alter 335,335 $
type parm,nddl,i,y, ounit1 $
type parm,,i,n, ounit1x $
message //' ounit2 before out2geom call= '/ounit2 $
IF ( POST<0 ) THEN $
IF ( OUNIT1>0 ) THEN $
OUNIT1X=OUNIT1 $
ELSE $
OUNIT1X=OUNIT2 $
ENDIF $
ENDIF $
IF ( ( ((DROPT=0 OR DROPT=4) AND APP1<>'BKL1')
OR APP='BKL0') AND
(OGEOM='YES' OR OGEOM=' ') AND
(POST=-2 OR POST=-4) ) CALL OUT2GEOM,
CSTM,GPL,GPDT,EPT,MPT,GEOM2,GEOM3,GEOM4,
GEOM1,DIT,DYNAMICS,EQEXIN,BGPDT,CASECC//
OUNIT1x/S,NP $
@
29704; 70.5; Performance - MPYAD - Large MPYAD has Excessive Run Time
It has been observed on a very large test problem (1.5M DOF) that the
matrix operation PhiT*M required more than 1000 passes, where M is
symmetric and nearly diagonal, and Phi has the same number of rows of M
and 70 columns. This type of operation is unlikely to occur in a standard
solution, but may occur in a user-dmap. This operation is, for example,
used in the checka.v70 sssalter. The MPYAD method selected was method 1T,
which unpacks the (mostly null) rows of M in memory. This is on a model
for which eigenvectors had already been computed, and the time spent up to
the start of the mpyad was less than the estimate for the multi-pass
solution.
Avoidance: Instead of solving for x=PhiT*M in one step, break the
operation into two steps as follows:
xt = M*Phi
x = trnsp(xt)
This way the sparse methods are used in the first multiply, which is done
in less than 10 passes.
@
29705; 70.5; Postprocessing - PARAM,POST,-2, SCR=MINI - Missing OUTPUT2 Data
When an input file is submitted with the scr=mini option and param post,-2
is used, no op2 output will be generated.
Avoidance: Do not use the scr=mini option. If the scr=mini option and
param,post,-2 are needed, then the following dmap alter may be used.
COMPILE SUPER3 SOUIN=MSCSOU NOLIST NOREF
ALTER 335,335
IF ( (OGEOM='YES' OR OGEOM=' ') AND
(POST=-2 OR POST=-4) ) CALL OUT2GEOM,
CSTM,GPL,GPDT,EPT,MPT,GEOM2,GEOM3,GEOM4,
GEOM1,DIT,DYNAMICS,EQEXIN,BGPDT,CASECC//
OUNIT2/S,NP $
@
29710; 70.7; Miscellaneous - PBARL,PBEAML, ECHO=NONE - Unrequested punch file
If both the beam library and echo=none are used in the same run, then a
punch file containing the equivalent beam properties will be created.
Avoidance: If this additional punch file is not desired, then remove the
echo=none case control command.
@
29723; 70.7; Buckling & Differential Stiffness - PARAM,FOLLOWK - SFM 3002 (TA
FF)
When performing buckling analysis or stress stiffened analysis (STATSUB)
with follower force effects (PARAM, FOLLOWK, YES), and a LOAD bulk data
entry references more than one PLOAD4 entry, then the following fatal
message will occur:
*** SYSTEM FATAL MESSAGE 3002 (TAFF )
EOF ENCOUNTERED WHILE READING DATA BLOCK (NONE) (FILE 0) IN
SUBROUTINE TAFMEG
 Avoidance:
1) Set PARAM, FOLLOWK, NO -- thus follower force effects will be ignored.
2) Do not reference more than one PLOAD4 card on the LOAD bulk data entry.
@
29724; 70.5; Miscellaneous - Incorrect OS Version on Fujitsu systems
The Operating System version reported by the nastran command and in the
MSC/NASTRAN F06 title block is incorrect. The correct version is obtained
with the 'uname -v' command. The result "1" indicates "V10L10"; "2"
indicates "V10L20"; and "3" is expected to indicate "V20L10".
The correct version may be specified with the "s.osv" keyword in any RC
file, for example
s.osv=V10L10
@
29736; 70.5; Optimization/Design Sensitivity - ANALYSIS=MFREQ - SFM 4276 EC 3
001 module DSAD
If FREQ3, FREQ4 or FREQ5 Bulk Data entries are specified in SOL 200 run
with ANALYSIS=MFREQ and a METHOD entry, the following message is
erroneously received:
USER WARNING MESSAGE 6645 (FRSRTD)
FREQ3,FREQ4, AND FREQ5 ENTRIES ARE PRESENT IN A DIRECT FREQUENCY RESPONSE
ANALYSIS.
and the following message is issued in the DSAD module:
BIOMSG: ERROR 3001 HAS OCCURRED IN ROUTINE GREAD , FILE INDEX (FILEX) =
0
STATUS = 0
0*** SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0
Avoidance: Insert the following DMAP alter:
compile presens
alter 'dopr3 '
type parm,nddl,logi,n,zuzrl1 $
if ( not zuzrl1 ) then $
delete /CONTAB,R1TAB,RESP12,RSP1CT, $
delete /frqrsp,CASEDS,ointds,pelsetds, $
endif $ not zuzrl1
alter 'enddo'(,-1)
if ( not zuzrl1 ) delete /objtab,,,, $
alter 'return'(,-1)
if ( not zuzrl1 ) delete /objtbg,,,, $
compile desopt
alter 'frlgen ','modacc '
alter 'CALL PRESENS,'(2,-1)
IF ( DESCYCLE=1 ) then $
ISOLAPP=6 $
FRLG CASErS,,,FRL,,,,/
,,,fOL,/
'modal'//'freqresp' $
PARAML FOL//'TRAILER'/1/S,N,NOFREQ $
 matgen ,/pjunk/7/1/nofreq $
modacc cases,fol,,pjunk,,/oli,,xxx,,/'freqresp' $
type parm,nddl,logi,n,zuzrl1 $
type parm,,logi,n,npl $
zuzrl1=true $ indicates 2nd call to presens
npl=gmbs $ save original value of gmbs
gmbs=true $ force 2nd call to presens
endIF $ DESCYCLE=1
alter 'CALL PRESENS,'(2)
gmbs=npl $ restore original value of gmbs
$ added to try to avoid FATAL error...
alter 188,188
@
29759; 70.5; Eigenvalue Analysis - EIGR entry with METH=SINV - Inefficient Pr
ocessing
Auto-omit processing is done unnecessarily with METHOD=SINV on the EIGR
entry. Auto-omit requires a decomposition to condense out the massless
dofs which produces a denser stiffness matrix.
Avoidance: Insert the following DMAP alter:

compile moders
alter 'noared'(2)
noared=-1 $
@
29772; 70.5; Documentation - QRG, DRESP1 - Wrong units
The units for eigenvalues is incorrect as described in remark 18 of the
Quick Reference Guide's description of the DRESP1 entry. The correct
units are (radian/time)**2 instead of (radian/time).
@
29778; 70.5; Plotting - PLTSET - Poor Performance
The PLTSET module for Version 70 has a severe performance degradation as
compared to Version 68.1. For one probelm, the two PLTSET calls in
Version 68.1 used less than 11 CPU seconds while Vhile 70 consumes 470 CPU
seconds. This problem has been corrected in Version 70.5
@
29779; ; P-Elements - h-p Interface, FEEDGE - Incorrect Results
When a model consists of both h- and p-elements, the edges of the
p-elements which are on the h-p boundary will not remain at p-level 1 if
these edges are defined with curved geometry. These edges will adapt to
higher p-level, giving incorrect results.
Avoidance: Do not define curved FEEDGE geometry on the h-p element
boundary.
@
29780; 70.5; P-Elements - Singular Stiffnes Matrix
Singular stiffness matrices or negative eigenvalues may occur in QUAD
p-elements for lower p-values when the p-distribution is nonuniform. This
is due to an error in bubble functions which are used for edges with p=1
and p=2. The error is fixed in V70.5
Avoidance: Turn bubble functions off with system cell 225, specify
NASTRAN SYSTEM(225)=1 in the input file.
@
29781; 70.5; Optimization/Design Sensitivity - PBARL, LINKER - Trap Encounter
ed
In an optimization run, if the PID (field 4) on the DVPREL1 entry
references a PBARL entry, and the FID (field 5) references an invalid
field on the PBARL entry, then the job will terminate in the f06 file
without any messages. In this case, the job terminates after generating
the equivalent PBAR entries. This error will result in the following
error messages in the log file.
Trap encountered:
SIGSEGV - Segmentation violation
Traceback:
Offset 0x000006f8 in procedure domcb_
Offset 0x000005dc in procedure domc_
Offset 0x0000028c in procedure dopr1f_
Offset 0x000010c0 in procedure dopr1_
Offset 0x00000284 in procedure xsem03_
Offset 0x00000934 in procedure mainal
Avoidance: Reference a valid PBARL FID.
@
29788; 70.5; DMAP - MATMOD(12) - Illegal Trailer, BIOMSG: ERROR 211
MATMOD(12) may output matrices with an incorrect trailer of the form of 7,
when it should be a form of 2. Some modules may accept this vector, while
others may crash in unpredictable manners. The ADD module, for example,
exits with the message below:
BIOMSG: ERROR 211 HAS OCCURRED IN ROUTINE BLDPK, FILE INDEX (FILEX) = 0
Avoidance: Use the MODTRL module to change the form to 2.
@
29797; ; Optimization/Design Sensitivity - DSAD, ANALYSIS=FLUTTER - SFM 3
007
Flutter optimization cannot be performed with superelements unless the
design model is in the residual. If attempted the run will fail and
issue the following messages:
^^^ FLUTTER ANALYSIS COMPLETED. DESIGN CYCLE NUMBER= 1
*** SYSTEM FATAL MESSAGE 3007 (DSAD )
ILLEGAL INPUT TO SUBROUTINE DSAD14
0FATAL ERROR
@
29798; 70.5; Optimization/Design Sensitivity - DRESP1, FRSPCF, TSPCF, DSADJ -
UFM 5423, 3007
If the DRESP1 entries requests spc forces (FRSPCF or TSPCF) in a frequency
response or transient response optimization task and all the constraints
are inactive then an incompatible matrix multiply may result leading to
the following message:
*** SYSTEM FATAL MESSAGE 3007 (DSADJ )
ILLEGAL INPUT TO SUBROUTINE DSDJ2D
0FATAL ERROR
 or
*** USER FATAL MESSAGE 5423---ATTEMPT TO ADD INCOMPATIBLE MATRICES,
(SADD5)
Avoidance: Use the DSCREEN entry to force a constraint in each subcase to
be active.

@
29800; 70.5; Virtual Mass - SOLs 3 and 7 through 12, MFLUID - SFM 3001 in MGE
N
In SOLs 3 and 7 through 12, if there is an ELIST1 entry and no free
surface nor plane of antisymmetry then
SFM 3001 is issued from the MGEN module.
Avoidance: Use SOLs 103 and 107 through 112.
@
29802; 70.5; Documentation - QRG, SPLINE3 - Incorrect Equation
The equation written in Remark 2 for the SPLINE3 Bulk Data Entry is
incorrect. The sign on the right-hand side of the equation should have a
minus sign. That is the processing of the SPLINE3 entry is similar to the
MPC entry and the equation should be written as:
u + sum(A * u ) = 0
d i i
@
29817; 70.0.5; Materials - MATT1, MATT2 - Temperature Dependent Materials May F
ail, SFM 3007
In Version 70, material table lookups can fail if there are points outside
the range of the supplied tables. This is true
for both linear and nonlinear solution sequences. Symptoms of this error
are the following messages:
*** USER WARNING MESSAGE 4698. STATISTICS FOR DECOMPOSITION OF MATRIX
KLLRH .
THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN

1.00000E+07 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL.

*** USER WARNING MESSAGE 3057 (NLITER)
MATRIX IS NOT POSITIVE DEFINITE.
*** SYSTEM FATAL MESSAGE 3007 (NLITER)
ILLEGAL INPUT TO SUBROUTINE NLINIT
Avoidance: Define the temperature dependent material tables over a wider
range.
@
29825; 70.5; Superelements - BEGIN BULK SUPER, GENEL - UFM 316, Partitioned S
uperelements
If GENEL elements are used in a partitioned superelement run, then "USER
FATAL MESSAGE 316 (IFPDRV) ILLEGAL DATA ON BULK DATA ENTRY GENEL xxx
SORTED ENTRY COUNT = yyy" will be incorrectly issued if any GENEL ids in
 the second and subsequent bulk data sections are greater than or equal to
any GENEL ids in a previous bulk data section.
Avoidance:
1) Do not use duplicate GENEL ids, and
2) Number your GENELs and order the bulk data sections such that the GENEL
ids of the second and subsequent bulk data sections are less than those of
the preceding bulk data sections.
@
29829; 70.5; Static Analysis - STRSORT - Stress Sorting may be Inefficient
In very large models, if stress sorting is requested, then the time spent
in the stress sorting module STRSORT may dominate the total runtime. This
is most apparent in less efficient I/O machines.
Avoidance: There is no known avoidance.
@
29837; 70.5; Differential Stiffness - STATSUB, SOLs 107, 108, 109, scr=no - U
FM 9058
If scr=no and the Case Control command STATSUB are specified in SOLs 107,
108, or 109 to select the static subcase for differential stiffness, then
the run will fail with the message:
USER FATAL MESSAGE 9058 (STATICS2) - THE SOLUTION FOR THE RESIDUAL
STRUCTURE AND THE CURRENT SUBCASE DOES NOT EXIST.
Avoidance: Submit the job with scr=yes or insert the following DMAP
alter:
compile phase1dr
alter 'not(statics) or kdiff',''
IF ( true ) then $

@
29851; 70.5; Installation - Not all the files could be Decompressed.
If the "GZIP" environment variable is set, mscsetup may fail with the
error message
Not all the files could be decompressed.
<req> MB free space are required.
<free> MB were available in the installation directory /some/dir/
<free> MB were available in the work directory /some/otherdir/
Please make more space available and redo this installation using the
command
/some/dir/mscsetup.pbk -t <work_directory>
where <work_directory> is an alternate directory with at least <req> MB
free
space that can be used for scratch purposes during the installation
process.
 Prior to this error message, you may also see lines like:
7 22:18 36976 120864 69.4% ./msc70/irix64/demo1 11 Jul
7.gz: No such file or directory
22:18.gz: No such file or directory
36976.gz: No such file or directory
120864.gz: No such file or directory
69.4%.gz: No such file or directory
./msc70/irix64/demo1.gz: 69.4% -- replaced with ./msc70/irix64/demo1
7 22:17 28057 74876 62.5% ./msc70/irix/demo2 7 Jul
7.gz: No such file or directory
22:17.gz: No such file or directory
28057.gz: No such file or directory
74876.gz: No such file or directory
62.5%.gz: No such file or directory
./msc70/irix/demo2.gz: 62.5% -- replaced with ./msc70/irix/demo2
Avoidance: Unset the GZIP environment variable:
unset GZIP # Bourne and Korn shell
unsetenv GZIP # C shell (ugh)

and rerun the installation.

@
29852; 70.0.5; Grid Point Force - GPFORCE - UWM 2354 (GPFDR)
Large TETRA models, about 35000 grids, with a GPFORCE request may fail
with the message:
. *** USER WARNING MESSAGE 2354 (GPFDR)
MODULE IS UNABLE TO CONTINUE AND HAS BEEN TERMINATED DUE TO ERROR
MESSAGE PRINTED ABOVE OR BELOW THIS MESSAGE.
THIS ERROR OCCURRED IN CODE WHERE THE VARIABLE -NERROR- WAS SET =
37
*** SYSTEM WARNING MESSAGE 3003 (GPFDR )
ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK
SCRATCH (FILE 302) IN SUBROUTINE GPFDFB
0FATAL ERROR
Avoidance: Increase memory.
@
29859; 70.7; Complex Eigenvalue - CEAD, EIGC, HESS - Traceback in hess1, UWM
6137
Complex eigenvalue analysis that selects the Hessenberg method may issue a
traceback in the hess1 routine with no other message whenever the mass
matrix is singular. Other runs may issue a warning message that the
input matrix is rank deficient or may explicitly state that the mass
matrix is singular and proceed and give unreliable answers. Examples of
the traceback and warning message are given below:
Offset 0x00000030 in procedure pack_
Offset 0x00000ec4 in procedure hess1_
Offset 0x00000c60 in procedure cead_
Offset 0x0000006c in procedure xsem11_
Offset 0x00000994 in procedure mainal
*** USER WARNING MESSAGE 6137 (DFMN)
 INPUT MATRIX IS RANK DEFICIENT, RANK = 159
USER ACTION: CHECK MODEL
*** USER INFORMATION MESSAGE 4158 (DFMSA)
----STATISTICS FOR SPARSE DECOMPOSITION OF DATA BLOCK CMDXX FOLLOW
NUMBER OF NEGATIVE TERMS ON FACTOR DIAGONAL = 6
MAXIMUM RATIO OF MATRIX DIAGONAL TO FACTOR DIAGONAL = 3.5E+16
AT ROW NUMBER 209
Avoidance: When using the Hessenberg method of complex eigenvalue
extraction, carefully inspect the results for all warning messages and use
aset/omit entries to remove massless degrees of freedom from the dynamic
set. Otherwise select the complex lanczos method on the eigc entry which
is not sensitive to massless degrees of freedom.
@
29890; 70.5; Frequency Response - DLOAD, RLOADi, NEC - Possible Wrong Answers
On the NEC, if the DLOAD entry references more than 12 RLOADi entries in a
frequency response analysis, only the last 12 RLOADi entries will be
included in the applied load. The other loads are ignored.
Avoidance: If more than 12 combined loadings are required, use a single
RLOADi entry which in turn references multiple DAREA entries with the same
ID.
@
29898; 70.5; Modal Augmentation - PARAM,RESVEC, SUPORTi - UIM (RESVEC) No Sta
tic Shapes
When a structure is free-free and a SUPORTi entry exists, there will be no
static shapes to augment the mode shapes when there are no constraints in
the model (i.e., no s-set, m-set or o-set). This condition leads to the
following message:
^^^ USER INFORMATION MESSAGE ( RESVEC )
^^^ NO STATIC SHAPES WERE FOUND TO AUGMENT MODE SHAPES
Avoidance: Add a disjoint grid point, with no elements attached to it.
Autospc will cause an s-set to exist.
@
29901; 70.5; Superelements - CBEND - UFM 2016
If a superelement contains a CBEND element that uses a MAT1 that is not
referenced by any elements type, the MAT may not get copied internally
into the bulk data associated with the superelement. This will results
in:
UFM 2016 MODULE = EMG
NO MATERIAL PROPERTIES EXIST
Avoidance: Make sure that each MAT1 that is used by a CBEND elements is
also used by an element type other than the CBEND. This may require a
dummy element such as CBAR be included in the appropriate superelements.
This dummy element does not need to be part of the structure and may be
fully constrained.
@
29909; 2001; Frequency Response - FREQ - SFE 4276, EC 32, Missing Output
In a frequency response analysis, if too many continuation entries are
used for a FREQ entry, then the job will either fail with the following
 error messages:
SYSTEM FATAL ERROR 4276.GETSTR ERROR CODE 32 PID= 0
or the output will not be generated for all the frequencies.
Avoidance: Use a combination of the FREQ3, FREQ4, and FREQ5 in a modal
frequency response analysis.
@
29919; 70.5; Documentation - SEEXCLD - Incorrect description
The description for field 3 (SEIDB) of the SEEXCLD entry should be changed
from
SEIDB Superelement identification. (Integer > 0 or Character = "ALL")
to
SEIDB Superelement identification. (Integer >= 0 or Character = "ALL")

@
29922; ; Rigid Elements - RSSCON, SECNSD - UFM 6692
When a highly skewed plate element is attached to a solid element with an
RSSCON entry, the run may fail with the message:
*** USER FATAL MESSAGE 6692 (SECNSD )
RSSCON NUMBER *** CANNOT CONNECT AN EDGE OF QUAD ELEMENT *** TO HEXA
ELEMENT ***
Avoidance: The only avoidance is to reduce the skewness of the plate
element or to use the GRID option on the RSSCON entry.
@
29932; 69; Installation - MSC/NASTRAN will not run on SGI IRIX 6.4
MSC/NASTRAN V68.2.3 (r2) for the SGI R8K may not run on IRIX 6.4 or later
systems. The failure will occur at program start, before any processing
has begun, with the following message:
MSC/NASTRAN started Mon Nov 17 08:45:14 PST 1997
Process Id = 27828
idalloc: address of dynamic common is too low
current brk value: 0x12700000 ( 309329920)
current address of XNSTRN: 0x121f0000 ( 304021504)
***memory allocation failed, program stop***
real 0m0.04s
user 0m0.02s
sys 0m0.02s
MSC/NASTRAN finished Mon Nov 17 08:45:14 PST 1997
The specific reason is that the amount of code and data put into the
user's address space by the runtime libraries and OS has become too large
and interferes with MSC/NASTRAN's open core common block /XNSTRN/.
Avoidances:
1) Do not run IRIX 6.4 or later or run a more current version of
MSC/NASTRAN.
 2) It appears that at the one site that could run on IRIX 6.4 they had
removed patch SG0002422. Thus, this may be an avoidance. However, be
aware that another site that did not have this patch installed still could
not run.
@
29944; 70.5; Static Aeroelastic Analysis - QSETi, SEQSETi - UWM 5229, SFMs 21
98, 2199, 3007
If QSETi entries are present in SOL 144, Static Aeroelastic Analysis, the
run will fail in subDMAP AESTATRS. The message issued from subDMAP
AESTATRS varies. If the a-set contains only q-set dofs, the following
messages are issued:
*** USER WARNING MESSAGE 5229 (DDCOMP)
INPUT FILE TO MODULE DECOMP PURGED. MODULE RETURNS WITH PURGED
OUTPUT.
*** SYSTEM FATAL MESSAGE 2198 MODULE = FBS
INPUT DATA BLOCK 101 HAS BEEN PURGED.

*** SYSTEM FATAL MESSAGE 2199 MODULE = FBS

SUMMARY, ONE OR MORE OF THE ABOVE FATAL ERRORS WAS ENCOUNTERED IN
SUBROUTINE DFBS .
If the a-set contains q-set and t-set dofs the the following message are
issued:
*** SYSTEM FATAL MESSAGE 3007 (UMERGE)
ILLEGAL INPUT TO SUBROUTINE MERGE
QSETi and SEQSETi entries are not permitted in SOL 144. However this is
undocumented.
Avoidance: Remove the QSETi and SEQSETi Bulk Data entries.
@
29965; 70.5; DBC - SWM 6065 - Undocumented Limit in DBC, SWM 6065 (DBC)
The DBC module always has had a limit of 9999 allowable output positions
available for any process list, ie time, frequency,
modes, etc. The enforcement of this limitation has not always been
consistent until V70. This limit may lead to the message:

^^^ SYSTEM WARNING MESSAGE 6065 (DBC)

^^^ DBCMSG ERROR MESSAGE 43 FROM SUBROUTINE DBCAOL
OUTPUT LIST SUPPLIED EXCEEDS INTERNAL LIMIT, SUPPLIED 20013
INTERNAL LIMIT 10000.
Version 70.5 uses a new dynamic access key technique which effectively
increases the limit for many problems. MSC/XL and other programs may no
be compatible with this new technique. SYSTEM(274)=1 (a.k.a. DBFACT=1)
is provided to access the old technique. This system cell may be set in
the system or user rc file as well as the individual test deck.
@
29976; ; Online Documentation - MNE - Irregular Fonts
The characters on the menus of the MSC/NASTRAN Online Encyclopedia (MNE)
appear as irregular fonts when displayed on a Digital UNIX V4.X.
 Avoidance: Change the Iview*fontList resource value from
-*-bold-r-*-*-14-* to -*-bold-r-*-*-12-*
@
29998; 70.5.2; Matrix Methods - MPYAD METHOD 4T - UFM 4276 EC -31
Certain problems cause MPYAD METHOD 4T to be selected when other methods
would solve the problem much more efficiently. The problem will take
longer to run than it should and may run out of disk space. The symptom
of the bad method selection is that MPYAD METHOD 4 T is selected with a
large number of passes. This can be seen in the f04 file, for instance,
as follows:
20:36:57 267:07 53987.9 626.3# 6843.0 188.9# FDRMGEN
MPYAD END
20:36:58 267:08 53998.1 0.5 6843.1 0.0 FREQRS 15
MPYAD BEGN
METHOD 4 T , NBR PASSES=****, EST.
CPU= 81.7, I/O= 40.0, TOTAL= 121.8
20:36:58 267:08 53998.1 0.0 6843.1 0.0 MPYAD BGN P=2930
20:36:58 267:08 53998.9 0.8 6843.2 0.1 MPYAD PASS= 1
20:36:59 267:09 53999.1 0.3 6843.3 0.1 MPYAD PASS= 2
20:36:59 267:09 53999.6 0.5 6843.4 0.1 MPYAD PASS= 3
20:36:59 267:09 53999.9 0.3 6843.5 0.1 MPYAD PASS= 4
20:36:59 267:09 54000.1 0.3 6843.6 0.1 MPYAD PASS= 5
20:36:59 267:09 54000.6 0.5 6843.8 0.1 MPYAD PASS= 6
Avoidance: The avoidance it to force another method by deselection of
method 4T at the MPYAD where the fatal message occurred. Deselecting
various methods is documented in the MSC/NASTRAN Version 68 Numerical
Methods User's Guide. For instance, for the run above, use the following
alter:
compile subdmap=freqrs
alter 14
putsys(128,66)
alter 15
putsys(0,66)
endalter

@
30005; 70.5; External Superelement - ASETi, SE - SEP1X, SFM 3007
In an external superelement generation run, if the DMIGPCH option is used
with the ASETi entries, and the external superelement contains upstream
superelements, then the punch file will not contain the physical boundary
matrices. The job may also fail with SFM 3007.
Avoidance: There are two possible avoidances:
1. Use BSETi instead of ASETi entries to represent the boundary points.
2. Use the DMIGDB, DMIGOP2, or MATRIXDB option instead of the DMIGPCH
option.
@
30007; ; Documentation/Nonlinear Analysis - Multiple Subcases, Different
Answers
A transient nonlinear analysis may yield slightly different results when
the problem is defined with multiple subcases rather than a single
 subcase. The difference occurs because loads are averaged over three
timesteps. The loads at the end of a subcase and the beginning of the
subsequent subcase may then be slightly different.
@
30009; 70.5; Installation - SIGNAL 11 (SIGSEGV) CODE 5, or CODE 6
The submittal script on MSC/NASTRAN Version 70 on Convex computers has an
error which may cause jobs that request more than 10 million words of
memory to fail with the message:
***** RECEIVED SIGNAL 11 - Segmentation fault (SIGSEGV);
CODE 5 - Invalid page table reference (SEG_INVPTP_TRAP)
JOB ABORTING...
Avoidance: Modify the nast70.lcl file in the msc70/bin directory as
follows:
%MSC_ARCH=convex:% $tcmd %a.solver% $asgf %mem%
rather than
%MSC_ARCH=convex:% $tcmd %a.solver% %a.rel=68:% $asgf %mem%
@
30010; ; Superelements - BEGIN BULK SUPER, RESTART - SFM 3001
A restart of a superelement problem will result in the fatal message shown
below, if the cold start contains superelements defined in the main bulk
data section as well as other superelements defined in the partitioned
bulk data sections.
*** SYSTEM FATAL MESSAGE 3001 (SEP1X )
THE EXECUTIVE FILE 0 DOES NOT EXIST.
PROGRAMMER INFORMATION: SUBROUTINE SEPMAB
Avoidance: Define all superelements in either the main bulk data section
or the partitioned bulk data section, but not both.
@
30011; 70.5; Optimization/Design Sensitivity - MFREQ, OFREQ - SFM 4276.GREAD
EC 3001
If a multidisciplanary optimization run uses MFREQ as one of the
disciplines, contains an OFREQ Case Control requests and has a value of
P2>=8 on the DOPTPRM entry to indicate that the responses are to be
printed, then the job will fail with the following fatal message in the
DOM10 Module:
SYSTEM FATAL ERROR 4276.GREAD ERROR CODE 3001 PID= 0
USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH DIAG
44 INSERTED IN THE EXECUTIVE CONTROL SECTION.
Avoidance: Remove the OFREQ request or lower the value of P2 on the
DOPTPRM entry.
@
30015; 70.5; ESTIMATE - Incorrect estimates of DOF on TRIAX6 elements.
The ESTIMATE utility may incorrectly count the degrees of freedom
associated with CTRIAX6 elements. If the singular degrees of freedom are
explicitly constrained, either on spc entries, grid entries or grdset
entries, then the ESTIMATE will assume that a constraint in the 2 degree
of freedom is removing one of physical degrees of freedom. This is not
correct as a CTRIAX element is define in the x-z plane not the x-y plane.
 As a result, ESTIMATE will undercount the DOF by one per CTRIAX6 element.
Avoidance: Do not explicitly constrain the singular degrees of freedom,
let autospc perform this function.
@
30020; 70.7; Matrix Methods - MPYAD 2 or 3 - Wrong Answers
When very low memory is requested such that MPYAD method 2 or 3 are
selected, and multiple passes are required, then the job will produce
wrong answers. Insufficient memory causes the sparse method to be
deselected and MPYAD METHODs 2 and 3 to spill in multiple passes. This
information is contained in the f04 file with DIAG 19.
Avoidance: Deselect method 2 and 3 by adding the following system cell to
the top of the input file or system rc file:
nastran system(66)=60
This may cause some jobs that require methods 2 or 3 to fail with the
message:
*** SYSTEM FATAL MESSAGE 3001 (*** )
THE SCRATCH FILE 399 DOES NOT EXIST.
PROGRAMMER INFORMATION: SUBROUTINE GOPEN
0FATAL ERROR
The most likely jobs to encounter this message are optimization jobs in
DOPR1. If the fatal message is encountered, set system(66) back to the
default of 0, turn on diag 19, and give the job plenty of memory. Then,
scan the Execution Summary (f04 file) for MPYAD and ensure that multiple
PASSES were not required for METHODs 2 or 3.
@
30023; 70.5.1; Miscellaneous - Ishell does not work on HP Systems
The ISHELL module will not work on the HP Version 70 system because the
allocated size of the stack and child thread was too small for HP machine.
The original cause of this error was fixed in Version 70.5.0 but another
error effectively renders the error not fixed on HP-UX 10.2 systems. The
error has been fixed in Version 70.5.1.
Avoidance: There is no avoidance.
@
30027; 70.5; Optimization/Design Sensitivity - SPCD, scr=no, auxiliary models
- Wrong Answers
In a shape optimization run with SPCD loads on auxiliary models and scr=no
the answers are wrong after the first cycle.
Avoidance: Use scr=yes or insert the following DMAP avoidance:
compile update
alter 'delete \/yg1,'
DBVIEW ygp=yg (WHERE DESITER=DESP) $
RESTART YGp,YG,//TRUE $ DETECT ENFORCED DISP CHANGES
endalter
@
30030; 70.5; Aeroelastic Analysis - SSG2 - Poor Efficiency
The ssg2 module reduces the spline matrix gtkg to several other
 components. The r-set component is not used anywhere, and this operation
takes a considerable amount of cpu time.
Avoidance: An alter which eliminates this computation is to remove this
component from the dmap call:
COMPILE AESTAT NOLIST NOREF $ V70
ALTER 'SSG2','' $ REMOVE GTKR FOR EFFICIENCY. NOT USED DOWNSTREAM.
$SSG2 USET,GM,,,GOA,DM,GTKG/GTKR,GTKO,GTKS,GTKA,GTKL $
SSG2 USET,GM,,,GOA,DM,GTKG/,GTKO,GTKS,GTKA,GTKL
@
30037; 70.5; Optimization/Design Sensitivity - DESOPT, TFL - UFM 3055
If the TFL command is specified and ANALYSIS=MFREQ or DFREQ in a SOL 200
run that performs adjoint sensitivity then the run will fail with UFM 3055
at statement 404 in subDMAP DESOPT.
Avoidance: Insert the following DMAP alter:
compile desopt
alter 'PARAML CASERS'
PARAML CASERS//'DTI'/1/ 15// S,N,TFL $

@
30041; 70; Installation/Security - No read access to authorize.dat or licen
se.dat Files
If the user running mscsetup has their umask set to a value that disables
group or other access, e.g., 026 or 027, mscsetup may create files that
are not readable by the user community. The specific files of concern
are:
<install-dir>/flexlm/licenses/license.dat and/or
<install-dir>/conf/authorize.dat.
Note, having a restrictive umask value is a necessary, but not sufficient,
condition for the error to occur. This error has been corrected on the
MSC/NASTRAN V70 (r4) CD.
Avoidance: Either of the following actions below will avoid the error.
1) Set the umask to 022 before running mscsetup.
or

2) Change the permissions of the affected files after installation, e.g.,

chmod a+r <install-dir>/flexlm/licenses/license.dat
chmod a+r <install-dir>/conf/authorize.dat
chmod a+r <install-dir>/conf/nast70rc.dat
@
30045; 70.5; Frequency Response Analysis - CDAMP1, CBUSH - Wrong SPCFORCE Ou
tput
If frequency dependent damping and no stiffness is specified on a
CBUSH-element then the SPC forces will be wrong.
 Avoidance: Insert the following DMAP alter:
compile freqrs
alter 'equivx bggf\/xggf',''
ADD bggf,/xGGf/iomega $
endalter

@
30054; 70.7; Eigenvalue Analysis - STATSUB, PARAM,INREL - UFM 9058 (SUPER3)
If a PARAM,INREL,-1 is erroneously placed in a SOL 103 run along with a
STATSUB command for differential stiffness, then the following error
message will be received:
^^^ USER FATAL MESSAGE 9058 (SUPER3) - THE SOLUTION FOR THE RESIDUAL
STRUCTURE DOES NOT EXIST.
This error needs to be trapped in order help the user correct the problem.
Avoidance:
Remove the PARAM,INREL,-1 since it is only valid for inertial relief in
SOL 101.
@
30058; 70.5.2; Documentation - Incorrect Specification of FLEXlm License Server
s on Windows NT
The MSC/NASTRAN Version 70.5 NT "Configuration and Operations Guide"
indicates the format for a list of alternate FLEXlm licensing files or
license server nodes is specified as
value:value:...
The correct format for NT machines is
value;value;...
Note, ANY list of values, such as a PATH, or the FLEXlm license spec, use
semicolons on Windows NT.
@
30073; 70.6; Nonlinear Analysis - APPEND - SFM 4276, EC 1039
In a nonlinear run, it is possible that the run may terminate in APPEND
with SFM 4276, EC 1039.
Avoidance: Set SYSTEM(158)=1000 and SYSTEM(159)=500

@
30083; 70.5; Transient Response Analysis - TLOAD, TRL ,- time: command termin
ated abnormally
If a transient response analysis has a DLOAD that references a nonexistent
TLOAD id, the job aborts without any messages except:
*** TERMINATING
/pw/devl/svr4/structures/t477/nastran/msc70/solaris/analysis.ult
ra
 *** Received signal 11 (SIGSEGV)
time: command terminated abnormally
in the log file.
Avoidance: Remove the reference to the nonexistent TLOAD id.

@
30101; 70.5; Superelements - Component Mode Synthesis - UFM 3007 (UMERGE)
If an insufficient number of q-set degrees of freedom are allocated
through the SEQSETi entry to hold all of the internally generated residual
flexibility vectors, the run will fail in the UMERGE module with the
message "USER FATAL MESSAGE 3007" indicating illegal input to a
subroutine.
Avoidance: Supply a larger number of SEQSETi degrees of freedom.
@
30104; 70.5; Documentation, QRG - TIME - Incorrect Default Value
The default value for the executive control section entry "TIME" changed
in V69. The value is now 1.892160e9 (seconds), not 60 minutes as
documented.
Avoidance: Specify a user defined value for TIME.
@
30109; 70.5; DMAP - NORM - SFM 4276 EC 7777
The norm module sets an output parameter to the size of the maximum
magnitude number in a matrix. On short word machines these numbers can be
as large as 1.E60+. The number is stored in an output parameter of single
precision whose maximum size is 1.E30+. Therefore, matrices which are
legal for other modules cause overflow in norm.
Avoidance: Short of not generating matrices with large numbers in them,
there is no avoidance.
@
30124; 70.0.3; Miscellaneous - GPWG, PARAM,GRDPNT,0 - time: command terminated
abnormally
Jobs with PARAM,GRDPNT on SUN SOLARIS Version 70.0.0 systems may fail in
the GPWG Module with the message:
*** TERMINATING /local/msc/msc70/solaris/analysis
*** Received signal 11 SIGSEGV
time: command terminated abnormally.
Avoidance: Remove the PARAM,GRDPNT from the data deck. The error could
not be repeated on Version 70.0.3 or 70.5.
@
30136; 70.5; Installation - "The following items are missing:"
If a user or site has replaced the standard tar(1) command with another
"equivalent" utility, e.g., pax(1), mscsetup may fail with the following
error:
The following items are missing:
msc70/access
 where msc70/access is any file to be loaded from the CD-ROM.
Avoidance: Make sure the tar command invokes the standard tar(1)
executable.

@
30147; ; Nonlinear Analysis - NLPCI, CRIS - Negative Load Factors
If the Crisfield arc length method (TYPE=CRIS) is used in a post buckling
analysis, then the job may not converge. A symptom of this problem is the
printing of large negative load factors as shown below.
SUBID 2 LOOPID 29 LOAD STEP -6.389 LOAD
FACTOR-7.38852477
SUBID 2 LOOPID 30 LOAD STEP -7.891 LOAD
FACTOR-8.89144518
SUBID 2 LOOPID 31 LOAD STEP -10.059 LOAD
FACTOR***********
Avoidance: Use the RIKS method.
@
30200; 70.5.2; Rigid Elements - RSSCON - UFM 6694, Incorrectly Worded Fatal Mes
sage
When two solid elements are connected to one shell element using the ELEM
option on the RSSCON entry, the following error message
is issued:
USER FATAL MESSAGE 6694 (GP4RS4)
SHELL GRID POINT xxx IS CONNECTED TO MORE THAN TWO SHELL ELEMENTS BY
RSSCON BULK DATA ENTRIES.
THIS IS NOT ALLOWED IF THE SOLID ELEMENT FACES CONNECTED ARE TRIANGLES.
Although the user made an illegal connection, the error message is
incorrectly worded. The problem is not that the shell is connected to a
shell but that it is connected to two solids. The RSSCON entry can not
connect more than one solid element to one shell element when the ELEM
option is used. If the shell to solid connection spans across more than
one solid element, the GRID option must be used.
@
30201; 70.7; Loads - FORCE - Unexpected Answers
A negative sign in the CID field of the FORCE entry may not be trapped and
the analysis may continue as if the field was left blank indicating the
loads are to be applied in the basic coordinate system.
Avoidance: Other than checking the oload output and entering proper data
in the field, there is no avoidance.
@
30204; ; Sensitivities/Optimization - MSGMESH - MSGMESH FE 9104
Running an input file with an auxiliary model generated with MSGMESH fails
with the fatal error:
1 CGEN QUAD4 2001 2000 1 3 2001
2054 INPUT **
*** MSGMESH FATAL ERROR 9104 *** N1=1 N2=5 N3=0
N4=0 ERROR **
 for any MSGMESH input entries.
Avoidance: Do not use MSGMESH entries in the auxiliary subcase.
@
30205; 70.5; Dynamics - PARAM,RESVEC - Wrong Answers with MAX Normalization
The Modal Augmentation (Residual Vector) Method introduced in Version 70
gives wrong answers if the modes are normalized to MAX.
Avoidance: Use MASS normalization of modes.

@
30212; 70.5; Superelements - SUPORTi, CSETi, GDR - SFM 3007 (UMERGE)
When performing component mode synthesis using generalized dynamic
reduction (GDR) and CSETi and SUPORTi entries are present the run may fail
in the GDR module with the following message:
^^^ USER INFORMATION MESSAGE ( RESVEC )
^^^ NO STATIC SHAPES WERE FOUND TO AUGMENT MODE SHAPES
*** SYSTEM FATAL MESSAGE 3007 (UMERGE)
ILLEGAL INPUT TO SUBROUTINE MERGE
0FATAL ERROR
Avoidance: Apply the following DMAP alter to avoid the error:
compile semr3
alter 'call resvec',''
call resvec pg,uset,gm,kfs,ys,bgpdts,scstm,cstms,medge,mff,
cmpho0,loo,koo,cmlama0,moo,got,koo,cmphrc/
cmpho,cmlama/
nozset/norc/seid/'o ' $
alter 'equivx cmpho0',''
CALL SEGOA uset,GOT,/GOAt/NOOSET/NOQSET/NOTSET $
EQUIVX CMPHO0/PHIVZ/NORC $
IF ( NORC>-1 ) Umerge1 USET,CMPHO0,CMPHRC,,/
PHIVZ/'V'/'O'/'C'/1 $
INREL USET,PHIVZ,,,,GOAT,/GOQ/0/EPSRC $
dotrip=true $
alter '/goq/' $
paraml goq//'trailer'/1/s,n,noqset1 $
if ( noqset1>noqset ) then $ Truncate GOQ to the true q-set size
nqq1=noqset1-noqset $
matgen ,/vq1qx/6/noqset1/noqset/nqq1 $
partn goq,vq1qx,/phx3,,,/1 $
equivx phx3/goq/-1 $
dotrip=true $
endif $
compile resvec
alter 'if ( nobset',''
if ( nobset>-1 ) upartn uset,got/,,gorc,/'t'/'b'/'c'/2 $
@
30229; 70.5; Installation - SGI R10000 misidentified during V70 installation
On multi-processor SGI system running IRIX 6.2 or later, mscsetup may
report a syntax error such as the following:
 ./mscsetup[24]: 10000^J10000^J10000^J10000^J10000^J10000^J10000^J10000:
syntax error
Avoidance:
Set the MSC_ARCH environment variable to "irix64" for any R8 or R10
processor running IRIX64, or "irix" for any other combination of processor
and OS. Note, specifically "O2" systems use the "irix" architecture.
For example, on an Origin2000 system, set
export MSC_ARCH=irix64 # Korn shell
setenv MSC_ARCH irix64 $ C-Shell
@
30235; 70.5; Superelements - SUPORT1, PARAM,SESEF - UFM 3055 (MPYAD)
The use of SUPORT1 entries with PARAM,SESEF,>=0 and superelements will
result in *** USER FATAL MESSAGE 3055 (MPYAD ) AN ATTEMPT TO MULTIPLY OR
MULTIPLY AND ADD NON-CONFORMABLE MATRICES TOGETHER WAS MADE IN MODULE
MPYAD.

Avoidance:
$V70
compile postreig nolist noref
alter 4 $
type parm,nddl,i,n,suport $
alter 25 $ after call setq
paraml casedr//'dti'/1/255//s,n,suport $
ENDALTER
@
30250; 70.5.2; Optimization/Design Sensitivity - REPCASE - SFM 3002 (DSADJ )
The message:
SYSTEM FATAL MESSAGE 3002 (DSADJ)
may occur when the adjoint method is invoked and there are REPCASE case
control commands.
Avoidance: The best avoidance is to remove the REPCASE commands. A more
complex and less robust avoidance is to force the direct sensitivity
method. This latter technique entails forcing the retention of an element
response.
@
30251; 70.5; Grid Point Force - GPFORCE, SUBCOM - GPFDR - SFM 4276, EC 7777
If a run contains SUBCOMs and a GPFORCE request, the run will fail on
Digital Alpha UNIX systems in the GPFDR module with the following message
and traceback:
SYSTEM FATAL ERROR 4276.tbk ERROR CODE 7777 PID= 0
USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH DIAG
44 INSERTED IN THE EXECUTIVE CONTROL SECTION.
Process Id = 9467
handler(0x11ffff138,0x0,0x120974d00,0x150fb59a4,0x1405fd6c0) L. 24
<<?? pc=0x12105c844>>(0x120974d00,0x150fb59a4,0x1405fd6c0,0x8,0xc00000007)
L. 24
gpfdr_(0x1409ac7b0,0x11ffff494,0x14057e9b0,0x14057e9b4,0x1409ac7b4) L.
 1057
xsem13_(0x12005727c,0x11ffff915,0x11ffff710,0x0,0x1) L. 94
mainal_(0x11ffff710,0x0,0x1,0x0,0x3ffc0082620) L. 349
<<?? pc=0x12005727c>>(0x140a40188,0x0,0x100000000,0x100000001,0x100000000)
L. 349
<<?? pc=0x120056fa0>>(0x1,0x11ffff918,0x0,0x11ffff939,0x11ffff94e) L. 349
10:57:44 NSEXIT: EXIT(0)
Avoidance: The only avoidance is to remove either the SUBCOM or GPFORCE
requests. Equivalent SUBCASES can be defined to replace the SUBCOMs.
@
30252; 70.5; Heat Transfer - RADMTX - Duplicate Continuation Entries in Punch
File
If a heat transfer analysis contains multiple cavities and each cavity
exceeds 100,000 elements the duplicate continuation entries for RADMTX
may be output to the punch file.
This problem only occurs when each cavity exceeds 100,000 elements.
If each cavity has more than 100,000 elements,then you will get the
duplicate continuation entries on the RADMTX bulk data entry. This error
will lead to:
*** USER FATAL MESSAGE 4392. CONTINUATION CARD ERRORS. EXPLANATIONS FOLLOW
LIST OF CARDS IN ERROR
when the punch file is used.
Avoidance: Manually edit the punch file. The continuation fields may be
left blank as long they are in order.

@
30254; 70.5; Complex Eigenvalue Analysis - CEAD - SFM 4276, EC 7777
If a complex eigenvalue analysis is missing a cmethod entry in the Case
Control the job may end with a traceback in the CEAD module.
Offset 0x00000020 in procedure raise
Offset 0x00004110 in procedure bioxxx_routine
Offset 0x00000134 in procedure bioxxx_
Offset 0x00000240 in procedure nsexit_
Offset 0x0000019c in procedure pexit_
Offset 0x00004048 in procedure gperr_
Offset 0x000002c0 in procedure mesage_
Offset 0x000010d8 in procedure cead_
Offset 0x00000094 in procedure xsem11_
Offset 0x000009b4 in procedure mainal
--- End of call chain ---
Avoidance: Add a cmethod entry in the Case Control.
@
30259; 70.5; Heat Transfer - TETRA, PENTA, CHBDYG - Poor Answers
The ten noded TETRA element and the 15 noded PENTA element produce poor
answers when the convection coefficient is large relative to the thermal
conductivity of the material. For example , when h*delta(x)/k is larger
than 10, where h is the convection coefficient, delta(x) is the element
 length and, the k is the thermal conductivity of the material poor answers
result.

Avoidance: Use other element types such as the 4 noded TETRA which uses
CHBDYG,AREA3 for convection.

@
30273; ; DMAP - CALL statement - UFM 72, Poor message
If a value or variable is not specified for all parameters on a CALL
statement then UFM 72 is issued with
misleading information. For example,
83 CALL FRRD2X KDDZ,BDDZ,MDDZ,,PDFZ,FOL,CASES,
EDT,SILD,USETD,/

UDVF1,FOLI1/ENFORCE///////FRRDSET//FIRSTBD0 $
0*** USER FATAL MESSAGE 72 (XGPIDG )
PARAMETER SYNTAX ERROR IN ABOVE STATEMENT. PARAMETER IS ENFORCE ---
ERROR CODE = 2
Avoidance: Remove the unspecified parameters or specify dummy variables
or values for the parameters.
@
30293; 2001; Optimization/Design Sensitivity - DEQATN - Undocumented Limit
The intrinsic functions MIN and MAX for the DEQATN entry have an open
ended list of arguments. There is an undocumented limitation that the
number of arguments for these functions has to be less than 100. This
problem is considered fixed in Version 2001 by raising the limit to 1000.
Avoidance: If it is desired to take the max (or min) of more than 100
arguments, several intermediate functions can be defined that find the max
of a number of arguments less than 100 and then the max of the
intermediate functions will result in the overall maximum.
@
30299; 70.5; Optimization/Design Sensitivity - BCDR, Superelements - SFM 3002
If all the design variables and responses are specified in upstream
superelements then the following messages may be issued in subDMAP PSLGDV
and module BCDR:
*** SYSTEM FATAL MESSAGE 3002 (BCDR )
EOF ENCOUNTERED WHILE READING DATA BLOCK (NONE) (FILE 101) IN
SUBROUTINE BCDR.
Avoidance: There are two possible avoidances:
1. Specify at least one design variable or response in the residual
structure.
2. Insert the following DMAP alter:
compile desopt
alter 'call descon'
seid=0 $
peid=0 $
equivx cases/casdsn/-1 $
 endalter
@
30322; 70.0.5; Heat Transfer Analysis - TETRA - Wrong Answers
In MSC/NASTRAN Version 70, a temperature look-up for a temperature
dependent material property for the TETRA element only occurs at
temperature=0.0. This causes wrong answers.
Avoidance: Use other element types such as HEXA or PENTA elements when
there are temperature dependent material properties in the problem.
@
30325; ; DMAP - Reserved Names
If a parameter or datablock on a type statement uses a name that has
"cond" as a prefix, for example, condnosq, then the results using these
parameters or datablocks in subsequent operations are unpredictable. For
example, in the following dmap operation,
type parm,,rs,n,condnosq $
condnosq = 1.0 $
message //condnosq $
does not produce a value of 1.0 for the condnosq parameter. The reason is
that "cond" is a dmap exec statement for branching operations.
Avoidance: Do not define a datablock, parameter, or label using "cond" as
a prefix for the name.

@
30336; ; DMAP - Building a New Solution Sequence - UIM 1179 (RESDBS), Poo
r Message
An attempt to LINK a "called SUBDMAP" (one with input and output arguments
on the SUBDMAP line) will result in the following error message, which
does not explain the problem:
*** USER INFORMATION MESSAGE 1179 (RESDBS)
A DB-DIRECTORY IMPROPERLY CLOSED IN THE PREVIOUS RUN IS BEING
RESTORED BY THIS RUN.
MAIN: "Access violation" (C0000005) exception encountered.
MAIN: For reference, A(MAINAL) = 408540, A(/SYSTEM/) = 14ef640
MAIN: Exception occurred at address 89050D.
MAIN: Program attempted to access data at location 19970738.
MAIN: Context Flags 0001001F
Avoidance: Do not LINK a "called SUBDMAP" only MAIN SUBDMAPs with no
arguments in the SUBDMAP line are to be linked.
@
30367; 70.5; Dynamic Analysis - TEMP(LOAD) - Incorrect Stresses
It is an undocumented limitation that thermal loads cause incorrect
stresses in dynamic analysis. This limitation will be documented under
TEMP(LOAD) case control entry in the Version 70.5 Quick Reference Guide.

@
30414; 70.7; Miscellaneous - SET - SFM 3008
 Long SET case control entries may cause NASTRAN fail with the message
SYSTEM FATAL MESSAGE 3008. The following SET entry resulted in this
fatal:
SET 1 = 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21
Avoidance: The above SET entry may be broken up as follows:
SET 1 = 1,2,3,4,5,6,7,8,9,
10,11,12,13,14,15,16,17,18,19,
20,21
@
30419; 70.5; Documentation - QRG - Description of Large Field Format Continua
tion Entries
In the discussion in the QRG on the "Format of Bulk Data Entries" there is
no mention made of how to continue a large field format Bulk Data Entry.
There is an example of a large field format entry, which shows the correct
use of the "*" in column 1 that marks field 1B as a continuation mnemonic
of the parent entry begun in field 1A. However, in the "Continuations"
discussion, there is no mention made of how to continue with child entries
that are also of the large field format. These types of continuations
require an "*" in column 1 of field 1B along with the continuation
mnemonic of the first half of the entry.
Additionally, it is stated that continuation fields can be auto-generated
(the recommended practice) within MSC/NASTRAN, provided that those
continuations immediately follow the parent and fields 1 and 10 of the
continuation lines are left blank. Under this usage, blank fields mean
small field format and errors can result from the first half of the entry
being in large field format and the second half being in small field
format.
In order to use auto-generated continuation fields on large field format
entries, field 6 of the first line, as indicated in the large field format
description must be blank and field 1B of the second line must contain an
"*" in column 1. All additional continuations using large field format
data must also conform to the above or they will be incorrectly classified
as small field data and processed as such.

@
30420; 70.5; Nonlinear - EIGB - SFM 6707, SFM 3007, Poor Message
In a nonlinear buckling analysis, if the method on the EIGB entry is SINV
or INV and no eigenvalues are found in the range, then the following
misleading messages will be printed in the .F06:
*** SYSTEM WARNING MESSAGE 3030 (OFP) THE OFP MODULE CANNOT PRINT THE
DATABLOCK LAMA USER ACTION: USE THE MATPRN OR TABPT MODULE. PROGRAMMER
INFORMATION: OFP LOC CODE = 40 FILE = 102
*** SYSTEM FATAL MESSAGE 6707 (PARAML) THE PARAML MODULE WITH P1=DTI
DETECTS A STRING RECORD. *** SYSTEM FATAL MESSAGE 3007 (PARAML) ILLEGAL
INPUT TO SUBROUTINE PARAML
or
 0*** SYSTEM FATAL MESSAGE 4540,
INPUT ESTNL DATA BLOCK IS NOT NONLINEAR FORMAT.
*** SYSTEM FATAL MESSAGE 3007 (NLITER)
ILLEGAL INPUT TO SUBROUTINE NLINIT

Avoidance: Choose a more appropriate range on the EIGB entry.

@
30425; ; Miscellaneous - (PUNCH) - Segmentation Fault in OFP Source OFPPU
N
Large punch jobs on the HP and SUN may fail with the following message:
Segmentation fault in OFP source OFPPUN
SIGSEGV in ofppun.f
and in the log file:
11:21:41 Beginning Analysis
Process Id = 23091
Process 23091 received signal 8
Process traceback ...
( 0) 0x0009700c sysrcv + 0x320 [/apps/msc70/msc70/hpux/analysis.pa2_0]
( 1) 0x00f83310 _sigreturn [/apps/msc70/msc70/hpux/analysis.pa2_0]
( 2) 0x00f48644 __convert_and_assign_data + 0x1464
[/apps/msc70/msc70/hpux/analysis.pa2_0]
( 3) 0x00f4a368 __cnv_fc + 0xb0
[/apps/msc70/msc70/hpux/analysis.pa2_0]
( 4) 0x00f335bc __a_c_t + 0xed4
[/apps/msc70/msc70/hpux/analysis.pa2_0]
( 5) 0x00f33ecc __a_c_t + 0x17e4
[/apps/msc70/msc70/hpux/analysis.pa2_0]
( 6) 0x00f342d8 __io_tn1 + 0x118
[/apps/msc70/msc70/hpux/analysis.pa2_0]
( 7) 0x00f3cb74 __F90_DO_IO_ITEM + 0xfc
[/apps/msc70/msc70/hpux/analysis.pa2_0]
( 8) 0x00f2fb58 FTN_DO_R4IOA + 0x30
[/apps/msc70/msc70/hpux/analysis.pa2_0]
( 9) 0x00418098 ofppun_ + 0x4a8
[/apps/msc70/msc70/hpux/analysis.pa2_0]
(10) 0x0016c4b4 ofp_ + 0x2f38 [/apps/msc70/msc70/hpux/analysis.pa2_0]
(11) 0x00093b88 xsem14_ + 0x294
[/apps/msc70/msc70/hpux/analysis.pa2_0]
(12) 0x0008ec44 mainal_ + 0x99c
[/apps/msc70/msc70/hpux/analysis.pa2_0]
Avoidance: Other than not requesting punch output there is no avoidance.
@
30453; 2001; Documentation - ALTER - String Based Alter, UFM 802
If a string-based DMAP alter uses the ! in the expression (find occurrence
at the beginning of a line), it is possible that MSC/NASTRAN will fail
with USER FATAL MESSAGE 802 stating that no match was found, even if there
should be a match.
Avoidance: If you encounter UFM 802, check to see that the desired line
is not preceded by a comment in the DMAP. The search logic using the "!"
assumes the line begins on the preceding comment. In order to avoid this,
remove the ! from the text string (be careful to verify that the DMAP
alter is inserted at the correct location, as this will then search for
 any occurrence of the text on a line, rather than only searching for lines
which begin with the text). If necessary a more unique search string will
have to be selected.
@
30459; 70.5; Estimate - Leading 0's in Integers - Poor Results
If an integer has leading 0's (e.g., 05 instead of 5) on a NASTRAN bulk
data entry, MSC/ESTIMATE may not interpret the data correctly.
Avoidance: Remove any leading 0's from the bulk data of a file to be
processed by ESTIMATE.
@
30462; 70.5; Estimate - Excessive Errors - Run Aborted
If 'memory=estimate' or 'buffsize=estimate' is specified on the nastran
command line or in an RC file, messages generated by the ESTIMATE utility
may overflow a buffer within the nastran command. When this overflow
occurs, the following message will appear:
*** SYSTEM WARNING MESSAGE (nastran.estimate_job_requirements)
a.estimate=<install-dir>/msc705/<arch>/estimate (program default)
ESTIMATE executable filename. This keyword may only be specified on
the
command line or in the command initialization file (see <0.ini>).
It appears that ESTIMATE has failed. Ensure that the program is
properly
installed. Assuming default value for buffsize.
Avoidance:
Either of the following avoidances will work:
1. Fix the errors in the input file and rerun with estimate.
2. Run estimate by itself. Use the suggested parameters to run
MSC/NASTRAN.
@
30468; 70.5; Miscellaneous - VECPLOT - SFM 3062
If a very large memory is requested in a run, then the job may fail in
vecplot with the following error messages:
0*** SYSTEM FATAL MESSAGE 3062 MODULE = VECPLOT
FAILURE OCCURRED IN UTILITY GETGPT CALLED IN SUBROUTINE VECPL1 .
PREXYZ FLAG ARGUMENT WAS SET WITH AN INCORRECT VALUE TO GET THE
REQUESTED DATA RETRIEVAL.
In this case, a memory of 400mw was requested.
Avoidance: Reduce the amount of memory request.
@
30496; ; Restarts - Delete Entry - Various User Fatal Messages
When a restart delete entry, using the small field format, is incorrectly
shifted a column to the left, the beginning and end values are misread
without a fatal message.
For instance the following delete bulk data entry:
/ 65266 65267
 is read as:
/ 5266 5267
Avoidance: Use free field, comma separated input for this entry.
@
30507; 70.5; Optimization/Design Sensitivity - SDR2 - SFM 3062
A shape optimization task may fail with a system fail message,
0*** SYSTEM FATAL MESSAGE 3062 MODULE = SDR2
FAILURE OCCURRED IN UTILITY MAPXTI CALLED IN SUBROUTINE SHLNRM .
REQUESTED EXTERNAL ID(S) NOT FOUND.
The failure may occur when the following conditions are met:
1. A shape optimization job;
2. Shell normal is invoked;
3. The identification id for grid points is larger than the total number
of grids in the model. Say the model has 10 grids but the ten grids are
identified by 100001 to 100010.
The possible avodance:

1. Always start grid ids from 1;

2. Set SNORM=0.0.
Notice that if shell normal can not be ignored, the avoidance # 1 should
be used.
@
30515; 70.5; Restarts - EPOINT - Inefficient Restart
If a restart is made from a solution 103 normal models run and epoints are
added, the modes will be recalculated even though they are unaffected by
the epoints.
Avoidance: Use the following alter, it removes the check on DYNAMIC(s)
and the restart skips the READ or REIGL module.
THIS AVOIDANCE PRODUCES INCORRECT RESULTS!
compile ifpl
alter 158,158
compile phase0
alter 590,590
endalter
@
30522; 70.5; Optimization/Design Sensitivity - PCOMP - SFM 5628 (DSAD12)
In design optimization, if there are composite responses (DRESP1 =
CSTRESS, CSTRAIN or CFAILURE) and force responses (DRESP1=FORCE), a USER
FATAL MESSAGE 5628 as shown below will be encountered unless the highest
element ID has a composite response.
0*** USER FATAL MESSAGE 5628 (DSAD12)
DRESP1 ID = 5363003 IS FOR RTYPE = FORCE AND ELEMENT ID =
5682.
THE RESPONSE COULD NOT BE FOUND, EITHER BECAUSE THE ELEMENT DOES NOT
EXIST OR
BECAUSE THE RESPONSE TYPE IS NOT AVAILABLE FOR THE ASSOCIATED ELEMENT
 TYPE.
0*** USER FATAL MESSAGE 5628 (DSAD12)
DRESP1 ID = 5363004 IS FOR RTYPE = FORCE AND ELEMENT ID =
5682.
THE RESPONSE COULD NOT BE FOUND, EITHER BECAUSE THE ELEMENT DOES NOT
EXIST OR
BECAUSE THE RESPONSE TYPE IS NOT AVAILABLE FOR THE ASSOCIATED ELEMENT
TYPE.
0FATAL ERROR
AVOIDANCE: It is not sufficient to make sure that the highest constrained
element id is a composite element since the design sensitivity analysis
spawns element ID's that are out of the user's control and will produce a
gread error in DSAL. Instead,
the only known avoidance is to remove the TYPE=FORCE dresp1 entries. They
could be replaced with equivalent STRAIN or STRESS
dresp1 entries.

@
30542; 70.5; Response Spectra Analysis - SUBCASE - Missing Output
When using Response Spectrum ananlysis in SOL 103, if you have multiple
SUBCASEs (solutions) and there are diffenrent output requests in various
SUBCASEs, only the requests which match the first SUBCASE will be honored.
Avoidance: Insert the following DMAP alter:
compile postreig
alter 'sdr2 ',''
SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,,,,BGPDTS,OLB,PHQG,
PHIGH,EST,XYCDBDR,,,,GPSNTS,,,/
,IQG1,IPHIG1,IES1,IEF1,/
'mmREIG' $
endalter
@
30550; 70.5; Documentation - QRG, INIT,SCRATCH - Incorrect Default
The default value for the SCRATCH dbset is stated as 175,000 gino blocks
in the Quick Reference Guide. The correct default value
is 250,000 gino blocks.
@
30560; 70.5; Restarts - scr=mini, SUPORT command - UWM 4410, Missing output
If SCR=MINI is specified on the command line and the SUPORT command is
specified in the Case Control, then subsequent data recovery restarts will
terminate with UWM 4410 and no data recovery output.
Avoidance: Insert the following DMAP alter:
compile sestatic
alter 'call setq'
PARAML CASES//'DTI'/1/255//S,N,SUPORT $
endalter
@
30563; 70; Transient Analysis - OTIME, op2, f06 - wrong answers
In Version 69.1, the OTIME command is ignored for element-type output
(e.g., forces, stresses). Furthermore, the time labels are in the correct
 sequencial order, but the label values are incorrect. The OTIME command
functions properly for grid-type output (e.g., disp, velo, acc). This
problem has been corrected in Version 70.
Avoidance: Either remove the OTIME command or upgrade to Version 70.
@
30575; GL; Optimization/Design Sensitivity - Limit on the Number of Design
Variables
There is a limit imposed by the optimization algorithm used for design
optimization that the number of design variables (NDV).
squared must be less than the number of machine precision words available.
For example, if a job is run with 40 million single precision words, the
number of design variables must be less than the sqrt(40,000,000/2) = 4472
on a short word machine and 6324 on a long word machine.
@
30578; ; Nonlinear Analysis - PARAM,LGDISP, PLOADX1, LOAD - Wrong Answers
If PLOADX1 (pressure loads for axisymmetric elements) is referenced by a
bulk data LOAD record, follower forces (in nonlinear solution sequences)
do not work. Pressure loads are computed based upon the initial geometry.

Avoidance: An avoidance is to write PLOADX1 records for every load

magnitude needed, and not use the LOAD records to combine loads.
@
30585; 70.5; Installation/Security - Unable to Obtain Licenses for MSC/NASTRA
N
There is an error in the FLEXlm libraries provided by Globertrotter that
result in the libraries ALWAYS examining the LM_LICENSE_FILE environment
variable even though a valid license file or server has been explicitly
named in the "authorize" keyword. As a result of this, the following
limitation is set:
If the LM_LICENSE_FILE environment variable is set, it shall point to
a valid license file.
This error is most easily seen by setting the "authinfo" value to 9 and
examining the LOG file output. If a license file not named by the
"authorize" keyword is present, it was obtained from the LM_LICENSE_FILE
environment variable.
Avoidance: Delete the LM_LICENSE_FILE environment variable if it does not
point to a valid FLEXlm license file.
@
30590; 70.5; Nonlinear Static Analysis - Superelements, Restarts - SFM 4540
When downstream superelements are processed in a SOL 106 coldstart by
using SEALL=n, where n is the superelement number, NASTRAN incorrectly
issues the following fatal message:
SYSTEM FATAL MESSAGE 4540,
INPUT ESTNL DATA BLOCK IS NOT NONLINEAR FORMAT.
Avoidance: If you are using partitioned superelements and you are going
to restart to process the residual, ignore the fatal message and restart
anyway. The fatal message is benign and restarts are possible.
@
30617; 70.5; Documentation - QRG, Force(corner), strain(corner) - Incorrect d
efault
The Quick Reference Guide incorrectly lists cubic as being the same as the
corner output for the force and strain output. The correct default corner
option for the force and strain corner output is BILIN, similar to the
stress output.
@
30631; ; Documentation - Reference Manual, GPSTRESS - Strain Not Supporte
d
In the second paragraph of Section 10.1 of the MSC/NASTRAN Reference
Manual it states "The calculation of strains may be sustituted for
stresses using the STRAIN command in place of the STRESS command in the
Case Control Section" inside the same paragraph that describes the
GPSTRESS features. This may lead users to incorrectly conclude that grid
point strain can be obtained by substituting the STRESS Case Control
Command with the STRAIN command. Grid point strain output is not
supported.
@
30638; 70.5; Grid Point Force Balance - GPFORCE, PLOAD - Wrong Answers
Grid point force balance output of element force contributions are not
correct when pressure loads are applied. For example, PLOAD1 on BAR, BEAM
and BEND elements will produce different element forces with the ELFORCE
data recovery than from that printed in the grid point force balance
output. The same is true of the plate elements under pressure loads.

Avoidance: There is no avoidance, this is currently a limitation of the

GPFDR module. The GPFDR module has as input the discrete load vector to
perform the data recovery operations. The discrete load vector does not
contain any of the distributed load effects and will calculate element
force contributions based on the discrete loads not the distributed loads.
The limitation has not been removed but is now documented under the
GPFORCE case control command.
@
30639; 70.5; Optimization/Design Sensitivity - PLOAD1/BEAM - Wrong Answers
A design optimization job may report incorrect element response value
(Stress/Strain/Force) for CBAR, CBEAM and CBEND elements subjected to
pressure loads because SOL200 does not account for distributed forces of
the PLOAD1 entry while recovering those element response quantities.
Avoidance: There is no avoidance.

@
30645; 70.5; DMAP - User Supplied subdmaps may Discard datablocks on type Sta
tements
If a user subdmap is of the form:
subdmap mine in1/out1
type db sils
mod1 sils/out1 etc.
etc.
. . .
return
end $
 and sils is a datablock in the nddl, which means it is stored in the
database, then assuming qualifiers are set correctly, module mod1 will
find sils and use it appropriately. However, at the end of the execution
of the subdmap the LTU logic will automatically delete the datablock.
This will not be discovered until the datablock is needed by a downstream
subdmap, when it will no longer be in the database. Consequences include
wrong answers or mysterious fatal error crashes such as "SILS NOT
AVAILABLE" for modules that allow purged inputs. These errors occur even
though SILS was in the database in prior subdmaps.
Avoidance: Do not write the type statement, but instead input the
datablock in a calling argument:
subdmap mine in1,sils,out1 $ etc.
@
30703; GL; Postprocessing - XDB Limitation - Maximum Number of Grids
There is a limit on the number of GRID points which can be processed by
the DBC module for the ".xdb" file. The limit is dependent on many
factors, one of the most significant factors is the solution sequence.
For dynamic and nonlinear solution sequences the limit is 214748 GRID
points. If more GRIDs than this exist in the model, some of the desired
data may not be processed and placed in the ".xdb" file.
Although there is still a limit on the total amount of data DBC can
process, Version 70.5 contains changes that take into account the actual
number of timestep or load steps actually solved. Thus, the program can
more accurately predict the amount of data that it must process. The
maximum number of grids is now determined dynamically and this change
effectively greatly increases the maximum number of grids allowed in most
dynamic and nonlinear models.
Avoidance: If your model has more than 214748 GRID points in it, then
divide it into superelements, each with less than 214748 GRID points in
it. If data recovery is only required for a few grid points, then
numbering these points first and referencing a SET with these points will
also avoid the error.
@
30704; 70.5; Nonlinear Heat Transfer - CHBDYG - UFM 2157
The backside on an ambient plate element cannot be used for thermal
radiation. If attempted the follow message results:

USER FATAL MESSAGE 2157, MATERIAL ID = 0 FOR ELEMENT ID = 1

DOES NOT APPEAR ON ANY RADM RADIATION MATERIAL DATA CARD
FATAL ERROR
Avoidance: Add a front face RADM entry.

@
30712; 70.5; Optimization/Design Sensitivity - PARAM,INREL, ADD - UFM 5423
The USER FATAL MESSAGE below may occur when performing an inertial relief
sensitivity analysis with multiple subcases and one or more subcases does
not have any retained responses. The message is:
0*** USER FATAL MESSAGE 5423---ATTEMPT TO ADD INCOMPATIBLE MATRICES,
 (SADD5)
COLS ROWS FORM TYPE NZWD DENS
TRAILER FOR PL IS 6 57 2 2 50 3596
TRAILER FOR PLR6Y IS 12 57 2 2 64 5614
TRAILER FOR PLIPA IS 12 57 2 2 34 2281
FATAL ERROR
Avoidance: Insert the following DMAP alter:
compile desopt
alter 'call descon'
solapp='stnmbk' $
call dbstore caseds,casdsn,,,//0/0/' '/0 $
compile inersen
alter 'plbase','qrbase'
call dbfetch /caseds,casdsn,,,/0/0/0/0/0 $
type parm,,i,n,icase0=1,icase1=1 $
type parm,,logi,n,loopflg=true,sync=true $
matgen ,/one/1/1 $ 1x1 identity matrix
matgen ,/zero/7/1/1 $ 1x1 null matrix
file vc=append $
do while ( icase0>0 ) $
paraml caseds//'dti'/s,n,icase0/1//s,n,sid0 $
if ( icase0>0 ) then $
if ( sync and icase1>0 ) then $
paraml casdsn//'dti'/icase1/1//s,n,sid1 $
if ( icase1>0 ) icase1=icase1+1 $
endif $
if ( sid0=sid1 ) then $
append zero,/vc/2 $
sync=true $
else
append one,/vc/2 $
sync=false $
endif $
icase0=icase0+1 $
endif $
enddo $
trnsp vc/vr $
append vc,/vc6///6 $
trnsp vc6/vr6 $
$
DBVIEW CRXbase0 = CRX WHERE (DELTA=FALSE) $
partn crxbase0,vr,vr6/crxbase,,,/1 $
$
DBVIEW QRbase0 = QR WHERE (DELTA=FALSE) $
partn qrbase0,vr6,/qrbase,,,/1 $
$
DBVIEW PLBASE0 = PL WHERE (DELTA=FALSE) $ GET FOR LAST 6 COLUMNS
partn plbase0,vr6,/plbase,,,/1 $
endalter
@
30721; 70.5; Optimization/Design Sensitivity - MODTRK - Eigenvectors in Wrong
Order
Mode tracking in design optimization seeks to place modes back into their
original order if any mode switching has occurred as a result of changes
made to the design. The eigenvalue summary, however, will not be
correctly correlated with the eigenvector output if modes have switched,
and the parameter, NASPRT is set to a positive integer to request data
recovery for the intermediate design cycles. The values are correct but
the order does not agree with the eigenvalue summary table.
Avoidance: The following is a DMAP alter for Version 70 that correctly
reorders the eigenvector summary output as a result of mode tracking.
This alter has been built into Version 70.5 of MSC/NASTRAN along with some
additional module MODTRK enhancements.
$
$ v70 avoidance for csr 30721:
$
compile fea souin=mscsou list noref
alter 267,268
if (feaopt='analysis') then $
message //' dmap information message 9051 (fea)' $
message //' normal modes analysis initiated for analysis ='/
analysis/' and design cycle number ='/descycle $
endif $

alter 284
$ before computing modes, determine if mode tracking is on:
PARAML CASEM//'DTI'/1/275//S,N,MODETRAK $
alter 285,285
IF ( NOT NOPH2 ) CALL MODEFSRS,
MR,USET,DM,CASEM,DYNAMIC,MAA,KAA,GPLS,SILS,EED,
EQEXINS,VAFS,PG,GM,GOT,KFS,YS,LLL,KLL/
PHA,LAMA,PHFA,LAMAF,PHSA,LAMAS,PHSA0,LAMAS0/
' '/NORSET/NOLSET/ASING/NOASET/FS/NOFASET/
NOSASET/METH/METHF/NASOUT/NOQSET/NSUBM/modetrak $
alter 286,286 $ remove misplaced paraml
alter 297
$ if modes cannot be tracked output latest lama table, even
$ though unordered:
message //' modes could not be tracked for design cycle '/
descycle $
message //' though the eigenvalue summary table for the '/
'latest design follows, be aware that the mode '/
'order corresponds to the untracked state.' $
ofp lamas,/ $
alter 338
$ for mode tracking, output reordered lama table on every design
cycle:
if ( modetrak>0 ) ofp lama,/ $
alter 405,405
CALL MODERS ,,USET,DM,CASEM,DYNAMIC,KDAA,KAA,
GPLS,SILS,EED,EQEXINS,,/
BPHA,BLAMA/
NORSET/NOLSET/'BUCKLING'/ASING/
FALSE/FALSE/NASOUT/NOQSET/FALSE/NSUBB/modetrak $
endalter
 compile modefsrs souin=mscsou list noref
alter 1,1
SUBDMAP MODEFSRS MR,USET,DM,CASES,DYNAMICS,MAA,KAA,GPLS,SILS,EED,
EQEXINS,VAFS,PG,GM,GOT,KFS,YS,LLL,KLL/
PHA,LAMA,PHFA,LAMAF,PHSA,LAMAS,PHSA0,LAMAS0/
APP/NORSET/NOLSET/ASING/NOASET/FS/
NOFASET/NOSASET/METH/METHF/NASOUT/NOQSET/NSUB/modetrak $
alter 6
type parm,nddl,i,n, modetrak $
alter 37,37
CALL MODERS MR,USET,DM,CASES,DYNAMICS,MSAA,KSAA,GPLS,SILS,EED,
EQEXINS,,VAFS/
PHSA0,LAMAS0/
NOSRSET/NOLSET/'MODES'/ASING/FALSE/SECND/
NASOUT/NOQSET/FS/NSUB/modetrak $
alter 59,59
CALL MODERS MR,USET,DM,CASES,DYNAMICS,MFAA,KFAA,GPLS,SILS,EED,
EQEXINS,,VAFX/
PHFA,LAMAF/
NOFRSET/NOLSET/'MODES'/ASING/FALSE/TRUE/
NASOUT/NOQSET/FS/NSUB/modetrak $
endalter

compile moders souin=mscsou list noref

alter 1,1
SUBDMAP MODERS MR,USET,DM,CASES,DYNAMICS,MMAA,MKAA,GPLS,SILS,EED,
EQEXINS,VACOMPR,VAFS/
PHA,LAMA/
NORSET/NOLSET/READAPP/ASING/
CYCLIC/SECND/NASOUT/NOQSET/FS/NSUB/modetrak $
alter 7
type parm,nddl,i,n, modetrak $
alter 126,126
$ output eigenvalues only if mode tracking is not active.
$ otherwise, reserve output until after modes have been sorted.
if ( nasout and modetrak=0 ) ofp lama,oeigs/ $
endalter

@
30738; 2001; Frequency Response - QUAD4, sort2, corner, - SFM 3001, Possible
wrong answers
In a frequency response analysis, if SORT2 output is requested, then the
following error message may be issued in data recovery after printing the
element force output.
*** SYSTEM FATAL MESSAGE 3001 (DPD )
THE OUTPUT DATA BLOCK IN POSITION 11 DOES NOT EXIST.
 Furthermore, on the SGI (and possibly other machines), a sort2, corner
output request for quad4 elements may produce incorrect results that are
evident by the printing of very large element forces (e.g. E+20) or "Inf".

Avoidance: Use a sort1 output request with frequency response. If corner

output is not being requested, and if random response is not being
performed, then the following alter may avoid the problem in Version 70.

compile sedrcvr souin=mscsou nolist noref $ V70

alter 'dpd'(1,-2),'dpd'(1,7) $ Delete the offending DPD module and
associated modules.
endalter
@
30740; 70.5; Aeroelastic Analysis - AMP, CAERO1, CAERO2 - UFM 3005, Subroutin
e AMPC
The message:
USER FATAL MESSAGE 3005 (AMP )
ATTEMPT TO OPERATE ON THE SINGULAR MATRIX SCRATCH IN SUBROUTINE AMPC
FATAL ERROR
will occur in the AMP module if multiple group ids are used in defining
the aerodynamic model and the model contains both lifting surface entries
(CAERO1) and body entries (CAER02).
Avoidance: Place all the aerodynamic panels and bodies in the same
interference group.
@
30766; 70.5; MNE - Missing Collection on V70 MNE
If you choose an installation option that does not install the PRIMARY
(V70) documentation set onto the hard disk, you must do the following
after executing the installation script:
1. Change the current directory to <install_dir>/mne70, where
<install_dir> is the directory in which you installed the MNE:
cd <install_dir>/mne70
2. Remove the existing erroneous symbolic link to v70qrg:
rm v70qrg
3. Add the correct link for v70qrg:
ln -s /<CDROM>/doc/mne70/v70qrg
where <CDROM> is the CD-ROM mount point.
@
30775; 70.5; Loads - PLOADi - Incorrect Follower Forces
There is an error in the follower force stiffness matrix (see PARAM
FOLLOWK) due to PLOAD, PLOAD2 or PLOAD4, if the global displacement system
is not in the basic coordinate system. This error occurs if the first
grid point listed on the PLOADi has a non-basic coordinate system
reference in field seven of the corresponding GRID entry.
@
30776; GL; Element Strain Energy - ESE - Incorrect for Buckling
The element strain energy output for modes found in SOL 103 using STATSUB
is incorrect. It includes only the linear portion
of the ESE; the differential stiffness contribution is ignored.
 Avoidance: The following alter will calculate the strain energy
contribution of the differential stiffness in addition to the linear
portion:
$ the following dmap is for V70 SOL 103 with STATSUB
$ after the solution completes, there are 2 ESE outputs
$ 1st: unstiffened ESE due to KELM
$ 2nd: stiffened ESE due to KDELM
$ combine these to get the total modal ESE
$ this has not been tested with superelements.
COMPILE SEKDR,NOLIST,NOREF
ALTER 'EMG.*KDELM' $
CALL DBSTORE KDELM,KDDICT,,,//-99/SEID/'DBALL'/0 $
COMPILE SEDRCVR,NOLIST,NOREF
ALTER 'OFP.*ONRGY1' $ after unstiffened gpfdr
MESSAGE //'DIFFERENTIAL ESE CONTRIBUTION FOLLOWS:' $
TYPE PARM,,I,N,OKKDGET $
CALL DBFETCH /KDELM,KDDICT,,,/-99/SEID/0/0/S,OKKDGET $
$MATPRN KELMX,KDICTX,KDELM,KDDICT// $
GPFDR CASEDR,UG,KDELM,KDDICT,ECTS,EQEXINS,GPECT,PG1,QG,
BGPDTS,SILS,CSTMS,VELEM,PTELEM/
ONRGY1D,OGPFB1D/APP2/TINY//CYCFLG $
OFP ONRGY1D//S,N,CARDNO $
@
30782; ; Superelements/Buckling - Wrong Answers.
Superelements give wrong answers for buckling.
Avoidance: Do not use superelements.

@
30797; 70.5; Utilities - ESTIMATE - Poor Diagnostics
The ESTIMATE program is currently limited to estimating resources for
models composed of GRID and connectivity entries. If no GRIDS are
present, ESTIMATE will return a USER FATAL MESSAGE or may indicate NaNQ
for memory or disk estimates. If no recognized connectivity bulk data
entries are present, ESTIMATE will return DOF=-2**(32-1).

Avoidance: Do not use the ESTIMATE program for models as described above.
Do not use either the "memory=estimate" or "buffsize=estimate" keywords on
a nastran command for models as described above.
@
30805; ; Nonlinear Analysis - Multiple Subcases - Error in Bisection logi
c, Infinite Loop
In a nonlinear analysis, in some instances, the load factor tolerences may
be in error. When this occurs the value of the load factor exceeds 1.0 and
the solution may enter an infinite loop or produce some other error later
on.
The problem is identified when one of the load factors > 1.0
Avoidance: On the NLPARM entry increase the ninc field by 1 and restart
into the offending subcase. This will change the tolerance
@
30810; 2001; Optimization/Design Sensitivity - SEP1A - Infinite CPU Loop
When the analytic boundary shape option is used for a shape optimization
task and a local coordinate system is defined for the auxiliary boundary
model, the job may fall in an infinite loop in the SEP1A module.
Avoidance: Do not define local coordinate systems in the auxiliary
boundary model.
@
30852; GL; Postprocessing - PARAM,POST,-4 - UIM 1207, UIM 1206
If PARAM,POST,-4 is used and FORMATTED output is requested in the OUTPUT2
file, error messages 1207 and 1206 will occur indicating that several
datablocks have not been written to the file. An example of these
message are:
USER INFORMATION MESSAGE 1207 (OUTPN2)
THE DATABLOCK KDICT /KDICT DEFINED AS NDDL TYPE KDIC IS NOT
SUPPORTED BY NEUTRAL FILE TRANSLATION.
THIS DATABLOCK WILL NOT BE TRANSLATED TO NEUTRAL FORMAT
0*** USER INFORMATION MESSAGE 1206 (OUTPN2)
THE DATABLOCK OPHIG /OPHIG IN AN OUTPUT2 REQUEST IS NOT DEFINED
IN THE MSC/NASTRAN DATA DEFINITION LANGUAGE (ND
This is due to the fact that these datablocks are not described in the
NDDL.
Avoidance: Use binary OUTPUT2 if PARAM,POST is set to -4.
@
30903; GL; Superelements - PARAM,AUTOSPC - Incorrect Constraint
If two or more superelements which have no internal constraints applied,
meet at a single GRID point, and no elements which are not in those
superelements connect to that GRID point, AUTOSPC may incorrectly
constrain that point. The reason for this is that the reduced stiffness
matrix at a single point for a determinate structure with no constraints
is a null matrix. This limitation is unlikely to affect real world models
since it is unlikely that the boundary between superelements will be
comprised of a single point.
Avoidance:
1. Put the DOFs you want as dynamic analysis variables on spcoff bulk
data entries.
2. Put in small springs to ground, large enough to turn off autospc but
not large enough to change dynamic analysis results. This is the most
practical method for superelement analysis.
@
30947; 70.5; Static Analysis - MATTi - SDR2, biomsg error 7777
For a large job on the hp (hpux), if temperature-dependent material
property exists in the model, then the job may fail in SDR2 with following
error messages:
***** RECEIVED SIGNAL 8, ABORTING ...
BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX) =
 0
Avoidance: Remove the MATTi entry and define the desired properties at
the appropriate temperature on the MATi entries since for a linear
analysis, the temperature-dependent lookup is only performed once in the
beginning of the run.
@
30959; ; Nonlinear Analysis - CGAP, PGAP - Possible Wrong Answers when U0
is Negative
Use of a gap element (tmax>=0.) with a negative value of U0 when GA and GB
are NOT coincident may lead to wrong answers. Consider two grids on the
x-axis with GA to the left of GB and the CID field blank and U0 specified
to a negative value. Then the gap will always try to push GA to the left
and GB to the right regardless of whether GA comes first or GB comes first
in field 4 on the CGAP entry. Thus, a press fit with the inner radius of
the outer disk at GA and the outer radius of the inner disk at GB would
always pull in the wrong direction when UO is negative.
Avoidance: To make the gap sensitive to the order of GA and GB with U0
negative, the CID entry of field 9 must be specified.
@
30969; 70.7; GPFORCE - SET, GPFORCE - Poor Message, SFM 4276, EC 3061, SWM 30
02
If a GPFORCE request is made for a SET that only contains grid points
attached to rigid elements then the deck will fail in the GPFDR module.
The exact message encountered may vary. The following messages have been
encountered:
SYSTEM FATAL MESSAGE 4276, ERROR CODE 3061
or
*** USER WARNING MESSAGE 2354 (GPFDR)
MODULE IS UNABLE TO CONTINUE AND HAS BEEN TERMINATED DUE TO ERROR
MESSAGE PRINTED ABOVE OR BELOW THIS MESSAGE.
THIS ERROR OCCURRED IN CODE WHERE THE VARIABLE -NERROR- WAS SET =
33
*** SYSTEM WARNING MESSAGE 3002 (GPFDR)
EOF ENCOUNTERED WHILE READING DATA BLOCK SCRATCH (FILE 303) IN
SUBROUTINE GPFDFA
Avoidance: GPFDR does not support rigid elements, request MPCFORCE output
instead.
@
30976; 70.5.1; Online Documentation - Installation
The UNIX installation aborts if you choose an installation option that
installs the V69.X documentation set on the hard disk. The old V69
documentation is duplicated on this CD only for those customers still
running MSC/NASTRAN V69 or V69.2. If you use MSC/NASTRAN V70, we recommend
not installing V69.X documentation onto your hard disk, since all the
documents applicable to V70 are already contained in the Standard
installation. If you occasionally need to refer to V69 documentation, it
is best to access it directly from the CD.
If you frequently need V69-specific documentation and you prefer to access
it from the hard disk instead of the CD, please follow the Installation
 Instructions on the pink sheet included with the delivery package.
@
30981; 70.5; Online Encyclopedia - INSTALL.txt on V70 Online Encyclopedia is
Incorrect
The INSTALL.txt file on the MSC/NASTRAN Version 70 Encyclopedia (MNE) is
incorrect. It is the file for MSC/NASTRAN Version 70 not the MNE.
Avoidance: Either view the INSTALL.htm file using a web browser or
request a printed copy of the file from MSC.
@
30987; GL; Eigenvalue Analysis - EIGRL, GMINTC, GMINTS - Possible Missing R
oots, SFM 5407
The Lanczos method of eigenvalue extraction does not work well when
interface elements (GMINTC or GMINTS) are present, roots may be missing in
the range. This may lead to the message:
SYSTEM FATAL MESSAGE 5407 (LNCS1D)
INERTIA (STURM SEQUENCE) COUNT DISAGREES WITH THE NUMBER OF MODES ACTUALLY
COMPUTED IN A (PARTIAL) INTERVAL

Avoidance: If interface elements are present, the SINV method is

recommended.
@
30995; 70.5; Acoustic Analysis - GE on MAT10, PARAM,ACSYM,YES - Wrong Answers
If GE is specified on the MAT10 entry and PARAM,ACSYM is YES (default)
then the answers will be wrong.
Avoidance: Insert the following DMAP alter:
compile gma
alter 'if ( app=\'tranresp\''(2),''
alter 'if ( app=\'tranresp\''(3,-1)
if ( app='tranresp' or soltyp='direct' or modacc>=0 ) then $
alter 'k4hh',''
if ( nok4dd>0 ) then $
if ( acsyml ) then $
add5 ifdd,,,,/mifdd/-1. $
MERGE isdd,,,mifdd,VAFSE,/idd $
mpyad k4dd,idd,/k4ddx $
SMPYAD PHDH,K4DDx,PHDH,,,/K4HH/3////1 $
else $
SMPYAD PHDH,K4DD,PHDH,,,/K4HH/3////1 $
endif $
endif $ nok4dd>0
endalter
@
31000; GL; Eigenvalue Analysis - READ - Initial Vectors Ignored
The READ module has provisions for initial vectors as input for the
Lanczos method. These vectors are ignored in the READ module.
Avoidance: Use the REIGL module.
@
31003; 70.5; DMAP - MATMOD Option 12 - Wrong Answers
 The form of the column vector output in MATMOD option 12 is 7 (labeled
"ROW VECTOR") and should be 2 (labeled "ROW RECTANG.").
Avoidance: Run MODTRL on the output; for example,
MATMOD I1,I2,I3,,,/O1,O2/12 $
MODTRL O2// / /2 $
@
31027; 70.5; Miscellaneous - PARAM,CHECKOUT and PARAM,PRTPG, SSG1 - SFM 3003
If PARAM,CHECKOUT,YES and PARAM,PRTPG,YES are specified, and thermal loads
are also applied, then SYSTEM FATAL MESSAGE 3003 will terminate the run in
the SSG1 module in the PHASE0 subDMAP.
Avoidance: Insert the following DMAP alter:
compile phase0
alter 'ssg1 ',''
SSG1 SLT,BGPDTS,CSTMS,SILS,EST,MPTS,ETT,EDT,MJJX,CASESX,
DIT,,DEQATN,DEQIND,GPSnts,cstm0,scstm/
PJX,PTELEM,NEWSLT/
LUSETS/1//'ALL '/ALTSHAPE/TABS/seid $
endalter
@
31037; 70.5; Superelements - SEMA, SEBNDRY, SECONCT, SEEXCLD - UFM 4192, UFM
7341, UFM 6787
Two errors relating to checking the input on the SEBNDRY, SECONCT and
SEEXCLD entries have been discovered.
1. The SEIDA field is not verified to be restricted to partitioned
superelements. The failure to check this may cause UFM 4192 to occur in
SEMA. This message may be followed by a SFM 3062 concerning the MAPBTI
utility. The true message is UFM 7341.
2. Conflicts in grid usage between the SECONCT and SEEXCLD is not properly
verified and should have result in UFM 6787.
Avoidance: Follow the rules for the entries delineated in the Quick
Reference Guide.
@
31051; 70.5.2; P-elements - GMSURF, GMCURV, GP0 - Segmentation Violation
If you use an EQUATION or Aries BSP for geometry evaluation, in other
words, if "MSCGRP1", or "MSCABS" is specified for the 'group name' field
on the GMCURV or GMSURF entries the job may fail with the following
traceback in GP0:

Trap encountered:
SIGSEGV - Segmentation violation
Traceback:
Offset 0x000003d4 in procedure get_edge_info
Offset 0x000000e8 in procedure abspnt
Offset 0x000000a8 in procedure ngpont_
Offset 0x00000b80 in procedure gp0dx_
Offset 0x00000680 in procedure gp0dc_
 Offset 0x000017c8 in procedure gp0da_
Offset 0x00000cf8 in procedure gp0_
Offset 0x000004b0 in procedure xsem02_
Offset 0x000009a8 in procedure mainal

--- End of call chain ---

Avoidance: There is no avoidance.
@
31054; 2001; Documentation - SEBULK, SECONCT, SEEXCLD, SELABLE, SELOC, SENQSE
T - UWM 9013
In the interest of upward compatibility, several of the bulk data entries
implemented in Version 69 will only work if PART superelements (BEGIN
SUPER) exist. These entries are: SEBULK, SECONCT, SEEXCLD, SELABEL, SELOC,
SEMPLN, SENQSET, SETREE.
Avoidance: Do not use these entries unless you have defined PART
superelements.
@
31067; 70.5.1; Miscellaneous - SFM 4276 EC 7777, signal 10
It is possible, but not probable, that a file run on the HP9000 using the
analysis.pa2_0 executable may fail with SFM 4276, EC 777, signal 10. This
is due to a kernel error on the PA2 machines.
Avoidance: There are several possible avoidances:
1) run using a different BUFFSIZE or memory
2) try using a different SMEM
3) run using the older 'analysis' executable (this will avoid the error,
but is slower)
4) change bpool
@
31069; ; Optimization/Design Sensitivity - DVPREL1, DVPREL2, EMG - UFM 20
42
In an optimization analysis, if a negative FID is specified on a DVPREL1/2
entry when PTYPE is not PBEAM, the job may result in the following error
message:
USER FATAL MESSAGE 2042 (EMG )
MISSING MATERIAL TABLE ******** FOR ELEMENT IIIIIIII
Avoidance: Always use a positive FID on a DVPREL1/2 entry when PTYPE is
not PBEAM.
@
31070; 70.5; Nonlinear Analysis - File indexing error
A file indexing error may cause a nonlinear analysis to fail in the NLTRD2
module.
Avoidance: Increase the bufferpool or smem.

@
31071; 70.5; DMAP - PARAML DTI option - Wrong Answers
 If the PARAML module DTI option is used to extract a character string then
the variable contains both the value from the selected word in the record
and also the value in the next word; e.g.,
PARAML TABLE//'DTI'/1/45////S,N,STRING $
the variable STRING will contain the character value from the 45th and
46th word.
Avoidance: Add the following statement after the PARAML module:
STRING=SUBSTRIN(1,4,STRING) $

@
31073; 70.6; DMAP - MATMOD OPTION(17) - UFM 6968
MATMOD OPTION(17) is designed to read the PARTN case control entry. When
PARTN is present, it sets p4 to 0, properly. When it is not present it
sets p4 to -1, also properly, but also outputs
*** USER FATAL MESSAGE 6968 (MTMD17)
NO GRID OR SCALAR POINTS IN SELECTED SET. PARTITIONING VECTOR N ULL.
However, execution continues.
Avoidance: Ignore the message, or test for the presence of a PARTN
command, and branch over the module if none is present. SSSALTER
evds103.f705 demonstrates this avoidance.
@
31076; 70.5; Nonlinear Transient Heat Transfer - PENTA(15) - Failed Convergen
ce
If there are temperature dependent material properties in the model and
you are running a transient thermal analysis using the PENTA(15) element
the run will fail to converge.
Avoidance: Use lower order PENTA elements.
@
31083; 70.7.2; Superelement - SET, THRU - Extra Output
In a superelement analysis, if output is requested with a SET command
using multiple "thru" ranges, then extra output that is not specified in
the range may appear in the output in addition to the requested output.
An example of the offending SET command is as follows:
SET 100 = 1 THRU 10, 80 THRU 90
DISP = 100
The displacements for grid points that are outside of the range may appear
in addition to output for the above grid points. The answers are,
however, correct.
Avoidance: If the extra output is not desired, then specify the SET
explicitly without using the THRU range. The above
example can be modified as shown below:
SET 100 = 1,2,3,4,5,6,7,8,9,10,
 80,81,82,83,84,85,86,87,88,89,90
disp = 100
@
31096; 70.5; P-Elements - GP4 - BIOMSG: ERROR 1044 OCCURRED IN ROUTINE FILPS2
A P-element analysis with coordinate systems may fail in the module GP4
with the message:
BIOMSG: ERROR 1044 HAS OCCURRED IN ROUTINE FILPS2 , FILE INDEX (FILEX)
= 0 STATUS = 0
Avoidance: Increase the memory request.
@
31101; ; P-Elements - PLOT, SDR2 - SFE 4276 EC 211
The nastran plotter does not support the plotting of p-element results.
If the nastran plotter is requested with p-elements, then the job will
fail with the following unclear messages.
BIOMSG: ERROR 211 HAS OCCURRED IN ROUTINE BLPK68 , FILE INDEX (FILEX)
= 0
STATUS = 0
0*** SYSTEM FATAL ERROR 4276.BLPK68 ERROR CODE 211 PID= 0
0*** USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH
DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION.
Avoidance: If plotting of p-element results is desired, use MSC/PATRAN.
@
31104; 70.5; Matrix Methods - MPYAD - SFM 4276, EC 7777
It has been observed on a very large problem, if the matrices are very
dense, then the job may fail in MPYAD. On some machines, the following
message may appear at the bottom of the f06 file.
***** RECEIVED SIGNAL 8, ABORTING ...
BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX)
=
0
STATUS = 0
On other machines, there is no messages in the f06 file, but the following
messages may appear in the log file:

Trap encountered:
SIGTRAP - Trace trap
Traceback:
Offset 0x00000024 in procedure gino_
Offset 0x000007bc in procedure mpytim_
Offset 0x00000708 in procedure mpyad_
Offset 0x000004f8 in procedure dmpyad_
Offset 0x0000029c in procedure xsem07_
Offset 0x000008f4 in procedure mainal
This is due to an overflow in the timing estimate routine.
Avoidance: The only avoidance is to break up the model into
 superelements.
@
31105; 70.5; Superelements - Duplicate Point ID's on Boundary Lead to UFM 679
4
When a model with list superelements has part superelements added to it,
duplicate grid points at the boundary of the list superelements are not
allowed, although they were allowed prior to Version 69. This condition
leads to following error message:
USER FATAL MESSAGE 6794 (SEPSEB)
DUPLICATE BOUNDARY POINTS DETECTED IN UPSTREAM SUPERELEMENT xxx WHILE
SEARCHING 0 AT GRID yy.
LIST FOLLOWS:
Avoidance: Remove the part superelements, or move one of the offending
grid points.
@
31114; 70.5; Database - DBUNLOAD - Signal 11 Writing LLL
If the NEUTRAL option is used in DBUNLOAD, the run will fail. The f06
will likely contain no message, the log file will contain a message of the
form:
***** RECEIVED SIGNAL 11, ABORTING ...

BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX)
= 0
STATUS = 0

Avoidance: Use the BINARY DBUNLOAD.

@
31118; 70.0.5; GPFORCE - GPFDR, TEMP(LOAD) - SFM 6218
If a model has thermal loadings and GPFORCE is requested, the run may
terminate with SFM 6218 in the GPFDR module.
Avoidance: Remove the request for GPFORCE or the thermal loads. Defining
SETs (which omit the GRID points connected to the element mentioned in the
message) for GPFORCE may avoid the problem.
@
31139; 70.5; DMAP - SAMEAS Keyword in NDDL - Poor Message and Wrong Answers.
If SAMEAS,recnam is specified in the NDDL and RECORD=recnam is specified
below SAMEAS,recnam then the following
message is issued but the run is not terminated.
*** SAMEAS CTETRA NOT FOUND
RECORD=recnam must appear above SAMEAS,recnam.
Avoidance: There is no avoidance.
@
31141; ; Eigenvalue Analysis - EIGRL, very large v2 - Possible Wrong Answ
ers
If the specified V2 value on the eigrl entry is substantially larger than
the last mode found and the absolute value of V2 is large (e.g., 99999.),
 then the eigenvectors may be incorrect. This can lead to incorrect
responses in a subsequent modal response analysis. This error is most
likely to occur if there are massless degrees of freedom and/or large mass
in the model and the above conditions exist. If the model is stable, a
possible symptom of this error is large rotational displacements in the
eigenvectors. This error does not occur with the new Lanczos method which
is the default method in Version 70.5.
Avoidance: There are two possible avoidances:
1. In v70.5, use the new Lanczos (the default method).
2. Based on the last eigenvalue calculated and the sturm sequence output,
reduce v2 on the eigrl entry accordingly.

@
31145; 70.5; Nonlinear Analysis - Restarts into Dynamic Analysis for Base Exc
itation
A method to perform automatic restarts from static nonlinear analysis (SOL
106) to linear dynamic analysis (SOLs 103-112) has been avaialble for
several versions. The resart runs have limitations on the type of data
that can be changed. A Technical Application Note (TAN) describing this
capability can be obtained from your MSC Technical Represenative.
@
31150; 70.7; Optimization/Design Sensitivity - PARAM,OPTEXIT - Incorrect Term
ination Message
When optexit is set to -4 (output design sensitivity information into an
external file) for a design optimization input file, the task terminates
with an incorrect message:
RUN TERMINATED DUE TO MAXIMUM NUMBER OF DESIGN CYCLES = IIIII.
The correct message should be RUN TERMINATED DUE TO PARAMETER OPTEXIT =
-4.
@
31154; 70.5; Superelements - Duplicate Element ID's not Detected
It is a rule of superelement usage that element ids must be unique across
all element types, including structural, rigid, and plot elements. When
part and grid list superelements are in one model, duplicated element ids
in the list superelements are not detected. The results are difficult to
predict. Some models have computed correct results, but it is possible
that others would either give wrong answers or take confusing fatal exits.
Avoidance: An avoidance is to perform a data check run where all of the
part superelements are removed. This will result in a different path
through the code which will identify all illegal duplicate element ids.
@
31155; ; Buckling, Nonlinear Analysis - CBEAM, CQUAD4, ZOFF, OFFSET - UFM
6174
Offsets should not be used for the beam and plate elements in a buckling
or nonlinear analysis. This is a documented limitation. However, the
action taken for the beam and plate element with offsets is inconsistent.
In the case of a beam element, the job terminates with UFM 6174; however,
in the case of the plate element, the job continues without any messages.
@
31189; 70.5.2; Nonlinear Analysis - NLITER, CREEP - SFM 4276, EC 7777
Nonlinear analysis with creep and temperature loads may fail in the NLITER
module with the message:
0*** SYSTEM FATAL ERROR 4276.tbk ERROR CODE 7777 PID= 0
The traceback will indicate the routine nmcrd. Other machines may produce
NaN's in the output.
Avoidance: The only avoidance is to use Version 70.5.2 where the error
has been fixed.
@
31209; GL; Eigenvalue Analysis - READ, REIGL - SFM 5401
If f1 is set on an eigrl entry to avoid computing rigid body modes, and f1
is much closer to the origin than the first flexible mode's frequency,
than SYSTEM FATAL MESSAGE 5401 may be encountered.
Avoidance: An avoidance is to first run with f1 blank. It should be
possible to compute the flexible modes, as well as the rigid body modyes.
On a second run, set f1 10% below the first flexible mode.
@
31218; 70.5; GPFORCE - SUBCOM, TEMP(LOAD) - Missing Forces of SPC
GPFORCE output for SUBCOMs with thermal loads will be missing the
contribution of the Forces of Single Point Constraint. Standard SUBCASE
processing will be correct.
@
31225; 70.5; Elements - PBAR/PBARL, PBEAM/PBEAML - UFM 4382, Wrong Answers
A job that includes PBARs and PBARLs with the same PID will issue USER
FATAL MESSAGE 4382 and then continue to run. The same behavior is seen
with PBEAM/PBEAML.
*** USER FATAL MESSAGE 4382 (IFP9 )
PBARL AND PBAR IDENTIFICATION xxx WAS DUPLICATED
@
31232; ; Transient Analysis - LOADSET - UWM 4582
If a static pre-load is used in a modal solution, then LOADSET entry may
not be used to define the dynamic loading. This combination will lead to
either UWM 4582 show below (if the number of LSEQ entries is not equal to
the number of static subcases) or wrong answers (if the number of LSEQ
entries is equal to the number of static subcases).
USER WARNING MESSAGE 4582---, LSEQ CARD SID= 100 REFERS TO A NON-EXISTENT
STATIC LOADS MATRIX COLUMN 2 ( NCOLS= 1 )
USER WARNING MESSAGE 4582---, LSEQ CARD SID= 100 REFERS TO A NON-EXISTENT
STATIC LOADS MATRIX COLUMN 3 ( NCOLS= 1 )
Avoidance: Do not use the LOADSET command when using static pre-loads.
@
31237; ; Online Encyclopedia - Installation
If MSC/NASTRAN Online Encyclopedia (MNE) is installed on a directory path
 name that is too long or contains a name with embedded space characters
(e.g. "Program Files"), the documentation browser will not execute
properly.
Avoidance: Use the suggested default directory <Disk>:\msc
@
31249; 70.5; Heat Transfer Analysis - SFM 3002 (TAHT)
If a control node is used in a very large heat transfer analysis model,
then the following fatal message may be encountered:
*** SYSTEM FATAL MESSAGE 3002 (TAHT )
EOF ENCOUNTERED WHILE READING DATA BLOCK NEWSLT (FILE 101) IN
SUBROUTINE TAHA
Avoidance: Run the job with more memory.
@
31256; 70.5; Optimization/Design Sensitivity - SDR2, SUBCOM - SWM 2092, ERROR
3021
If a design optimization design task contains subcoms and one or more of
the subcom related subcases has no retained responses,
the following fatal message may occur:
SYSTEM WARNING MESSAGE 2092 MODULE = SDR2
SDR2 FINDS A SYMMETRY SEQUENCE LENGTH = xx
AND AN INSUFFICIENT NUMBER OF VECTORS AVAILABLE = yy WHILE ATTEMPTING TO
COMPUTE STRESSES AND FORCES.
ALL FURTHER STRESS AND FORCE COMPUTATION TERMINATED.
BIOMSG: ERROR 3021 HAS OCCURRED IN ROUTINE GREAD , FILE INDEX (FILEX) =
0
STATUS = 0
In a related error, if a subcom contains subseq coefficients that are 0.0
and the subcase associated with this zero coefficient has no retained
responses, the above fatal message may occur.
Avoidance: Either remove the subcase that has no retained responses or
force the retention of at least one response in the subcase.
@
31262; 2001; Documentation - PBEAML - Warping Coefficients are Always Set to
Zero.
The documentation for the PBEAML does not state that the equivalent PBEAM
created will not contain warping coefficients.
Avoidance: Provide the terms using a PBEAM entry.
@
31263; 70.5; Documentation - QRG - PARAM,ACSYM
The description for the parameter "acsym" in the Quick Reference Guide
states that the parameter only works for coupled fluid-structure analysis
in direct frequency response. This parameter also works for coupled
fluid-structure analysis in modal frequency response.
@
31266; 70.5; Documentation - QRG, DRESP1
 The Quick Reference Guide states that the PKNL method can be used in the
evaluation of a DRESP1 with RTYPE=FLUTTER if the
word `PKNL' is placed in the ATTB field of the DRESP1 entry. This is an
error, it should be placed in the PTYPE field of
the DRESP1 entry.
Avoidance: Move PKNL from the ATTB to the PTYPE field.
@
31271; 2001; Documentation - DECOMP, PARAM,AUTOSPC,NO - Unexpected Answers
If the automatic grid point singularity check is turned off
(param,autospc,no), and there are singularities in the model, then the run
may continue and produce large deflections. The next release of the Quick
Reference Guide will warn the user that if param,autspc,no is set, then
the user should take extra caution analyzing the results in the GRID POINT
SINGULARITY TABLE, EPSILON output as well as all warning messages.

@
31278; 70.5; Superelements - DBLOCATE, SERST,SEMI - UFM 4188
A superelement job that contains PARAM,SERST,SEMI and attempts to add the
bulk data for a new superelement fails with the message:
USER FATAL MESSAGE 4188 QUAD4 ** REFERENCES UNDEFINED GRID POINT **.
Avoidance: Apply the following alter:
COMPILE IFPS1 nolist noref
alter 60 $
putsys(1,109)
endalter
@
31304; ; Transient Analysis - PARAM,ADSTAT,YES with TEMP(LOAD) - Wrong An
swers
In a transient analysis, if "param,adstat,yes" is used to include a
preload effect in the transient analysis data recovery, and the preload is
due to thermal loads, then the transient results will be incorrect.
Avoidance: Use solution 129.
@
31308; 2001; Element Strain Energy - ESE(PUNCH), CELASi - Missing Punch Outpu
t
In MSC/NASTRAN Version 70, if ESE(PUNCH) is requested and there are CELAS1
elements in the model, then there will be no output in the punch file.
Avoidance: Insert NASTRAN SYSTEM(210)=1 at the top of the MSC/NASTRAN
input file.
@
31312; 70.5; Aeroelastic Analysis - CAERO2 - Possible Wrong Answers
If the aerodynamic coordinate system is not aligned with the basic
coordinate system, the geometry of CAERO2 slender bodies IN THE BASIC
COORDINATE SYSTEM (BGPA or BGPDT where MODLTYPE=aerostruc) may be
incorrect. These data are used in Spline calculations, so that, if the
geometry is affected in such a way as to shift the location of the body
relative to the spline computational coordinate system, erroneous results
 may be obtained. The geometry in the aerodynamic coordinate system is not
affected, so the "rigid aerodynamic" data are not affected--only the
spline matrix MAY BE affected.
It has been observed, however, that certain erroneous transformations in
combination with TYPEs of CAERO2 bodies (Z, Y or ZY) can yield identical
answers even if the transformation error has moved the body relative to
the spline computational coordinate system. This results because the
relevant degrees of freedom of the body are transverse to the translation
caused by this error.
Avoidance: Aerodynamic models with CAERO2 bodies in which the aerodynamic
coordinate system is rotated relative to the basic coordinate system will
need to be modified and checked.
@
31318; 70.5; Composite Analysis - SEDRCVR - SFM 4276, EC 7777
Large composite models that request large amounts of data recovery may
fail on CRAY systems with SYSTEM FATAL MESSAGE 4276, ERROR CODE 7777 in
the SEDRCVR module.

@
31322; 70.5; Heat Transfer - RSSCON and PARAM,HEATSTAT - SWM 3001 in SEP1X, W
rong answers
If PARAM,HEATSTAT,YES is specified in SOL 101, then SWM 3001 will be
issued by the SEP1X module. Theis warnings
may safely be ignored unless there are RSSCON elements present, in which
case, wrong answers may result.
Avoidance: Insert the following DMAP alter:
compile sestatic
alter 'hsflag\=false'
peid=-1 $
seid=-1 $
endalter
@
31323; ; Miscellaneous - SFM 3063, Memory Allocation Error
If a run is submitted using the MEM= option and the memory allocated is
above or near the limit of what is available, it is possible that the run
may fail with SFM 3063 indicating that there is no more dynamic memory
available.
Avoidance: Always leave at lease 8mW of memory available for the system
when using the MEM keyword. For example, if you are using a queue which
has a 128mW limit, use a MEM value of 120mW or less.
@
31327; 2001; Documentation - CSUPER - Should Reference EXTDR, EXTDROUT, EXTOU
T and EXTUNIT
The Version 70.5 Quick Reference Guide documentation for the CSUPER entry
should be modified to reference parameters EXTDR, EXTDROUT, EXTOUT and
EXTUNIT in section 6 of the QRG.
@
31334; 2001; Documentation/Optimization - DVPREL1, DVPREL2 - QRG Not Updated
for V70.5
 Version 70.5 has simplified the interface for the design of constant
section beams. The revised user interface requires changes
in the Quick Reference Guide's description of the DVPREL1 and DVPREL2
entries. These changes were not completed in time before the Version 70.5
QRG was sent to press.
Avoidance: The revised descriptions can be found in Appendix B of the
Version 70.5 Release Notes.
@
31351; 70.5; Optimization/Design Sensitivity - SUBCOM - SFM 4276.GETSTR ERR
OR CODE 31
The message:
SYSTEM FATAL MESSAGE 4276, GETSTR ERROR CODE 31
may occur if the user employs SUBCOMs and the adjoint sensitivity method
is selected.
Avoidance: Since the adjoint method is not selected when there are
p-elements, an easy avoidance is to tell the module that determines
whether the adjoint method can be selected that there are p-elements in
the model. This is done with the following dmap alter:
compile descon $
alter 'dsad'(1,-1) $
type parm,,logical,n, hold $
hold=pexist $
pexist=true $
alter 'dsad' $
pexist=hold $
@
31364; 70.5; Aeroelastic Analysis - AEROS, ZY Bodies - Wrong Answers
IF there are zy bodies in the model and the aerodynamic coordinate system
is rotated about the x-axis relative to the aerodynamic reference system,
wrong answers can occur in the stability derivative prints, and hence, the
trim solution.
Avoidance: If ZY bodies are included in the analysis, make sure the
aerodynamic coordinate system and the reference coordinate systems
appearing on the AEROS entry are not rotated relative to each other.
@
31374; 70.5; Elements - PBEAM, IFP - SFM 469
If a PBEAM entry contains an integer in a real field, then the job may
fail in IFP without a meaningful message.
Avoidance: The error was only experienced on CRAY systems. If you have
access to another system then run the job on that system to identify the
invalid entries and correct them.
@
31381; ; Miscellaneous - INPUTT2 - UFM 1019, SFM 6551
If the .dat file that is used to read the INPUTT2 file has the same prefix
as the referenced INPUTT2 file--for example, test.dat is the .dat file and
test.op2 is the referenced INPUTT2 file--then depending on the value of
STATUS=x, the job may fail with one of the following error messages:
 ASSIGN INPUTT2='test.op2',UNIT=12
USER FATAL MESSAGE 1019 (OPNPFL)
FORTRAN UNIT OPEN ERROR
IOSTAT = 6, No such file or directory(errno=2)
LOGICAL = INPUTT2 FILE = test.op2
SYSTEM FATAL MESSAGE 6551 (INPTX2)
AN END-OF-FILE WAS DETECTED ON UNIT 12 DURING BINARY INPUTT2 PROCESSING
PROGRAMMER INFORMATION:
1. A KEY VALUE WAS BEING READ
2. THE LAST KEY VALUE READ: 0
FATAL ERROR
0 NO. FORTRAN RECORDS READ = 0
NO. FORTRAN WORDS READ = 0

Avoidance: Modify the name of either the .dat file or the INPUTT2 file.
@
31392; 70.5; SSSALTERS - GPFORCE, trncompa - SFM 5423, Incompatible Matrix Mu
ltiply
If GPFORCE output is requested when using the trncompa alter, then the job
may fail with the following error messages:
*** SYSTEM FATAL MESSAGE 5423 (DMPYAD)
ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES
COLS ROWS FORM
TYPE NZWD DENS
TRAILER FOR QG2 IS 101 21 739 2
2 12 77
TRAILER FOR UHDISP IS 102 21 11 2
2 18 7792
TRAILER FOR (NONE) IS 0 0 0 0
0 0 0
TRANSPOSE FLAG IS 0
0FATAL ERROR
There are two possible avoidances:
1. Add "param,ddrmm,-1" to the input file

or
2. Replace the following dmap statement
mpyad qg,uhdisp,/qgfull/ $
with
equivx qg/qgfull/always $
@
31399; 2001; Documentation - PARAM,USETPRT, PARAM,USETSEL - Incorrect Decimal
Equivalent
the Version 70 and 70.5 Quick Reference guide parameter section indicates
that the decimal equivalent of set U6 (to be specified on PARAM,USETSEL
for PARAM,USETPRT) is 76108864, it should be 67108864.
@
31422; 2001; Documentation/QRG - XYPLOT, PARAM,CURVPLOT - UWM 979
If XYPLOT commands are specified for plotting in transient or frequency
response analysis with grid points on the abscissa, as defined under
PARAM,CURVPLOT, then USER WARNING MESSAGE 979 is issued and no plots are
output. PARAM,CURVPLOT is applicable only in static and normal modes
analysis; i.e., SOLs 1, 3, 5, 14, 15, 16, 101, 103, 105, 106, 114, 115,
144, and 200 even though Table 1 of Section 6 in the Quick Reference Guide
indicates that it is applicable in all the structured solution sequences.
Table 2 is documented correctly for CURVPLOT. Section 13.3.6 of the
Reference Manual is also documented incorrectly with regards to solution
sequence applicability.
@
31423; ; Interface Elements - GMBNDC - GMCURV Option Results in UFM 6403
The GMCURV option on the GMBNDC entry may incorrectly result in the
following fatal message:
*** USER FATAL MESSAGE 6403 (CIFCRV)
THE GMBNDC ID = 2 REFERENCED BY GMINTC ID = 5001 IS NOT
CONTINUOUS AT GRID ID= 1.
Avoidance: The exact cause of the problem is not yet known. The FEEDGE
and GRID options may be used instead.
@
31427; 70.5; Miscellaneous - Access violation (C0000005) exception encountere
d
When there is an illegal bulk data entry "0" as shown below, then the job
may fail in various ways. NT systems have failed with the message "Access
violation (C0000005) exception encountered"
0 4602 4603 4604 4605
0 4606 4607

Avoidance: Correct the offending bulk data entries.

@
31430; 2001; Documentation/Elements - CMASS1, CMASS2, GPWG - QRG Unclear
The CMASSi elements have provisions for defining a pair of DOFs, although
they are often used for point masses. To input a point mass, define only
a single DOF. When two DOFs are used a coupled mass matrix is produced.
If not input correctly it may cause internal constraints. No checks are
made for internal constraints in such elements. The following remarks
should be added to the CMASS1 and CMASS2 Quick Reference Guide
Descriptions:
7. Scalar elements input coupled mass matrices when the second pair of
fields is entered. When uncoupled point masses are desired input only the
first pair of fields. When a coupled mass matrix is requested the
submatrix added has M on the diagonal, and -M on the off-diagonal. The
element is not checked for internal constraints, which is the user's
responsibility if desired. There are instances where elements with
internal constraints are desired, although not frequently. To identify
the presence of internal constraints caused by coupled mass, inspect GPWG
 output, OLOAD output due to GRAV loads, and rigid body modes of free
structures. Some forms of coupled mass will cause coupling of rigid body
translational mass terms in GPWG output, and poor rigid body modes in
modal analysis.

@
31436; 2001; Documentation - PBARL, PBEAML - Missing Documentation
There are undocumented limits in the use of the beam library. For the
PBEAML:
The beam can have no more than 14 dimensions per station.
The total number of dimensions at all stations must be less than 200.
The transfer of data with the beam server is limited to 4000 words.
None of these limits are exceeded with the MSC beam library, but a user
defined beam library could. The data transfer limit would most likely be
exceeded when the number of dimensions for a PBARL exceeds 235 (4000 / 17
properties per bar).
There is a further limit that the NSM values input on the PBARL and PBEAML
must be >= 0. No other property types have this limit.
@
31447; 70.5; Aeroelastic Analysis - EPOINT, SOL 145, Multiple Eigenvalue Subc
ases - UFM 1126
In aeroelastic flutter analysis (SOL 145 only) UFM 1126 as shown below is
issued when there are multiple normal mode boundary conditions and the
model contains epoints.
UMERGE1 IS ATTEMPTING TO OUTPUT DATABLOCK NAME = GMD which already exists
Avoidance: Insert the following DMAP alter:
compile sefluttr
alter 1
file gmd=save,ovrwrt/god=save,ovrwrt/k4dd=save,ovrwrt $
endalter
@
31452; ; P-Elements - GMINTC and GMINTS - Non-Zero Rigid Body Stresses
Models that connect curved surfaces with the GMINTC and GMINTS interface
elements may experience non-zero stresses due to rigid body motion.
These non-zero stresses are caused by the differences in cubic geometry on
the dissimilar mesh.
Straight interfaces have no such problem, since the geometry is identical.

Avoidance: The tolerance TOL on the PINTC and PINTS property entries will
generate a warning message for differing geometry, and may be set lower to
be more conservative.
@
31461; 2001; Documentation - PARAM,NQSET - Missing QRG Documentation
For part superelements, generalized coordinates can be defined
automatically by senqset entries, as documented in the qrg, or by the
 param, nqset entry. The param method was introduced in Version 69 but has
yet to be documented. The documentation follows:
NQSET Default=0
If NQSET>0, and the PARAM entry is in case control, all part superelements
that do not contain qset entries, or are not referenced by senqset entries
in the main bulk data deck, have NQSET generalized coordinates assigned to
them. These qset variables are functionally equivalent to those generated
by senqset entries.
@
31463; 70.5; Optimization/Design Sensitivity - ANAL=STATICS - UFM 9058
In solution 200 with ANALYSIS=STATICS, if the static loads are null then
UFM 9058 will be issued during the sensitivity analysis:
THE SOLUTION FOR THE RESIDUAL STRUCTURE DOES NOT EXIST
Avoidance: Insert the following DMAP alter:
compile pslgdv
alter 'call sela1'(,-1)
dbstatus egxp//s,n,noegxp $
if ( noegxp<=0 ) matgen ,/egxp/7/lusets/nopgs $
endalter
@
31469; GL; Superelements - SECONCT - UFM 4156
If PART superelements are used in a model and manual connections are
specified through the SECONCT entry and there are coincident exterior
points in the PARTs, the run will fail when attempting to connect the
reduced matrices with UFM 4156 indicating an incorrect matrix trailer as
shown below:
USER FATAL ERROR 4156--- DIMENSIONS OR TYPE OF DATA BLOCK XAAV ,
SUPERELEMENT xx
TRAILER=( 1278 1278 6 2 58 52) ARE INCONSISTENT WITH
SUPERELEMENT MAP.
FATAL ERROR
Avoidance: Coincident points are not supported when connecting PART
superelements. It is possible that the error could be avoided by only
connecting one of the points to the other PARTs in the model and using an
RBAR or CBUSH to connect the coincident points in the PART. If the
automatic connection logic were used, the program would fail with a fatal
message indicating that there were coincident points in the PART.
@
31472; 2001; P-Elements - PBEAM - UFM 3005 (SNGERR)
BEAM p-elements fail in an adaptive loop if J(A)>0 is not specified on
the PBEAM entry. The following fatal message is issued.
*** USER FATAL MESSAGE 3005 (SNGERR)
ATTEMPT TO OPERATE ON THE SINGULAR MATRIX D IN SUBROUTINE ADPF1D
0FATAL ERROR
Avoidance: Specify J(A)>0 in the 8th field of the PBEAM entry.
@
31473; 2001; Documentation - CVISC - Stress Plot Codes not Applicable
The Appendix A in the Quick Reference Guide documents stress plot codes
for the CVISC element. The CVISC element only supports force output.

@
31475; 70.6; Optimization/Design Sensitivity Analysis - Multi-Level Superelme
nts - UFM 9058
"USER FATAL MESSAGE 9058" may appear in a superelement sensitivity
analysis if the model contains multilevel superelements and an upstream
superelement contains a design model but more than one of its downstream
superelements does not.
Avoidance: Place dummy design information in downstream superelements.
@
31479; 70.5; P-Elements - TETRA, GMLOAD, GMCORD - UFM 6497 on SGI and SUN Mac
hines
TETRA p-elements with GMLOAD and GMCORD may lead to the following fatal
message:
*** USER FATAL MESSAGE 6497 (TRNJFD)
THE POINT AT: X= 0.346875D+03 Y= 0.500000D+03 Z=-0.203698D+03, IN THE
BASIC COORDINATE SYSTEM
COULD NOT BE LOCATED ON FEFACE ID= 1.
*** USER FATAL MESSAGE 2025 MODULE = SSG1
UNDEFINED COORDINATE SYSTEM 1
0FATAL ERROR
Avoidance: Either:
- mesh with HEXA elements.
- use a system other than GMCORD with GMLOAD
- use PLOAD4 instead of GMLOAD
@
31488; 70.5; ESTIMATE - May Ignore Entry After TITLE or SUBTITLE
Under certain circumstances, ESTIMATE will ignore the entry following a
TITLE, SUBTITLE, or LABEL entry. This may cause ESTIMATE to ignore a
METHOD, MPC, SPC, STRAIN, or STRESS CC entry if it immediately follows a
TITLE, SUBTITLE, or LABEL that contains a comma "," character.
Avoidance: One of the following avoidances may be used:
1) Enter a comment following the TITLE, SUBTITLE, or LABEL.
or
2) Specify the missing METHOD, MPC, or SPC entry on the command line.
@
31498; ; Restart - DBALL offline - UFM 1127, SFM 6357, plus others
If scr=no is specified in a cold start, but the data base path and name
is not assigned by the dbs keyword or the ASSIGN statement, then a restart
run in a different directory with only an ASSIGN statement for the MASTER
dbset, and not DBALL,
will fail with UFM 1127. This problem does not occur if the dbs keyword
 specifies the location of the database on the command line. Some of the
possible error messages are shown below:
*** USER FATAL MESSAGE 1127 (GNFIST)
DMAP MODULE PLTSET IS ATTEMPTING TO ACCESS DATA BLOCK 103 NAME
= GEOM2 WHICH IS ON AN UNAVAILABLE DBSET
0FATAL ERROR
*** SYSTEM WARNING MESSAGE 3001 (SEP1X )
THE INPUT DATA BLOCK NAMED GEOM1QS AT POSITION 2 DOES NOT
EXIST.
*** SYSTEM FATAL MESSAGE 6357 (SEP1A)
NO GRID POINTS FOUND.

Avoidance: There are three avoidances:

1. Submit the restart from the same directory as the cold start
2. Specify ASSIGN statements for all the DBSETs
3. Use the dbs keyword with the full pathname

@
31537; 2001; Documentation - QRG, PARAM,AUTOSPC - Incorrect Default
Starting with Version 69, the QRG states that the default for
PARAM,AUTOSPC is NO in SOL's 1 thru 12, when in fact it is YES for all of
these solutions except SOL 4.

@
31540; 69.1; Postprocessing - Incorrect Approach Code in XDB
Solution 1 in Version 69 will incorrectly set the approach code to 4 in
the xdb file. The correct value is 1.
Avoidance: Either insert the following DMAP alter or use solution 101:
compile sol1
alter 'dbc '(2),''
DBC CASECC,,,,,,,,,,,,,,,,,,,//
'CASECC'////////////////////
-1/DBCPATH/S,N,CP/soltype//V,Y,GEOMU/V,Y,LOADU/V,Y,POSTU/
V,Y,DBCDIAG/V,Y,DBCCONV/V,Y,DBCOVWRT $
endalter

@
31568; 70.7; Restarts - DBLOC - UFM 723 (LCDEFW)
The WHERE clause on the DBLOC command does not work consistantly with the
VERSION and PROJECT qualifaiers with the DBDICT entry. The following
statement:
DBLOC DATABLK=(QKKS) WHERE (VERSION=* AND PROJECT=*) LOGICAL=QTODB5
 caused the diagnostic message:
0*** USER FATAL MESSAGE 723 (LCDEFW)
PROJECT ID EXPRESSION HAS AN UNBALANCED NUMBER OF SINGLE QUOTES
Other variations of the WHERE clause with these qualifiers may cause data
to not be found.
@
31578; ; Nonlinear Analysis - Load Factor Greater than 1.0 causes Infinit
e Loop
In a nonlinear analysis, on some instances, the load factor tolerances may
be in error. When this occurs the value of the load factor exceeds 1.0 and
the solution may enter an infinite loop or produce some other fatal
message later on.
The problem is identified when one of the load factors > 1.0
Avoidance: On the NLPARM entry increase the NINC field by 1 and restart
into the offending subcase. This will change the tolerance calculation.
@
31582; 2001; Miscellaneous - PARAM,DESPCH1 - Punch output with small field fo
rmat not allwed
Bulk Data paramter DESPCH1 can be used to request the format type for the
punched output. A negative DESPCH1 requests the small field format while
the positive DESPCH1 requests the large field format. However, when an
optimization deck specifies the negative DESPCH1 and also specifies the
second parameter DESPCH>0, the small field format is ignored.
Avoidance: Apply the dmap alter below:
compile nddl=nddl
include 'nddl.dat' $ where the nddl.dat is the modified nddl
list that includes the following two lines:
param,despch =0, type=I, PATH=GEOM,LOCATION=master
param,despch1 =6, type=I, PATH=GEOM,LOCATION=master
compile desopt
alter 'despch'
type parm,,i,y,despch1
compile exitopt
alter 'despch1','despch1'
type PARM,,I,Y,cnvflg,despch,omaxr,despch1,xyunit

@
31583; ; Acoustic Analysis - METHOD(STRUCTURE), SDFREQ - UFM 4276 EC 1305
In modal frequency response analysis, solution 111, with fluids, if the
METHOD(STRUCTURE)=n case control command is missing, the following error
will occur:
USER INFORMATION MESSAGE 5222 (SQFREQ)
COUPLED SOLUTION ALGORITHM USED.
BIOMSG: ERROR 1305 HAS OCCURRED IN ROUTINE ENDPUT , FILE INDEX (FILEX) =
0
STATUS = 0
********* NASTRAN FILE TABLE *********
INDEX LOGNAME STREAM_ID NAME
2 OBJSCR 00000002 /work/30314_42.OBJSCR
 3 DBA 00000003 /work/fnd1.dba
4 MAST 00000004 /work/fnd2.mast
6 SCRA 00000006 /home/no4/emc/emc3fnd/fnd5.scra
7 SCRU 00000007 /tmp/fnd6.scru
8 MASTER 00000008 /work/30314_42.MASTER
9 SC2A 00000009 /work/30314_42.SC2A
10 MASTERA FFFFFFF6 /msc/msc70/aix/SSS.MASTERA
11 MSCOBJ FFFFFFF5 /msc/msc70/aix/SSS.MSCOBJ
0*** SYSTEM FATAL ERROR 4276.ENDPUT ERROR CODE 1305 PID= 0
0*** USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH
DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION.
1 * * * END OF JOB * * *

This user input error need to have a valid error message given.
Avoidance: Put a METHOD(STRUCTURE)=n case control command in the input
deck.
@
31590; ; Nonlinear Analysis - Missing Intermediate output on a nonlinear
restart run
If additional subcases are added in a restart of a solution 106 nonlinear
job then there may be missing intermediate output.
Avoidance: The problems is under investigation. Other than performing
the job as a cold start there is no avoidance.
@
31591; 70.5.2; Postprocessing - DBC, PARAM,POST PARAM,GEOMU,0 - NASTRAN Stops i
n DBC Module
If PARAM,GEOMU,0 is used to suppress the writing of geometry into XDB
(i.e. PARAM,0) is in the deck, then the job will stop without message in
the DBC module. The log file may contain a message of the form:
1525-001 The READ statement on the file ./s31591.xdb cannot be completed
because the end of the file was reached. The program will stop.
@
31608; ; File Management Section - ASSIGN, DELETE - DBALL is not Deleted
The fms statement ASSIGN includes the DELETE qualifier, which specifies
that an existing data base at the start of a run should be deleted, and
replaced with a new version. If this qualifier is applied against a
MASTER database, it properly deletes the MASTER database. However,
MASTERs point to one or more data base members, such as DBALL. The index
for these databases are stored on MASTER, so they are useless after MASTER
is deleted and they too should be deleted but are not.
@
31622; 70.5.1; Heat Transfer Analysis - CHBDYE, TA1A - Segmentation Error
Heat Transfer models with CHBDYE elements may fail with a segmentation
error.
Avoidance: There is no avoidance.

@
31636; ; Optimization/Design Sensitivity - DESSUB - UFM 6293 incorrectly
issued
 If an attempt to place limits on a global response (such as weight) is
applied using a DESSUB command, the program will incorrectly issue UFM
6293.
Avoidance: Global responses (such as WEIGHT) may only be constrained
using the DESGLB command. They may not be constrained by the DESSUB
command.
@
31655; 70.5.2; Optimization/Design Sensitivity - DOM9 - SFM 4276, EC 3002, EC 1
409, EC 31
The message:
SYSTEM FATAL MESSAGE 4276, ERROR CODE 1409
can occur in the DOM6 or DOM9 module if there are DRESP2 responses and two
or more of the DRESP1 id's invoked by a single DRESP2 are duplicates.
Avoidance: Remove the duplicate DRESP1 id's from the DRESP2 entry. It is
most likely not the user intent to have duplicated the id's, but it will
be permitted once this error is corrected.
@
31662; 70.7; Miscellaneous - DBVIEW - Infinite CPU Loop
If the view name and the datablock name are identical on the DBVIEW
statement, and the datablock is local scratch, then an
infinite CPU loop will occur in GNPREP when the view name is used as input
to a module. The following input deck demonstrates the
problem:
sol AERBG
diag 8,24$
compile AERBG
$
SUBDMAP AERBG
dbview uset = uset ( where peid=1)
tabprt uset
END $
$
CEND
BEGIN BULK
ENDDATA
@
31664; 70.7; Rigid Elements - RBE3, GP4 - EXIT from GP4, No Fatal Message
The um fields of the RBE3 element contain grid point and component data.
A GID-component pair may appear only once in the um fields. When a pair
appears more than once, the gp4 modlule takes an exit without producing
any diagnostics.
The most likely clue that this has happened is to inspect the .f04 file.
If the last several modules to run are gp4, append, bcdr,prstum, and exit,
a bad rbe3 may be the cause.
Avoidance: An avoidance is to check the um fields of all rbe3 entries.
@
31665; ; P-Elements - MATGPR - GRIDN Labels Incorrect
 When using the MATGPR module to print a matrix for p-elements, the
components of the higher-order degrees of freedom represented by the
GRIDNs are incorrect. They are labeled T1,T2,T3,R1,R2,R3 followed by
garbage, since the number of components may be much higher than 6. They
should simply be labeled with integers.
@
31668; 70.5.1; Optimization/Design Sensitivity - PBEAM - Wrong Answers
The Version 70.5 Release Notes state that it is now possible to design
properties for the constant cross section PBEAM by
referencing the END A fields only on the PBEAM bulk data entry.
(Previously, one had to reference the EPT and explicitly
design END A and END B). The implementation of this new capability was
incomplete in that it does not change the DVPREL2 invoked designed
properties at END B. As a consequence, sensitivities of END B responses
are most likely incorrect. If an optimization
task in undertaken with END B responses, the redesigned structure will
provide incorrect analysis results at END B and lead to invalid designs.

Avoidance: Either do not use the new feature or do not ask for data
recovery at END B in the design task.
@
31671; 2001; Documentation - MODES - Inefficient Processing, Excessive Output
The MODES case control command may be used in all solution sequences as a
subcase replicator. The most common and obvious use is in eigenvalue
solutions and is demonstrated under the Quick Reference Guide's
description of this entry. This command may be of no use in other
solution sequences such as transient response and may cause the data
recovery processing to be repeated and printed multiple times.
@
31674; 2001; Documentation - PARAM, CB1, CB2, CK1, CK2, CK3, CP1, CP2, CM1,
CM2 - UFM 6011
Parameters CB1, CB2, CK1, CK2, CK3, CP1, CP2, CM1, and CM2 are incorrectly
indicated to be real parameters with a default of 1.0 when
they are complex parameters with a default of (1.0,0.).
Avoidance: If you use these parameters, define both the real and
imaginary terms.
@
31686; 70.7; Elements - CQUAD4, CQUAD8, CTRIA3, CTRIA6 - UFM 316
The MCID field on the CQUAD4, CQUAD8, CTRIA3 and CTRIA6 entries cannot be
greater than 999999 or else UFM 316 will be generated. However, on the
PSOLID and CGAP, coordinate system IDs > 999999 are accepted.
@
31689; ; Nonlinear Static Analysis - PGAP - Unnecessary Restriction for
Non-Adaptive Gap
The adaptive gap element requires that MU2 <= MU1; however, this
restriction is unnecessarily imposed on the non-adaptive gap element.
Avoidance: This limitation can be avoided by proper choice of the
orientation vector.
@
31694; 70.7; Dynamics - SOLs 103-112, PARAM - Wrong Answers
In SOLs 103, 105, and 107 through 112, if the Case Control section
contains multiple dynamic subcases or a static subcase for differential
stiffness, then PARAM commands specified in the dynamic subcase(s) are
ignored.
Avoidance: Specify PARAMs in the Bulk Data section or insert the
following DMAP alter:
compile sekdr
alter 'pvt ',''
pvt pvts,casebuck/pvtxxx $
endalter
@
31703; 70.5.1; Aeroelastic Analysis - AELIST, ADG - UFM 6944
If an AELIST set is referenced by more than one AESURF entry, then the job
will fail with the message:
USER FATAL MESSAGE 6944 (ADG)
AELIST ENTRY ID = xxx THAT IS REFERENCED ON AESURF ENTRY ID =yyy DOES NOT
EXIST.
The id's specified in the message are also scrambled in that aesurf id's
are given in the aelist position and vice versa.
Avoidance: Create another AELIST entry with a unique ID but with the same
set of aero boxes.
@
31717; 70.7; DMAP - SOLVE - Inefficient Processing
The internal resequencing option for decomposition in the SOLVE module is
not performed. This may lead to inefficient processing.
Avoidance: Use the DCMP and FBS modules.
@
31721; 70.5.1; Miscellaneous - Missing Release Guide Test Problems
The MSC/NASTRAN Version 70.5 Release Guide as well as the README file in
the Version 70.5 relnotes directory refer to the problems ar29be01.dat,
ar29src1.dat, tpsnp.dat, and tpsp.dat. These problems are missing on the
UNIX R0 cd.
Avoidance: The input decks are listed below separated by $fname.
$ ar29be01.dat
$
$*********************************************
$ Test of nonlinear shock absorber
$ under large deformation
$********************************************
TIME 1000
SOL 129
$********************************************
CEND
$
TITLE = Shock absorber and spring loaded with sine pulse
LABEL = CBUSH1D with SHOCKA option and SPRING option
ECHO = SORT
SET 1= 1,2
DISP(SORT2,PLOT) = 1
SET 2 = 1, 100
STRESS(SORT2,PLOT) = 2
OLOAD(SORT2,PLOT) = 1
$
SPC = 1
$
SUBCASE 10
DLOAD = 100
TSTEPNL = 10
$
SUBCASE 20
DLOAD = 100
TSTEPNL = 20
$
OUTPUT(XYPLOT)
XAXIS = YES
YAXIS = YES
XGRID LINES = YES
YGRID LINES = YES
XTITLE = TIME
$
YTITLE = Applied load
TCURVE = Applied load at GRID 2 in vertical direction, T3, z
XYPLOT OLOAD 10,20/2(t3)
$
YTITLE = DISPLACEMENT
TCURVE = Vertical displacement at beam end, GRID 2, T3, z
XYPLOT DISP 10,20/2(T3)
$
$************************************************************************
*
BEGIN BULK
$
PARAM, POST, -1
PARAM,AUTOSPC,YES
PARAM,MAXRATIO,1.E+8
PARAM LGDISP 1
PARAM LANGLE 2
$
$ CBUSH1D element (SHOCKA option with TABLE for S(u)
$ and TABLE for SPRING constant)
$
CBUSH1D, 100, 101, 8, 19
$
$ CVT = 0.2 damping for tension
$ CVC = 0.4 damping for compression
$ S(u) = 1.0 constant scale factor
$ EVT = 1.0 exponent of velocity for tension
$ EVC = 0.5 exponent of velocity for compression
$
PBUSH1D, 101, 10.0, 0.4, , , , , ,+PB1
+PB1, SHOCKA, TABLE, 0.2, 0.4, 1.0, 0.5, 999, , +PB2
+PB2, SPRING, TABLE, 998
$
TABLED1, 999, , , , , , , , +TB999
+TB999, -1000., 1.0, 0.0, 1.0, 1000., 1.0, ENDT
$
TABLED1, 998, , , , , , , , +TB998
+TB998, -1000., -10000., 0.0, 0.0, 1000., 10000., ENDT
$
PLOTEL, 9999, 8, 19
$
$ Loading - Unit amplitude half sine
$
$TLOAD2 ID DAREA DELAY TYPE T1 T2 F P
TLOAD2 100 10 0 .0010 .5010 1.0 -90.0
$
$ Point of application, direction and magnitude of load
$
$DAREA ID P1 C1 A1
DAREA 10 2 3 900.0
$
$ Time step integration. Set MAXBIS to a negative number,
$ terminate if convergence not reached.
$
$TSTEPNL, ID, NT, DT , NO, METHOD, KSTEP, MAXITER, CONV
TSTEPNL, 10, 1200, 0.0005, , ADAPT , , , ,+T10a
+T10a, , , , , , , , ,+T10b
+T10b, -9
$
TSTEPNL, 20, 800, 0.0010, , ADAPT , , , ,+T20a
+T20a, , , , , , , , ,+T20b
+T20b, -9
$
$ Geometry
$
GRID 1 0.0000000.0000000.000000
GRID 2 1200.0000.0000000.000000
GRID 3 100.00000.0000000.000000
GRID 4 200.00000.0000000.000000
GRID 5 300.00000.0000000.000000
GRID 6 400.00000.0000000.000000
GRID 7 500.00000.0000000.000000
GRID 8 600.00000.0000000.000000
GRID 9 700.00000.0000000.000000
GRID 10 800.00000.0000000.000000
GRID 11 900.00000.0000000.000000
GRID 12 1000.0000.0000000.000000
GRID 13 1100.0000.0000000.000000
GRID 19 -92.82040.000000399.9991
$
$ Connectivity
$
CBEAM 1 1 7 80.000000-1.000000.000000
+E61
+E61 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 2 1 1 30.000000-1.000000.000000
+E62
+E62 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 3 1 3 40.000000-1.000000.000000
+E63
+E63 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 4 1 4 50.000000-1.000000.000000
+E64
+E64 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 5 1 5 60.000000-1.000000.000000
+E65
+E65 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 6 1 6 70.000000-1.000000.000000
+E66
+E66 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 7 1 8 90.000000-1.000000.000000
+E67
+E67 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 8 1 9 100.000000-1.000000.000000
+E68
+E68 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 9 1 10 110.000000-1.000000.000000
+E69
+E69 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 10 1 11 120.000000-1.000000.000000
+E6A
+E6A 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 11 1 12 130.000000-1.000000.000000
+E6B
+E6B 0 00.0000000.0000000.0000000.0000000.0000000.000000
CBEAM 12 1 13 20.000000-1.000000.000000
+E6C
+E6C 0 00.0000000.0000000.0000000.0000000.0000000.000000
$
$ Beam property and material
$
PBEAM 1 2600.0000180000.05000.0000.0000000.0000000.000000
$
MAT1 22.00E+6 7.69E+050.3000007.80E-09
$
$ Boundary conditions
$
SPC 1 1 1234 0.000000
SPC 1 2 246 0.000000
SPC 1 3 246 0.000000
SPC 1 4 246 0.000000
SPC 1 5 246 0.000000
SPC 1 6 246 0.000000
SPC 1 7 246 0.000000
SPC 1 8 246 0.000000
SPC 1 9 246 0.000000
SPC 1 10 246 0.000000
SPC 1 11 246 0.000000
SPC 1 12 246 0.000000
SPC 1 13 246 0.000000
SPC 1 19 1234 0.000000
$
ENDDATA
$ ar29scr1.dat
$
$ Name: ar29src1
$
$ A swinging pendulum is modeled. The pendulum is loaded with
$ gravity load. The analysis goes through one and a half cycles,
$ starting in horizontal position. Large deformation is turned on.
$
$ The pendulum is modeled with two seperate systems.
$ 1. CROD and CVISC, CROD rotates, CVISC stays in horizontal
$ orientation.
$ 2. CBUSH1D with SPRING and DAMPER, spring and damper rotate.
$
TIME 10 $
SOL 129 $
CEND
$
TITLE= ar29src1 swinging elastic pendulum
SUBTITLE= demonstrate rotating damping
ECHO=BOTH
SET 1 = 22, 42
SET 2 = 104, 108
DISP=1
STRESS=2
spc= 4
$
SUBCASE 1
DLOAD=100
TSTEPNL=100
$
OUTPUT(XYPLOT)
XTITLE = TIME IN SECS
XGRID LINES = YES
YGRID LINES = YES
YTITLE = X DISPLACEMENT
TCURVE = DISPL of ROD and VISC compared to CBUSH1D
XYPLOT DISP RESP/22(T1), 42(T1)
YTITLE = Z DISPLACEMENT
TCURVE = DISPL of ROD and VISC compared to CBUSH1D
XYPLOT DISP RESP/22(T3), 42(T3)
YTITLE = STRESS/FORCE
TCURVE = STRESS/FORCE of ROD compared to CBUSH1D
XYPLOT STRESS RESP /104(2), 108(2)
$
BEGIN BULK
$
param, lgdisp, 1
$
$ GEOMETRY
$
grid,21, , 0., 2., 0.
grid,22, , 1., 2., 0.
grid,41, , 0., 4., 0.
grid,42, , 1., 4., 0.
$
$ Boundary conditions
$
spc1, 4, 12356, 21, 41
spc1, 4, 246, 22, 42
$
$ mass
$
conm2, 53, 22, , 1000.
conm2, 55, 42, , 1000.
$
$ rod with viscous damping
$
crod, 104, 201, 21, 22
$
prod, 201, 222, 1.
mat1, 222, 1.e+7
mats1, 222, , plastic, 0., 1, 1, 1.e+12
$
cvisc, 114, 310, 21, 22
pvisc, 310, 1000.
$
$ CBUSH1D spring and rotating viscous damper
$
cbush1d, 108, 208, 41, 42
pbush1d, 208, 1.e+7, 1000.
$
$ $ $ $ $ $ $ $ $ $
$ SOLUTION STRATEGY
$
TSTEPNL, 100, 400, .01, , adapt, , , , +TSP100
+TSP100, , , , , , , , , +TSP100
+TSP100, , 0
$
$ LOADING
$
DAREA 100 22 3 10000.
DAREA 100 42 3 10000.
$
TLOAD1 100 100 0 100
TABLED1 100
+ABC
+ABC 0. 1. 10. 1. ENDT
$
ENDDATA
$ tpsnp.dat
$ DEC/CMS REPLACEMENT HISTORY, Element HA144E.DAT
$ *2 6-JUL-1994 14:07:00 A_BOYADJIAN "68 PLUS/G/ CHANGE DBSDIR: TO
TPLDIR: FOR INCLUDE CARDS"
$ *1 5-JUL-1994 16:51:13 A_BOYADJIAN "68 PLUS/G/ NEW FOR V68 AERO_SS
BOOK"
$ DEC/CMS REPLACEMENT HISTORY, Element HA144E.DAT
ID MSC, HA144E $ E_JOHNSON V68 5-JUL-1994
$ID MSC, HA144E
$$$$$$$$ HANDBOOK FOR AEROELASTIC ANALYSIS EXAMPLE HA144E $$$$$$$$
$ $
$ MODEL DESCRIPTION FULL SPAN 30 DEG FWD SWEPT WING $
$ WITH AILERON, CANARD AND AFT SWEPT $
$ VERTICAL FIN AND RUDDER. $
$ BAR MODEL WITH DUMBBELL MASSES. $
$ $
$ SOLUTION QUASI-STEADY AEROELASTIC ANALYSIS $
$ USING DOUBLET-LATTICE METHOD $
$ AERODYNAMICS AT MACH NO. 0.9. $
$ $
$ OUTPUT PLOTS OF THE STICK MODEL AND AERO $
$ GRID, LISTS OF RESTRAINED AND $
$ UNRESTRAINED SYMMETRIC AND ANTI- $
$ SYMMETRIC STABILITY DERIVATIVES, $
$ AERODYNAMIC FORCES AND PRESSURES $
$ PLUS STRESSES AND DEFLECTIONS FOR $
$ LEVEL FLIGHT AND SEVERAL UNSYM- $
$ METRICAL MANUEVERS. $
$ $
$ $
$$$$$$$$ $$$$$$$$
TIME 5 $ CPU TIME IN MINUTES
SOL 144 $ STATIC AERO
CEND
TITLE = EXAMPLE HA144E: 30 DEG FWD SWEPT WING WITH 3 CONTROLS HA144E
SUBTI = UNSYMMETRIC FLIGHT CONDITIONS, DOUBLET-LATTICE AERO
LABEL = FULL-SPAN MODEL
ECHO = BOTH
SPC = 1 $ SYMMETRIC CONSTRAINTS
SET 1 = 1 THRU 999999
DISP = 1 $ PRINT ALL DISPLACEMENTS
STRESS = ALL $ PRINT ALL STRESSES
FORCE = ALL $ PRINT ALL FORCES
AEROF = ALL $ PRINT ALL AERODYNAMIC FORCES
APRES = ALL $ PRINT ALL AERODYNAMIC PRESSURES
SUBCASE 1
TRIM = 1 $ HIGH SPEED LEVEL FLIGHT
SUBCASE 2
TRIM = 2 $ HIGH SPEED ROLLING PULLOUT
SUBCASE 3
TRIM = 3 $ HIGH SPEED PULLOUT WITH ABRUPT ROLL
SUBCASE 4
TRIM = 4 $ HIGH SPEED SNAP-ROLL ENTRY
SUBCASE 5
TRIM = 5 $ HIGH SPEED CLIMBING TURN
OUTPUT(PLOT)
PLOTTER = NASTRAN
SET 1 = ALL
FIND SCALE, ORIGIN 1,SET 1
PLOT SET 1
PLOT STATIC DEFORMATION 0, ORIGIN 1, SET 1, OUTLINE
BEGIN BULK
$
$ DEC/CMS REPLACEMENT HISTORY, Element FSWTWO.DAT
$ *1 5-JUL-1994 16:46:14 A_BOYADJIAN "68 PLUS/G/ NEW FOR V68 AERO_SS
BOOK"
$ DEC/CMS REPLACEMENT HISTORY, Element FSWTWO.DAT
$*** * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * ***$
$ $
$ * * * STRUCTURAL DATA * * * $
$ $
$ (LB-FT-SEC SYSTEM) $
$ $
$ * * GRID GEOMETRY * * $
$ $
$ GRID 90 - 100 (T3) FUSELAGE POINTS $
$ GRID 110 - 122 (T3) WING POINTS $
$ GRID 310 - 312 (t3) FIN POINTS $
$ $
$ * FUSELAGE GRID * $
$ $
$ THE GRID ENTRY DEFINES THE LOCATION OF A STRUCTURAL GRID $
$ POINT. LISTED ARE ITS COORDINATE SYSTEM ID, ITS LOCATION, $
$ THE ID OF THE COORDINATE SYSTEM IN WHICH ITS DISPLACEMENTS $
$ ARE DEFINED, ITS PERMANENT SINGLE-POINT CONSTRAINTS AND $
$ ITS ASSOCIATED SUPERELEMENT ID. $
$ $
$ ID CP X1 X2 X3 CD PS SEID
GRID 90 15. 0. 0.
GRID 97 0. 0. 0.
GRID 98 10. 0. 0.
GRID 99 20. 0. 0.
GRID 100 30. 0. 0.
$ $
$ * RIGHT WING GRID * $
$ $
$ ID CP X1 X2 X3 CD PS SEID
GRID 111 24.61325 +5. -2.
GRID 110 27.11325 +5. 0.
GRID 112 29.61325 +5. 2.
GRID 121 18.83975+15. 2.
GRID 120 21.33975+15. 0.
GRID 122 23.83975+15. -2.
$ $
$ * LEFT WING GRID * $
$ $
$ ID CP X1 X2 X3 CD PS SEID
GRID 211 24.61325 -5. -2.
GRID 210 27.11325 -5. 0.
GRID 212 29.61325 -5. 2.
GRID 221 18.83975-15. 2.
GRID 220 21.33975-15. 0.
GRID 222 23.83975-15. -2.
$ $
$ * FIN GRID * $
$ $
GRID 310 32.88675+0. 5.
GRID 311 30.38675+0. 5.
GRID 312 35.38675+0. 5.
$ $
$ * * STRUCTURAL STIFFNESS PROPERTIES * * $
$ $
$ * FUSELAGE STRUCTURE * $
$ $
$ THE CBAR ENTRY DEFINES A SIMPLE BEAM ELEMENT. LISTED ARE $
$ ITS PROPERTY ENTRY ID, THE TWO GRID POINTS JOINED BY THE $
$ BEAM AND COMPONENTS OF A VECTOR FROM THE FIRST POINT. $
$ THIS VECTOR DEFINES THE DIRECTION OF THE STRUCTURAL DE- $
$ FLECTION OF THE POINT AND ITS POSITIVE SENSE. $
$ $
$ EID PID GA GB X1,GO X2 X3
CBAR 101 100 97 98 0. 0. 1.
CBAR 102 100 98 90 0. 0. 1.
CBAR 103 100 90 99 0. 0. 1.
CBAR 104 100 99 100 0. 0. 1.
$ $
$ THE PBAR ENTRY DEFINES GEOMETRIC PROPERTIES OF THE BEAM. $
$ LISTED ARE ITS ASSOCIATED MATERIAL ENTRY ID, ITS CROSS SEC- $
$ TIONAL AREA, AREA MOMENTS OF INERTIA, TORSIONAL MOMENT $
$ OF INERTIA AND NON-STRUCTURAL MASS PER UNIT AREA. THE $
$ OPTIONAL CONTINUATION ENTRY CONTAINS STRESS RECOVERY $
$ COEFFICIENTS, I.E., Y,Z COORDINATES WHERE STRESSES ARE $
$ TO BE COMPUTED. K1 AND K2 ARE AREA FACTORS FOR SHEAR $
$ STIFFNESS (DEFAULT IS BLANK; THEN SHEAR STIFFNESS IS $
$ INFINITE, I.E., SHEAR FLEXIBILITY IS ZERO. I12 IS THE $
$ AREA PRODUCT OF INERTIA. $
$ $
$ PID MID A I1 I2 J NSM
PBAR 100 1 4.0 .347222 .30 1.0
+PB1F
$ C1 C2 D1 D2 E1 E2 F1 F2
+PB1F 1.0 1.0 1.0 -1.0 -1.0 1.0 -1.0 -1.0
+PB2F
$ K1 K2 I12
+PB2F 0.0
$ $
$ THE MAT1 ENTRY DEFINES THE MATERIAL PROPERTIES. LISTED $
$ ARE ITS ID, ITS ELASTIC MODULUS, SHEAR MODULUS, POISSONS $
$ RATIO, MASS DENSITY, TEMPERATURE EXPANSION COEFFICIENT, $
$ REFERENCE TEMPERATURE AND A STRUCTURAL DAMPING COEFFICIENT. $
$ $
$MAT1 MID E G NU RHO A TREF GE
+MT
MAT1 1 1.44+9 5.40+8 0.0
$
$
$ * * MASS AND INERTIA PROPERTIES * * $
$ $
$ * FUSELAGE MASSES * $
$ $
$ THE CONM2 ENTRY DEFINES A CONCENTRATED MASS. LISTED ARE $
$ ITS ID, GRID LOCATION, COORDINATE SYSTEM TO LOCATE THE $
$ CENTER OF GRAVITY, THE MASS VALUE AND THE LOCATION OF $
$ THE CENTER OF GRAVITY RELATIVE TO THE GRID LOCATION. $
$ $
$ EID G CID M X1 X2 X3
CONM2 97 97 0 3000.0
CONM2 98 98 0 3000.0
CONM2 99 99 0 3000.0
CONM2 100 100 0 3000.0
$ $
$ * * STRUCTURAL PARAMETERS * * $
$ $
$ THE PARAM,GRDPNT,XX ENTRY CAUSES THE GRID POINT WEIGHT $
$ GENERATOR TO BE EXECUTED USING GRID POINT XX AS THE REF- $
$ ERENCE POINT. THEN THE INERTIA MATRIX, THE TRANSFER MATRIX $
$ FROM BASIC TO PRINCIPAL AXES AND OTHER PERTINENT INERTIA $
$ DATA ARE PRINTED. $
$ $
PARAM GRDPNT 90
$ $
$ THE PARAM,WTMASS,GINV CAUSES ALL THE STRUCTURAL MASSES AND $
$ MASS DENSITIES TO BE MULTIPLIED BY GINV (I.E., BY ONE OVER $
$ THE ACCELERATION OF GRAVITY). THE DYNAMIC PRESSURE SUPPLIED $
$ FOR AERODYNAMIC FORCE CALCULATIONS WILL NOT BE MULTIPLIED $
$ BY GINV. $
$ $
PARAM WTMASS .031081
$
$ THE PARAM,AUNITS,GINV PERMITS THE ACCELERATIONS ON THE TRIM
$ ENTRY TO BE SPECIFIED IN UNITS OF LOAD FACTOR (I.E., IN G'S)
$
PARAM AUNITS .031081
$ $
$ * * STRUCTURAL CONSTRAINTS * * $
$ $
$ THE SPC1 ENTRY CONSTRAINS THE LISTED GRID POINTS IN THE $
$ SPECIFIED DOF COMPONENTS. $
$ $
$ SID C G1 G2 G3 G4
SPC1 1 1 90
$ $
$ THE SUPORT ENTRY IDENTIFIES A GRID POINT OR A SCALAR POINT $
$ AND SPECIFIES THE DOF COMPONENTS IN WHICH THE USER DESIRES $
$ REACTIONS TO BE APPLIED TO PREVENT RIGID BODY MOTION. IT $
$ THUS INVOKES THE SOLUTION OF THE BALANCE EQUATIONS TO DETER- $
$ MINE THE REACTIONS. IN THE STATIC AEROELASTIC SOLUTION $
$ THE DOF COMPONENTS MUST BE CONSISTENT WITH THE UNDEFINED $
$ VARIABLES ON THE TRIM ENTRIES. $
$ $
$ ID C
SUPORT 90 23456
$ $
$ THE OMIT1 ENTRY IDENTIFIES GRID POINTS TO BE OMITTED FROM $
$ THE REMAINDER OF THE ANALYSIS. $
$ $
$ ID G G G G
OMIT1 4 110 120 210 220 310
$ $
$ * * * AERODYNAMIC DATA * * * $
$ $
$ (LB-FT-SEC SYSTEM) $
$ $
$ * * ELEMENT GEOMETRY * * $
$ $
$ THE AEROS ENTRY IS UNIQUE TO THE STATIC AEROELASTICITY $
$ SOLUTION, SOL144. ACSID IDENTIFIES THE AERO COORDINATE $
$ SYSTEM, RCSID IDENTIFIES THE REFERENCE COORDINATE SYS- $
$ TEM FOR RIGID BODY MOTION. REFC IS THE REFERENCE CHORD. $
$ REFB IS THE REFERENCE SPAN. REFS IS THE REFERENCE WING $
$ AREA. SYMXZ AND SYMXY ARE SYMMETRY KEYS. $
$ $
$ ACSID RCSID REFC REFB REFS SYMXZ SYMXY
AEROS 1 100 10.0 40.0 400.0
$ $
$ THIS CORD2R ENTRY DEFINES THE AERO COORDINATE SYSTEM $
$ FLAGGED BY THE AEROS ENTRY. THE ORIGIN IS AT THE CANARD $
$ QUARTER CHORD. LISTED ARE THE ORIGIN, A POINT ALONG THE $
$ Z AXIS AND A POINT IN THE X-Z PLANE, ALL IN THE RID $
$ COORDINATE SYSTEM. $
$ $
$ CID RID A1 A2 A3 B1 B2 B3
CORD2R 1 0 12.5 0. 0. 12.5 0. -10.
+CRD1
$ C1 C2 C3
+CRD1 20. 0. 0.
$ $
$ THIS CORD2R ENTRY DEFINES THE NACA COORDINATE SYSTEM TO $
$ WHICH ALL THE STABILITY DERIVATIVES AND TRIM CONDITIONS $
$ WILL BE REFERENCED. $
$ $
$ CID RID A1 A2 A3 B1 B2 B3
CORD2R 100 0 15.0 0.0 0.0 15.0 0.0 -10.0
+CRD100
$ C1 C2 C3
+CRD100 0.0 0.0 0.0
$ $
$ $
$ * WING AERODYNAMIC MODEL * $
$ $
$ THE CAERO1 ENTRY IS USED FOR DOUBLET-LATTICE AERODYNAMICS. $
$ LISTED ARE ITS PAERO ENTRY ID AND THE COORDINATE SYSTEM $
$ FOR LOCATING THE INBOARD AND OUTBOARD LEADING EDGE POINTS $
$ (1 AND 4). NSPAN AND NCHORD, OR LSPAN AND LCHORD, ARE $
$ USED TO PARTITION THE WING INTO AERODYNAMIC PANELS, $
$ THE FORMER FOR UNIFORMLY SPACED PANELS AND THE LATTER $
$ FOR NON-UNIFORMLY SPACED PANELS. IGID IS THE ID OF ITS $
$ ASSOCIATED INTERFERENCE GROUP. THE CONTINUATION ENTRY $
$ DEFINES POINTS 1 AND 4, THE ROOT CHORD AND TIP CHORD. $
$ THE BOXES FORMED BY THE GRID LINES WILL BE NUMBERED $
$ BEGINNING WITH EID SO CHOOSE A NUMBER THAT IS UNIQUE, $
$ AND IS GREATER THAN ALL STRUCTURAL GRID, SCALAR AND $
$ EXTRA POINT IDS. $
$ $
$ * RIGHT WING *
$ EID PID CP NSPAN NCHORD LSPAN LCHORD IGID
CAERO1 1100 1000 8 4 1
+CARW
$ ( FWD LEFT POINT ) CHORD ( FWD RIGHT POINT ) CHORD
$ X1 Y1 Z1 X12 X4 Y4 Z4 X14
+CARW 25. 0. 0. 10. 13.45299+20. 0. 10.
$ $
$ THE PAERO1 ENTRY IS REQUIRED EVEN THOUGH IT IS NON-FUNCTIONAL $
$ (BECAUSE THERE ARE NO ASSOCIATED BODIES IN THIS EXAMPLE). $
$ $
$ PID B1 B2 B3 B4 B5 B6
PAERO1 1000
$ $
$ * LEFT WING *
CAERO1 2100 1000 8 4 1
+CALW
+CALW 13.45299-20. 0. 10. 25. 0. 0. 10.
$ $
$ * CANARD AERODYNAMIC MODEL * $
$ $
$ * RIGHT SIDE *
CAERO1 1000 1000 2 4 1
+CARC
+CARC 10. 0. 0. 10. 10. 5. 0. 10.
$ $
$ * LEFT SIDE *
CAERO1 2000 1000 2 4 1
+CALC
+CALC 10. -5. 0. 10. 10. 0. 0. 10.
$ $
$ * FIN AERODYNAMIC MODEL * $
$ $
CAERO1 3100 1000 4 4 1
+CA1FI
+CA1FI 30.7735 0. 10. 10. 25. 0. 0. 10.
$ $
$ * * SPLINE FIT ON THE LIFTING SURFACES * * $
$ $
$ * BEAM SPLINE FIT ON THE WING * $
$ $
$ THE SPLINE2 ENTRY SPECIFIES A BEAM SPLINE FOR INTERPOLAT- $
$ ION OVER THE REGION OF THE CAERO ENTRY (ID1 AND ID2 ARE $
$ THE FIRST AND LAST BOXES IN THIS REGION). SETG REFERS $
$ TO A SET1 ENTRY WHERE THE STRUCTURAL GRID POINTS ARE $
$ DEFINED. DZ AND DTOR ARE SMOOTHING CONSTANTS FOR LINEAR $
$ ATTACHMENT AND TORSIONAL FLEXIBILITIES. CID IDENTIFIES $
$ THE CORD2R ENTRY THAT DEFINES THE SPLINE AXIS. DTHX AND $
$ DTHY ARE ROTATIONAL ATTACHMENT FLEXIBILITIES (-1. SPECIFIES $
$ NO ATTACHMENT). $
$ $
$ * RIGHT WING * $
SPLINE1 1601 1100 1100 1131 1100 0. TPS
SET1 1100 99 100 111 110 112 121 120
122
$ $
$ * LEFT WING * $
SPLINE1 2601 2100 2100 2131 2100 0. TPS
SET1 2100 99 100 211 210 212 221 220
222
$ $
$ * BEAM SPLINE FIT ON THE CANARD * $
$ $
$ * RIGHT SIDE * $
SPLINE2 1501 1000 1000 1007 1000 0. 1. 1
+SPRC
+SPRC 1. -1.
$ $
SET1 1000 98 99
$ $
$ * LEFT SIDE * $
SPLINE2 2501 2000 2000 2007 1000 0. 1. 1
+SPLC
+SPLC 1. -1.
$ $
$ * BEAM SPLINE FIT ON THE FIN * $
$ $
SPLINE2 3100 3100 3100 3115 3100 0. 1. 300
+SP2FI
+SP2FI -1. -1.
$ $
SET1 3100 99 100 311 312
$ $
$ THE CORD2R ENTRY DEFINES THE COORDINATE SYSTEM IN WHICH THE $
$ BEAM SPLINE EXTENDS ALONG THE WING Y-AXIS. IT LISTS THE $
$ ORIGIN, A POINT ALONG THE Z-AXIS AND A POINT IN THE X-Z $
$ PLANE. $
$ $
$ * RIGHT WING SPLINE AXIS * $
$ $
$ CID CS A1 A2 A3 B1 B2 B3
CORD2R 2 0 30. 0. 0. 30. 0. 10.
+CRD2RW
$ C1 C2 C3
+CRD2RW 38.66025+5.0 0.
$ $
$ * LEFT WING SPLINE AXIS * $
$ $
$ CID CS A1 A2 A3 B1 B2 B3
CORD2R 20 0 30. 0. 0. 30. 0. 10.
+CRD2LW
$ C1 C2 C3
+CRD2LW 38.66025-5.0 0.
$ $
$ * FIN SPLINE AXIS * $
$ $
 CORD2R 300 0 30.0 0. 0. 30.0 10.0 0.
+CRD2FI
+CRD2FI 20.0 0.0 5.7735
$ $
$ * CONTROL SURFACE DEFINITION * $
$ $
$ THE AESURF ENTRY DEFINES AN AERODYNAMIC CONTROL SURFACE. $
$ LISTED ARE THE ALPHANUMERIC NAME OF THE SURFACE, THE ID $
$ OF A COORDINATE SYSTEM THAT DEFINES THE HINGE LINE AND $
$ THE ID OF AN AELIST ENTRY. $
$ $
$ ID LABEL CID1 ALID1 CID2 ALID2
AESURF 505 ELEV 1 1000 1 2000
AESURF 517 AILERON 110 1100 210 2100
AESURF 518 RUDDER 301 3000
$ $
$ THE AELIST ENTRY LISTS AERODYNAMIC BOXES THAT LIE ON THE $
$ CONTROL SURFACE. $
$ $
$ SID E1 E2 E3 ETC
AELIST 1000 1000 THRU 1007
AELIST 2000 2000 THRU 2007
AELIST 1100 1119 1123 1127 1131
AELIST 2100 2103 2107 2111 2115
AELIST 3000 3103 3107 3111 3115
$
@
31724; ; Optimization/Design Sensitivity - OFREQ - Wrong Answers
If sensitivity is requested in a frequency response analysis and there is
an OFREQ request to limit output, the sensitivity results will be in
error.
Avoidance: Remove the OFREQ request from case control.
@
31738; 70.7; Optimization/Design Sensitivity - DRESP1 - Possible Infinite Loo
p
In optimization analysis, if none of the connectivity cards reference
existing property entries and a dresp1 selects an existing property card
for an element response, it is possible to enter into an infinite loop in
subroutine DOMLOC of module DOPR3.

Avoidance: Correct the missing property entries or the incorrect

connectivity references.
@
31751; 2001; Nonlinear Analysis - CBUSH - Wrong Answers
This error points out two problems:
1. A code problem where the CBUSH has a secant logic for load deflection
which can not properly track the curve.
2. A user problem with use of the field 9 CID. The CID sets the element
axes in cement. The element then becomes sensitive to the order of the
grids when computing an internal force. The relation is Ke * (Ub - Ua) If
Ua > Ub a compressive force will result. This is unlike the G0 option
which flips the element axies if the grids are interchanged.
@
31757; 70.5.1; Transient Analysis - TRLG - UFM 2025
The following fatal message may be incorrectly encountered in the TRLG
module:
*** USER FATAL MESSAGE 2025 (TRLG)
UNDEFINED COORDINATE SYSTEM ID = 1065353216
0FATAL ERROR
Avoidance: The error exists in Version 69.1, 70. and 70.5. The best
avoidance is to request MSC/NASTRAN Version 70.5.1. The correction of
this error has been integrated in to Version 70.0.5 for NASTRAN for
Windows. The fixed version remains at 70.5.1 since this error was never an
issue for the prior release of NASTRAN for Windows which was based on
Version 68.2.
@
31759; 2001; Documentation - RELEASE - UFM 6755
The documentation for the RELEASE entry does not state that it is not
supported if the model has PART superelements. If a model containing PART
superelements contains a RELEASE entry, the program will issue UFM 6755
and terminate the run.
@
31763; 70.5.1; Eigenvalue Analysis - EIGRL, READ - SFM 7340 (LNNHERR), LNNDRVD
In the rare case that Lanczos shifts directly on a mode, the new lanczos
shift logic in MSC/NASTRAN Version 70.5 may encounter an error which may
cause the job to fail with messages similar to those below instead of
performing a new shift.
*** USER WARNING MESSAGE 4698 (DCMPD)
STATISTICS FOR DECOMPOSITION OF MATRIX (NONE) .
THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER
THAN
4.85302E+11 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL.
*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
process error reported by subroutine LNNP2CD
see LNNP2CD abstract (ier = -729)
*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
warning reported by subroutine LNNDRVD
see LNNDRVD abstract (ier = 729)
Avoidance: The simplest, most guaranteed avoidance is set system(273)=1
to select the Version 70 Lanczos shift logic.

@
31769; 70.5; Online Encyclopedia - RANDPS - RANDPS is Missing from Index
The RANDPS index is missing from the Table of Content section of the
online documentation. The actual RANDPS entry is in the online CD.
Avoidance: Click on the RANDT1 button and scroll one page up to find the
RANDPS entry.
@
31775; 70.5.1; Utilities - PLOTPS - Fail with UNRECOVERABLE library error
On all Cray platforms, PLOTPS will fail with the following error message
 lib-1171 plotps: UNRECOVERABLE library error
An output list item is incompatible with its data edit-descriptor.
Encountered during a sequential formatted WRITE to an internal file
(character variable)
Error initiated at line 647 in routine '_wrfmt'.
Abort
Beginning of Traceback:
Interrupt at address 74547a in routine '_lwp_killm'.
Called from line 16 (address 75571a) in routine 'raise'.
Called from line 92 (address 15347c) in routine 'abort'.
Called from line 59 (address 352112a) in routine '_ferr'.
Called from line 647 (address 424651c) in routine '_wrfmt'.
Called from line 232 (address 343053c) in routine '_xfer_iolist'.
Called from line 396 (address 423231c) in routine '_FWF'.
Called from line 1483 (address 12245c) in routine 'SYMBOL'.
Called from line 620 (address 4641b) in routine 'DRWBCD'.
Called from line 1225 (address 10656a) in routine 'PLTFIL'.
Called from line 358 (address 3532c) in routine 'PLOTPS'.
Called from line 316 (address 36334b) in routine '$START$'.
End of Traceback.
Abort(coredump)
Avoidance: Change line 1529 in <install-dir>/msc705/util/plotps.F from
50 format(4i6,' s',i2.2)
to
50 format(f6.0,3i6,' s',i2.2)
@
31778; ; Nonlinear Analysis - DECOMP - UWM 4698 Does not Identify Degrees
of Freedom
When running sparse decomposition in solution 106, the message:
USER WARNING MESSAGE 4698, STATISTICS FOR DECOMPOSITION OF MATRIX ***.
THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN
***
OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL"
does not list the grids/dofs corresponding to the negative diagonal terms.

Avoidance: Run the job with SPARSE=1 to identify the negative diagonal
terms.

@
31784; 70.5.2; Nonlinear Analysis - CBUSH, PBUSHT - Missing CBUSH STRESS/FORCE
Output
In a SOL 106 run, if a CBUSH element is present using the nonlinear
force-deflection feature (KN on the PBUSHT entry) and there is no other
nonlinear material present, then the STRESS/FORCE output for the CBUSH
element may be missing. This has been observed on some, but not all
problems.
Avoidance: Add param,lgdisp,1 to the input file. A safer avoidance would
 be to define a dummy rod or quad4 element with
an associated MATS1 entry
@
31787; 70.5.1; Heat transfer Analysis - CHBDYG, CHBDYP - Wrong Answers
If the user has nonunique CHBDYP element ID that are the same as a CHBDYG
element ID, the NASTRAN program does not know which CHBDYx element ID
should be used in the convection calculation. The fix will allow the
program to flag the non-unique CHBDYP and CHBDYG id's in the gp2 module.
Avoidance: Change the conflicting CHBDYP and CHBDYG element id's so that
they are unique.
@
31796; 70.5.1; P-Elements - ADAPT, EPSTOL - Wrong Answers
When a shell or beam p-element is omitted from the error analysis because
of the EPSTOL parameter on the ADAPT bulk data entry, all subsequent shell
and beam p-elements will also be omitted.
Avoidance: An avoidance is to set the EPSTOL parameter to zero on the
ADAPT entry so that no elements are omitted.

@
31801; 70.5.1; P-Elements - ADAPT, EPSTOL - Poor Answers
If a shell or beam p-element has a low stress at one grid, the element may
be omitted from the error analysis because of the EPSTOL parameter on the
ADAPT bulk data entry. The element should not be omitted unless all grids
have a low stress.
Avoidance: An avoidance is to set the EPSTOL parameter to zero on the
ADAPT entry so that no elements are omitted.
@
31804; 70.5.2; Thermal Loads - PSHELL - Wrong Stresses
If the T1 and T2 fields are specified on the TEMPP1 entry and the MID1
field on the PSHELL record is left blank, then the calculated element
stress will be wrong.
Avoidance: Specify the MID1 field on the PSHELL entry.
@
31814; 70.5.1; Matrix Methods - MPYAD Method 1 NT/T Storage 1 - Memory Allocati
on Error
Runs on aix platforms with large multiply-adds, in which Method 1 NT/T
Storage 1 is selected, may unexpectedly run out of virtual memory.
Encountering the error depends on such variables as the size of the
problem and the amount of system paging space available on the machine.
When the error occurs, the run terminates with the following error
message:
SYSTEM ERROR MESSAGE (gotsig33):
SYSTEM RAN OUT OF PAGING SPACE.
Avoidance: Note that when this error is encountered, increasing the
memory allocation with the NASTRAN mem parameter will not generally avoid
the problem. The avoidance is to increase the system paging space.
@
31817; 70.5.2; Dynamics - PARAM,VMOPT,1 - Inefficient Processing
If PARAM,VMOPT,1 is specified then the MGEN module is executed twice.
Also, if component modes are not requested then the fluid mass should not
be computed at the g-set level; in other words, PARAM,VMOPT,1 should be
ignored.
Avoidance: Insert the following DMAP alter:
compile semrm
alter 'mgen '(,-1)
if ( vmopt=0 or not((DYNRED=-1 or METHOD=-1) and noqset>-1) ) then $
alter 'vmopt'(2),''
alter 'call mgen1'(,1)
endif $ NOMGEN>-1
alter 'PURGEX \/mvaa',''
dbstatus mej//s,n,nomej $
if ( nomej>0 ) then $
alter 'mpyad meo'(,-1)
CALL SEGOA uset,GOT,GOQ/GOA/NOOSET/NOQSET/NOTSET $
alter 'mpyad meo'(,1)
else $
purgex /mea,,,,/-1 $
endif $ nomej>0
alter 'ENDIF \$ NOMGEN',''
compile semg
alter 'vmopt'(2),''
type db uset0 $
PARAML CASES//'DTI'/-1/150//S,N,DYNRED $
PARAML CASES//'DTI'/-1/5 //S,N,method $
PARAML USET0//'USET'//////'q'/S,N,NOqSET $
IF ( vmopt=1 and (DYNRED=-1 or METHOD=-1) and noqset>-1 ) THEN $
endalter
@
31832; ; Miscellaneous - PARAM,USETPRT - Wrong Sort
If there is a user defined degree of freedom set (USET, USET1), and row
sorting is selected (PARAM,USETPRT,0 or 10), then the uset print tables
will not be sorted in ascending order, but will be in a random order.
Avoidance: Select 'Column'(val=1 or 11) or 'Row and Column'(val=2 or 12)
sort on the PARAM,USETPRT entry.
@
31842; 70.7; Miscellaneous - INCLUDE, SOL - UFM 443
If the SOL executive statement points to a user-written dmap, and the dmap
is inserted in front of the SOL statement using the include command, then
the job may fail with the following messages on some machine.
*** USER FATAL MESSAGE 443 (DBALOC)
THE SUBDMAP MAIN REFERENCED BY THE SOL, LINK OR COMPILE STATEMENT WAS
NOT FOUND ON THE MSCOBJ DBSET.
Avoidance: Either move the SOL statement before the include command or
insert the dmap into the actual file instead of using the include command.
@
31844; 70.7; Aeroelastic Anaslysis - AEKX, AESMETH=ITER - Slow Convergence
When using AESMETH=ITER or AESMETH=AUTO in SOL 144 and the displacement
and force splines are not the transpose of one another (e.g., if the USAGE
 field is not BOTH on all SPLINEi entries), the iterative aeroelastic
solution may converge slowly or even fail to converge. This is caused by
an error in the AEKX subdmap in which it was mistakenly assumed that the
splines were transposes of one another (as they had been required to be
prior to Version 70.5). If the force and displacement splines are
transposes of one another, this problem does not occur.
Avoidance: An avoidance is to use identical FORCE and DISP splines (which
was a requirement prior to Version 70.5). Another avoidance is to apply
the following alter to Version 70.5:
compile aekx list noref $
$ fix errors due to mqg' and g being different matrices.
$ COMPUTE P = KS*U + MQG*Q*G*U = KS*U + KA*U,
$ AVOIDING COMPUTING OF Ka FOR REASONS OF EFFICIENCY
$ IF TRANSP=1, ka' IS USED IN PLACE OF ka
$
alter '/ksu','/p ' $
if (transp=0) then $
MPYAD G,U,/UK $
MPYAD Q,UK,/PK $ Q IS K BY K AND DENSE
MPYAD MQG,PK,/kau $
else $
mpyad mqg,u,/uk/1 $ ka'*u = g'*q'*mqg'*u
mpyad q,uk,/pk/1 $
mpyad g,pk,/kau/1 $
endif $
MPYAD KS,U,kau/p $ end of aekx.v705 alter
endalter
@
31849; 70.5.1; Submittal - nastran.ini on Windows NT - Application Error
If the last line of the nastran.ini file (found in the
$MSC_BASE\msc705\<arch>\ directory) is not terminated, then the NASTRAN
command will fail with the following message generated by Windows:
An application error has occured and an application error log is being
generated nastran.exe
Exception: access violation (0xc0000005), Address: 0x0041c608

This error is easily created with Notepad.

Note that the ini file is now only used to set a very small number of
keywords, e.g., msc_base, and version. In most other cases the rcf files
are now used. If other keywords are set in the nastran.ini file they are
ignored.
Avoidance: Put an end-of-line at the end of the last line in the
nastran.ini file.
@
31855; ; Documentation - CGAP, PARAM,G - Unreasonably Large Damping
With penalty CGAP (default), W3 damping can become excessive because of
large CGAP penalty stiffness. This is a limitation that should be
documented under the CGAP entry.
 Avoidance: When gap elements are used in nonlinear transient analysis,
param,g damping should be avoided. Damping should be specified on the
MATi entries and PARAM,W4 set.

@
31884; 70.7; Optimization/Design Sensitivity - ADD5, PARAM, DSNOKD - UFM 5423
The message "USER FATAL MESSAGE 5423" indicating an incompatible matrix
multiple is output from the ADD5 module in SOL 200 when PARAM DSNOKD is
non-zero and there are multiple subcases that have active buckling design
constraints.
Avoidance: Set DSNOKD to 0.0 or restrict the buckling design task to a
single subcase.
@
31885; 70.7; Rigid Elements - RBE3 - UWM 7343
Diagnostic USER WARNING MESSAGE 7343 for the RBE3 element appears,
properly, when an RBE3 element has inadequate connectivity and is
approaching singularity. It may also appear for a stable element when
there is a large distance between the grid points to which it attaches.
This is because the test used to measure singularity is not independent of
the units of measure of the model. A test model of an rbe3 connecting 4
grid points in a square pattern to a point at their center does not cause
the message when the points are 100 units apart, but does cause the
message when they are 10,000 units apart on certain computer types.
Avoidance: The criterion used to print this message is that the
computational zero ("error value") associated with its use is greater than
1.E-10. This conservative value was chosen because a related fatal
message (new in Version 70.5) and client feedback had not been obtained
with its use. If the error value is less than 1.E-6 it may be safely
ignored. The warning message can generally be avoided if the grid points
to which the element is attached are moved, or one is added, which allows
them to approach a description of a plane, rather than a condition
approaching a line. If the grid points are widely separated the only
method known to avoid the message is to use a different unit of measure
for grid point locations.

@
31890; 70.5.2; Rigid Elemets - RBE3 - UFM 2038, Only one illegal RBE3 found per
run
All bad rigid elements should be found and identified in a single
execution. However, the RBE3 is an exception. It causes the program to
stop after the first bad element is found.
This is a particular problem with Version 70.5 because improved logic for
detecting singular RBE3 elements was introduced in this version, and the
program may fail RBE3 elements that earlier versions passed.
Avoidance: The error has been corrected in Version 70.5.2. Another
avoidance is to run the model in Version 70 with param, checkout yes
present. This makes tests on the model including equibrium tests on the
rigid element constraint matrix. Elements which will pass tests in
Version 70 but fail with singularity messages in Version 70.5 are likely
to show poor equilibrium properties in the param, checkout tests. Look
for output of the EMH matrices. It will identify dependent DOFs that do
 not meet equilibrium checks, which may be tied back to RBE3 elements. All
elements which fail will be identified in Version 70. This technique will
not work in Version 70.5.0 because the program will take a fatal exit
before the checkout operations begin.
@
31891; 2001; Documentation - GPSTRESS - No Printout for Midside Nodes
MSC/NASTRAN does not printout GPSTRESS for midside nodes. This limitation
is documented in the Application Manual but since that manual is no longer
current a remark should be added to the Quick Reference Guide.
@
31898; 70.5; Frequency Response - FREQ - UWM 6498
In a frequency response analysis run (SOL 108 and 111), if the FREQ Case
Control command does not exist, then the job will terminate with the
following unclear error messages on some machines.
*** USER WARNING MESSAGE 6498 (MsgGetStr)
Message number is not found
PROGRAMMER INFORMATION:
Check call to MSGPOP from FRLG with the message number 3031
Severity is USER WARNING
Parameter 1 is 0
Parameter 2 is FRL
Parameter 3 is FRRD
USER ACTION: Please contact MSC Client support
0FATAL ERROR
On other machines, the job may terminate with a signal 11 error in the log
file.
Avoidance: Add the appropriate FREQ case control command.
@
31906; 70.5.2; Nonlinear Analysis - DBC - UFM 4551
If a nonlinear analysis fails to converge and PARAM,POST,0 is set, the job
will issue fatal messages indicating that the analysis has not converged
but then will continue and will fail in the DBC module with a machine
dependent message. An example of the messages are:
*** USER INFORMATION MESSAGE 6194,
*** STOPPED ITERATIONS DUE TO REACHING MAXIMUM ITERATION LIMIT
WITHOUT CONVERGENCE ***
*** USER INFORMATION MESSAGE 6193,
*** MAXIMUM NUMBER OF BISECTIONS OR MINIMUM LOAD STEP HAS BEEN
REACHED.
*** USER FATAL MESSAGE 4551,
*** STOPPED PROBLEM DUE TO FAILED CONVERGENCE
^^^ USER INFORMATION MESSAGE 9005 (NLSTATIC)
^^^ THE SOLUTION FOR THE LAST CONVERGED LOOPID= 85 IS SAVED FOR
RESTART
***** RECEIVED SIGNAL 8, ABORTING ...
BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX)
= 0
STATUS = 0
*** SYSTEM FATAL ERROR 4276.RECOVR ERROR CODE 7777 PID= 0
*** USER INFORMATION MESSAGE: TO OBTAIN A NASTRAN DUMP, RESUBMIT JOB WITH
 DIAG 44 INSERTED IN THE EXECUTIVE CONTROL SECTION.
On other computers the f06 file ends abruptly after UFM 4551, the f04 file
ends abruptly after a DBC call and a system message occurs in the log file
such as the following from a SUN computer:
*** TERMINATING /app/msc/msc70/solaris/analysis
*** Received signal 11 (SIGSEGV)
time: command terminated abnormally.

Avoidance: The failure in dbc does not affect the rest of the run. The
analysis needs to be modified to achieve converge.
@
31919; 70.7; Installation - Sun/Solaris - fatal : libsunmath.so.1: can't open
file errno = 2
For Sun SuperSPARC V70.5 and v70.6 and Sun UltraSPARC V70.0.3 Clients
only, If you do not have a Fortran compiler on your system, you may
receive the following error message in the LOG file when you attempt to
run MSC/NASTRAN:

ld.so.1: <install-dir>/msc705/solaris/analysis:
fatal : libsunmath.so.1: can't open file errno = 2
This error occurs when the system cannot find the "libsunmath" dynamic
library.
This error is corrected on the MSC/NASTRAN V70.5.1 (r2) CD-ROM.
You can also obtain the missing libraries from your MSC representative or
directly from MSC's Web site at the URL:

ftp://ftp.macsch.com/pub/msc-products/NASTRAN/V70.5/solaris/libsunmath.so
1
Once you have obtained the file, please copy it to the solaris
architecture directory, i.e.,
<install-dir>/msc705/solaris
where <install-dir> is your MSC/NASTRAN V70.5 installation base
directory.

Avoidance:
Download the libsunmath.so.1 file from the ftp site.
@
31931; 70.6; Acoustics - ACMODL, CTETRA, CTRIA3, GP5 - SFM 6169
Some models with TRIA3 structural elements and TETRA fluid elements at the
fluid/structure interface using IDENT method on the ACMODL bulk data entry
fail in the GP5 module with a message similar to:
*** USER FATAL MESSAGE 6169, FOUND TWO OR MORE STRUCTURAL GRID POINTS
WITH IDENTICAL COORDINATES
WHICH LIE ON THE FLUID/STRUCTURE BOUNDARY.
USER ACTION: CHECK GRIDS AND/OR SET1 CARDS WHICH
BELONG TO PANEL OR ACMODL CARDS.
 GRID ID = 12
GRID ID = 101979
^^^ USER WARNING MESSAGE 9031 (ERRPH1) - NOGO ENCOUNTERED IN SUBDMAP
SEMG
This message is in error since one of the referenced structural grids is
really a fluid grid. Abitrary changes to the model can make it run or not
run.
Avoidance:
Use the DIFF method on the ACMODL bulk data entry.
@
31933; ; Documentation - MATGEN Option 5 - DMAP Module Dictionary
The documentation for MATGEN option 5 states that the parameter p5 may be
used to select pseudo-random or non-repeating random numbers, but the
equation used is incorrect. It states that if P5<0, a matrix of random
numbers that do not repeat from call to call is produced. By implication,
values outside this range produce pseudo random numbers. The usual
definition of this term is that while the numbers produced should have
reasonable distributions for a set of random numbers, the matrix produced
on each call should be the same, both on one run, and on different runs.
The correct formula is "if p5 is less than or equal to 0" the
non-repeating ("seeded") type of random numbers are used. Note that this
changes the action for the default value of p5, which is zero.

@
31934; 2001; Documentation - RFORCE - Clarification of the CD Field
The content description of the CID and R1, R2, R3 fields for the RFORCE
entry currently states CID must be a Caresian system. It should state
that if a non-rectangular system is specified, RFORCE will use the CID as
if it were a rectangular system and unexpected answers may result. The
desrcription of CID, R1, R2, R3 should read as follows and remark 14
should be added.
CID Coordinate system used to define the components of the r
otation
vector.
(integer >= 0, Default = 0)
R1, R2, R3 Rectangular components of the rotation vector R. The ve
ctor
will pass through G,
please see remark 14. (R1**2 + R2**2 + R3**2 > 0.0)
Remark 14. If CID is not a rectangular coordinate system, RFORCE will
treat it as if it were a rectangular system
when defining the components of the rotation vector.
@
31958; ; Memory Management - REIGL, GPFDR - SFM 4275 and SFM 4276
On a large model, it has been observed that the job may fail in REIGL or
GPFDR with the following unclear messages:
 *** SYSTEM FATAL MESSAGE 4275. DURING A READ OPERATION, UNIT = 26,
DATA BLOCK=NUMERIC
, A BLOCK NUMBER CHECK CONDITION OCCURRED. ( IONAST
CODE= 304)
0*IONAST* err 304 pid,name,b+1,z4(b+1),nxtblk,pbknbr= 0 118
62402415 0 3377 5912
BIOMSG: ERROR 304 HAS OCCURRED IN ROUTINE IONAST , FILE INDEX (FILEX)
= 0
STATUS = 0
The exact cause of this error is currently unknown. MSC has been unable
to obtain the input file that shows this error due to the confidentiality
of the input file.
Avoidance: Increasing memory cured this particular problem.
@
31966; 70.7; Nonlinear Analysis - CBUSH, CELAS1, CELAS2, MATHP - UWM 6286 in
ZREAD
If nonlinear stress output is requested for CBUSH, CELAS1, and CELAS3
elements which reference PBUSHT and PELAST in nonlinear
analysis then UWM 6286 will be issued. This also occurs for hyperelastic
element grid point output.
Avoidance: Insert and/or replace the following changes into nddl.dat
under DATABLK OES:
OR,40,{ Bush1d }
EITHER,BACK(FCODE)=0,{ Real }
FE,RS,{Element Force}
UE,RS,{Axial Displacement}
VE,RS,{Axial Velocity* }
S,RS,{Axial Stress*}
E,RS,{Axial Strain*}
EP,RS,{Plastic Strain*}
FAIL,I,{Failed Element Flag}
OR,1,{ Real / Imaginary }
FER,RS,{Element Force}
UER,RS,{Axial Displacement}
SR,RS,{Axial Stress*}
ER,RS,{Axial Strain*}
FEI,RS,{Element Force}
UEI,RS,{Axial Displacement}
SI,RS,{Axial Stress*}
EI,RS,{Axial Strain*}
OR,216,{ Nonlinear form of hyperelastic 4 NODE TETRA }
TYPE,CHAR4,{ }
ENTRY,
ID,I,{ }
SX,RS,{ }
SY,RS,{ }
SZ,RS,{ }
SXY,RS,{ }
SYZ,RS,{ }
SZX,RS,{ }
PRESSURE,RS,{ }
VOLSTR,RS,{ }
EX,RS,{ }
 EY,RS,{ }
EZ,RS,{ }
EXY,RS,{ }
EYZ,RS,{ }
EZX,RS,{ }
ENDENTRY,COUNT=004,
OR,217,{ Nonlinear form of hyperelastic 3 NODE Strain }
TYPE,CHAR4,{ }
ENTRY,
ID,I,{ }
SX,RS,{ }
SY,RS,{ }
SZ,RS,{ }
SXY,RS,{ }
PRESSURE,RS,{ }
VOLSTR,RS,{ }
EX,RS,{ }
EY,RS,{ }
EZ,RS,{ }
EXY,RS,{ }
ENDENTRY,COUNT=003,
OR,218,{ Nonlinear form of hyperelastic 20 NODE }
TYPE,CHAR4,{ }
ENTRY,
ID,I,{ }
SX,RS,{ }
SY,RS,{ }
SZ,RS,{ }
SXY,RS,{ }
SYZ,RS,{ }
SZX,RS,{ }
PRESSURE,RS,{ }
VOLSTR,RS,{ }
EX,RS,{ }
EY,RS,{ }
EZ,RS,{ }
EXY,RS,{ }
EYZ,RS,{ }
EZX,RS,{ }
ENDENTRY,COUNT=008,
OR,219,{ Nonlinear form of hyperelastic 9 NODE Strain }
TYPE,CHAR4,{ }
ENTRY,
ID,I,{ }
SX,RS,{ }
SY,RS,{ }
SZ,RS,{ }
SXY,RS,{ }
PRESSURE,RS,{ }
VOLSTR,RS,{ }
EX,RS,{ }
EY,RS,{ }
EZ,RS,{ }
EXY,RS,{ }
ENDENTRY,COUNT=004,
OR,220,{ Nonlinear form of hyperelastic 15 NODE }
TYPE,CHAR4,{ }
ENTRY,
ID,I,{ }
SX,RS,{ }
 SY,RS,{ }
SZ,RS,{ }
SXY,RS,{ }
SYZ,RS,{ }
SZX,RS,{ }
PRESSURE,RS,{ }
VOLSTR,RS,{ }
EX,RS,{ }
EY,RS,{ }
EZ,RS,{ }
EXY,RS,{ }
EYZ,RS,{ }
EZX,RS,{ }
ENDENTRY,COUNT=006,
OR,221,{ Nonlinear form of hyperelastic 10 NODE }
TYPE,CHAR4,{ }
ENTRY,
ID,I,{ }
SX,RS,{ }
SY,RS,{ }
SZ,RS,{ }
SXY,RS,{ }
SYZ,RS,{ }
SZX,RS,{ }
PRESSURE,RS,{ }
VOLSTR,RS,{ }
EX,RS,{ }
EY,RS,{ }
EZ,RS,{ }
EXY,RS,{ }
EYZ,RS,{ }
EZX,RS,{ }
ENDENTRY,COUNT=004,
OR,222,{Nonlinear form of hyperelastic 3 NODE Axisym }
TYPE,CHAR4,{ }
ENTRY,
ID,I,{ }
SX,RS,{ }
SY,RS,{ }
SZ,RS,{ }
SXY,RS,{ }
PRESSURE,RS,{ }
VOLSTR,RS,{ }
EX,RS,{ }
EY,RS,{ }
EZ,RS,{ }
EXY,RS,{ }
ENDENTRY,COUNT=003,
OR,223,{ Nonlinear form of hyperelastic 9 node Axisym }
TYPE,CHAR4,{ }
ENTRY,
ID,I,{ }
SX,RS,{ }
SY,RS,{ }
SZ,RS,{ }
SXY,RS,{ }
PRESSURE,RS,{ }
VOLSTR,RS,{ }
EX,RS,{ }
EY,RS,{ }
 EZ,RS,{ }
EXY,RS,{ }
ENDENTRY,COUNT=004,
OR,224,{ Nonlinear ELAS1 }
F,RS,{ Force }
S,RS,{ Stress }
OR,225,{ Nonlinear ELAS3 }
F,RS,{ Force }
S,RS,{ Stress }
OR,226,{ Nonlinear BUSH }
FX,RS,{ Force X }
FY,RS,{ Force Y }
FZ,RS,{ Force Z }
STX,RS,{ Stress Translational X }
STY,RS,{ Stress Translational Y }
STZ,RS,{ Stress Translational Z }
ETX,RS,{ Strain Rotational X }
ETY,RS,{ Strain Rotational Y }
ETZ,RS,{ Strain Rotational Z }
MX,RS,{ Moment X }
MY,RS,{ Moment Y }
MZ,RS,{ Moment Z }
SRX,RS,{ Stress Rotational X }
SRY,RS,{ Stress Rotational Y }
SRZ,RS,{ Stress Rotational Z }
ERX,RS,{ Strain Rotational X }
ERY,RS,{ Strain Rotational Y }
ERZ,RS,{ Strain Rotational Z }
@
31970; ; Superelements - CSUPEXT Ignored if no Elements Exist in a Supere
lement
If there are no elements inside a superelement, but there are interior
points, and CSUPEXT entries are used to define the exterior points, the
CSUPEXT entries will be ignored.
Avoidance: Add PLOTEL elements between interior and exterior points.
@
31980; 70.5.1; Nonlinear Analysis - NLCOMB - Traceback
If a problem has BEAM, BAR, or BEND elements and thermal loads in a
nonlinear solution sequence (106 and 129), the job may fail in module
NLCOMB. This error first occurred in Version 70.5. This has been observed
only on the Cray, but might occur on other systems. On the CRAY the
following traceback was encountered:
Beginning of Traceback:
Called from line 42 (address 15517150b) in routine 'TRBK'.
Called from line 39 (address 132053d) in routine 'RECOVR'.
Called from line 112 (address 140623d) in routine 'sysrcv'.
Called from address 15237006a in routine '__handlr'.
Called from address 5616165c in routine 'NLCWRK'.
Called from line 126 (address 5607145c) in routine 'NLCOMB'.
Called from line 90 (address 10632226d) in routine 'XSEM15'.
Called from line 363 (address 1734c) in routine 'MAINAL'.
Called from line 306 (address 3647a) in routine '$START$'.
End of Traceback.
 Avoidance: There is no known avoidance.
@
31982; 70.6; SSSALTERs - checka.v70 - UFM 13
In the checka.v70 and checka.v705 alters, there are two lines which exceed
80 columns. If these alters are pasted into the data file, the line will
wrap in many editors, and cause the following fatal message:
0*** USER FATAL MESSAGE 13 (XGPIDG )
ERROR IN DMAP INSTRUCTION Q-SET INSTRUCTION NO. 50
DMAP INSTRUCTION NOT IN MODULE LIBRARY.
Avoidance: Shorten the comments following the DMAP instructions "type
db,aagds,ptga $", and "upartn uset0,aagds/ptga,,,/'G'/'A'/'S'/1 $".
Another option is to avoid pasting the file in, but instead use the
include statement.

@
31986; 70.5.2; Superelements - DBDN DBset Too Large
The DBDN DBset is larger in V70 and V70.5 than in earlier versions of
MSC/NASTRAN. The increased size is due to several data blocks which are
generated during dynamic reduction being incorrectly stored on DBDN. The
only effect of this error is an increase in the DBDN Dbset disk usage.
Avoidance: Insert the following DMAP alter:
compile semr3
alter 'call cmpmode'
type parm,nddl,char8,n,dbupx,dbdnr1 $
type parm,,char8,n,dbtmp $
dbtmp=dbdnr1 $
dbdnr1=dbupx $
alter 'call sesum'
dbdnr1=dbtmp $
compile sekr
alter 1
type parm,nddl,char8,n,dbupx,dbdnr1 $
type parm,,char8,n,dbtmp $
dbtmp=dbdnr1 $
dbdnr1=dbupx $
alter 'end \$'(,-2)
dbdnr1=dbtmp $
compile semrm
alter 'mvaa'(2,-1)
type parm,nddl,char8,n,dbupx,dbdnr1 $
type parm,,char8,n,dbtmp $
dbtmp=dbdnr1 $
dbdnr1=dbupx $
alter 'end \$'(,-2)
dbdnr1=dbtmp $
endalter
@
31990; 2001; Documentation - PBUSH1D
The MSC/NASTRAN Quick Reference Guide for Version 70.5 incorrectly states
that SHOCKA, CVT, IDTS, IDETS, IDETSD, TYPE, IDT, IDTTU, IDTDV are
required on the PBUSH1D entry. These fields may be required based on
 values in other fields and options selected, but are not mandatory.
@
32011; 2001; Documentation - PBUSH - Element Property Direction Not Described
The documentation for PBUSH does not say which coordinate system the
"directions 1 through 6" refer to. These directions refer to the element
coordinate system.

@
32014; 70.5.2; Transient Analysis - NLRGAP, TRD1 - Poor Answers, Traceback
Input files using the NLRGAP in SOL 109 may fail with a floating point
exception on NT platforms or may issue a traceback in the TRD1 module,
routine tabld on SGI R8000/R10000 systems. The error is caused by the
table look up for the coefficient of friction for the NLRGAP. All 32 bit,
short-word, machines will return an incorrect value for the coefficient of
friction; however, this is not likely to have a significant effect on the
results of most problems.
Avoidance: Run the problem in solution 129. A different error with
similar symptoms may also be encountered in the same area of the code,
please see error number 32685.
@
32015; 2001; Documentation - Release Guide - NLRGAP Does Not Work in Solution
112.
The MSC/NASTRAN Release Guide for Version 70.5 incorrectly states that the
NLRGAP element works in SOL 112. The NLRGAP does not work in this
solution, but it does work in SOL 9.
@
32016; 2001; GAP Constraints - PARAM,CDITER - UWM 9096, Wrong Answers
If the iterative solution for GAP constraints in SOL 101 does not
converge, a warning message is issued (9096), indicating that the solution
has not converged; however, the results for displacements, etc. are
still output.
^^^ USER WARNING MESSAGE 9096 (CSTRDISP)
^^^ CONSTRAINED DISPLACEMENT ITERATIONS FAILED, ITERATION LIMIT LOAD CASE
= 1
Avoidance: If warning message 9096 is issued, do not use the results.
Increasing the value of CDITER may help the problem to converge.
@
32022; 70.5.1; Miscellaneous - NLRGAP - Incorrect TPL Problem
The tpl test deck nlrgap.dat referred is not the problem described in the
Version 70.5 Release Guide. The tpl deck nlrgap.dat has been replaced
with the correct input deck on the Version 70.5.1 r2 cd. A listing of the
correct input follows:
$ shaft1x.dat
$ radial gap, .002 clearance, contact stiffness = 1.e6 constant
$ example: rotating load on cantilevared shaft
$ 40" shaft, circular x-section, 2" diameter shaft
$ 10" diameter rotor
$ 15 hz. imbalance
 sol 129
time 80
cend
title= 15 hz. rotating imbalance, radial gap
set 1 = 9,10
set 2 = 1,8
set 3 = 1001,1002
$sdisp=3 $ epoint displacements (relative displacements)
disp(plot) = 1$ NLRGAP grid displacements
force(plot) = 2
nlload=1 $ NLRGAP load output
nonlinear = 200 $ NLRGAP set ID
tfl=900 $ transfer function ID for relative displacement check
dload = 1
tstepnl = 21 $ 0 - .4 seconds
output(xyplot)
xgrid= yes
ygrid= yes
xtitle = Time (seconds)
ytitle = x relative displacement epoint 1001
xyplot sdisp /1001(t1)
ytitle = y relative displacement epoint 1002
xyplot sdisp /1002(t1)
ytitle = X Nonlinear loads Grid 9
xyplot nonlinear /9(t1)
ytitle = Y Nonlinear loads Grid 9
xyplot nonlinear /9(t2)
$ytitle = RZ Nonlinear loads Grid 9
$xyplot nonlinear /9(r3)
$
begin bulk
param coupmass1
param grdpnt 0
param,post,-1
param,w4,419.0
$
$ dynamic loads simulating 15 hz. 3 lb. centripetal imbalance
dload 1 1. 1. 10 1. 20
$ x direction
darea 40 9 1 3.0
darea 41 9 2 3.0
$ cosine imbalance
tload2 10 40 0.0 100.0 15.0
$ sine imbalance
tload2 20 41 0.0 100.0 15.0 -90.0
$ 0-.03 s
tstepnl 21 300 0.0001 1 adapt
+
+ +
+ 0
$MAT1 MID E G NU RHO A TREF GE
mat1 100 3.0e7 .3 .7-4 1.0
$PBAR PID MID A I1 I2 J NSM
$pbar 1 100 1.0 .0491 .0491 .0982
+
$+ .5 0. 0. .5 -.5 0. 0. -.5
pbar 1 100 3.14 .8 .8 1.6
+
+ 1. 0. 0. 1. -1. 0. 0. -1.
$CBAR EID PID GA GB X1,G0 X2 X3
cbar 1 1 1 2 1.
cbar 2 1 2 3 1.
cbar 3 1 3 4 1.
cbar 4 1 4 5 1.
cbar 5 1 5 6 1.
cbar 6 1 6 7 1.
cbar 7 1 7 8 1.
cbar 8 1 8 9 1.
$
grid 1 0.0 0.0 0.0 123456
grid 2 0.0 0.0 5.0
grid 3 0.0 0.0 10.
grid 4 0.0 0.0 15.
grid 5 0.0 0.0 20.
grid 6 0.0 0.0 25.
grid 7 0.0 0.0 30.
grid 8 0.0 0.0 35.
$ shaft grid and rotor mass
grid 9 0.0 0.0 40.
$ 10" diameter, 1" thick disk
conm2 109 9 .0057
.036 .036 .071
$ Housing grid and housing mass
grid 10 0.0 0.0 40. 345
conm2 110 10 1.
1.
$ ground point for housing
grid 11 0.0 0.0 40. 123456
$CELAS2 EID K G1 C1 G2 C2 GE S
celas2 10 1.0+6 10 1 11 1 100.
celas2 11 1.0+6 10 2 11 2 100.
celas2 12 1.0+6 10 6 11 6 100.
$
$*****************************************************************
$NLRGAP SID GA GB PLANE TABK TABG TABU RADIUS
nlrgap 200 9 10 XY 961 965 967 5.0
$
$ Constant contact stiffness vs time
tabled1 961
0. 1.+6 10. 1.+6 endt
$
$ Constant gap clearance vs time
tabled1 965
0. 0.002 10.0 0.002 endt
$
$ gap friction vs time, positive value assumes rotation
$ about dof 3 from axis 1 toward axis 3
tabled1 967
0. 0.1 10. 0.1 endt
$*****************************************************************
$ gap dof relative displacements for verification
epoint,1001
epoint,1002
$
tf 900 1001 1.0
10 1 1.0
9 1 -1.0
tf 900 1002 1.0
10 2 1.0
 9 2 -1.0
enddata
@
32047; 70.7; Optimization/Design Sensitivity - Wrong Answers with Material Da
mping
Material damping defined on a material bulk data entry is ignored in an
adjoint sensitivity analysis.
Avoidance; If damping is a constant across all materials, param,g may be
used. Otherwise the adjoint method needs to be avoided by forcing the
direct sensitivity method using the following alter:
COMPILE DESCON
alter 'dsad'(,-1) $
type parm, ,logi,n,hold $
hold = pexist $
pexist = true $
alter 'dsad'
pexist = hold $
@
32070; 70.5.2; Differential Stiffness - SOLs 103, 107-112 - Wrong answers
If QSETi, SEQSETi, or SENQSET entries are used to compute component modes
in SOLs 103 and 107-112 and differential stiffness is requested with the
STATSUB Case Control command, then the differential stiffness will not
include the effects of the q-set degrees-of-freedom.
Avoidance: Insert the following DMAP alter which is valid for Version
70.5 only:
compile sekdr
$ this alter is for Version 70.5 only
alter 'call segoa',''
type db goq $
CALL SEGOA USET,GOT,goq/GOAT/NOOSET/NOQSET/NOTSET $
endalter
@
32073; 70.5.2; Superelements - Partitioned SE - Wrong Answers or Fatal Messages
If scalar points are defined in a partitioned superelement and the scalar
points are also defined as exterior to the superelement (e.g.; QSETi) then
the run may produce wrong answers and/or issue a fatal message. This
error can also occur using the SENQSET entries.
Avoidance: Insert the following DMAP alter:
compile phase0
alter ',SILS,VGFS'
delete /vgfs $
alter ',SILS,VGFS'(2)
delete /vgfs $
endalter
@
32074; 70.5.2; Superelements - Data Recovery for Externals in SOLs 110-112 - UW
M 9098
 Performing data recovery using external superelements in SOLs 110, 111,
and 112 with SEBULK type external superelements will result in user
warning message 9098 indicating that the data necessary to perform data
recovery is not available. The run will then fail in an unexpected manner.
Avoidance : Include the following DMAP alter in the run which performs
the residual structure solution
compile super3
alter 'call extdrout'(4),''
if ( extdrout<>' ' ) then $
if ( app1='mmreig' ) then $
call extdrout uls,ol1m,eqexins,maps/
dum/app1/' '/true $
call extdrout uh2,ol1r,eqexins,maps/
dum/app/soltyp/true $
else $
call extdrout uls,ol1m,eqexins,maps/
dum/app1/soltyp/true $
endif $
endif $ extdrout<>' '
endalter $

@
32080; 70.7; Complex Eigenvalue Analysis - CEAD - SFM 4276, EC 1605
On the SGI, complex Lanczos eigenvalue analysis may produce the unexpected
system fatal message 4276 error code 1605. This appears to be caused by
insufficient memory for the CEAD module.
Avoidance: In most if not all cases experienced to date, increasing the
memory has allowed the runs to run successfully.
@
32087; 70.7; Frequency Response - TLOADi - Wrong answers
TLOADi entries are ignored in direct frequency response (SOLs 8, 11, and
108).
Avoidance: Insert the following DMAP alter in SOL 108:
compile sedfreq
alter 'CASE CASErs'
type parm,,logi,n,fourier $
paraml dynamics//'bulk'//////'tload1'/s,n,tload1/
'tload2'/s,n,tload2 $
PARAML CASEF//'DTI'/-1/38//S,N,TSTEP $
fourier=(tstep=-1 and (tload1=-1 or tload2=-1)) $
alter 'call freqrs',''
CALL freqrs CASEF ,USETD ,DLT ,FRL ,trl ,
rpd ,DIT ,EST ,cstms ,mpts ,
bgpdts ,gpsnts ,GPECT ,SILS ,edts ,
ckdd ,cmdd ,cbdd ,k4dd , ,
, , ,sild ,ditid ,
vafs ,rgdfd ,rsgg ,rmgg ,ppf ,
fol , ,vgfd ,vgfs /
UDF ,qmgf ,qsgf ,pdf /
soltyp /freqdep /noa /fourier /false /
1 /false /0. /dompcffd/dospcffd/
fs /noue /0 $
 alter 'call vdr1'(,-1)
IF ( Fourier ) then $ Fourier transform
delete /ppf,psf,,, $
app='tranresp' $
endif $
compile freqrs
alter 'ift '(,-1),''
IF ( not(delta) and Fourier ) THEN $
endalter
Insert the following DMAP alter for SOL 8 or 11, making sure to specify
"SOL8" or "SOL11", respectively, on the COMPILE statement:
COMPILE SOL8 or SOL11
alter 'lblgo'(,-1)
paraml dynamics//'bulk'//////'tload1'/s,n,tload1/
'tload2'/s,n,tload2 $
PARAML CASExx//'DTI'/-1/38//S,N,TSTEP $
IF ( tstep=-1 and (tload1=-1 or tload2=-1) ) THEN $
type parm,,i,y,iftm=0 $
IFT Uhv,CASExx,TRL,olst/
UHT1,TOL1/IFTM $
EQUIVX TOL1/olst/-1 $
EQUIVX UHT1/Uhv/-1 $
delete /pp,ps,pd,, $
soltype='tranresp' $
ENDIF $
endalter
@
32091; 70.7; Nonlinear Analysis - PCOMP, STRESS, STRAIN, PARAM,POST,0 - Missi
ng XDB Output
If PARAM,LGDISP,>0 and PARAM,POST,0 is specified in SOL 106, then laminar
stress and strain output is missing from the .xdb file.
Avoidance: Insert the following DMAP alter:
compile nlstatic
alter 'oescp,oestrcp'
file oescpr=append/oestrcpr=append $
append oescp,/oescpr/2 $
append oestrcp,/oestrcpr/2 $
alter 'OUTPUT2 oesnlxr'
IF ( POST=0 ) then $
DBC OEScpr,,,,,,,,,,,,,,,,,,,//
'OESC'/'CASECC'/'OL'//////////////////
-1/DBCPATH/S,N,CP/app//GEOMU/LOADU/POSTU/
DBCDIAG/DBCCONV/DBCOVWRT $
DBC OEStrcpr,,,,,,,,,,,,,,,,,,,//
'OESC'/'CASECC'/'OL'//////////////////
-1/DBCPATH/S,N,CP/app//GEOMU/LOADU/POSTU/
DBCDIAG/DBCCONV/DBCOVWRT $
endif $ post=0
endalter
@
32093; 70.5.1; Aeroelastic Analysis - CSSCHD - UFM 311 (IFPDRV)
If a subcase invokes multiple CSSCHD bulk data entries, the user fatal
message shown below incorrectly states that the CSSCHD ID's must be
 unique.
*** USER FATAL MESSAGE 311 (IFPDRV)
NON-UNIQUE FIELD 2 ON BULK DATA ENTRY CSSCHD 10
SORTED ENTRY COUNT = 153
Avoidance: There is no avoidance other than to limit the scheduled
control surface functionality to a single control surface for each
subcase.
@
32094; 70.7.2; Optimization/Design Sensitivity - Nonlinear Trim Analyis - BIOMS
G: ERROR 100
Version 70.5 added a basic capability to perform nonlinear static
aeroelastic trim analysis. This capability is not available in SOL 200
for sensitivity analysis and its use results in a nondescriptive fatal
message such as
BIOMSG: ERROR 100 HAS OCCURRED IN ROUTINE PACK
Avoidance: Either linearize the static aeroelastic trim analysis or do not
attempt to perform this type of sensitivity analysis. Version 70.7.2 has
been corrected to provide a meaningful message.
@
32098; 2001; Documentation, QRG - PSHELL, remark # 14 - typographical error
There is a typographic error for Remark 14 of the PSHELL entry in the
Version 70 and 70.5 Quick Reference Guide. The term (Ts)10**8(Ts) should
simply be 10**8 instead.
@
32099; ; Postprocessing - PARAM,POST,-2 - Multiple Geometry Written to OU
TPUT2 File
If multiple boundary conditions are specified in a normal modes analysis
and param,post,-2 is set, the geometry data blocks are written to the
output2 file in each subcase rather than just once. This unnecessarily
increases the size of the output2 file.
Avoidance: Specify PARAM,OGEOM,NO in all but the last subcase or insert
the following DMAP alter:
compile super3
alter 'call out2geom'(,-1)
type parm,nddl,i,n,ncase $
if ( ncase<=1 ) then $
alter 'call out2geom'
endif $
endalter
@
32101; 70.7; Plotting - XYPLOT of Grid Point Stress in SOL 101 - UWM 979
If an XYPLOT of grid point stress (param,curvplot,1) in solution 101 will
lead to the following message:
*** USER WARNING MESSAGE 979 (XYTRAN)
AN XY-OUTPUT REQUEST FOR POINT OR ELEMENT ID xxx
-STRE- CURVE IS BEING PASSED OVER. THE ID COULD NOT BE FOUND IN DATA
BLOCK
Avoidance: Insert the following alter:
 COMPILE SEDRCVR
$ALTER 'SORT2\=','' $ for V70
ALTER 'SORT2\='(2),'' $ for V70.5
SORT2=NOT(ANDL(NOSORT2,NOSTR2) OR CURVPLOT>-1) $
ENDALTER
@
32112; 70.7.2; Miscellaneous - MSCACT Missing Summary Output
The MSCACT utility delivered with MSC/NASTRAN Version 70.5 on the UNIX and
Windows NT systems has an error that disables the "standard" output
summary. If one of the alternate summaries is selected via the "summary"
keyword, the correct detail lines will be generated, but the total line
will be missing.
This error can be corrected by changing the following lines in the
<install-dir>/msc705/util/mscact.c source file and rebuilding the MSCACT
utility as described in Section 6 of the "Configuration and Operations
Guide" for you operating system.
NOTE: This error does not affect the accounting logger functions.

The changes are:

Change line 1339 from
if( total == -1 && !strcmp( p, TOTAL_LABEL ) )
to
if( total == EOL && !strcmp( p, TOTAL_LABEL ) )
Change line 1341 from
else if( null == -1 && !strcmp( p, NULL_LABEL ) )
to
else if( null == EOL && !strcmp( p, NULL_LABEL ) )
Change line 1386 from
if( null > -1 )
to
if( null != EOL )
Change line 1390 from
if( total > -1 )
to
if( total != EOL )

@
32113; 2001; Miscellaneous - DMIG, DMI - Poor Messages
The messages encountered when there is an error on a DMIG or DMI entry do
not point to the location of the error. This can make it almost impossible
to locate the error in a large matrix input.
@
32116; 70.5.1; Miscellaneous - ISHELL - Program Called by ISHELL Not Found
If the program called by the ISHELL module has a name with fewer than 8
characters without including the extension, the module will not search
properly for the extension. For example, the file named "EFM.EXE" will be
found if the name "EFM.EXE" is entered, but will not be found if the name
"EFM" is entered.
 Avoidance: Use 8 characters in the filename if you do not include the
extension.
A related error that may occur on Windows NT if a 16-bit executable is to
be run. If ISHELL processing fails with an "UNABLE TO CREATE A NEW
PROCESS" error, specifying the standard NT command processor as the
processor to be used may solve the problem. That is, setting the
MSC_ISHELLEXT environment variable to exe#cmd, either in the
System/Environment Control Panel tab or by a command line set option ("set
MSC_ISEHHEXT=exe#cmd") will usually correct the problem, especially if the
program can be run from a command prompt. Note: The MSC_ISHELLEXT
capability will only be available starting with MSC/NASTRAN V70.5.1 (R2
CD).

The short name problem has been corrected in MSC/NASTRAN V70.5, Round 11
@
32125; 70.5.1; Submittal - DBINIT - Long filenames on SGI
If a very long filename is used, either via the job datafile pathname, the
dbs or out keywords, or an ASSIGN statement, the job may fail without any
messages in the F06, F04, or LOG files. A clue to this error may be the
termination of the job in the DBINIT module (as indicated in the F04 file)
or incomplete reports for the "SUMMARY OF FILE ASSIGNMENTS" or "SUMMARY OF
PHYSICAL FILE INFORMATION" or "SUMMARY OF PHYSICAL FILE I/O ACTIVITY"
tables.
Avoidance: Reduce the lengths of the longer pathname components. A
pathname component is either a directory name (i.e., the text between
slashes) or the actual filename part.
@
32127; 70.5.1; Aeroelastic Analysis - AELINK, CSSCHD - Wrong Answers
If AELINK and CSSCHD entries appear is the same subcase in a static
aeroelastic analysis, wrong answers are produced.
Avoidance: There is no avoidance.
@
32134; 70.5.1; Installation - SGI IRIX64 Only
When trying to run Version 70.5 on SGI IRIX 64 systems the following error
message may appear in the command window:

3708:/msc/msc705/irix64/analysis: rld: fatal error: cannot successfuly map

soname 'libcomplib.sgimath.so' under any of the filenames < then a list of
library files follows >.
Avoidance: If the client does not have a complete set of sgi math
libraries they will need to get a /usr/lib64/mips4/libcomplib.sgimath.so
from MSC. This can be found at the following locations:
On the web at:
ftp.macsch.com/pub/msc-products/NASTRAN/V70.5/irix64
filename: libcomplib.sgimath.so
or the following computer file on MSC's net:
 /net/oberon/oberon/disk1/msc705/unix/msc705/irix64/libcomplib.sgimath.so
@
32135; 70.6; Complex Eigenvalue Analysis - CEAD - Missing Left Hand Vectors
The new CEAD module option for the CLAN method in Version 70.5 will
produce both left and right hand eigenvectors. However, if there are less
than three DOFs in the problem only the right hand vectors are output.
Avoidance: An avoidance is to add a third DOF, uncoupled, if necessary.
@
32136; 2001; Documentation/QRG - PARAM,CURVE - Incorrect Equation
The documentation for PARAM,CURV in the Quick Reference Guide indicates
that if:
| n dot lm | >= .4, then the projection of the y-axis on the surface of
the element is taken as the reference axis.
The left hand side of the inequality should be squared.
( n dot lm)**2 >= .4.
@
32143; ; P-Elements - TETRA - Incorrect Mass Matrix
The mass for TETRA p-elements at p-level 1 is incorrectly calculated.
Avoidance: Run the analysis at a p-level of 2 or higher when using TETRA
p-elements are present.
@
32145; 70.5.2; Transient Analysis - TRD1 - Received Signal 10
Transient analysis may fail in the TRD1 module with no message. The log
file may contain a machine dependent message of the form:
*** Received signal 10 (SIGBUS)
time: command terminated abnormally
Avoidance: The error is machine dependent and is caused by a boundary
alignment problem. The error will only occur on HP, SUN and SGI systems
(Digital will print warning messages but will complete successfully). The
error should be avoided by increasing memory by exactly 2 words on HP and
SUN system and by 6 words on SGI systems.

@
32157; ; Online Documentation - Installation - Not a Valid WorldView Docu
ment
The MSC/NASTRAN Online Encyclopedia issues a "Not a valid WorldView
document:" error message when any hotlink button on a book selected from
the main books menu is picked. Iview configuration file entries for
fulltextPath and welcomeDoc are valid and the documents could be selected
manually from the Open/File menu picks.

Resolution/Avoidance:
This problem could occur on any system where:
 1) the mne collections are moved from one location to another
2) the mne installation directory is nfs mounted or linked to a
differently named location
As part of the installation process, several *.cfg files are created on
the fly and locations are hardcoded with pathname of the installation
directory. If the mne collection directory is changed, it is not
sufficient to update the Iview*fulltextPath list in the Iview
configuration file - these *.cfg files must be edited also.
Example 1: Moving mne collection location
-MNE initially installed to /disk1/msc/mne705.
-The entire collection is later copied/moved over to /disk2/msc/mne705
and /disk1/msc/mne705 is removed.
-The Iview*fulltextPath list in the Iview configuration file is updated
to reflect the disk1 -> disk2 change.
Avoidance: edit the *.cfg files to accomdate the new location (disk1 ->
disk2)

Example 2: Installation dir not same as nfs mount point

-MNE installed on server1:/disk1/msc/mne705
-That directory is nfs mounted on client2 as /msc/mne705.
-The Iview*fulltextPath list in the Iview configuration file on client2
is updated to use the "/msc/mne705" path.
Avoidance: Use a path that will be common to ALL workstations that will
nfs mount the collection (e.g. /msc/mne705)
AND IS ALSO ON THE SERVER NODE (/msc/mne705 should also exist
there). Edit the *.cfg files to update to the
common location.
The *.cfg files that must be updated are:
docmenu/docmenu.cfg v705all/v705all.cfg v70all/v
70all.cfg
v69all/v69all.cfg v705prim/v705prim.cfg v70prim/
v70prim.cfg
v69prim/v69prim.cfg recomend/recomend.cfg

@
32168; 70.5.2; Matrix Methods - DCMP - SFM 4276 EC 7777
It is possible that a job that should fail in DCMP with a max ratio error,
identifying the problem degrees of freedom, instead fails with a traceback
in sparse decomposition. The following is an example of the traceback
from an hpux system.
Process traceback ...
( 0) 0x0009700c sysrcv + 0x320
[/cae_k_apps/msc/msc70/hpux/analysis.pa2_0]
( 1) 0x00f83310 _sigreturn [/cae_k_apps/msc/msc70/hpux/analysis.pa2_0]
( 2) 0x005efba0 dfmsj_ + 0x130
[/cae_k_apps/msc/msc70/hpux/analysis.pa2_0]
( 3) 0x0022b7b4 dfmsa_ + 0x2b20
[/cae_k_apps/msc/msc70/hpux/analysis.pa2_0]
 ( 4) 0x004c355c sdcomp_ + 0xb8
[/cae_k_apps/msc/msc70/hpux/analysis.pa2_0]
( 5) 0x00220eb8 dcmpa_ + 0x2b8
[/cae_k_apps/msc/msc70/hpux/analysis.pa2_0]
( 6) 0x000c0fe8 dcmp_ + 0x130
[/cae_k_apps/msc/msc70/hpux/analysis.pa2_0]
( 7) 0x0009232c xsem07_ + 0x5c4
[/cae_k_apps/msc/msc70/hpux/analysis.pa2_0]
( 8) 0x0008ecb4 mainal_ + 0xa0c
[/cae_k_apps/msc/msc70/hpux/analysis.pa2_0]
Avoidance: If it is possible run the analysis on another machine to
identify the problem degrees of freedom and then correct the modeling
problems. Running the deck with sparse decomposition tuned off by setting
SPARSE=1 on the NASTRAN command may also allow the problem degrees of
freedom to be identified.
@
32171; 70.7; Composites - PARAM,POST,0 - Missing output
Laminar stress and strain output is not written to the xdb file in SOL 103
and 105.

Avoidance: Insert the following DMAP alter:

compile sedrcvr
alter 'DBC OSTR1'
endIF $ SORT2
alter 'endIF $ SORT2'(3),''
endalter
@
32174; 70.5.1; Installation - xnastran not delivered
The XNASTRAN utility is missing from the Version 70.5 (R1) CD. The
executable file, <install-dir>/msc705/<arch>/Xnastran, can be copied from
any earlier version.
@
32176; 2001; Documentation - EIGR - Point Normalization
In modal analysis (sol 103), when POINT normalization is selected on the
EIGR bulk data entry, and the defined component is zero, NASTRAN does not
revert to MAX as specified in the Quick Reference guide for the EIGR bulk
data entry:
NORM Method for normalizing eigenvectors. (Character: "MASS," "MAX,"
or
"POINT"; Default+"MASS")
.
.
POINT Normalize to a positive or negative unit value of the component
defined in fields 3 and 4
(defaults to "MAX" if defined component is zero).
Documentation of NORM POINT should read:
.
.
POINT Normalize to a positive or negative unit value of the component
defined in fields 3 and 4
 (defaults to "MASS" if defined component is zero).
@
32177; 2001; Documentation - QRG, INIT FMS - Incorrect Default values
Under the description of the INIT command in the Quick Reference Guide,
the default values for DBALL and SCR300 are stated incorrectly as 25000
and 175000 blocks. The correct default value is 250000 blocks for each of
the DBSETs.
@
32193; 70.6; Nonlinear Analysis - CELAS1, CELAS3, PELAST, NLITER - SFM 2085
If the CELAS1 or CELAS3 element references the PELAST entry in a nonlinear
analysis where the TKNID field is used to define the nonlinear force vs.
displacement relationship, the job will fail with the following message:
*** SYSTEM FATAL MESSAGE 2085 (NLITER)
ELEMENT TYPE 11 NOT SUPPORTED FOR NONLINEAR ANALYSIS.
*** SYSTEM FATAL MESSAGE 3007 (NLITER)
ILLEGAL INPUT TO SUBROUTINE NLEMG
Avoidance: Use the CBUSH entry and define the nonlinear force vs.
displacement relationship through the KN field on the PBUSHT entry.
@
32218; 2001; Optimization/Design Sensitivity - CBUSH, DVGRID - SFM 3007
In shape optimization, if the CBUSH element is present, (even when not
referenced by a DVGRID), the job fails with the following fatal message,
"element type 102 is not supported *** SYSTEM FATAL MESSAGE 3007 (DOPR6)
ILLEGAL INPUT TO SUBROUTINE Dopr6a".
Avoidance: Don't use the CBUSH element in a shape optimization run.
@
32220; ; Transient Analysis - TSTEP - UFM 4562
In a modal transient analysis, the following error messages may be
erroneously issued after the completion of the job.
0*** USER FATAL MESSAGE 4562,
TSTEP (TIME STEP) DATA IS MISSING
Avoidance: This message can be safely ignored if all the requested output
is obtained.
@
32225; 2001; Documentation - QRG, System Cell 206 - Missing Option
Option 9 of system cell 206 which selects the better of METIS and MMD, is
not documented in the Version 70.5 Quick Reference Guide.

@
32259; 70.7; Aeroelastic Analysis - Incorrect Answers for Unrestrained Hinge
Moment Coeff.
Two errors have been detected that relate to the printing of unrestrained
hinge moment coefficients in SOL 144 (static aeroelasticity):
1. The unrestrained hinge moment coefficient values printed for the
"INTERCEPT" coefficients are in error if they are nonzero.
 2. If there are multiple subcases with the same Mach and q (dynamic
pressure) conditions, the unrestrained hinge moment coefficient values
printed for the second and subsequent subcases are in error. The
coefficients for the first subcase are deemed correct with the exception
of the INTERCEPT value mentioned previously.
Avoidance: Make the following DMAP change:
compile aestatrs list
alter 'totu','kbrzx'
dbstatus ipzf//s,n,noipz $
if ( noipz=-1 ) matgen ,/ipzf/7/norset/1 $
add ipzf,/mipzf/-1.0
append mipzf,kr1zx/totu $
alter 'kbr1zx','urddi'
solve mirr,totu,,,/urddi//-1 $
DBSTATUS M4rr//S,N,NOM4rr $
IF( NOM4RR=-1 ) THEN $
MPYAD M1RL,AMLR,/M3RR//-1 $
FBS L2RR,U2RR,M3RR/M4RR $
endif $ nom4rr=-1
@
32264; 69; Parallel Processing - EMA - Wrong Answers
Due to an error in parallel EMA, any system running parallel EMA (even if
they are only using one processor), could get wrong answers.
Avoidance: Use the serial version.
@
32282; 70.7.2; Utilities - XNASTRAN - xmonast Started with the Wrong File Names
The XNASTRAN utility does not properly handle an alternate input data file
type, e.g., "example.bdf", when directly starting XMONAST to monitor a
job. It will attempt to directly start XMONAST on the files
"example.bdf.f04", "example.bdf.f06", and "example.bdf.log". Instead, it
should be setting the nastran command's "xmon" keyword and relying upon
the nastran command to correctly monitor the job.
Avoidance: Set "xmon=yes" in an RC file or as an "advanced argument" and
do not use XNASTRAN's ability to start XMONAST.
@
32286; 70.6; Superelement - STATSUB, multiple subcases - SFM 6710 (TAFF)
When more than one static subcase exists in the bulk data for forming the
differential stiffness matrix for a normal modes analysis (STATSUB), and
the model contains superelements, then the job will fail with the
following unclear messages:
*** USER INFORMATION MESSAGE 7325 (SEP2DR)
PROCESSING OF SUPERELEMENT 10 IS NOW INITIATED.
^^^ USER INFORMATION MESSAGE 9049 (SEKDR)
^^^ SUPERELEMENT DIFFERENTIAL STIFFNESS GENERATION, ASSEMBLY, AND
REDUCTION.
*** SYSTEM FATAL MESSAGE 6710 (TAFF)
THE THIRD INPUT PARAMETER TO MODULE TAFF REFERENCES AN UNDEFINED LOAD
SET ID.
0FATAL ERROR
This job will fail even if there is only one static subcase callout, but
 there are multiple loadings in the bulk data section that are not used.
Avoidance: There are two possible avoidances:
1. Do not use superelements
or
2. Only keep a single static and a single normal modes subcase and remove
the extra loading entries from the bulk data section.
@
32294; 70.5; Miscellaneous - ISHELL - Insufficient Memory on CRAY
Use of the ISHELL utility on a CRAY computer may cause the job to fail due
to insufficient memory.
Avoidance: Specify twice the memory that the job would normally take.

@
32296; ; Differential Stiffness - STATSUB, Superelements - UFM 5423 (MPYA
D)
In SOL 105, or if differential stiffness is requested (STATSUB Command)
in SOLs 103 and 107 through 112, when superelements are present and no
SUPER command, or SUPER=ALL, is specified, then static loads will be
generated for the second or dynamic subcase and the job will fail with the
following error messages.
*** USER FATAL MESSAGE 5423 (MPYAD)
ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES
COLS ROWS FORM
TYPE NZWD DENS
TRAILER FOR GOAT IS 106 12 50 2
2 60 4333
TRAILER FOR UAS IS 103 1 12 2
2 8 3333
TRAILER FOR UOOX IS 104 2 50 2
2 40 2000
TRANSPOSE FLAG IS 0
Avoidance: Specify a separate subcase for the upstream superelements.

@
32303; ; Nonlinear Transient Heat Transfer - SUBCASE, SET - Missing Outpu
t
In nonlinear transient heat transfer (SOL 159), if different subcases
request different outputs, only the request for the first subcase will be
honored and no output will be obtained for subsequent subcases. In the
following example thermal results for grid 1 and 2 will be output for
subcase 1 but no thermal output will be obtained for subcase 2.
SUBCASE 1
TSTEPNL = 100
DLOAD = 200
SET 1 = 1,2
THERMAL =1
SUBCASE 2
 TSTEPNL = 100
DLOAD = 200
THERMAL = ALL

Avoidance: The only avoidance is to request the same output for all
subcases above the first subcase. In the above example, request
THERMAL=ALL above the first subcase.
@
32320; 70.7; Inertia Relief - PARAM,ALTRED,YES - Wrong Answers
If PARAM,INREL,-1 and PARAM,ALTRED,YES are specified then wrong answers
are produced.
Avoidance: Insert the following DMAP alter:
compile phase0
alter 'LFIXEDB OR INREL\=-1',''
IF ( LFIXEDB ) ALTRED='NO' $
alter 'SEID\=0 and (LFIXEDB',''
IF ( SEID=0 and (LFIXEDB or cditer>0)
) INREL0=false $
compile sedisp
alter 'MPYAD LAO,UA,UOX','IF ( FIXEDB\>-1 ) CALL SEGOA,'
if ( inrel0 ) then $
IF ( SEID=0 ) THEN $
IF ( inrel=-1 ) then $
MPYAD UOx,CRX,/UOXx $
else $
equivx uox/uoxx/-1 $
endif $
else $
MPYAD UOx,CRj0,/UOXx $
endif $
else $
equivx uox/uoxx/-1 $
endif $
MPYAD LAO,UA,UOXx/UOY/1 $
FBS LOO,,UOY/UOO///-1 $
ENDIF $ NOOSET>=0
endif $
IF ( APP<>'CYC' ) THEN $
alter 'MPYAD UOO,CRX,\/UOOX',''
if ( altred<>'yes' ) then
MPYAD UOO,CRX,/UOOX $
else $
EQUIVX UOO/UOOX/ALWAYS $
endif $
alter 'MPYAD UOO,CRJ0,\/UOOX',''
if ( altred<>'yes' ) then
MPYAD UOO,CRJ0,/UOOX $
else $
EQUIVX UOO/UOOX/ALWAYS $
endif $
alter 'sdr1'(3,-1)
IF ( altred<>'yes' and FIXEDB>-1 ) CALL SEGOA,
uset,GOt,/GOAt/NOOSET/NOQSET/NOTSET $
endalter
@
32321; 70.7; Static Analysis - PARAM,ALTRED,YES and MPCFORCE - Missing output
 MPCForce output is not computed if PARAM,ALTRED,YES is specified.
Avoidance: Insert the following alter:
compile sedisp
alter 'sdr1'(3,1)
ENDIF $ ALTRED
alter 'ENDIF \$ ALTRED',''
endalter
@
32361; ; Online Documentation, MNE - Links in DMAP Module Dictionary Do N
ot Work
In the Version 70 MSC/NASTRAN Online Encyclopedia, the chapter contents
links in the first page of each chapter of the DMAP Module Dictionary do
not work.
Avoidance: Use the main TOC in the quick access bar located at the top of
every page.
@
32362; 2001; Documentation, Interface Elements - GMBDC, GMBNDS - DMAP Fatal M
essage 9050
The Quick Reference Guide bulk data definitions for the GMBNDC and GMBNDS
entries needs the following remark added:
The interface elements may generate high or negative matrix/factor
diagonal ratios. If there are no other modelling errors, these messages
may be ignored and PARAM,BAILOUT,-1 may be used to continue the run.

@
32364; 70.5.2; Postprocessing - TABPRT, OUTPUT2 - Wrong Answers, SWM 1204
Incorrect data may be output by a TABPRT and/or an OUTPUT2 request using
the neutral option. Many occurences of SYSTEM WARNING MESSAGE 1204 may
be an indication of a problem.
Avoidance: For the OUTPUT2 neutral option, increase mem such that the
value is at least
f2 x size of largest record + 2 x nasbuff + compiled length of the NDDL of
that datablock
For TABPRT NDDL-output, increase mem such that the value is at least
4 x size of largest record + 1 x nasbuff + compiled length of the NDDL
of that datablock
@
32365; 2001; Documentation/QRG - SESUP - Problems with free-free Dynamic Redu
ction
Models with free-free component modal superelement reduction give poor
answers (or may fail). This limitation should be documented under the
SESUP entry by adding the following remark:
5. SESUP bulk data records are not allowed for Part (partitioned bulk
data) Superelements. Use the SUPORT bulk data records to identify
component rigid body modes.
@
32375; 70.5.2; P-Elements - Transverse Shear Strain Output Mis-Labeled
The p-shell element's two transverse shear strain outputs for the default
curvature option are switched. Otherwise the values are correct.
Avoidance: There is no avoidance.
@
32376; 2001; Documentation/P-Elements - OUTRCV - Missing P-Shell Output
When the PROJ=F option is used on the OUTRCV bulk data entry for p-shell
elements, no output is produced in the .f06 file. The option should be
used if a postprocessor requires results in terms of 3D vectors or
tensors, for example, stress tensors with 6 stress components. The option
does not produce output in the print file *.f06.
@
32388; ; Nonlinear Transient Analysis - TSTEPNL, ADD5 - UFM 4276 GETSTR E
RROR CODE 31
If a nonlinear transient analysis has trouble converging and output is not
requested for every timestep, the job mail fail with one of the following
type messages in ADD5.
0*** SYSTEM FATAL ERROR 4276.NXTCOL ERROR CODE -31 PID= 0 ->
CELAS2
0*** SYSTEM FATAL ERROR 4276.GETSTR ERROR CODE 31 PID= 0 ->
CBUSH1D
Avoidance: Request output for every timestep.
@
32390; 70.7; Cyclic Symmetry - READ, REIGL - UIM 6137, SFM 3001
The READ module may issue the message: UIM 6137 INPUT MATRIX TO
DECOMPOSITION HAS xx SINGULARITIES, followed by *** SYSTEM FATAL MESSAGE
3001 (READ). This is due to null columns in the matrices. For cyclic
symmetry AUTOSPC does not remove unused degrees of freedom if they are for
grid points listed on CYJOIN records.
Avoidance: For solid element models, add a bulk data record
GRDSET,,,,,,456 to remove unused rotation degrees of freedom. Setting
NASTRAN SYSTEM(273)=1 (Use earlier version Lanczos method) has been shown
to work as well in NAST705. The following ALTER may be applied to version
70.5:
COMPILE MODERS
ALTER 'MATMOD *CASES'
SETVAL //S,N,METHTYP/0 $
MESSAGE //'DO A-SET REDUCTION FOR MASSLESS DEGREES OF FREEDOM'
@
32394; 70.5.1; Complex Eigenvalue Analysis - EIGC - SFM 3008 (UTMCOR)
The new block complex lanczos method in Version 70.5 (selected by setting
system(108)=4) does not accurately predict the memory requirement. For
large problems this may lead to the following fatal message:
 *** SYSTEM FATAL MESSAGE 3008 (UTMCOR)
MODULE = CEAD
INSUFFICIENT CORE AVAILABLE FOR SUBROUTINE CLDRV
USER ACTION: INCREASE OPEN CORE BY AT LEAST 1130847 WORDS
PROGRAMMER INFORMATION: LOCATION KEY = TSLSPILL
Avoidance: The fatal message may not be able to be avoided by increasing
memory. The analysis may have to be performed with another method.
@
32395; 70.5.2; Eigenvalue Analsysis - EIGRL, READ - SFM 7340 (LNNHERR)
The new lanczos in MSC/NASTRAN Version 70.5 may fail with some variation
of the following message:
*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
process error reported by subroutine LNNDRVS
see LNNDRVS abstract (ier = -1)

The message that should have been output is:

FACTORIZATION ERROR DURING FINITE INTERVAL ANALYSIS AT FLEFT. FLEFT
= -3.9478D-05
*** SYSTEM FATAL MESSAGE 5299 (LNCS1D)
FACTORIZATION ERROR AT A BOUNDARY SHIFT.
POTENTIAL FIXES:
IF LOWER BOUND IS ZERO USE A SMALL NEGATIVE VALUE.
IF LOWER BOUND IS A SMALL NEGATIVE VALUE DECREASE IT.
ELIMINATE MASSLESS MECHANISMS THE MODEL MAY CONTAIN.
Avoidance: Follow the direction of SFM 5299 above.
@
32411; 70.7; Aeroelastic Analysis - APD, RESTART LOGICAL - UFM 1245
If a read-only restart (RESTART LOGICAL FMS statement) is performed in
SOLs 144, 145, or 146, then the run will fail in the APD module with the
following message:
0*** USER FATAL MESSAGE 1245, (BIO)
AN ALLOCATION ERROR HAS OCCURRED FOR A DATABASE FILE, FILX = 10.
WRITE ATTEMPTED ON FILX WHICH WAS OPENED FOR READONLY.
Avoidance: Use RESTART with the COPY option.
@
32426; 70.7; P-Elements - GP0 - Distorted Element Edges May Lead to Traceback
If a p-element edge is too distorted, the program may fail with the
traceback that follows in GP0 instead of exiting with a meaningful
message.
SIGTRAP - Trace trap FP division by zero
Traceback:
Offset 0x00001380 in procedure evlpnd_
Offset 0x00000c54 in procedure gp0dc_
Offset 0x00001710 in procedure gp0da_
Offset 0x00000cd4 in procedure gp0_
 Offset 0x000004b0 in procedure xsem02_
Offset 0x000009dc in procedure mainal
--- End of call chain ---
Avoidance: Other than correcting the distorted geometry, there is no
avoidance.
@
32438; 70.5.2; Submittal - Wrong executable may be selected on HP-UX 11
Apparently HP will make both 32- and 64-bit versions of HP-UX 11.0
available. As a result of this, if the client is running the 32-bit
version of HP-UX 11.0 on a PA-RISC 2.0 system, the nastran command will
incorrectly choose the V-class executable.
This error can only occur with MSC/NASTRAN V70 and MSC/NASTRAN V70.5 (r1).
The bit size of the OS is determined with the following command:
$ getconf KERNEL_BITS
The command will print 64 on a 64-bit kernel, e.g., on a V-Class system;
and 32 on a 32-bit kernel.
Avoidance: Set the following keyword in the either system (e.g.,
<install-dir>/conf/nast705rc) or node (e.g.,
<install-dir>/conf/net/<node>/nast705rc) RC files.
On a PA-RISC 1.1 system:
proc=
On a PA-RISC 2.0 system:
proc=pa2_0
The processor type can be determined using the nastran command, e.g.,
msc705 nastran system
@
32457; 70.5.2; P-Elements - Beam Error Estimator - Poor Answers, Inefficient Pr
ocessing
The error estimators for the beam p-element are incorrect by a factor
based on the length of the beam. This error may lead to over adapting or
under adapting. That is the error may lead to poor answers or it may
lead to an unnecessarily high p-level.
Avoidance: There is no avoidance.
@
32458; 70.6; ESTIMATE - P-Elements - Insufficient Memory
MSC Estimate underestimates memory for input files with P-elements. This
limitation will be removed with version 71 for P-element runs without
interface elements (GMINTC/GMINTS). The limitation still exists for runs
with interface elements.
Avoidance: Do not use estimate to determine memory and disk requirements
of runs with p-elements prior to Version 71, and do not use estimate to
determine memory and disk requirements of any run with interface elements.
@
32467; 70.5.5; Parallel Processing - Incorrect CPU Times
When parallel processing is selected, the f04 per processor cpu times may
show negative cpu times on Version 70.5 NT systems. This problem has been
corrected in the Version 70.5.5 release.
@
32476; ; Utilities - RCOUT2, BGPDT - rcout2 utility does not transfer dat
a correctly
The rcout2 utility will not transfer the BGPDT correctly on short-word
UNIX systems.

@
32477; 70.6; Estimate - SPC/SESET THRU 9999999 - performance
If an input file has an SPC entry specifying 1 THRU 9999999, then estimate
will take a lot of CPU time. This performance error will be removed in
version 71.
Avoidance:
Specify the last grid id instead of 9999999.
@
32482; 2001; DMAP - DRMS1 - SFE 4276 EC 1305 in ENDPUT.
The DRMS1 module may fail with SFE 4276 EC 1305 in ENDPUT.
Avoidance: Reduce the number of subcases (or modes) in the input data
blocks to the DRMS1 module.
@
32489; ; Aeroelastic Analysis - AEROF - Incorrect Aerodynamic Box ID's
In SOL 146 for Harmonic Gust (Frequency Response), the AEROF=<setid|ALL>
option is supported. However, if the ALL option is used in Version 70.5,
the f06 output listing incorrectly lists the Aerodynamic Box ID's as 0
(zero) for all boxes. Whereas, if a specific SET is used, the correct
results are obtained. The complex forces printed in both cases, are
correct.
@
32490; 70.7; Inertia Relief - UPARTN, PARAM,INREL,-1 - Poor Message, UFM 3007
If PARAM,INREL,-1 is specified without any SUPORTi entries then the run
will fail in the UPARTN module with the message:
*** SYSTEM FATAL MESSAGE 3007 (UPARTN)
ILLEGAL INPUT TO SUBROUTINE PARTN
Avoidance: Remove PARAM,INREL,-1 or specify SUPORTi entries.
@
32495; 70.6; Nonlinear Statics - PARAM,NMLOOP,-1 and POST,0 - Missing .xdb Ou
tput
If PARAM,NMLOOP,-1 and PARAM,POST,0 are specified along with multiple
subcases in SOL 106, then only the eigensolution results for the last
subcase are written to the .xdb file.
Avoidance: Insert the following DMAP alter:
 compile nlstatic
alter 'call super3'(2),''
CALL SUPER3 CASECC ,casei ,PHA ,OLB , ,
, ,olb , , ,
, , , , ,
,PCDB ,XYCDB ,POSTCDB ,FORCE ,
EMAP ,MAPS ,EQEXINS ,PVTS ,CASES ,
SLT ,ETT ,GOT ,GOQ , ,
, , , ,GPLS ,
USET ,SILS , ,EDTs , ,
GM , ,KFS ,KSS ,QR ,
CMPHO ,CMLAMA ,CMPHA ,MAR ,MEA ,
DYNAMICS,BGPDTS ,CSTMS ,MPTS ,DIT ,
EST ,GEOM2S ,GEOM3S ,ECTS ,EPTS ,
INDTA ,KELM ,KDICT ,GPECT ,VELEM ,
,ditid ,pvtbc ,scstm , ,
,GEOM4S , , , ,
,EPT ,MPT ,GEOM1 ,GEOM2 ,
GEOM3 ,GEOM4 ,PTELEM ,MELM ,MDICT ,
casei , , , , ,
,gpsnts ,pvt ,DEQATN ,DEQIND ,
, , ,KSF ,KMG ,
MMG ,MSF , , , ,
, , , , ,
, , , , ,
geom1s , , , ,ugni ,
/
PHG ,PHQG ,phqmg ,LAMA1 , ,
, , , , ,
, , , /
APP /APP /RSONLY /FALSE /FALSE /
FALSE /0 /FALSE /S,PFILE /S,CARDNO/
0 /FALSE /0 /FALSE /0 /
FALSE /false /false /0 /false /
' '/false /false $
endalter
@
32496; ; Documentation - PSDF, ATOC, RALL - Random Requests not yet Suppo
rted in PATRAN
The documentation for the new random analysis data recovery request
options PSDF, ATOC and RALL for DISP, VELO, ACC, FORCE, and STRESS
documented in the Version 70.5 Quick Reference Guide is misleading. These
options are not yet supported in PATRAN and thus are of no use and should
not be selected. Presently if they are selected, the random output is
written to output2 file and no random output is written to the f06 file.
Currently, nothing is written to the xdb graphics data base as might be
implied by the Quick Reference Guide remarks. Furthermore, if
PARAM,POST,-1, is set, and the OUTPUT2 file is FORMATTED and the new
format of the output requests for RANDOM results is used, such as
DISP (PSDF) = all
then the run will issue UWM 1206 indicating that the datablocks created by
the RANDOM module are not written to the OUTPUT2 file.
Avoidance: Currently the avoidance is simply do not use these output
requests since they are not supported in PATRAN.
@
32502; ; Optimization/Design Sensitivity - Superelements - UFM 9068, Poor
Message
If superelements are used in design optimization and a DRESP2 entry
references DRESP1 entries which select responses in different
superelements, the run may fail with UFM 9068 incorrectly stating that
there are no DRESP1 or DRESP2 entries for the DESOBJ selected. The program
should issue a message indicating that the DRESP1 entries incorrectly
spans superelements.
Avoidance: Remark 2 in the Quick Reference Guide description of the
DRESP2 entry states that DRESP1 entries referenced on a DRESP2 may not
span superelements. However, the error message issued is misleading.
Simply place all of the referenced items into a single superelement.
@
32504; ; Restarts - TEMPD - Inefficient Processing
If a bulk data entry used to define a temperature loadings
(TEMP,TEMPD,etc) is changed on a restart, the restart logic will delete
the stiffness matrix and create a new one. This results in inefficient
restarts.
Avoidance: Either:
1) Place all TEMPi entries in the initial input file, or
2) Insert the following DMAP in the restart run;
$ DMAP alter to remove checking for changed in temperature
$ BE CAREFUL WHEN USING THIS, AS IT REMOVES SOME OF THE RESTART LOGIC
compile ifpl
alter 'restart.*geom3,geom3r',''
compile phase0
alter 'restart.*geom3s,geom3sr',''
@
32514; 2001; Documentation - LOADSET - Only One LOADSET Command is Allowed pe
r Superelement
The Quick Reference Guide description of the LOADSET command does not
explicitly state that only one LOADSET command may be used for each
superelement (or, in the case of no superelements, only one per run). This
command should occur in the first SUBCASE for each superelement (or above
the SUBCASEs). Any LOADSET commands in subsequent SUBCASEs will be
ignored.
@
32522; 70.5.2; External Superelements - CSUPER, SPOINTS - Wrong Answers
If an external superelement is brought in using the CSUPER entry and
SPOINTs are used for some of the attached dof, it is possible that the
matrices will be attached incorrectly. In the problem demonstrating this,
if all GRID points were listed before the SPOINTs on the CSUPER, the
matrices were attached correctly. If the SPOINTs were listed, then the
GRID points, then the matrices were incorrectly attached. (NOTE - the
matrices are attached in the order listed on the CSUPER and care must be
exercised to verify that the correct order is listed - do not simply
change the order on the CSUPER to correct this problem)
Avoidance: If possible, use only GRID points on the CSUPER, or be careful
to create the matrices for the external superelement in such a way that
the terms for the GRID points occur first. Another possible avoidance is
 to use all SPOINTs on the CSUPER and then to use MPCs to connect to the
GRID points. Bringing in the external superelement as a PART using the
EXTRN entry also works.
@
32526; 70.5.2; Matrix Methods - MPYAD - SFM 4276, EC 3012 or Wrong Results
On the SUN computer, if a MYPAD operation uses the DOT3RD kernel, then the
answers may be incorrect or the job may terminate with one of the
following error messages:
BIOMSG: ERROR 3012 HAS OCCURRED IN ROUTINE QOPEN , FILE INDEX (FILEX) =

process error reported by subroutine LNNP2CD

see LNNP2CD abstract (ier = -728)
*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
Avoidance: We have been unable to describe in modeling terms when the
erroneous method is chosen. To date the error has only been observed in
two similar models using superelements while calculating the system modes
at the residual. To prevent the incorrect method from ever being
selected it is recommended to add the following system cell to the top of
the input file or to the system rc file for all runs on Version 70.5 SUN
Ultra SPARC systems.
nastran system(252)=0
Note that the SUN Solaris system rc file on the Version 70.5.1 R3 cd which
added the Fujitsu UXP/V system was modified to include this avoidance.
@
32550; ; Static Analysis - SUBCASE, SDR2 - Missing Output, SFM 3002
If a subcase is repeated in a model with many subcases and multiple
boundary conditions, the job may issue the message:
*** SYSTEM FATAL MESSAGE 3002 (SDR2)
EOF ENCOUNTERED WHILE READING DATA BLOCK UG2 (FILE 112) IN
SUBROUTINE SDR2
or may be missing output for some subcases.

Avoidance: Remove the duplicated subcases.

@
32551; 2001; Optimization/Design Sensitivity - DESVAR, IFP - SFM 4276, Error
Code 7777
If integer values are used for XINIT, XLB, or XUP on DESVAR entries,
MSC/NASTRAN may fail in the IFP module with either SIGNAL 8, error 7777,
or SFE 4276, EC 7777.
Avoidance: Correct the format on the DESVAR entries.
@
32557; 70.7; Differential Stiffness - STATSUB, Superelements - UFM 4562 in TR
LG
When performing transient analysis (SOL 109) with superelements and
differential stiffness, with STATSUB, then UFM 4562 may
 be issued.
Avoidance: Insert the following DMAP alter:
compile super3
alter 'equivx CASERS'
else if ( nocasrs1>-1 ) then $
equivx CASERS1/CASE0/-1 $ for non-SOL 200 with dynamic looping
endalter
@
32561; 70.5.2; Frequency Response/Random Response - RANDPS - Time Command Termi
nated Abnormally
An underflow error may occur on some machines when the random module
squares stress values that are numeric zeroes. If this error is
encountered the job will stop in the RANDOM module, the f06 file will not
contain a message, and the log file may contain and overflow/underflow
message. The DEC ALPHA OSF/1 system issued the message:
Time command terminated abnormally.

Avoidance: Do not request random xy output for very small output

quantities. If the quantity is near 1.e-19, when it is squared for random,
it can cause an underflow to occur. Most machines will catch this and
output a zero in the results. The ALpha stops with no meaningful messages
@
32572; 70.6; Superelements - SUBCASE,SUPER - UFM 5423 in SDR1 or UFM 7324, m
issing output
If there exists more subcases for a superelement than the residual
structure, then the run may fail with UFM 5423. If the load sequence
number is specified on the SUPER commands for the superelement subcases,
then UFM 7324 is issued.
However, the run continues and output is missing for the extra
superelement subcases.
Avoidance: Remove the extra superelement subcases or add residual
structure subcases.
@
32608; 70.6; Inertia relief - DMIG, UACCEL - UFM 6143
If less than 6 degree of freedom are specified on a single grid point on
the DMIG,UACCEL entry, then the job will fail with the following messages:

*** USER FATAL MESSAGE 6143, (MERGE1)

INPUT MATRICES AND VECTORS ARE INCOMPATIBLE
Avoidance: Insert the following dmap alter:
$
$ for v70, v70.5
$
compile statrs
alter 'UPARTN USET,UACCEL',''
IF ( NOUACCEL>=0 ) UPARTN USET,UACCEL/URAx,,,/'G'/'R'/'S'/1 $
alter 'dbstatus',''
dbstatus urax//s,n,noura $
alter 'solve '
endif $
 alter 'endif \$ nullqrr',''
endalter

@
32609; 2001; Documentation - PARAM,UNSYMF - Missing Documentation
Complex eigenvalue analysis may be performed in nonlinear static analysis,
SOL 106, with PARAM,UNSYMF,YES. This feature is documented in the Version
70.5 Release Guide but the PARAM,UNSYMF is not documented in the Version
70.5 Quick Reference Guide. The parameter should be documented and
referred to from the cmethod case control command in the QRG.
@
32611; 70.5.2; P-Elements - Beam Error Estimator
If there are beam p-elements in a model with no strain, such as rigid body
modes in modal analysis, the error estimator may crash. The EPSTOL
parameter does not work with beam p-elements so it will not help avoid
this error.
Avoidance: Exclude beam p-elements from the error estimate in such models,
or turn off the error estimator.
@
32612; ; P-Elements - BEAMP - Wrong Answers at P=2
A BEAMP p-element has bad residual bending flexibility when p=2. The
error is more apparent when small values for the modulus "E" are used.
Avoidance: Set NASTRAN SYSTEM(226)=1, to turn off the residual bending
flexibility.
@
32615; ; Submittal - A "$" Sign in Command Line Argument Causes Argument
to be Truncated
If a command line argument contains a dollar sign ('$'), the command line
argument is truncated at the dollar sign. This probably should only occur
for NASTRAN cards encountered in RC file processing.
Avoidance: There is no avoidance for Windows NT. The error can be
avoided on UNIX by quoting the argument and escaping the dollar sign,
however, a dollar sign is a very unusual UNIX filename character.
@
32626; ; Matrix Methods - DECOMP - UFM 9050 (SEKRRS) Excessive Pivot Rati
o
If negative terms occur on the factor diagonal matrix during matrix
decomposition using the SPARSE (default) solver, UFM 9050 will correctly
be printed, but the dof containing negative factors on the factor diagonal
matrix will not be identified. Terms which have factor diagonal ratios
greater than the value of PARAM,MAXRATIO will be properly identified.
Avoidance: If the problem is run using NASTRAN SPARSE=1, the degrees of
freedom with negative terms in the factor diagonal matrix will be properly
identified in the listing.
@
32633; 70.6; ACCESS - smplr - Program to Read .xdb Does not work Properly.
A dynamic factor was added for the access key in MSC/NASTRAN Version 70.5.
 Thus, the program smplr.f delivered with MSC/NASTRAN Version 70.5 does
not work properly.

Avoidance: An illustration of the protocode required is:

Call READR (file,ia,0,iret) <- when output object is open
attempt to locate entry with key=0
if ( iret.eq.0 ) then
factor = ia(2) <- obtain dynamic
factor for access key
else
if ( approach=statics ) then <- conventions before
V70.5 more or less
factor = 1
else
factor = 10000
endif
key = intid*factor + offset <- intid from
EQEXING/E and offset from TOL,FOL, etc.

@
32639; 70.6; Complex Eigenvalues - EIGC, HESS - wrong answers
The HESS method may produce one pair of roots that are wrong for diagonal
input matrices. The symptoms indicate that the shift may not be removed
from the last pair calculated. The solchk alter in the sssalter libray
may be used when it is suspected that roots may be wrong. It computes a
normalized residual for each eigenvector. Similar code is built into the
complex eigensolver for systems after v70.5. It has been confirmed that a
new HESS method implemented there has solved the two problems discovered
to date that have wrong answers.
Avoidance: An avoidance is to use the CLAN method instead. In Version
70.6 and beyond this error occurs only when the old HESS method is
selected by system cell 108. An avoidance is to remove the system cell
request from the NASTRAN card. The new default method is the QZ HESS
method, which does not have this error.
@
32642; 70.7; Complex Eigenvalue Analysis - CEAD - Floating Point Exception
An error in the spill logic on "small endian" computers which include
Digital and Windows NT systems may cause a floating point exception to
occur in the complex eigenvalue solver. The message obtained is machine
dependent, one example is:
"The FPU has been reset after floating point exception"
Avoidance: Increase the available amount of memory until spill is not
necessary. Memory on Intel NT systems can be increased above 25 mw only
in Version 70.5.
@
32647; 70.6; P-Elements - EMA - Wrong Answers
A wrong stiffness matrix may occur if both h- and p-elements occur in the
same model, the h-element grid points use a non-basic coordinate system
for displacements, and multiple passes are required for matrix assembly.
Since some columns are output using the Basic coordinate system rather
than the local system, the matrix will be unsymmetric. The wrong answers
 should be evident by high epsilon values and the spcforce resultants will
not balance the applied loads.
Avoidance: Either increase memory to avoid multiple passes or avoid the
use of displacement coordinate systems.
@
32660; 70.5.2; Installation - time: command terminated abnormally
MSC/NASTRAN will fail to run if the hard limit for the system's data
segment size, see the "data(kbytes)" line of the "ulimit -H -a" command,
has been set to any value other than "unlimited".
Avoidance: Reset the hard limit to unlimited. This can only be done by
root, it will require the user to exit the affected shell.
@
32664; 70.6; Random Analysis - RANDPS - UFM 5418, UFM 316
The RANDPS bulk data entry has an error that limits the number of RANDPS
entries that are processed. This error occurs only for the special case
of uncoupled PSDF (see remark 8 in the QRG). The error may results in
various fatal messages. Those experienced include USER FATAL MESSAGE 5418
and USER FATAL MESSAGE 316. The text of these messages indicate that
there are illegal RANDPS entries and NO GRID POINTS IN THE MODEL. It is
speculated that there is a transition area somewhere between 100 and 320
bulk data entries where different types of bulk data entries may be
discarded, leading to many different error messages.
This error does not apply to other types of random analysis input, where
the number of RANDPS entries that are allowed depends only on memory
requirements.
Avoidance: There is no error for the coupled case. Add a dummy RANDPS
entry, where J<K are distinct subcase identifiers, and both X and Y are 0.
@
32670; 70.7; P-Elements - QUADP - Spurious Zero Frequency Modes
The QUADP element may have spurious modes (more zero frequency modes than
expected) for certain values of p along the edges.
Avoidance: An avoidance is to set SYSTEM(225)=1, which turns off the
bubble functions.
@
32677; 2001; Documentation - PARAM,EIGD - PARAM,EIDG Does Not Exist
In the Version 69, 70, and 70.5 Quick Reference guides the parameter
description for EIGD says:
"PARAM,EIDG,YES requests ..." But, it should say:
"PARAM,EIGD,YES requests ..." (note EIDG vs EIGD)
@
32685; 70.5.2; Transient dynamics - NLRGAP with SOL 109 - TRD1, Signal 10, Bus
error
When using the NLRGAP element in SOL 109, the job may terminate with no
error messages in the TRD1 module on some machines. Depending on the
machine, the log file may contain messages similar to the following:
Receiving signal number 10: Bus error
 The FPU has been reset after floating point exception.
Avoidance: Look at the "TOTAL MSC/NASTRAN MEMORY LIMIT" in the Execution
Summary (f04 file) and increase the memory by exactly 2 words on SUN
systems and 6 words on SGI systems by specifying mem= on the nastran
submittal command. See also error 32014 before proceeding.
@
32705; 2001; Documentation - RSPLINE - Maximum of 100 Grids
The RSPLINE has an undocumented limit of 100 grid points. A remark will
be added to the next release of the Quick Reference Guide.

@
32707; 70.5.2; Acoustic Analysis - CTETRA - UFM 4683 (READ)
Acoustic Analysis with the 10 noded CTETRA element fails with the message:
*** USER FATAL MESSAGE 4683 (READ)
MASS MATRIX NEEDED FOR EIGENVALUE ANALYSIS
Avoidance: Other than using the 4 noded TETRA element there is no
avoidance.
@
32717; ; Complex Eigenvalue Analysis - CEAD, EIGC - signal number 10: Bus
error
If the INV complex eigenvalue method is used in version 70+ and the model
contains transfer function callout (tfl) and p-elements, then the job may
fail with one of the following error messages:
Receiving signal number 10: Bus error
Trap encountered:
SIGSEGV - Segmentation violation
Access violation
Avoidance: Use the complex lanczos method instead of the inverse method.
@
32725; ; Postprocessing/Component Mode Synthesis - PARAM,FIXEDB - Missing
Output
If PARAM,POST is set to 0 to generate an MSC/XL database and
PARAM,FIXEDB,-1 is used, the ".xdb" file will not contain the component
modal results.
Avoidance: There is no avoidance.
@
32728; 70.5.2; Cyclic Symmetry - Incorrect UWM 6231 is given is SOL 114
An erroneous message of the following form may appear in SOL 114:
*** USER WARNING MESSAGE 6231 (CY1CCT)
BOUNDARY CONDITIONS IN SUBCASEKKLL ARE DIFFERENT THAN THOSE IN
THE FIRST SUBCASE.
SUBCASE 1 SPC= .001 AND MPC=
SUBCASE .001 SPC= AND MPC= .001
 USER INFORMATION: ONLY ONE BOUNDARY CONDITION IS PERMITTED FOR ALL
SUBCASES IN CYCLIC SYMMETRY ANALYSIS.
THEREFORE, THE CONDITIONS IN THE FIRST SUBCASE WILL
BE USED FOR ALL SUBCASES.
Avoidance: Ignore the message.
@
32731; 70.5; SSSALTERs - sebloada.v705 - Possible Wrong Answers
Due to an error in the sssalter sebloada.v705, output from the alter in
transient solutions may not be printed correctly. Rather than printing a
time history of OLOAD (boundary force), it may print only a few points
with an incorrect time step.
Avoidance: Change the alter:
change the line:
EQUIVX OL1/OUTPLIST/ALWAYS $ OUTPUT LIST
to:
EQUIVX OL1r/OUTPLIST/ALWAYS $ OUTPUT LIST
@
32733; ; External Superelements - PARAM,EXTOUT - Ignores differential sti
ffness
When creating an external superelement using PARAM,EXTOUT, the first
residual structure SUBCASE is used. This means that it is not possible to
include a static preload.
Avoidance: Apply the following DMAP alter when creating the reduced
superelement:
$
$ alter to allow static pre-load when creating reduced external
superelement matrices
$
$ For of the case control:
$
$ SUBCASE 1
$ LOAD = 99
$ SUBCASE 2
$ STATSUB = 1
$ METHOD = 3
$
compile phase1dr
alter 1
$ remove external superelement creation
alter 'IF ( EXTOUT.*) THEN','ENDIF $ EXTOUT' $
$
compile semodes, list
$
$ insert external superelement creation
$
alter 'xadd5.*kaa,kdaa'
$
type parm,,i,n,gotit $
type db,gpak,gdak
type parm,nddl,char8,y,extout
IF ( EXTOUT<>' ' ) THEN $
CALL EXTOUT MKAA,MAA,BAA,K4AA,PA,GPAK,GDAK,EQEXINS,USET/
DUM/LUSETS $
 CALL PRTSUM /RSTJUNK $
EXIT $
ENDIF $ EXTOUT<>'
$
$ fix for error 32070
$
compile sekdr
$ this alter is for Version 70.5 only
alter 'call segoa',''
type db goq $
CALL SEGOA USET,GOT,goq/GOAT/NOOSET/NOQSET/NOTSET $
endalter
@
32748; 70.5.2; Documentation - Network Mounted Database Files Affect Elapsed Ti
me Performance
Database files that are remotely mounted will affect elapsed time
performance of MSC/NASTRAN jobs. There is no explicit documentation or
user warning to this fact.
Depending on the I/O characteristics of the job, the relative performance
of the processor, and the relative performance of the network, the penalty
can be extreme. One user saw a 24:1 ratio of elapsed time to CPU time
with NFS files and 2:1 with local files. Higher ratios would not be
surprising.
@
32751; 70.5.2; Nonlinear Transient Heat Transfer - EMG, MATT4 - SFM 4276, EC 7
777
Nonlinear transient heat transfer analysis that involves a phase change in
MSC/NASTRAN Version 70.5 may fail in the EMG module with the following
traceback:
Trap encountered:
SIGTRAP - Trace trap FP division by zero
Traceback:
Offset 0x00001250 in procedure tab_
Offset 0x0000059c in procedure hmat1_
Offset 0x0000aeb4 in procedure ehexd_
Offset 0x0000054c in procedure emgbrd_
Offset 0x00000b68 in procedure emgpro_
Offset 0x000004a0 in procedure emg_
Offset 0x000000fc in procedure xsem09_
Offset 0x00000948 in procedure mainal
--- End of call chain ---
Avoidance: User avoidance is to add an additional point at the end of
the TABLEMi table.
@
32760; ; Utilities - estimate program, replication entries - UFM (estimat
e.generate_)
If the input file uses the nastran replicator to generate continuation
entries, the ESTIMATE program may fail with the following error messages:
*** USER FATAL MESSAGE (estimate.generate_bulk_entry)
In file "run1.dat", at line 26:
 *1, *2, *2
Field 10 is blank.
Avoidance: Echo the input file into a sorted nastran punch file and run
the ESTIMATE program on the new file.
@
32773; 70.5.2; Rigid Elements - RSSCON with ELEM option - UFM 6695 (SEGRDD)
When using the ELEM option of the RSSCON entry, the following error may
result when the shell element has a significant acute angle causing
attachment to the wrong edges of the solid:
*** USER FATAL MESSAGE 6695 (SEGRDD)
RSSCON 12 CANNOT CONNECT SHELL GRID POINT 44 : USER
ACTION, CHECK THE GEOMETRY
*** USER FATAL MESSAGE 6695 (SEGRDD)
RSSCON 13 CANNOT CONNECT SHELL GRID POINT 43 : USER
ACTION, CHECK THE GEOMETRY
@
32783; 70.7; Installation/Security - flexlm, rcf - Incorrect Default
On NT systems using flexlm, the authorization code specified on the System
environment variable MSC_LICENSE_FILE (under settings/control
panel/system/environment) overwrites the rcf file but not the command
line. If you want the RC file to override MSC_LICENSE_FILE, you must lock
the keyword in the RC file. For example,
auth=what_ever_the_filename_is
lock=auth
Any attempt to set a new auth value will be silently ignored.
Unfortunately, an error in prior versions of the nastran command allowed
the environment to override a locked keyword. This error has been
corrected for v70.7.
Avoidance: Either specify the authorization file on the command line or
remove the MSC_LICENSE_FILE variable from the system environment.
@
32784; 2001; Documentation - EIGC - Incorrect OMEGAAJ Value
Under the Quick Reference Guide description of the EIGC entry, field 3,
the OMEGAAJ-field, of the first continuation entry should be +5.6 instead
of -5.6
@
32793; 70.7; Transient Analysis - LOADSET, LSEQ - Segmentation Violation in T
RLG
In SOL 129, various computer types may fail with a segmentation fault of a
"signal" error if LOADSET-LSEQ is used to define the dynamic loadings and
the static load ids on the LSEQ entries are identical to the associated
DAREA ids.
Avoidance: Always use unique ids for the static loadings and DAREA ids
@
32828; ; Parts superelements - SECONCT - Wrong Answers
External superelements at present require the presence of a seconct entry
 in the main bulk data deck, and an extrn entry in the part superelement.
If the number of DOFs listed on the seconct entry is less than the number
on the extrn entry, then what is called the g-set option is being used.
There is an error in the map matrix which resequences the boundary point,
leading to incorrect assembly of the external superelement in the assembly
run.
The error does not exist when the a-set option is used. For this option
the number of DOFs listed on the seconct entry is exactly equal to the
number on the extern entry.
Avoidance: If it is required to release some of the boundary dofs this
can be done by coupling them to an extra grid point in the downstream
superelement, then connecting this extra point to the actual downstream
boundary point with an rbar element which connects only the desired DOFs.
@
32829; 70.5.2; Beam Library - Missing bmsrv.o file for build
The main program object file for the Beam Server is missing from the
MSC/NASTRAN V70.5 CD's. This file is only needed by clients building a
beam server. If a build is attempted messages similar to the following
will be encountered:
f77 -o beamserv /big_fish/msc/msc705/hpux/bmsrv.o bsmsg.o
bsbrc.o bsbrg.o bsbri.o bsbrp.o bsbrt.o bscon.o bsgrq.o mevbr.o msbri.o
msbrg.o msbrc.o brtup.o brtui.o brtug.o brtuc.o
/big_fish/msc/msc705/hpux/libbmsrv.a -Wl,-a,archive +O3 +U77
+Oparallel_env +Onolimit +DS770 -v -lM
f77: Cannot read file /big_fish/msc/msc705/hpux/bmsrv.o
f77: Errors detected, no link
*** Error exit code 1
Stop.
Avoidance: The files will soon be available from the following URL:
ftp://ftp.macsch.com/pub/msc-products/NASTRAN/V70.5/
Select the appropriate directory for your system, and download the
"bmsrv.o" file.
@
32837; 70.5.2; Submittal - Remote Execution Via node Command will Fail.
Remote execution of a nastran job using the "node" keyword will fail. A
typical error message that will appear on the terminal or window is:
ksh: 128: not found.
Avoidance: Edit <install_dir>/bin/nast705.rmt and change all four (4)
occurrences of the text "0.name" to "rcmd".
@
32846; ; External Superelements - SECONCT - Wrong answers
The g-set option allows connecting less DOFs to the downstream
superelement than exist in the boundary matrices. If the number of DOFs
listed on the SECONCT entry is less than the number on the EXTRN entry,
then what is called the g-set option is being used. The method, used to
discard unwanted DOFs, is to discard their rows and columns from the
boundary matrices, which is analogous to spcing them. This will make
internal constraints in the model, that is, spcs whose affects are visible
 implicitly in missing rigid body modes, load and element force unbalance
at some of the grid points, etc. but the internal spc forces do not appear
in spcforce output. Also be aware of error 32828.
Avoidance: An avoidance is to use the a-set option only. See the
discussion in error 32828 on how to convert to this option. When using
the a-set option, boundary DOFs may be disconnected from the downstream
superelement by connecting the unwanted boundary point to an extra grid
point in the downstream superelement which is not connected to any other
structure.
@
32852; ; Security - FLEXlm - Can't locate Multi-processors
When using FLEXlm, if the number of processors specified by the PROC entry
in the license file is less than the total number of processors on a
machine, MSC/NASTRAN will use only one processor even if more are
requested.
Avoidance: Obtain a license file that allows the maximum available (or
unlimited) number of processors.
@
32854; 2001; Documentation/Superelements - SETREE, SELABEL - Entry Ignored, I
ncorrect SE Tree
The SETREE entry will only work if PART (BEGIN SUPER) superelements exist
in the model. If there are no PARTs in the model, the SETREE entries will
be ignored. A remark should be added to the SETREE and SELABEL entries
stating this limitation.
Avoidance: Either:
1) Change BEGIN BULK to BEGIN SUPER to force new superelement logic
or:
2) use DTI, SETREE
@
32858; 2001; Documentation/Aeroelastic Analyis - SPLINE2, GI - Poor Message
If an aeroelasticity run, contains a SPLINE2 element that incorrectly
references only one grid point, the GI module issues a FATAL ERROR message
with no explanation of the cause of the error.
Avoidance: Increase the number of grids to the required two or more.

@
32860; 70.6; Differential Stiffness - LOADSET, superelements, STATSUB - Wrong
answers or UFM
If the LOADSET command is used in SOLs 107 through 112 with superelements
and differential stiffness (STATSUB Case Control command) then wrong
answers may occur. For example, if LOADSET is specified above the
subcase, the static loads in the static subcase are wiped out. If the
LOADSET command is specified in the dynamic subcase following the static
subcase, the dynamic load
via LSEQ disappears.
Avoidance: Specify two LSEQ Bulk Data entries with the same SID and two
different DAREA IDs; one for the static preload and one for the
dynamically applied static load. The LSEQ with the smaller DAREA ID must
 correspond to the static preload. The LSEQ with the larger DAREA ID must
correspond to the dynamically applied static load. Also, insert the
following DMAP alter avoidance:
compile statics2
alter 'bcdr ','bcdr '(,2)
LCGEN CASES,SLT,ETT/CASESX/S,N,DUMMY $
BCDR CASESx//SEID/SOLAPP/
S,N,NSKIP/S,N,NLOADS/S,N,BCFLAG/S,N,SPC/S,N,MPC/
S,N,SUPORT/S,N,LOAD/S,N,LSEQ///1 $
PARAML CASESx//'DTI'/nskip/7//S,N,tempLD $
PARAML CASESx//'DTI'/nskip/6//S,N,deform $
endalter
@
32861; 70.5; Utilities - DBC.XDB missing from NT delivery
The DBC.XDB schema database was not included with the NT Version 70.5
delivery. This will cause the ddlqry and ddlprt utilities to fail or not
have access to the Version 70.5 description information.
@
32864; 70.5.2; Superelements - EXTRN, PARAM,EXTOUT, SMPYAD - SFM 3001
If external superelement matrices are used in SOL 101 or 103, then the run
may fail with SFM 3001 in SMPYAD.
Avoidance: Insert the following alter:
compile extin
alter 'smpyad xmap,x_k','mpyad xmap,x_p'
type parm,,char8,n,name $
trnsp xmap/xmapt $
smpyad xmap,x_k ,xmapt,,,/kxx/3////////6 $
call fname /mxx,ZUZR11/S,NAME
if ( name<>' ' ) smpyad xmap,x_m ,xmap,,,/mxx/3////1////6 $
call fname /bxx,ZUZR11/S,NAME
if ( name<>' ' ) smpyad xmap,x_b ,xmap,,,/bxx/3////1////6 $
call fname /k4xx,ZUZR11/S,NAME
if ( name<>' ' ) smpyad xmap,x_k4,xmap,,,/k4xx/3////1////6 $
mpyad xmap,x_p,/px $
@
32870; 70.5.2; Nonlinear Transient Heat Transfer - CTRIAX6 with MATT4 - UFM 455
1
Nonlinear transient heat transfer models that use the 6 noded CTRIAX6
axi-xymmetric element will not converge and will fail with the message:

*** USER FATAL MESSAGE 4551 (NL2CON)

Avoidance: Use the 3 noded axisymmetric element instead.
@
32873; 70.5.2; P-Elements/Plotting - Wrong VUGRID ID for TRIAP elements
The VUGRID id for TRIAP elements may be output as zero. This will cause
missing points in deformation plots.
 Avoidance: There is no avoidance.
@
32884; 70.7; Nonlinear Transient Heat Transfer - CBEND with MATT4 - Wrong Ans
wers
The CBEND element produces wrong answers in nonlinear transient heat
transfer analysis.
Avoidance: Use the CBEAM element.

@
32890; 70.6; Aeroelastic Analysis - GI, SPLINE2 - Unnumbered Fatal Message
Spline checking improvements made in V70.5 can result in erroneous and
unlabeled "FATAL ERRORs" out of the GI module. These errors occur if a
SPLINE2 is attached to a single grid point (a situation that can produce a
singular spline matrix). Due to the code error, ANY SPLINE2 with a single
structural grid point will generate the singularity fatal, whereas those
attached to a single grid point for both translation and rotations will
produce the correct results. Unfortunately, a secondary error causes the
message to be lost; hence, the unlabeled FATAL.
Avoidance: In the case of a single grid point accounting for both
translation and rotations, in order to avoid the erroneous fatal message,
RBAR another GRID to it. The other GRID has no use in the analysis, but
just allows the model to avoid the fatal message. Correct results will
occur.
@
32899; 70.6; Eigenvalue Analysis - EIGRL, READ, PARAM,AUTOSPC,NO - SFM 3001,
SFM 3002
A model which contains degrees of freedom with no mass or stiffness may
cause the Lanczos eigensolver to fail with one of the messages below. The
most likely cause for this is submitting a model with PARAM,AUTOSPC,NO.
This error is also likely in cyclic problems with solid elements and
points on the axis of symmetry since these points are unaffected by
PARAM,AUTOSPC,YES and are singular in the rotational degrees of freedom.
*** SYSTEM FATAL MESSAGE 3001 (READ)
THE INPUT DATA BLOCK IN POSITION 13 DOES NOT EXIST.
USER INFORMATION: THIS ERROR IS CAUSED BY ONE OF THE FOLLOWING:
1. THE DATA BLOCK IS UNSPECIFIED ON THE DMAP MODULE.
2. THE DATA BLOCK IS SPECIFIED ON THE DMAP MODULE AND ON THE
SUBDMAP STATEMENT
BUT NOT ON THE CORRESPONDING CALL STATEMENT.
*** SYSTEM FATAL MESSAGE 5299 (LNCS1D)
UNRECOVERABLE TERMINATION FROM LANCZOS ITERATION
INTERNAL ERROR CODE = -99 USER ERROR RETURN = ***
*** SYSTEM FATAL MESSAGE 3002 (READ)
EOF ENCOUNTERED WHILE READING DATA BLOCK SCRATCH (FILE 314) IN
SUBROUTINE REIGL3
0FATAL ERROR
Avoidance: Do not use PARAM,AUTOSPC,NO unless you are sure that there are
no singularities in your matrices. The avoidance for solid element cyclic
problems is to constrain the boundary grids in dof 456 or to select the
 old Lanczos with system(273)=1.
@
32901; 70.5.2; Nonlinear Transient - NLRGAP, AUTO, TSTEP, TABK - Possible Wrong
Answers
When using the NLRGAP element, if the gap stiffness varies as a function
of time, and either the AUTO or TSTEP method is selected on the TSTEPNL
entry, then the answers may be incorrect.
Avoidance: Use the default ADAPT method.

@
32906; ; Aeroelastic Analysis - SPLINEi Referencing Non-existent GRID - P
oor Message
If an aeroelastic model is created in which a SPLINEi entry references
non-existent aerodynamic grid points (boxes), a misleading error message
is issued in which the missing box id's are correctly listed, but the the
message associates them to a SET1 or SPLINE3 entry that is, itself,
nonexistent. Furthermore, the SET1/SPLINE3 association is only applicable
to missing structural grids and, in this case, it is the aerodynamic grids
that are missing. These data are on the SPLINEi entry and not on the SETi
entry.
@
32909; ; Nonlinear Transient Heat Transfer - PARAM,AUTOSPC - Wrong Answer
s
The default for PARAM,AUTOSPC is NO for the nonlinear solution sequences,
this includes solution 159 for nonlinear transient heat transfer. This is
because AUTOSPC only acts on the linear portion of the kgg matrix and thus
if it was YES it could lead to wrong answers. To prevent possible wrong
answers, Version 70.5 ignores a YES setting for AUTOSPC. The correct
action would be for AUTOSPC to act on the sum of the kggl and the kggnl
matrices. The correction of this error will be to take this action.
Avoidance: Do not use the PARAM,AUTOSPC,YES in transient heat transfer
analysis. Identify any isolated/disjoint grids and constrain them.

@
32917; 2001; Superelements/Plotting - SET, INCLUDE - All Elements Plotted ins
tead of Subset
When using superelements, any plot request which references a set
definition using the INCLUDE option to select elements will plot the
entire superelement.
Avoidance: Use the EXCEPT feature. Example:
SET 1 = ALL EXCEPT QUAD4 $ will plot all elements except QUAD4s.
@
32929; 2001; Documentation - P-Elements, PLOAD2, GP3 - UFM 2105
The MSC/NASTRAN Quick Reference Guide does not state that the PLOAD2 may
not be applied to p-elements. If a PLOAD2 references a p-element the
message:
*** USER FATAL MESSAGE 2105 MODULE = GP3
 PLOAD2 CARD FROM LOAD SET xxx REFERENCES MISSING OR NON-2-D ELEMENT
OR HYPERELASTIC PLANE ELEMENT yyy
will be encountered.

receive UFM 2105 stating that the loading references a missing or non 2-d
element or hyperelastic element.
Avoidance:
Use a PLOAD4 to load p-shells.
@
32931; ; Restarts - MATi, RESTART - Restarts Ignore Material Damping, Wro
ng Answers
If a restart is performed from solution 106 (nonlinear static analysis) to
solution 108 the GE field of the MAT1 entry is ignored.
Avoidance: Specify PARAM,G damping instead.
@
32937; 70.6; Optimization/Design Sensitivity - UWM 9058, UFM 4574
If the direct sensitivity method is used in one iteration and then the
adjoint method is selected by the program in a subsequent iteration, then
the run will fail with the following messages:

^^^ USER WARNING MESSAGE 9058 (FEA)

^^^ THE PSEUDO SOLUTION AND THE PSEUDO LOADS FOR THE RESIDUAL
STRUCTURE AND THE CURRENT SUBCASE ARE ZERO.
*** USER FATAL MESSAGE 4574,
DATA RECOVERY STARTED BUT NO DISPLACEMENT VECTORS AVAILABLE
Avoidance: Insert the following DMAP alter:
compile pslgdv
alter 'isolapp\=1'
psload=false $
endalter
@
32938; 70.5.2; P-Elements - TEMP, TEMPF - Wrong Answers
If a model with p-elements has both TEMPF and TEMP entries, a solid
element without a TEMPF entry will be assigned the default temperature,
rather than using the nodal TEMP temperatures.
Avoidance: A workaround is to use all TEMPF or all TEMP entries to assign
temperatures.
@
32944; 2001; Documentation - PARAM,FLUIDSE - Missing Documentation
The parameter fluidse is only documented in the Version 70.5 Release
Guide, it is not documented in the Version 70.5 Quick Reference.
Prior to Version 70.5 the fluid-structure analysis capability required
that all fluid elements and all points (structure and fluid) at the
fluid-structure interface be in the residual structure. With this
parameter the structure points may now be assigned to upstream
 superelements.
@
32945; ; Miscellaneous - BEGIN BULK - Wrong answers
The BEGIN BULK command may contain the words SUPER=seid or
AUXMODEL=auxmid. If any other word is input, then the BEGIN BULK command
is not recognized as a delimiter and no related fatal or warning message
is issued.
Avoidance: There is no avoidance.
@
32951; ; Nonlinear Analysis - CGAP - Poor Answers when GAPs used with Mid
-Side Nodes
GAP elements do not work well with elements with midside grid points.
Irregular stresses will be produced in elements near the GAPs. This
problem is severe when quadrilateral faces are involved (HEXA-20) and
moderate when triangular faces are used (TETRA-10).
Avoidance: Use elements without midside nodes.
@
32960; 70.5.2; Restarts - GP2 and MCE2 - Poor Performance
In the static solution sequences; e.g., SOLs 101, 114, 106, MCE2 may be
reexecuted unecessarily during a restart. In SOLs 101-200, GP2 is
reexecuted unecessarily.
Avoidance: Insert the following alter in all runs including the restart:
compile sekmr
alter 'ENDIF \$ SEMR'(,-1)
else if ( statics ) then $
purgex /goq/-1 $
compile phase0
alter 'gp2 ',''
GP2 GEOM2,EQEXIN,EPT,geom2a,epta/ECT,ecta/S,N,ACOUSTIx $
alter 'fscoup\=',''
FSCOUP=(ASCOUP='YES' AND ACOUSTIx=2) $
endalter
@
32961; 70.6; Aeroelastic Analysis - AMP, RESTART - SFM 4276 EC 1044
A restart into SOL 146 from SOL 103 may fail with SYSTEM FATAL MESSAGE
4276, ERROR CODE 1044 in the AMP module.
Avoidance: If this error occurs, it is because the spline matrices are
not being generated on the restart, insert the following DMAP alter.
compile phase1a
alter 'aspline'(,-1)
delete /gpjk,gdjk,,, $
@
32972; 70.5.2; P-Elements - PARAM,PEDGEP, FEEDGE - Parameter Ignored
If PARAM,PEDGEP,1 is specified in the bulk data file, the POINTs on the
FEEDGEs are not re-parameterized. This parameter only applies to cubic
edges; for quadratic edges, the parameter is ignored.
 Avoidance: There is no avoidance; however, it is unlikely that this
inconsistency would negatively affect answers.
@
32973; ; SSSALTERS - alter2ga, alter2ha - Potential Inefficient Restarts
In some instances, a restart from a run using alter2ga.vxx or alter2ha.vxx
may try to "re-process" the superelement brought in using the alter. The
cause of this is under investigation.
Avoidance: Insert the following DMAP into the run which is failing (not
the run with alter2ga or alter2ha). BEWARE - this DMAP will cause the
program to ignore any changes made in the model during a run which uses
it.
compile chkgoaq
alter 3,3 $
$
compile phase0
$ alter 581,585 $ V70
alter 694,697 $ V70.5
@
32979; 2001; Documentation - Static Aeroelasticity - Missing Remark from PAER
O2 Entry
The following remark needs to be added to the Quick Reference Guide for
the PAERO2 bulk data entry as follows:
Remark:
8. For half models, the theta arrays LTH1 and LTH2 should encompass a
full 360 degree range.
@
32981; 70.5.2; P-Elements - PSHELL, NSM - Wrong Answers
If NSM (non-structural mass) is used on shell p-elements wrong answers may
result. The NSM is interpreted as mass per unit volume instead of mass
per unit area. This leads to wrong answers in the Grid Point Weight
Generator, eigenvalues and static response due to gravity loads.
Avoidance: The only avoidance is to adjust NSM to take into account the
thickness of the element.

@
32983; 70.7; Superelements - SEBULK - SFM, EC 1044 IN FILPS2
When a sebulk entry references a seid that is not on a begin super entry,
the following system error occurs:
BIOMSG: ERROR 1044 HAS OCCURRED IN ROUTINE FILPS2
A log file message also appears, with a machine-dependent message.

@
32987; 70.5.2; Nonlinear Analysis - CBUSH1D, SDR2 - SFM 6985
The BUSH1D element when used in Nonlinear Statics (106) will abort in the
SDR2 module when STRESS data recovery is requested. The fatal message
 is shown below.
*** SYSTEM FATAL MESSAGE 6985 (SDR2RD)
PREVIOUS ELEMENT-ID = 101
DID NOT REQUEST CORRECT NUMBER OF PHASE-1 OUTPUT WORDS.
0FATAL ERROR
Avoidance: An avoidance is to request the BUSH1D data recovery with the
NLSTRESS Case Control Command and to EXCEPT the range or BUSH1D elements
from the STRESS Case Control Selection.
@
32991; 70.0.5; Nonlinear Analysis - PARAM,FOLLOWK, GRAV, RFORCE - SFM 3002, TAF
F
SOL 106 Follower forces. EOF ENCOUNTERED WHILE READING DATA BLOCK
(NONE) (FILE 0) IN SUBROUTINE TAFMEG
In nonlinear analysis, solution 106, when follower forces are used and
there are GRAV and/or RFORCE inertial loads, the job may fail in the TAFF
module with the message:
SYSTEM FATAL MESSAGE 3002
EOF ENCOUNTERED WHILE READING DATA BLOCK (NONE) (FILE 0) IN SUBROUTINE
TAFMEG.
Avoidance: A DMAP avoidance is under investigation. Currently there is
no practical avoidance other than inserting PARAM,FOLLOWK,NO or removing
the inertial loads. Although this error exists in Version 70.5.0 (it was
fixed in 70.5.2) the error correction was taken to Version 70.0.5 for
NASTRAN for Windows. Therefore, the fixed version has been changed to
70.0.5 under the assumption that all 70.5 clients should have upgraded to
70.5.2.

@
33006; 70.5.2; Frequency Response Analysis - MPYAD, SOLs 108, 111, 146, 200 - P
oor Performance
The MPYAD module in subDMAP FREQRS for frequency response analysis that is
used to compute the PDF matrix does not take advantage of the tendency of
the dynamic load matrix to be very sparse and this leads to poor
performance.
Avoidance: Insert the following alter:

compile freqrs
alter 'mpyad '(,-1)
trnsp ppf/ppft $
matmod ppft/vppf/12/s,n,nullppf/1 $
if ( nullppf>-1 ) then $
partn ppf,,vppf/ppf1,,,/1 $
partn rpx,,vppf/rpx1,,,/1 $
mpyad rpx1,ppf1,/pdf/1 $
else $
alter 'mpyad '
endif $
endalter
@
33012; 2001; Superelements - SPCs on Boundary Grids - SFM 6804
 If a boundary (exterior) point for a PART (BEGIN SUPER = i) superelement
is constrained, then the fatal message below will be issued if different
displacement coordinate systems are used for the coincident GRID points.
SYSTEM FATAL MESSAGE 6804 (BNDSP3)
THE DISPLACEMENT COORDINATE SYSTEM (CD FIELD ON GRID ENTRY) SPECIFIED FOR
BOUNDARY GRID ID 4 B
IN SUPERELEMENT 1 IS NOT CONGRUENT WITH DISPLACEMENT COORDINATE SYSTEMS IN
OTHER CONNECTING SUPERELEMENTS.
Avoidance: Either apply the constraints in the downstream superelement
(the one which has the coincident point as an interior point) or use the
same displacement coordinate system on all of the points which are
coincident.
@
33016; 2001; Documentation/QRG - SEEXCLD
The Quick Reference Guide description for the SEEXCLD entry states that
SEIDA(field 2) must be larger than zero. A value of 0 (zero) is also
allowed. Also, the thru range is indicated as 61<62 it should be
GIDA1<GIDA2
@
33023; ; Superelements - SEPLOT - No Contour Plot Output
Currently, the SEPLOT module does not support contour plots of
superelements.
Avoidance: There is no avoidance.

@
33026; 70.7; Nonlinear Buckling Analysis - UFM 3007 (UPARTN) from CALCV
A nonlinear buckling analysis will fail in the NLSTATIC module with the
message:
*** SYSTEM FATAL MESSAGE 3007 (UPARTN)
ILLEGAL INPUT TO SUBROUTINE CALCV
if the run contains superelements and there are no spc or mpc sets in the
residual structure.
Avoidance: Apply the following dmap avoidance:

compile nlstatic
alter 'upartn.*dugni.*dulv'(2),'upartn.*dpgni.*dpl0'
paraml usetnl//'uset'//////'G'/s,n,Gsize/'L'/s,n,Lsize $
message //'debug: from UsetNL, G-size is '/Gsize
message //'debug: from UsetNL, L-size is '/Lsize
if(Gsize = Lsize) then
message //' fyi: copying dugni --> dulv, dpgni --> dpl0'
equivx dugni/dulv/-1 $
else
message //' fyi: doing normal upartn on DUGNI and DPGNI'
UPARTN USETNL,DUGNI/DULV,,,/'G'/'L'/'M'/1 $
endif
endalter
@
33027; 2001; Optimization/Design Sensitivity - DEQATN - Incorrect ID
On CRAY systems DEQATN id's may be printed incorrectly in error message
output.

@
33045; 2001; Nonlinear Analysis - CGAP, TEMP(LOAD), ASETi - Wrong Answers at
Omitted DOF
Nonlinear analysis in Version 70 and 70.5 systems with multiple subcases,
thermal loads, asets and gaps will give erroneous results at omitted
degrees of freedom in the first printed load increment in the second
subcase. If intermediate load increment results are not requested, the
errors will occur in the final subcase.
Avoidances:
1) Use a single subcase, which should work since nonlinear elements are
not allowed to be omitted, or,
2) use superelements instead of ASETs, or,
3) printout intermediate load increments and ignore the first increments
results of the second subcase, or,
4) use a version prior to 70.
@
33075; 70.7; Superelements - SECONCT - Wrong Answers
If coincident grid points exist in a superelement and a SECONCT entry is
used to connect one of them to a grid point in another superelement, it is
possible that the SECONCT entry will not work properly and incorrect
answers will result. If no SECONCT is used, then the program will
correctly issue UFM 6794
Avoidance: Review the "BOUNDARY SEQUENCE ASSIGNMENT TABLE" for grids that
are attached to multiple downstream GRIDS instead of the usual one to one
correspondence. Do not create coincident grid points in a superelement
where BEGIN SUPER is used.
@
33076; 70.5.2; Buckling - EIGRL, READ - UFM 4276 7777 signal 10
It is possible that the new Lanczos eigensolver may terminate with a
machine-dependent error (signal 10 on an HP) followed by
BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR in the READ module.
Avoidance: For the model that demonstrated this error, adding V1 and V2
on the EIGRL entry in addition to the ND specification allowed the problem
to be solved. Setting SYSTEM(273)=1 to select the old Lanczos method will
allow the problem to be completed.
@
33078; 70.7; DMAP - CALL DBSTORE - Fails with Signal Trap
If CALL DBSTORE and DIAG 47 are used then the run will fail with a hard
crash in the PARAML module.
Avoidance: Remove DIAG 47 or insert the following DMAP alter:
compile dbstor
 alter 't6',''
type parm,,i,n,t6 $
endalter
@
33080; 70.7; Nonlinear Transient Analysis - NLTRD2 - BIOMSG: ERROR 1044
If a nonlinear transient analysis has only stiffness, but no mass or
damping the job may fail with the message
BIOMSG: ERROR 1044 HAS OCCURRED IN ROUTINE QOPEN , FILE INDEX (FILEX) =
0
STATUS = 0
in NLTRD2.
Avoidance: Add appropriate damping and mass to the model.
@
33101; 2001; Nonlinear Transient Analysis - TRLG - UWM 6498
Module TRLG issues the following message if a nonlinear transient analysis
has a DLOAD command which refers to non-existant TLOADi entries:
*** USER WARNING MESSAGE 6498 (MsgGetStr)
Message number is not found
PROGRAMMER INFORMATION:
Check call to MSGPOP from TRLG with the message number 3031
Severity is USER WARNING
Parameter 1 is 111
Parameter 2 is DLT
Parameter 3 is TRLG
USER ACTION: Please contact MSC Client support
Avoidance: Remove references to non-existant loadings on the DLOAD entry.
@
33108; 2001; Documentation/Buckling - EIGRL - UFM 4405
The Lanczos buckling shift logic in MSC/NASTRAN Version 70.5 was
enahanced. The enhanced code allowed many outstanding difficult problems
to be solved without resorting to adjustments of the EIGRL entry.
However, a very unusual case has been found that gives the Lanczos new
shift logic trouble, wheras the old Lanczos shift logic solved the
problem. The problem case occurs when there are on the order of thousands
of eigenvalues below 1. and there are on the order of thousands of
eigenvalues in the frequency range of interest. The old Lanczos shift
logic would shift near V1 first, thus if ND was specified and there were
ND roots near V1, the old shift logic would solve the problem. The new
shift logic has a tendency to shift near 1. first or V2 if V2 is less than
1. The new logic solves the more common problem of users specifying 0.
for V1 even though the first root is not near zero. Thus, if there are
thousands of roots between V1 and V2 or below 1, the code finds those
nearer to V2 or 1. but is never able to find the lowest ND roots.
Avoidance: Several avoidances are possible. First it is unusual to pose
a problem that has thousands of buckling modes below 1 so the model and
the loading should be reviewed. A smaller eigenvalue range of interest
could be specified or the SHFSCL field on EIGRL entry can be specified to
force the first shift near the lower end of the V1, V2 range.
@
33113; 70.6; Documentation/Superelements - CSUPER - BIOMSG: ERROR 3021 from
FILPS2
Although the documentation does not specifically state so, a superelement
which is a copy (mirror or otherwise) of another superelement may not be
copied or mirrored. Version 70.6 addresses this error by issuing a more
meaningful fatal message.
Avoidance: If additional copies are required, make more copies of the
original superelement.
@
33118; 70.5.2; Installation - IRIX 6.5; libblas.so missing
MSC/NASTRAN Version 70.5 does not run on IRIX 6.5 with a standard
operating system configuration. Error messages such as those shown below
will be encountered:
MSC/NASTRAN V70.5 (Silicon Graphics IRIX64 6.5) Fri Aug 21 20:11:26 1998
MSC/NASTRAN beginning job test. PID: 4669.
jever 5% 4735:/msc/msc705/irix64/analysis: rld: Fatal Error: Cannot
Successfully Map soname 'libblas.so' under any of the filenames
/msc/msc705/irix64/libblas.so:/usr/lib64/libblas.so:/usr/lib64/internal/l
bblas.so:/lib64/libblas.so:/opt/lib64/libblas.so:/msc/msc705/irix64/libbl
s.so.1:/usr/lib64/libblas.so.1:/usr/lib64/internal/libblas.so.1:/lib64/li
blas.so.1:/opt/lib64/libblas.so.1:
Avoidance: If the client does not have a complete set of sgi math
libraries they will need to get a /usr/lib64/mips4/libcomplib.sgimath.so
from MSC. This can be found at the following locations:
On the web at:
ftp.macsch.com/pub/msc-products/NASTRAN/V70.5/irix64
filename: libcomplib.sgimath.so
filenames < then a list of library files follows >.

@
33119; 70.6; Elements - TRIAX6, TEMP(LOAD) - Wrong Answers
The TRIAX6 element gives incorrect stresses if THETA (see CTRIAX6 entry)
is nonzero, and there is thermal expansion. The deflections are not
affected. The THETA field is most likely used for MAT3 orthotropic
materials.
Avoidance: The only known avoidance when there is thermal expansion, is
to input the material properties in the r-z system, so THETA is zero.
@
33121; 70.5.2; P-Elements - BEAMP - Element Coordinate System, Wrong Answers
If the curvature of a beam p-element exceeds 90 degrees, the element
coordinate system may reverse itself. For a beam with a symmetric cross
section, this will probably only affect the data recovery points.
Avoidance: Use beam p-elements that have a curvature of less than 90
degrees.
@
33131; ; Iterative Solver - P-Elements, Solvit - UFM 4157
 The iterative solver may fail without issuing a meaningful message if the
model contains adaptive p-elements. USER FATAL MESSAGE 4157 may be issued
indicating insufficient memory; however, the message may not indicate the
amount of memory required or it may indicate an incorrect amount.
Avoidance: The failure may be the result of insufficient memory. Thus,
increasing memory may solve the problem. If increasing memory does not
solve the problem, then do not use the iterative solver.
@
33171; ; Elements - CBUSH1D, SHOCKA - No Warning Message
By design intent NASTRAN allows the unmodified CBUSH1D element to be used
in a linear static analysis where the linear initial stiffness is used.
However, the SHOCK option has no stiffness and NASTRAN does not issue a
warning message.
@
33181; 70.5.2; Performance - SDR2 - Poor Performance on NEC V70.5.1
The SDR2 module in V70.5.1 on the NEC system may be up to 60% slower than
in V70.0.
Avoidance: The problem has been fixed in Version 70.5.2.

@
33186; ; Nonlinear Analysis - RESTARTS - SFM 4252
Nonlinear analysis restarts may fail with the message:
^^^ PHASE 1 - SUPERELEMENT GENERATION, ASSEMBLY AND REDUCTION.
*** SYSTEM FATAL MESSAGE 4252 (SEMA)
MATRIX XAAV FOR UPSTREAM SUPERELEMENT 10 DOES NOT EXIST.
USER INFORMATION: THE CURRENT SUPERELEMENT REQUIRES ALL OF IT'S
UPSTREAM SUPERELEMENT MATRICES
FOR ASSEMBLY INTO THE CURRENT SUPERELEMENT. THE
PROBABLE CAUSES ARE:
1. THE UPSTREAM SUPERELEMENT'S DBSET OR DATABASE
IS NOT ATTACHED.
2. THE UPSTREAM SUPERELEMENT DID NOT COMPLETE
REDUCTION IN A PREVIOUS RUN.
Avoidance: This problem is still under investigation but is believed to
occur when temperature loads are applied in the cold start with TEMP(INIT)
but are removed in the restart. Please see general limitation 4609.
@
33192; 2001; Documentation - STATSUB in SOL 200 only supports ANALYSIS=BUCKLI
NG
Remark 1 of the MSC/NASTRAN Version 70.5 Quick Reference Guide's
description of the STATSUB command indicates that STATSUB is supported in
SOLs 103, 105, 107 through 112 and 200. However, the support in solution
200 is limited to ANALYSIS=BUCKLING.
@
33202; 2001; Modal Response Analysis, Complex Eigenvalues. - GKAM - Wrong Ans
wers
It has been found that a very low memory request may result in poor
complex eigenvalues being produced when using the modal method or poor
answers in a modal response. In the case that demonstrated the problem
the memory was set to one million words, well below the default of four
 million words. The one million word value was barely above the minimum
needed for subroutine xspace. This error is not expected to occur if more
than 1mw of open core memory is available to nastran, which is the case if
the default memory and buffsize are used. The error appears to have
occurred in the GKAM module which assembles the modal matrices.
Avoidance: Make sure that sufficient open core memory is available.
@
33233; 70.6; Inertia Relief - SUPORT, PARAM,INREL - Wrong Answers
If an inertia relief analysis in solution 101 or 200 specifies less than
six suport degrees of freedom in Version 70.5 of MSC/NASTRAN, then wrong
answers may result.
Avoidance: Apply the following dmap alter or use solution 1.
COMPILE STATRS
ALTER 'upartn ','fbs '
UPARTN USET,DRG/D6T,,,/'G'/'T'/'S'/2 $
TRNSP D6T/DT6 $
MPYAD DT6,DT6,/D6T2/1 $
MODTRL D6T2////6 $
DECOMP D6T2/LD6T2,,////////1.+8/16 $
MATGEN ,/IRR/1/NORSET $
UMERGE USET,DM,IRR/DML/'T'/'L'/'R' $
MPYAD DT6,DML,/DT6TDML/1 $
FBS LD6T2,,DT6T/CRYY $
compile inersen
alter 'upartn ','fbs '
UPARTN USET,DRG/D6T,,,/'G'/'T'/'S'/2 $
TRNSP D6T/DT6 $
MPYAD DT6,DT6,/D6T2/1 $
MODTRL D6T2////6 $
DECOMP D6T2/LD6T2,,////////1.+8/16 $
MATGEN ,/IRR/1/NORSET $
UMERGE USET,DM,IRR/DTR/'T'/'L'/'R' $
MPYAD DT6,DTR,/DT6R/1 $
FBS LD6T2,,DT6R/CRYY $
Endalter
@
33239; 70.5.2; Nonlinear - GPSTRESS, QUAD4 - Incorrect results
In a nonlinear analysis with PARAM,LGDISP,1, the grid point stress
(GPSTRESS) output for grids attached to CQUAD4 elements with a
STRESS(CORNER) request will be incorrect. Values of 1.0E+30 are printed
for all components.
Avoidance: Do not request STRESS(CORNER) and GPSTRESS in the same run.
@
33242; 70.6; Aeroelastic Analysis - SOL 145 restart - UFM 5423 in AMP module
Restarts into SOL 145 or 146 may fail with UFM 5423 in the AMP module.
Avoidance: Insert the following DMAP alter:
compile phase1dr
alter 'call phase1a'(,-1)
delete /gpjk,gdjk,gdak,gpak, $
delete /gpgk,gdgk,,, $
 compile pfaero
alter 'DELETE \/D1JK,D2JK',''
DELETE /D1JK,D2JK,wskjf,, $
endalter
@
33272; 70.6; Eigenvalue Analysis - READ, REIGL - Indefinite Mass Matrix, Wron
g Answers
If the mass matrix is indefinite, then the roots coming out of Lanczos may
not be correct. A symptom of this error is the issuance of the following
error messages:
INERTIA (STURM SEQUENCE) COUNT DISAGREES WITH THE NUMBER OF MODES ACTUALLY
COMPUTED IN A (PARTIAL) INTERVAL
FOUND xxx MODES BETWEEN aaaa HZ AND bbbb HZ BUT THERE SHOULD BE ONLY
yyy MODES.
Note that in almost all cases, an indefinite mass matrix is caused by
modeling errors. Starting in Version 70.6, the job will terminate with
the fatal message below if an indefinite mass matrix is encountered.
USER FATAL MESSAGE 4646 (REIG)
THE MASS MATRIX INPUT TO THE REAL EIGENSOLVER IS INDEFINITE.
USER ACTION: REMOVE THE NEGATIVE MASS TERMS.
If GDR is being used (DYNRED) it is possible that poor numerics (very
small negative eigenvalues) may lead to the above fatal message in Version
70.6. For these cases the fatal message may be avoided by adjusting the
zero tolerance with SYSTEM(60). The default value 6, which gives a zero
value of 1.0E-6. Thus a value of 4 will relax the zero tolerance by two
orders of magnitude.
Avoidance: Modify the model to remove the indefinite mass matrix.

@
33274; 70.6; Coupled Fluid/Structural Analysis - Fluid Superelements - SFM 54
23
If PARAM,FLUIDSE,>0 and any structural grid points or SPOINTs are assigned
interior or exterior to the fluid superelement then SFM 5423 may be
issued.

Avoidance: Ensure that no structural points appear in the interior and

exterior point list for the fluid superelement. This can be verified in
the printout of the superelement map. This includes generalized
coordinates used for component mode synthesis on the fluid superelement.
The seqset1 should be input on fluid points (grid, CD=-1) rather than
scalar points.
@
33285; ; Plotting - Aeroelastic Elements - Missing Plot Output
Beginning in Version 70.5 the NASTRAN plotter no longer supports
aeroelastic elements.
Avoidance: The only avoidance is to use Version 70 or PATRAN Version 8.1
which is due out by the end of this year.
@
33299; 70.7; P-Elements - QUAD4 - Poor Answers for p=1
The QUAD4 p-element may be too stiff when p=1 on some edges. This is most
noticible for coarse meshes. An avoidance is to specify p >= 2.

@
33307; 70.6; Aeroelastic Analysis - AELINK, TRIM - UFM 5267
The above message was produced by a client whose length units was
millimeters and mass units were kilograms. The trim matrix became
singular in this case, apparently because of the dynamic range within the
equations to be solved.
*** USER FATAL MESSAGE 5267 (ASGLD)
ZXX MATRIX IS SINGULAR. TRIM AND AELINK ENTRIES SPECIFY AN
INADEQUATE CONSTRAINT SET.
Avoidance: In this case, the problem was avoided by scaling the ZZX and PZ
matrices going into the trim module by 1.0e-4. This avoidance is likely
to be problem dependent, so other factors may be required in other cases.
The alter to perfrom the scaling is:
compile aestatrs
alter 'asg'(,-1)
add zzx,/zzxs/1.0e-4 $
add pz,/pzs/1.0e-4
delete /zzx,pz,,,/ $
equivx zzxs/zzx/always $
equivx pzs/pz/always $
A more general avoidance is to scale the ZZX matrix so the magnitude of
its largest term is unity. Scale the PZ matrix by the same term. This will
work in more cases:
compile aestatrs list noref
alter 'asg ','asg '
type,parm,,rs,n,zzxnorm,xnorm,inorm $
type,parm,,i,n,ncol,nrow $
type,parm,,cs,n,cnorm $
file junk=ovrwrt/zzx2=ovrwrt/pz2=ovrwrt $
norm zzx/junk/s,n,ncol/s,n,nrow/s,n,zzxnorm/1 $
if ( zzxnorm >< 0.0 ) inorm = 1.0 / zzxnorm $
cnorm = cmplx(inorm) $
add zzx,/zzx2/cnorm $
add pz,/pz2/cnorm $
asg casea,zzx2,pz2,xlist,edt,tr,krzx,dit,erhm/
ux,uxdat $
mpyad zzx,ux,pz/balance//1/-1 $
$matprn zzx,zzx2,pz,pz2,ux $
$matprn balance,,,, $
norm balance/junk/s,n,ncol/s,n,nrow/s,n,xnorm/1 $
message //' xnorm of balance is '/xnorm $
@
33308; ; Sequencing - SEQP - Memory Allocation Error
Very large problems that use the metis method of sequencing
(system(206)=8) may fail with a memory allocation error as follows:
 *** SYSTEM INFORMATION MESSAGE 6916 (DFMSYN)
MALLOC FAILED INSIDE METIS. DECREASING THE MEMORY GIVEN TO THE JOB
WILL INCREASE THE MEMORY AVAILABLE FOR MALLOC.
DECOMP ORDERING METHOD CHOSEN: METIS, ORDERING METHOD USED: MMD
Dynamic memory starts at address 000000008BE56940
*** Opencore memory reduced to 49791588 ***
Metis Error: ***Memory allocation failed for CreateCoarserGraph: gdata.
Requested size: 21237992 bytes
time: command terminated abnormally.
Avoidance: There is no guaranteed avoidance. Increase the memory or use
another sequencing method. Requesting system(206)=72 requests the metis
method with a different pre-phase that typically uses 30% less memory.
@
33309; 2001; External Superelements - PARAM, EXTOUT, DMIGPCH - UFM 327
When using the external superelement capabilities with the
"param,extout,dmigpch" option in a SOL 103 run, the header entry for the
load PAX will also be written out, even though it is not used. This will
cause the issue of the following error messages in a subsequent attachment
run.
*** USER FATAL MESSAGE 327 (IFS2P)
BAD DATA OR FORMAT OR NON-UNIQUE NAME. DMIG PAX SORTED
CARD COUNT = 161
^^^ USER FATAL MESSAGE 9002 (IFPS)
^^^ ERROR(S) ENCOUNTERED IN THE MAIN BULK DATA SECTION
^^^ SEE MESSAGES ABOVE. ERROR ENCOUNTERED IN MODULE IFP,
Avoidance: Remove the offending PAX header. This entry is not used in a
SOL 103 run.

@
33334; 2001; Documentation - DEQATN - Increased Limit
The number of allowable non-blank characters for the DEQATN entry was
increased from 12,500 to 32,000 in Version 70; however, the Quick
Reference was not updated to reflect this.
@
33367; 70.6; Optimization/Design Sensitivity - DESVAR - UFM 316
The message:
*** USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA ENTRY DESVAR xx SORTED ENTRY COUNT = yy
was produced when a DESVAR punched at the end of an optimization task had
an XINIT that was very slightly less than the XLB value. When a new run is
attempted, an equality test is used to see if the input satisfies XINIT
.ge. XLB and the entry is rejected even though it had been produced in a
previous run.
Avoidance: Manually reset any punched DESVAR values to satisfy the
relations XLB .LE. XINIT .LE. XUB
@
33401; ; Case Control - SEALL, SEMG, SEKR, etc. - UFM 9015, Poor Message
If the SEALL, SEMG, SEKR, SELG, SELR, or SEMR command is specified in a
subcase which is not the first subcase then the operation requested by
that command is not performed and various messages or wrong answers may
result. For example, with SEALL or SEMG then UFM 9015 will be issued.
USER FATAL MESSAGE 9015 (SEMODES)
THE USET TABLE IS NOT PRESENT.
Avoidance: Specify SEALL, SEMG, SEKR, SELG, SELR, or SEMR above all
SUBCASE commands.
@
33424; 70.6; Aeroelastic Analysis - SPLINEi, AESTATRS - UFM, Incompatible Ma
trix Multiply
An incompatible matrix multiply can occur in calculating the intercept
hinge moment coefficients if the aerodynamic forces are splined only to
dependent degrees of freedom (e.g., to m-set and r-set points).
Avoidance: A DMAP workaround is to include the following alter:
compile aestatrs $
alter 'add *uintl','add *intl' $
type parm,,i,n,nointl $
dbstatus uintl//s,n,nointl $
if ( nointl < 1 ) then $
PARAML ALX//'TRAILER'/2/S,N,NOLSET $
MATGEN ,/MUNTL/7/NOLSET/1 $
MATGEN ,/MINTL/7/NOLSET/1 $
else $
ADD UINTL,/MUNTL/-1.0 $
ADD INTL,/MINTL/-1.0 $
endif $
endalter $
@
33427; 70.6; Aeroelastic Analysis - AELINK, RESTART, MPYAD - SFM 5423
An error in the dependencies results in restarts into solution 144 where
only the AELINK entries have been changed failing with SYSTEM FATAL
MESSAGE 5423 indicating an attempt to multiply incompatible matrices:
SYSTEM FATAL MESSAGE 5423 (DMPYAD)
ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES
COLS ROWS FORM TYPE NZWD DENS
TRAILER FOR WSKJ IS 101 40 80 2 2 4 250
TRAILER FOR FFAJI IS 102 7 40 2 2 80 10000
TRAILER FOR FFINT IS 103 2 80 2 2 0 0

Avoidance: Do not attempt restarts when only changing AELINK bulk data
entries. A simultaneous change in the aerodynamic model geometry will
enable the restart, but this would also entail recalulation of the
aerodynamics. Alternatively, the following DMAP alter may be inserted as
long as there are no other changes to the aero model.
compile ifpl
alter 'RESTART EDT,EDTR',''
endalter
@
33454; 2001; Documentation - EIGC, SYSTEM(108) - Block Complex Lanczos is not
the Default
The more modern and reliable Hessenberg which supports spill was made the
default in Version 69 by switching the meaning of system(108)=0 and 1.
However, the documentation was never updated to reflect this change.
Furthermore, because of poor performance for some problems the new Block
Complex Lanczos was not made the default in Version 70.5. Although this
decision is reflected in the release notes the decision was made too late
to be reflected in the Quick Reference Guide's description of system(108)
under the NASTRAN command and EIGC entry.
@
33476; 70.6; Miscellaneous - Elapsed Time Incorrect
The elapsed time will be incorrect if more that 24 hours are spent in any
one module.

@
33480; 70.6; Postprocessing - SOL 1 through 16, PARAM,POST - Missing output
No results will be written to the .op2 or .xdb file if PARAM,POST is
specified in SOLs 1 through 16 where SORT2 is the default output format or
in all SOLs if SORT2 is explicitly requested.
Avoidance: Insert the following DMAP alter in any SOL 1 through 16:
compile sol__ $ insert SOL number for __
alter 'IF (POST=0) THEN'(,-1)
LABEL LNOSORT1 $
alter 'LABEL LNOSORT1',''
endalter
@
33481; 70.6; Estimate - SPC - Undefined GRID referenced
When an SPC references an SPOINT, the estimate utility may indicate that a
undefined GRID has been referenced.

@
33491; 70.6.1; Optimization/Design Sensitivity - CAABSF - UFM 7115
The acoustic absorber element (PAABSF) is not supported in optimization
and design sensitivity (SOL 200). If a PAABSF entry is included in the
bulk data, even if it is not one of the design variables, the following
message is issued, " *** USER FATAL MESSAGE 7115 (DOM2B) PTYPE 1502 15 IS
NOT SUPPORTED".
Avoidance: Do not include PAABSF entries in an optimization run.

@
33495; ; Postprocessing/Random Analysis - PARAM,POST,0 - Missing Data in
xdb
The results from a response spectrum spectrum analysis are not written
into the ".xdb" file when PARAM,POST is set to 0.
Avoidance: Set PARAM,POST,-1 and they are written into the OUTPUT2 file.
@
33499; ; Elements - PSOLID - Spurious Modes
Spurious hourglass modes have been observed when solid elements refer to
an anisotropic smeared material property that has a modest shear modulus,
but almost zero extensional modulus.
Avoidance: Put ISOP=1 on the PSOLID record for these elements.
@
33500; ; Static Aeroelastic Analysis - Wrong Answers for Multiple Interfe
rence Groups
If multiple interference groups are used with bodies in any interference
groups except the last one and control surfaces exist, the DJX matrix
generated by ADG (adgdjx routine) is corrupted.
Avoidance: Specify a single interference group.
@
33517; ; Installation - Missing FLEXlm Installation Documentation
The Version 70.5 Configuration and Operation Guide for NT does not discuss
the installation of the FLEXlm license manager. This software is provided
on the Version 70.5 NT CD but must be installed separately. Please see
the "install.htm" on the NT-CD itself for installation instructions.
@
33527; 70.6; Iterative Solver - ITER, JACOBI - SWM 6715
An error in the iterative solver for real-unsymmetric problems using the
default Jacobi preconditions may lead to poor convergence and the message
below:
*** SYSTEM WARNING MESSAGE 6715 (SUCGRD)
THERE IS A ZERO DIAGONAL TERM IN COLUMN 1 OF DATA BLOCK CEE WHICH WILL
BE REPLACED BY 1.0
IN THE PRECONDITIONER.
USER ACTION: IT MAY BE BETTER TO USE CHOLESKY PRECONDITIONING
(PRECOND=C ON ITER BULK DATA ENTRY) OR
THE DIRECT SOLVER (ITER=NO ON NASTRAN STATEMENT) FOR THIS
PROBLEM.
Avoidance: The error is caused because instead of using the diagonal
terms of the stiffness matrix as the Jacobi preconditioner (as intended)
the iterative solver is running with an identity matrix as the
preconditioner which is the same as running with no preconditioner at all.
Use Cholesky preconditioning by setting PRECOND=C on the ITER bulk data
entry.
@
33529; 70.6; DBC - CTRIAR, PARAM,POST - UFM 6498
If a run with a ctriar element is execute back to back and the xdb file is
not removed, the job will fail with the following message unclear message:
*** USER FATAL MESSAGE 6498 (MsgGetStr)
Message number is not found
PROGRAMMER INFORMATION:
Check call to MSGPOP from with the message number 0.0
Severity is INVALID ERROR TYPE (0)
Parameter 1 is 3
 Parameter 2 is CTRIAR
Parameter 3 is CTRR
Parameter 4 is 0
USER ACTION: Please contact MSC Client support
*** USER FATAL MESSAGE 6498 (MsgGetStr)
Avoidance: Delete the xdb and re-submit the job.
@
33533; ; Superelements - Component Modes, SECSET - Possible Wrong Answers
If free-free component modes are used (e.g., secset), then the
eigenvectors may be incorrect. A symptom of this error is indicated by
the discrepancies between the eigenvalues printed in the eigenvalue table
versus the eigenvalues printed at the top of the eigenvector output.
Avoidance: Do not use free-free components, use the default fixed-fixed
component modes. See related error 32365.
@
33543; 70.6; Estimate - Recommends Setting SPARSE=1
The estimate program may incorrectly recommend turning sparse
decomposition off (i.e. setting SPARSE=1) for tet models. Sparse
decomposition is virtually always better than standard decomposition for
decomposing stiffness matrices for real world models. Sparse should never
be turned off.
@
33559; 2001; Documentation - PSHELL, STRESS(MAXS) - MAXS is not an Equivalent
Stress
A remark should be added to the Quick Reference guide description of the
STRESS case control command that MAXS for shell elements is not to be
confused with an equivalent stress as is the case for von Mises stress or
the octahedral stress.
@
33566; 2001; Static Analysis - PARAM,EXTOUT,DMIGPCH - UFM 327 (IFS2P)
If param,extout,dmigpch is used in but there are no loads (for example in
an eigenvalue analysis) a header record for the "PAX" load matrix is
printed to the punch file. When this punch file is included in the bulk
data of a subsequent run, the job will fail with the following message:
*** USER FATAL MESSAGE 327 (IFS2P)
BAD DATA OR FORMAT OR NON-UNIQUE NAME. DMIG PAX SORTED
CARD COUNT = 52
^^^ USER FATAL MESSAGE 9002 (IFPS)
^^^ ERROR(S) ENCOUNTERED IN THE MAIN BULK DATA SECTION
^^^ SEE MESSAGES ABOVE. ERROR ENCOUNTERED IN MODULE IFP,
Avoidance: Edit the punch file and delete the DMIG entry for PAX.
@
33568; 70.6; Optimization/Design Sensitivity - DSAD, PARAM,NOCOMP, PCOMP - SF
M 3007
When param,nocomps,0 is used in an optimization run with composite stress
or strain responses the following message will occur.
*** SYSTEM FATAL MESSAGE 3007 (DSAD )
ILLEGAL INPUT TO SUBROUTINE DSAD12
Avoidance: Use param,nocomps,1 which is the default.
@
33572; 70.6; Installation - MSC_LICENSE_FILE env var supersedes RC file "auth
" keyword
The precedence of the "authorize" keyword hierarchy in the "MSC/NASTRAN
Configuration and Operations Guide", Section 3.2, is incorrect. The
actual hierarchy is:
1) "authorize" keyword on command line.
2) MSC_LICENSE_FILE environment variable.
3) "authorize" keyword in an RC or RCF file.
4) <install-dir>/flexlm/licenses/license.dat
5) <install-dir>/conf/authorize.dat
<install-dir>/conf/authoriz.dat # Windows NT
6) LM_LICENSE_FILE environment variable.
Note, the first three items are consistent with all other nastran command
keywords.
@
33573; 70.6; Optimization/Design Sensitivity - DOPTPRM - Mislabeled Output
The P2 parameter on the DOPTPRM entry is used to control the output of
design optimization results. The output that is produced for composite
responses is scrambled in that CFAILURE output appears as CSTRESS, the
CSTRESS appears as CSTRAIN and the CSTRAIN appears as CFAILURE.
Avoidance: The numerical values are correct. It is only the labeling
that is in error.
@
33588; 70.6; Optimization/Design Sensitivity - DSAD, CSTRESS, CSTRAIN, CFAILU
RE - SFM 3007
The message:
*** SYSTEM FATAL MESSAGE 3007 (DSAD) ILLEGAL INPUT TO SUBROUTINE DSAD13
can be produced if there is a combination of CSTRESS, CSTRAIN, and
CFAILURE response types on DRESP1 entries and they invoke different sets
of elements.
Avoidance: Either only use one composite response type or apply the
multiple response types to exactly the same set of elements.
@
33600; 70.6; Aeroelastic Analysis - RESTART, MKAERO - Wrong Answers
If a restart is selected with additional MKAERO data, only the new MKAERO
pairs will be used in the Flutter or Dynamic Aeroelastic analysis and the
original data will be ignored.
Avoidance: For SOL 145, apply the following dmap alter in the restart run:

compile flutter
alter 'cntr'(2,-1)
endif $
alter 'endif'(2),'endif'(2)
For SOL 146, the appropriate alter is:
compile seaero
 alter 'cntr'(2,-1)
endif $
alter 'endif'(7),'endif'(7)
@
33606; 70.7; SSSALTERs - checka.v705 - SFM 3007, Illegal Input to Subroutine
CALCV
If the sssalter checka.v705 is used on a deck that has no spcs or rigid
elements then the message:
SYSTEM FATAL MESSAGE 3007, ILLEGAL INPUT TO SUBROUTINE CALCV
may be encountered.
Avoidance: Put in a dummy grid point and constraint.
@
33626; 70.6; Postprocessing - Nonlinear Analysis, PARAM,POST,0 - SWM 6066 (DB
C)
In nonlinear analysis with superelements, MSC/NASTRAN Version 70 and 70.5
may fail with messages similar to those below when it attempts to write
results to the XDB file.
*** SYSTEM WARNING MESSAGE 6065 (DBC)
*** DBCMSG ERROR MESSAGE 44 FROM SUBROUTINE DBCAOL
OUTPUT LIST DOES NOT MATCH EXISTING IN THE DATABASE, MERGE/RESTART
OPERATION IMPOSSIBLE.
MSC/XL DATA BASE CONVERSION FINISHED
*** SYSTEM WARNING MESSAGE 6066 (DBC)
*** DBCMSG ERROR MESSAGE 34 FROM SUBROUTINE DBC
THE DBC MODULE HAS BEEN TERMINATED BECAUSE OF THE ABOVE FATAL DBC
ERROR.
USER INFORMATION: SUBSEQUENT DBC CALLS WILL NOT BE EXECUTED, AND THE
DATABASE CONVERSION IS INCOMPLETE.
*** DBCMSG ERROR MESSAGE 14 FROM SUBROUTINE DBCROL
UNABLE TO OPEN INPUT RELATION NLLF RETURN CODE = 1.
*** SYSTEM WARNING MESSAGE 6065 (DBC)
*** DBCMSG ERROR MESSAGE 14 FROM SUBROUTINE DBCROL
UNABLE TO OPEN INPUT RELATION NLLF RETURN CODE = 1.
Avoidance: If PATRAN is being used for post processing the op2 option may
be used. Requesting sort2 rather than sort1 output may also solve the
problem.
@
33628; ; Superelements - CSUPEXT - UWM 4155
If an image superelement has SPOINTs for component mode synthesis the the
following warning message may be incorrectly encountered:
USER WARNING MESSAGE 4155---GRID POINTS FOR SECONDARY SUPERELEMENT 82 ARE
COLINEAR. CONFORMITY CHECKS NOT MADE.

@
33629; ; Documentation/P-Elements - CBEAM, CQUAD4, CTRIA3 - SFM 6352
If beam and plate p-elements are used in random analyses, the following
 uninformative message is printed in the .f06 file:
*** SYSTEM FATAL MESSAGE 6352 (SDRP36)
LOGIC ERROR: KERROR= 7.
USER ACTION: CONTACT MSC CLIENT SUPPORT.
PROGRAMMER INFORMATION: SUBCASE ID = 101 IN RECORD NUMBER
29 OF OUGV1 DATA BLOCK DOES NOT MATCH
THE SUBCASE ID = 102 IN RECORD NUMBER
2 OF CASEDR DATA BLOCK.
0FATAL ERROR
Avoidance: There is no avoidance, p-elements are not supported in random
analysis.
@
33644; 70.7; Transient Heat Transfer - Multiple subcases with DLOAD - Wrong A
nswers
Different DLOADs may not be applied in different subcases of a transient
heat transfer analysis. If attempted the load applied in the first subcase
will be lost in the second subcase.
Avoidance: Loads should be combined in a single DLOAD. If multiple
subcases are necessary to reference different tstepnl entries, than the
DLOAD case control request should be put above the subcase level.
@
33647; ; Follower Force - FORCE1 - Wrong Answers, Poor Convergence
In all linear solutions, the follower stiffness is erroneous when FORCE1
loads are applied. The convergence will be degraded in solution 106.
Avoidance: Use PARAM, FOLLOWK, NO to turn off the follower stiffness or
apply the loads using another type of force entry.
@
33705; ; Nonlinear Analysis - CELASi - Missing Punch Output
A request for elforce(punch) output for celas elements in nonlinear
analysis produces no output.
Avoidance: Request stress(punch) instead.

@
33725; ; Matrix Methods - MPYAD - System Fatal Message
If insufficient scratch space is available for a large mpyad, the job may
crash and a meaningful message may not be output.
Avoidance: Increase the amount of scratch space available ot the job.
@
33726; 70.7; Optimization/Design Sensitivity - Aeroelastic Analysis, FLUTTER
- SFM 7088
The message:
*** SYSTEM FATAL MESSAGE 7088 (DOPR3I)
MODE NO. xx ASSOCIATED WITH FLUTTER DRESP1 ID = yy EXCEEDS 0, THE
NUMBER OF MODES
AVAILABLE FROM THE ANALYSIS.
 may be produced in error when NVALUE is not explicitly specified on the
FLUTTER entry. Since there is a default value for NVALUE, the input may
be valid. The message is also incorrect in that it is
(a) listed as a system fatal message when it should have been a user fatal
message and
(b) the run does not stop despite the supposedly fatal message.
Avoidance: If the default value has been used for NVALUE, the message can
be ignored.
@
33749; 2001; Composites - PCOMP - Max Strain Ply Failure Index
When STRAIN ply results are requested, failure indices are not calculated.
Avoidance:
Supply a STRESS=n also, where n is the set of elements that will have
failure indices calculated, but you get the extra ply stress results in
addition to the ply strain results..

@
33753; 2001; Documentation/QRG - Replication example in QRG results in UFM 67
21
The replication example in the Version 70.5 Quick Reference Guide on page
427 results in USER FATAL MESSAGE 6721. The example:
GRID, 101, 17, 1.0, 10.5,,17,3456
=(4),*(1),=,*(0.2),==$
Should be changed to:
GRID, 101, 17, 1.0, 10.5,,17,3456
=,*(1),=,*(0.2),==$
=3
@
33770; 70.7; Documentation - Submittal - submit - Poor Message
A Bourne shell cannot be used to process the job file in a user's submit
command. If a Bourne shell is used, the following error may occur:
-sh: MSC_ASG=/Stemp/TT7293_19.asg: is not an identifier
The error message:
MSC_ASG=/Stemp/TT7293_19.asg: is not an identifier
was written by the Bourne shell, /bin/sh, and was complaining about an
export command.
Avoidance: Do not specify a Bourne shell. Section 3.11 of the
"Configuration and Operations Guide" does specifically state that a Korn
shell must be used. However, the example should be changed from "-s
/bin/sh" to "-s /bin/ksh". Note, the use of "/bin/sh" is correct on a
POSIX-compliant system, but could be misleading to an inexperienced user.
@
33775; ; Optimization/Design Sensitivity - MODES - Missing Output
When an optimization job with ANALYSIS=MODES uses more than one MODES
statements in Subcase Control section, the results are printed in *.f06
file only for the first subcase.
Avoidance: Do not use the MODES command.
@
33776; 70.6; Aeroelastic Analysis - AEROF - Null AEROF Output
The AEROF output from solution 146 is all zeroes in Version 70.5. For
frequency response analysis, the AEROF case control command should produce
force and moment data for the the aerodynamic boxes of the DLM.
Avoidance: A workaround is to supply a Case Control SET having the
desired ids and referencing that SET in the output request: AEROF=<setid>
@
33779; 70.6; Grid point stress - PSOLID, GPSTRESS, CORDM - Wrong Answers
When higher order solid elements are used in Grid Point Stress Volumes,
and their PSOLID references a material coordinate system, the local grid
point stress will be incorrect. The invariants, such as Pressure or Von
Mises are correct.
Avoidance: Use the default basic coordinate system.
@
33797; 2001; Aeroelastic Analysis - GUST - Documentation Clarification
The Version 64 Programmer's Manual indicates that the X0 field of the GUST
bulk data entry will change the phase relationship (it can be viewed as a
time delay) whenever a harmonic gust is used in SOL 146. Also the
streamwise location of aerodynamic elements introduces a phase change (a
delay) in the aerodynamic forces caused by a harmonic gust. The statement
on page 606 paragraph 2 is not completely accurate. The X0 and the
aerodynamic element center streamwise location are terms in the equation
that calculates the real and imaginary parts downwash for each aerodynamic
element. The phase relation represents a time delay for an excitation at
X0 to arrive at an aerodynamic box streamwise position.
@
33809; 70.6; Aeroelastic Analysis - SPLINEi, SUPORT - Incorrect Unrestrained
Hinge Moments
The unrestrained hinge moments are in error in static aeroelastic analysis
if the aerodynamic and structural model coupling splines the displacements
from the suport points to the aerodynamic mesh. This error is caused by
omitting the suport point displacements from the computation of the
aerodynamic mesh displacements for aeroelastic feedback.
Avoidance: A DMAP avoidance is:
compile aestatrs list noref
alter 'mpyad *gdkl.*ukpi',''
UMERGE USET,ULPI,URPI/UAPI/'A'/'L'/'R' $
mpyad gdka,uapi,/ukpi $ capture the r-set motion, too

@
33818; 70.7; Buckling Analysis - large problem - READ, Segmentation Fault
 On a very large buckling problem, 2.6 million degrees of freedom in this
case, depending on the machine, the job may fail with one of the following
error messages either in the log or f06 files. The error is caused by
trying to solve THE BUCKLING SHIFT SCALE CALCULATION FBS OF a very large
problem in core instead of out of core.
Trap encountered:
SIGSEGV - Segmentation violation
Traceback:
Offset 0x0000007c in procedure dbsmr1_
Offset 0x00000488 in procedure gbkmgr_
Offset 0x00000b50 in procedure ionast_
Offset 0x00000040 in procedure girdbk_
Offset 0x000003f8 in procedure sfupfc_
Offset 0x0000041c in procedure ssmf1d_
Offset 0x00000598 in procedure ssm5d_
Offset 0x0000039c in procedure bucksd_
Offset 0x000010c4 in procedure lnnrigl_
Offset 0x0000252c in procedure reig_
Offset 0x000000f4 in procedure xsem06_
Offset 0x00000958 in procedure mainal
--- End of call chain ---
-----------------------------------------------------------
***** RECEIVED SIGNAL 11, ABORTING ...
BIOMSG: ERROR 7777 HAS OCCURRED IN ROUTINE RECOVR , FILE INDEX (FILEX)
= 0
Avoidance: Possible avoidances are to set shift scale on the eigrl entry
or increase memory.

@
33841; 70.6; Nonlinear Analysis - Inaccurate Stresses using Drucker-Prager Yi
eld Criterion
The Mohr-Coulomb and Drucker-Prager yield criteria (selected on the MATS1
entry) entry create an inaccurate solution. The error is the result of an
incorrect correction factor to the yield stress, which accounts for
effects of the internal friction angle. Incorrect factors underestimate
the yield stresses, resulting in premature yielding in tension.
Avoidance: There is no avoidance, but the error is negligible if the
strain is very small.
@
33860; 70.6; Miscellaneous - Execution Summary - "nan" in the f04 Output
If MSC/NASTRAN is run on a computer which has more than 68 CPUs. Then the
F04 file may have "nan" in the output.
15:44:41 0:00 0.0 0.0 nan nan SEMTRN BGN
Avoidance: There is no avoidance but the "nan" ouput may be safely
ignored.
@
33861; 70.7; Optimization/Design Sensitivity - DSAJ, DVGRID - abnormally term
inated
Shape optimization jobs may fail in the DSAJ module on machines that
enforce boundary alignment. This includes SUN, HP and Digital Alpha. The
log file will contain a message of the form;
*** TERMINATING /software/msc/nastran705/msc705/solaris/analysis.ultra
*** Received signal 10 (SIGBUS)
time: command terminated abnormally.
and then a machine dependent traceback:
from SUN:
t@1 (l@1) signal BUS (invalid address alignment) in unpack_ at 0x8fcc58
unpack_+0x1d8: std %f2, [%o1]
(dbx) where
=>[1] unpack_(0x275, 0x1, 0x1dbb92c, 0x1a494e0, 0x1a49430, 0x0), at
0x8fcc58
[2] dsaj4_(0x18a88d4, 0x244598, 0x18a8680, 0x18a86d8, 0x1f6d010,
0x18a8638), at 0x244810
[3] dsaj_(0x18a8680, 0x18a8690, 0x1358, 0x18a8574, 0x0, 0x18a8578), at
0x243510
[4] xsem02_(0x123, 0x18852b0, 0x133eec4, 0x32, 0x32, 0x1870e70), at
0x11292c
[5] MAIN(0x186d400, 0x6b414, 0x0, 0x2, 0x0, 0x7), at 0x6bb68
(dbx)
From HP:
( 0) 0x0019b1cc sysrcv + 0x19c [/msc/msc705t2/sppux/analysis.pa2_0]
( 1) 0x011bc1d0 _sigreturn [/msc/msc705t2/sppux/analysis.pa2_0]
( 2) 0x01606ed0 P__FLUSHLINE + 0x437b38
[/msc/msc705t2/sppux/analysis.pa2_0]
Stack_Trace: error while unwinding stack
From Dec/Alpha I get lots of:
Unaligned access pid=8453 <analysis> va=0x145001cfc pc=0x1202e0d74
ra=0x1202e0e14 inst=0x9fe70000
Unaligned access pid=8453 <analysis> va=0x145001d04 pc=0x1202e0d7c
ra=0x1202e0e14 inst=0x9fe70008
Avoidance: Once the error is encountered, it may be avoided by adding a
dummy design variable to the deck.
@
33863; 70.7; Buckling Analysis - PARAM,FOLLOWK, TAFF - Poor Performance
If follower force stiffness is turned on (which is the default) in a
buckling analysis, and there are many load entries, then the calculation
of the follower force stiffness will be extremely slow.
Avoidance: If the effect of the follower force stiffness is not
important, add the following parameter to the bulk data section:
param,followk,no
@
33874; 2001; Documentation - EIGRL, SYSTEM(146) - Incorrect Description
The descriptions for system cell 146 in the Quick Reference Guide under
the EIGRL entry and the NASTRAN statements are incorrect. The system cell
 values are no longer a percentage increase of the memory reserved for the
sparse method, rather, it is calculated as (3 * x * buffsize), where x is
the value specified for system cell 146.
@
33896; ; Dynamic Analysis - TEMP, Thermal loads - UFM 4017
If multiple LSEQ entries exists in a model and the LSEQs reference
different TIDs, then the job will fail with the following unclear
messages:
*** USER FATAL MESSAGE 4017 (GETEMP)
THERE IS NO TEMPERATURE DATA FOR ELEMENT x IN SET yyy
0FATAL ERROR
Avoidance: Reference the same TID on all the LSEQ entries.

@
33911; 70.6; Restarts - GKAM - Inefficient Data Recovery Restart for Modal So
lutions
In a data recovery restart run in SOLs 110, 111, 112, 145, and 146, the
GKAM module is unnecessarily reexecuted.
Avoidance: Insert the following DMAP alter in the restart run:
compile semfreq
alter 'call gma',''
CALL GMA CASERS ,MATPOOLs,EQDYN ,TFPOOL , ,
, ,MAA ,GOD ,GMD ,
USETD ,KAA ,BAA ,K4DD ,PHsA ,
LAMAs ,DIT ,VAFS ,PHFA ,lamaf ,
AGG ,kddd /
, ,CMDD ,CK2DD ,CB2DD ,
MHH ,BHH ,KHH ,PHDH ,VPHFS ,
K2MG ,M2MG ,B2MG ,K2SF ,M2SF ,
B2SF /
LUSETD /SOLTYP /APP /NOUE /NOA /
FALSE /FMODE /FALSE /FALSE /FS /
NOMSET /fonly $
endalter
@
33913; 2001; Documentation - PARAM,CHECKOUT, PARAM,GPL - Missing Output
The parameter GPL described under PARAM,CHECKOUT in the Quick Reference
Guide should be PARAM,PRTGPL.

@
33936; 70.5; Aeroelastic Analysis - Mislabeled Force Output from a Trim Analy
sis.
If there are multiple interference groups (IGID on the CAEROx entry) and
the element ID's are not in ascending order with respect to the IGID's,
the trim force output can have incorrectly labeled element id's since it
uses monotonically ascending values starting from the lowest element ID.
Avoidance: Remove the multiple IGID's or make sure that element id's
ascend in order with the IGID's.
@
33938; 2001; RESTARTS - Inefficient Restart between binary compatible machine
s.
If a database is created on one machine and read into NASTRAN on a
different binary compatible machine, the RESTART logic may detect that
there are minute differences in the model and everything may be
re-calculated.
Avoidance:
compile phase0
alter 'RMGR,RMG,/','RMGR,RMG,/'
The above alter should be used only if your are certain there are no
changes to the model.
or superelement manual restart logic could be used. See SERST parameter.
@
33942; ; Documentation/Differential Stiffness - EIGR, EIGRL - Poor Rigid
Body Modes
If the pressurized modes of a free-free structure are desired, the rigid
body modes may not be very clean and may only be separated from the
flexible modes by 3 orders of magnitude. Negative values may also be
observed. This condition may occur in any model with follower stiffness
or rigid elements.
Avoidance: No avoidance is necessary. Even though eigenvalue analysis
without follower forces typically produces 6 orders of magnitude
separation between the flexible and rigid body modes the answers should be
acceptable.
@
33943; ; Miscellaneous - PARAM - No Warning Message
If a PARAM is entered in the input data but is not used (for example when
the param does not exist), no warning message is issued. Depending on the
intent of the user this may lead wrong answers.
@
33952; 70.7; Component Mode Synthesis - PARAM,METHCMRS Does not work in SOL 1
07
If residual structure cms is requested and METHCMRS is used to select the
eigenvalue solution method in SOL 107, the program will perform a static
reduction and will ignore the request for cms. Table 1 in section 6 of
the Quick Reference Guide incorrectly indicates that this parameter is
supported in solution 107.
Avoidance: Use the METHOD command in the CASE CONTROL to request cms for
the residual structure.
@
33953; ; Component Mode Synthesis - QSET - Wrong Answers
If residual structure cms is used, unused Q-set dof are not automatically
removed by AUTOSPC. This may result in unexpected or incorrect answers.
Avoidance: Either specify the correct number of Q-set dof (this may be
difficult, as it requires prior knowledge), or use PARAM,SMALLQ,0., which
will remove unused Q-set dof from the solution.
@
33954; 2001; Documentation - PARAM,SMALLQ - Missing Documentation
PARAM,SMALLQ is not documented in the Quick Reference Guide. The purpose
of this parameter is to remove unused Q-set degrees of freedom from the
solution set. If you are performing residual structure qms, then it is
recommended to set PARAM,SMALLQ,0. in order to remove the unused Q-set
dof.
@
33955; ; Elements - BEAM, BAR, TEMP(INIT), TEMP(LOAD) - Incorrect Beam St
resses
In Version 70.5 if TEMP(INIT) references TEMPRB or TEMPP1, then
TEMP(LOAD) must reference TEMPRB and TEMPP1 for the same elements,
conversely if TEMP(LOAD) references TEMPRB or TEMPP1, then TEMP(INIT) must
reference TEMPRB or TEMPP1 for the same elements, else wrong answers will
results.
@
33957; 70.7; Composites - CQUAD8, PCOMP - Poor Results
The bulk data description of the CQUAD8 in the Quick Reference Guide
needs to be changed as follows:
- Remove Figure 1 and change the title of Figure 2 to Figure 1,
- The description of the MCID field needs to be changed as follows:
"MCID
...................MCID coordinate system (defined by the
CORDij entry or zero for the basic coordinate system) onto
the surface of the element (see Remark 3). MCID is
ignored ........."
- Change Remark 3 as follows:
"3.
........the line of constant xi."

- Add the following paragraph to Remark 3:

"The definition of the material coordinate system by
projection is used to calculate an angle THETA.
Please note that since xi changes direction throughout
the element based on element shape, the material
coordinate system varies similarly. Because of this an
orthotropic or anisotropic material will cause the
CQUAD8's stiffness to be biased by both it's shape
and grid ordering. Use the QUAD4 element if a
constant material coordinate system direction is
desired."

Editors comment: Note that the "xi"s above need to be replaced with the
appropriate greek letter.

@
33960; ; Documentation/Aeroelastic UG - SOL 110 with Lagrange Multiplier
Technique
Section 8.11 in the Aeroelastic User's Guide demonstrates the use of the
Lagrange Multiplier Technique with TPL problem ha110a.dat. Currently,
solution 110 problems can not be solved reliably with the LMT approach and
this example should be removed.
@
33973; 70.6; DMAP - MATMOD Option 25 - Wrong Answers
The form of the column vector outputs in MATMOD option 25 is 7 (labeled
"ROW VECTOR") and should be 2 (labeled "ROW RECTANG.").
Avoidance: Run MODTRL on the output; for example,
MATMOD I1,I2,I3,,,/O1,O2/25 $
MODTRL O1// / /2 $
MODTRL O2// / /2 $
@
33994; 70.6; Optimization/Design Sensitivity - scr=no, DVBSHAP - UFM 1126
In Version 70.5, if shape optimization is attempted with the command line
option 'scr=no', then the following fatal message occurs:
*** USER FATAL MESSAGE 1126 (GNFIST)
DMAP MODULE MDCASE IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME
= CASESTAT, WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE CASESTAT=OVRWRT
Avoidance: Insert the following DMAP alter:
compile desopt
alter 1
file casestat=ovrwrt $
endalter
@
33997; 2001; Elements - Large Field PSOLID Entry - Stress Results at GAUSS Po
ints, not GRID
A missing PSOLID large field format continuation line will incorrectly
cause stresses to be returned at the GAUSS points instead of the GRID.
This will change the output values to those calculated at the GAUSS points
instead of the GRID points. This could cause optimistic results to be
inadvertently used for design.
Avoidances:
1) Include the blank continuation line, (you may want to consider
specifying the defaults so someone doesn't delete the blank line thinking
it is a useless line),
2) Use small field format.
@
34005; 70.6; Optimization/Design Sensitivity - Aeroelastic Analysis - Wrong A
nswers
The Version 70.5 release of MSC/NASTRAN enabled the ability to
simultaneously perform static and static aeroelastic analysis and design.
It has been found that, if the static aero and static subcases both have
load requests, the load calculation can become confused and incorrect or
no loads can be applied in the static aeroelastic subcases.
 Avoidance: The only known avoidance is to not use LOAD requests in static
aero subcases in situations where there are also static subcases.
@
34015; 70.7; Rigid Elements - RSPLINE - UFM 4204
The word "WARNING" in "USER WARNING MESSAGE 4204" which warns the user
that there are RSPLINE elements with no dependent degrees of freedom was
inadvertently changed to "FATAL" in Version 70.5. The run continues and
the results are unaffected.
@
34028; 70.5; Elements - CBAR, EMG, PARAM,COUPMASS - SFM 4276, EC 7777
If coupled mass is used with CBAR elements with pinflags, the run may fail
with the following traceback in the EMG module:
Traceback:
Offset 0x00000010 in procedure zerod_
Offset 0x00000130 in procedure ebarkd_
Offset 0x000003ac in procedure emgbrd_
Offset 0x00000a08 in procedure emgpro_
Offset 0x000003c8 in procedure emg_
Offset 0x00000108 in procedure xsem09_
Offset 0x0000097c in procedure mainal
--- End of call chain ---
Avoidance: Do not use coupled mass.

@
34038; 2001; Optimization/Design Sensitivity - DESVAR, DSAJ - SFM 4276, EC 31
When a solution 200 run has both shape and sizing variables and the shape
variables have smaller ids than those for the sizing variables, the job
may fail in the DSAJ module with "SYSTEM FATAL MESSAGE 4276, ERROR CODE
31".
Avoidance: Renumber the DESVAR ids so that the shape DESVAR have larger
ids than the sizing variables.
@
34042; 70.6; Postprocessing - GPSTRESS, PARAM,POST - Missing xdb Output
If your run contains multiple subcases in nonlinear static analysis and
you have requested strfield, output will only be obtained for the first
subcase and subsequent subcase output will be missing.
Avoidance: There is no avoidance other than performing RESTARTs for each
subcase.
@
34043; 2001; Documentation - ITER - Iterative Solver Support of SOL 153 not D
ocumented
The Quick Reference Guide description of the ITER bulk data entry fails to
list solution 153 as a supported solution sequence.
@
34060; 70.6; Superelements - PARAM,METHCMRS - UFM 3032
If superelements are present and PARAM,METHCMRS is specified for the
residual structure, then UFM 3032 will be issued stating that the program
is unable to find the selected set (EIGRL entry).
Avoidance: Insert the following alter:
compile cmpmode
alter 'methc\=',''
if ( seid=0 and methcmrs>0 ) methc=methcmrs $
alter 'read ',''
read cmkxx,cmmxx,,,dynamics,uset,cases,vyxw,sils,uset,,,eqexins/
cmlama,cmphixz,cmmi,cmoeigs,eigvmats,/
'mode'/s,n,nozset//methf/cmset/methc $
endalter
@
34084; 70.7; Aeroelastic Analysis - AMP, SPC, METHOD - SFM 4276 EC 1044 QOPEN
SYSTEM FATAL MESSAGE 4276. QOPEN ERROR CODE 1044 is issued in SOL 145 if
BOTH the METHOD and SPC change (or are not included) for each subcase.

Avoidance: Include the same METHOD and SPC callouts in all residual
subcases or place them above the subcase level.
Also include the following DMAP alter:
compile sefluttr
alter 'modefsrs',''
$ change cases to casem
call MODEFSRS MR,USET,DM,CASEm,DYNAMICS,MAA,KAA,GPLS,SILS,EED,
EQEXINS,,pg,gm,got,kfs,ys,lll,kll/
,,,,PHA,LAMA,PHA0,LAMA0/
app/NORSET/NOASET/false/
0/NOASET/METH/0/true/NOQSET/nsub/0 $
alter 'call setq',''
CALL SETQ CASE1//
SEID/PEID/S,MTEMP/S,K2GG/S,M2GG/S,B2GG/S,MPC/
S,SPC/S,LOAD/S,DEFORM/S,TEMPLD/S,P2G/S,DYRD/s,METH/
S,MFLUID $
endalter
@
34089; 70.7; Composites - RESTART, PCOMP, DBLOCATE - Infinite CPU Loop
If PCOMPs are changed in a RESTART and there is more than one PCOMP entry
in the model, then the run may enter an infinite cpu loop in the DBLOCATE
module.
Avoidance:
You can avoid the error with the following DMAP alter.
malter 'ept,eptr','ept,eptr'
restart ept,,/list/ $
malter 'epts,eptsr','epts,eptsr'
restart epts,,/hists/ $
This alter will redo the entire analysis. Only use this alter if there is
 a
difference in the number of plies between the restart pcomp card and
the new pcomp card.

@
34114; 70.6; Optimization/Design Sensitivity - PARAM,OPTEXIT,3 - Missing Outp
ut
PARAM OPTEXIT 3 is used to exit SOL 200 after evaluating the constraints.
However, the results of this evaluation are not printed, making this
option of marginal utility. These prints could be of particular benefit
if the user has synthesized a response that is not currently available
with standard output requests. The user could see the values of the
response, but avoid costly sensitivity/optimization calculations.
Avoidance: The following dmap alter provides the P2 output (see the
DOPTPRM bulk data entry) when PARAM OPTEXIT= 3
compile desopt $
alter 'optexit=3'(,-1) $
if ( descycle=1
) DOM10 DESTAB,XINIT,,CNTABRG, CVALRG, ,
DVPTAB,PROPI,,R1TABRG,R1VALRG,,
RSP2RG,R2VALRG,,OPTPRMG,OBJTbg,DRSTBLg,
tolv,folv,frqrprg,dbmlib,bcon0,bconxi,wmidg//
-1/DESMAX/OBJIN//eignfreq $
alter 'dom10','dom10' $
@
34139; 2001; Dynamic Analysis - LSEQ, STATSUB - Missing UWM 3204
If a dynamic response analysis specifies a static subcase but then in the
dynamic subcase the dynamic loads are incorrectly applied such that the
LID field of the LSEQ entry references non-existant static load entries,
no warning message will be issued and unexpected answers will result. The
following user warning message should be issued:
*** USER WARNING MESSAGE 3204 (SSG1)
UNABLE TO FIND SELECTED SET ( yy) IN TABLE (SLT ) IN SUBROUTINE
(SSG1).
Avoidance: Other that ensuring that the static load entries exist, there
is no avoidance.

@
34151; 2001; Miscellaneous - Unknown Error Message, Lost some, 2206 of 2209 e
liminated
When the new reordering methods are used, it is possible that the message
"Lost some, 2206 of 2209 eliminated" (the numbers may not be the same) may
be written to the f04 and the job will stop. This appears to happen when
the EXTREME reordering is used and insufficient memory is available. This
has only been observed in two client runs.
Avoidance: If the message above is encountered, disable the EXTREME
method by setting SYSTEM(206) to 1, 8, or 9
@
34153; 70.7; Optimization/Design Sensitivity - PCOMP - UFM 6482, Response Typ
es Mislabeled
 If a composite type response (CSTRESS,CSTRAIN, or CFAILURE) is invoked in
a subcase that does not support this type of response (e.g,
ANALYSIS=MODES). The designated response type is scrambled; CFAILURE
becomes CSTRESS, CSTRAIN becomes CFAILURE, and CFAILURE becomes CSTRAIN.
Further, RTYPE CFAILURE is misspelled CFAIURE in USER FATAL MESSAGE 6482.
Avoidance: The FATAL message is legitimate, but one needs to infer the
correct response type.
@
34200; 70.6; Geometric Nonlinear Analysis - CHEXA, CQUAD4, CTRIA3, TEMP(LOAD)
- Wrong Answers
In sol 4, ssg1 puts out an incorrect load vector for CHEXA, CQUAD4 and
CTRIA3 elements when thermal loads are present along with a global output
system.
Avoidance: Use solution 106 or 64.
@
34208; 70.7; Optimization/Design Sensitivity - TA1 - SFM 3001
When a design variable overrides an analysis property in a p-element
sizing optimization job, it results in the following error message:
SYSTEM FATAL MESSAGE 3001 (TA1)
THE INPUT DATA BLOCK NAMED EHT AT POSITION 10 DOES NOT EXIST.
Avoidance: Ensure that property values for both the analysis and design
model are identical or apply the following dmap alter.
compile desopt
alter 'luset= '
DBVIEW eht0=eht (WHERE DESITER=0) $
EQUIVX eht0/eht/always $
DBVIEW ehta0=ehta (WHERE DESITER=0) $
EQUIVX ehta0/ehta/always $
DBVIEW gdntab0=gdntab (WHERE DESITER=0) $
EQUIVX gdntab0/gdntab/always $
DBVIEW medge0=medge (WHERE DESITER=0) $
EQUIVX medge0/medge/always $
DBVIEW mface0=mface (WHERE DESITER=0) $
EQUIVX mface0/mface/always $
DBVIEW pval0=pval (WHERE DESITER=0) $
EQUIVX pval0/pval/always $

@
34211; 70.6; Shape Optimization - DBLOC, DVSHAP, MATGEN - UFM 6962
When performing shape optimization using an external database for basis
vectors, if the dblocated basis vectors (UGD) do not have same GSET size
as the analysis model, then the following fatal message will occur:
USER FATAL MESSAGE 6962 (MATGEN)
ILLEGAL VALUE FOR PARAMETER 3 = ****
Avoidance: Use an external model displacement vector (UGD) compatible with
the current analysis model GSET size.
@
34236; 70.5; Installation/Authorization - FlexLM Failing on NEC SX4 with SIGA
LARM error -5038
There is a problem on the NEC SX4 machine using the FlexLM network
licensing manager that may cause large jobs to fail with the message:
SIGALARM - Alarm call 5038 error.
Avoidance: There is no avoidance, use a node lock authcode instead.
@
34238; 70.7; Elements - PBEAM, PBAR, IFP - SFM 4276, EC 7777
HP systems may have fail with a traceback in IFP when it attempts to read
bulk data inputs that contain PBAR and/or PBEAM input with very large
property values.
Avoidance: There is no avoidance.
@
34239; 2001; Documentation - TLOADi - Missing Documentation
A remark under the description of the TLOADi entries in the Quick
Reference Guide indicates that if time dependent loads are applied in a
frequency response solution, that a Fourier transform is performed to
transform the loads to the frequency domain and that the results are
converted and reported in the time domain. This remark should refer the
user to the Section 5.4 in the Advanced Dynamics User's Guide to obtain
more information on the process.
@
34254; 70.7; Iterative Solver - ITER, BIC - Poor Performance, Poor Convergenc
e
If an initial guess is given to the iterative solver with BIC
preconditioning (default) in MSC/NASTRAN V70.5 it will be ignored.
avoidance: To be able to use an initial guess with the Version 70.5 set
PRECOND=BICWELL or PRECOND=BICILL on the ITER Bulk Data entry and set
SYSTEM(241)=2000 on the NASTRAN entry.
@
34259; 2001.0.1; Acoustic Analysis - Superelements - UFM 9101
When a fluid only model with a fluid superelement is run in solution 103,
the job may fail with the following messages:
^^^ BULK DATA AND CASE CONTROL PARTITIONING, RESTART CHECKING, AND
UNDEFORMED PLOTTING IS
^^^ INITIATED FOR SUPERELEMENT 0
^^^ USER FATAL MESSAGE 9101 (PHASE0)
^^^ FLUID ELEMENTS ARE DETECTED IN THIS SUPERELEMENT.
^^^ USER ACTION: ENSURE THAT ALL FLUID ELEMENTS ARE ASSIGNED TO
SUPERELEMENT yy
Avoidances:
1) Do not use a superelements for this analysis, it is not needed. This
may require stripping the superelement designation from all of the fluid
grids if the model comes from a fluid/structure analysis.
2) Add a dummy structural element similar to the following:
 grid, 10000 $ token grid point in se 0
celas2, 10000, 1.0 10000, 1 $ token element
param, ascoup, no $ try to avoid error

@
34264; 70.6; Estimate - File Names may not Contain Blanks
Blanks in pathnames are not handled properly when "memory=estimate" is
specified. Only the part of the directory name after the blank is shown
and estimate issues the message "This file does not exist." The job
continues, but the default memory or buffsize is used.
Avoidance: Specify a pathname that does not include blanks.
@
34265; 2001; Documentation/Superelements - PARAM,MINIGOA - Undocumented Param
eter
In MSC/NASTRAN Version 70, an undocumented parameter (MINIGOA) was added
to allow reduction in the amount of disk space used when using
superelements. When this parameter is used, the transformation matrix
(GOA) will contain terms only for the degrees of freedom defined in the U5
(USET, USET1, SEUSET, SEUSET1) set. As the GOA matrix is often one of the
largest matrices stored in the database, this can allow for a significant
reduction in the amount of disk space used. The limitation of using this
approach is that data recovery will be available only for these degrees of
freedom and the elements connected to them.
@
34266; 2001; Documentation - EIGRL - SYSTEM(273) Not Documented in the QRG
In Version 70.5, setting SYSTEM cell 273 to 1 allows the selection of the
old (pre-v70.5) Lanczos eigensolver. This is mentioned in the Version 70.5
Release Guide, but is not documented in the QRG.
@
34270; ; P-Elements - Wrong Answers for p-BEAM Element at p=1
BEAM p-elements with p-order of 1 are stiffer in bending than BEAM
h-elements due to incorrect residual bending flexibility in BEAM
p-elements.
Avoidance: The avoidance is to start at a p-order of 3.

@
34271; 70.6; Optimization/P-Elements - P-BEAM - Wrong Results in OES1 Table
Wrong ENDB max and min stresses are written to the OES1 table for the
p-Beam element types by Module SDRP. However, the OFP printout of the ENDB
max and min stresses is correct since the OES1 table values are not read
but computed by the OFP module. This error will affect design optimization
results and postprocessors which report these values.
Avoidance: There is no known avoidance.
@
34287; 70.6; Aeroelastic Analysis - Incorrect Stability Derivative Sensitivit
ies
The UXU and UXR matrices produced by DSARLP may be in error.
Consequently, this can produce wrong answers for the sensitivities of
 stability derivatives and will produce unpredictable downstream errors.
Avoidance: There is no known avoidance. This capability should be used
with caution. Zero or unreasonable stability derivative sensitivities may
be an indication that this error has been encountered.
@
34290; 2001; Documentation/QRG - TABLES1 - Discontinuity is not Allowed, UFM
316
The Quick Reference Guide description for the TABLES! entry should
explicity state that discontinuities are not allowed and will lead to a
fatal message.
@
34321; ; Installation - FLEXlm on NT needs lmutil.exe program
The client installation of FLEXlm on NT needs to also install the
lmutil.exe program (and the lmgr325c.dll file) so the user can check the
availability and usage of Nastran licenses. The current installation does
not allow this. Under the current installation procedure the user would
have to get these files from the FLEXlm license server computer. This
should be automatically done with the installation.

Avoidance: The flexlm directory in the standard installation should have

lmutil.exe. Also, the msc705\i386 directory contains lmgr325c.dll. The
installer may have to copy the .dll from the i386 directory to the flexlm
directory for lmutil to be able to find it.
@
34325; 70.7; P-Elements - GMLOADs on FEEDGEs with p=1 may give bad answers
GMLOADs on FEEDGEs with p=1 will give incorrect loads if another edge with
p>1 was processed.
Avoidance: There is no known avoidance.

@
34362; 70.7; P-Elements - Incorrect p-shell Output with PROJ=F Option
If the PROJ=F option is chosen on the OUTRCV entry for p-shells, the
moments may be incorrect. In addition, if a coordinate system is
specified with the CID=n option, the transformation may be incorrect.
Note that the PROJ=F option does not output any data into the f06 file,
but it does affect the output in the OP2 and XDB files. The option was
created primarily for use by MSC/ARIES.

@
34393; 70.7; Utilities - xdb, DBC - msc2ptc Failed to work for NASTRAN Versio
n 70.5
The msc2ptc utility does not run due to a problem with the nast705 script
and missing m2p.dsc file. An attempt to convert the xdb file fn.xdb
results in the message:
fn: executable for fn not found or not executable.
Verify that the fn executable has been installed.
The full pathname of the bracket executable is:
/nastran705/msc705/solaris/fn
 Avoidance: The updated files are posted at
http://www.macsch.com/support/support.html
@
34415; 70.6; Superelements - EPOINT - Wrong Answers
In part superelements with extra points, the sequence of grid points
provided in the usetd table, produced from the uset table by adding extra
point data, may differ. This will cause plausible but wrong answers. In
one example discovered in testing the load vectors were in a different
sequence than the corresponding terms in the stiffness matrix.
Avoidance: An avoidance which prevents the error is to put in at least
one SEQGP entry for a residual structure grid point. It can be input in a
manner that will cause no extra resequencing of the model. For example,
if the main bulk data file includes grid point 47, the following SEQGP
entry will suffice:
SEQGP, 47, 47.1
This avoidance may be used with any model, of any size, with negligible
effect on cost. The presence of an SEQGP entry forces the code to avoid
the branch which contains the error.
@
34424; 2001; File Management Section - DBDICT May Fail to List All Keys
NDDL datablocks are uniquely defined by their qualifier values. In
addition to the NDDL datablock name, a path, or list of applicable
qualifiers, is thus used to "point" to a
particular datablock. The DBDICT FMS statement (which can also be called
in DMAP) can be used to list qualifier values for datablocks stored on the
database, and does so by printing out "keys," or unique lists of
qualifiers and their values. A code error may prevent some of the keys
from being printed out.
Avoidance: If this problem is encountered, the DBDIR FMS statment
(also callable in DMAP) can be used to obtain a complete qualifier
listing.
@
34429; 70.6; Miscellaneous - Overflow "*****" in the f04.
If a job requires more than 999999.9 CPU seconds, then the F04 will print
********. If a job requires more than 999999.9G of I/O then ******** will
be printed.
Avoidance: There is no avoidance.
@
34459; 70.7; Heat Transfer - Missing RADLST Information in Punch File
If large CHBDYG id's exist, then the punch file which contains the RADLST
entries will be incomplete and contain a line such as:
RADLST 12 1****************
Avoidance: 1) Do not use element ID greater than 10000001
2) Use restart for the subsequent run rather than including the
 RADLST punch file because all the radiation matrix
information is in the NASTRAN database.
@
34462; 70.6; P-Elements - HEXA - Poor Answers, Element too stiff when modelin
g thin plate
The HEXA element with p=1 on most edges is too stiff. This was discovered
when analyzing the thin flat plate test problem of the MacNeal Harder test
set. For most problems the effect is small, but for the thin flat plate
model the effect is dramatic.
Avoidance: Use a higher p-order.
@
34467; 70.7; P-Elements - CHEXA, CPENTA, CTETRA - Wrong Answers for Transform
ed Stresses
If stresses and strains are being transformed using the CID= keyword in
the OUTRCV bulk data entry, then wrong answers may result.
Avoidance: The only avoidance is to request the stresses in the basic
(default) coordinate system.
@
34474; 70.6; Security - FLEXlm License may be Refused by the nastran Command
The nastran command will refuse to accept an otherwise valid FLEXlm
license spec for redundant servers written as:
port@host1,port@host2,port@host3
Avoidances:
1) Do not use redundant servers. They are high maintenence with little
gain. Multiple servers, specified via "port@host1:port@host:..." on UNIX
and "port@host1;port@host;..." are more powerful and less work.
2) Only specify one of the hosts. Each of the hosts in the quorum knows
about the other two, so the comma specification is redundant. You could
even specify all three hosts as if they weren't redundant, using the ":"
or ";" separators as appropriate.
@
34480; 70.6; Eigenvalue Analysis - PARAM,RESVEC,YES - Missing Eigenvalues
By simply specifying PARAM,RESVEC,YES, incorporated residual flexibility
in the solution sequences making the the resflex sssalter unnecessary.
There is an error in the capability which may cause missing modes. The
error in the results is obvious because there will be fewer rather than
more eigenvalues when you compare the eigenvalues "BEFORE AUGMENTATION OF
THE RESIDUAL VECTORS" to those "AFTER AUGMENTATION OF THE RESIDUAL
VECTORS".

Avoidance : Insert the following DMAP alter:

compile resvec
alter 'if ( noxset'(3),''
if ( noxset>-1 or append ) then $
endalter
@
34482; 70.6; DMAP - Database, DBSTORE - DBSTORE Does not Function on RESTART
The automatic restart logic does not schedule modules to run on restart
that have no output data blocks. The dbstore subdmap, and the dbstor
subdmap which it calls, have this attribute.
Avoidance: Use putsys(1,109) before the call statement, or use an alter
in the v70.5 sssalter library that includes dummy output data blocks for
both modules.

@
34491; 70.7; Delivery Database - eie layer unable to extend file
On Cray systems, if a user has the Delivery Data Base (SSS*) files on a
device that does not have a sector size that is a multiple of 4096 words,
and if the user specifies ff_io_defaults=eie.sds.diag, then the following
message is printed:
eie layer unable to extend file
"/apps/mscnast/v7052/msc705/unicosts/SSS.MASTERA" to a multiple of 65536
bytes
/apps/mscnast/v7052/msc705/unicosts/SSS.MASTERA : open failed : errno
=5048 : Unknown error
ffopen(/apps/mscnast/v7052/msc705/unicosts/SSS.MASTERA) using alternate
layers system
opening layer system
/apps/mscnast/v7052/msc705/unicosts/SSS.MASTERA : using layers : syscall
The job will then continue to run if the the BUFFSIZE is 4097. However,
if BUFFSIZE is set to a different value, then the following error is
printed:
** ERROR ** UNABLE TO VALIDATE DATABASE FILE FORMAT. DATABASE MAY NOT
HAVE BEEN BUILT ON A COMPATIBLE PLATFORM.
BIOPNR ERROR ** TIMESTAMP/RECL VALIDATION FAILED. FILX = 6, LOGNAME
= MASTERA , NSBUF3 = 32768
FILE = /apps/mscnast/v7052/msc705/unicosts/SSS.MASTERA
RECL SHOULD BE 32768, ACTUAL VALUE IS 4096.
Avoidance:
1. Change BUFFSIZE to 4097
2. Turn off FF_IO
@
34517; 70.7; P-Elements - STRAIN, SHELLS & SOLIDS -- Incorrect Principal and
von mises strain
If a problem has either shell or solid p-elements, then the principal, von
Mises, and mean pressure strains may be incorrect.
Avoidance: There is no avoidance.

@
34520; 70.7; Acoustic Analysis - METHOD(STRUCTURE) - Wrong Answers
In fluid/structure analysis, in modal frequency response (SOL 111), if the
fluid modes are calculated using the modal method, i.e. the case control
 command METHOD(FLUID)=n, but no METHOD(STRUCTURE) is specified, incorrect
results occur.
Avoidances:
1. Use SOL 108, or,
2. use both a METHOD(FLUID) and a METHOD(STRUCTURE), or,
3. put the fluid in a superelement and process it separately, or
4. use the following alter for Version 70.5:
compile gma
alter 295
partn k1dd,vafse,/k1shh,,, $
partn kddd,vafse,/kdhh,,, $
igh=ig $
call xadd5 k1shh,kdhh,,,/gk1shh/igh/cu/cz/cz/cz $
endalter $

@
34523; 2001; Optimization/Design Sensitivity - DEQATN - SFM 4276
Optimization runs with large DEQATN entries that are near or exceed the
limit of 32000 non-blank charachters may fail on Cray computers with the
following message:
lib-1190 /u/app/nastran/msc/msc705/unicosc90/NAST705: UNRECOVERABLE
library error
A numeric input field contains an invalid character.
Encountered during a sequential formatted READ from an internal file
(character variable)
Error initiated at line 656 in routine '_rdfmt'.
SIGNAL - Signal SIGABRT(6) caught, job aborting.
See signal(2), CRI manual SR-2012 for further information.
Beginning of Traceback:
Called from line 42 (address 15624627c) in routine 'TRBK'.
Called from line 39 (address 132322c) in routine 'RECOVR'.
Called from line 112 (address 15073635b) in routine 'sysrcv'.
Called from address 15335217c in routine '__handlr'.
Called from address 15327333d in routine '_lwp_killm'.
Called from line 16 (address 15331121a) in routine 'raise'.
Called from line 92 (address 15174456c) in routine 'abort'.
Called from line 59 (address 15674675d) in routine '_ferr'.
Called from line 656 (address 15732746d) in routine '_rdfmt'.
Called from line 346 (address 15665655d) in routine '_xfer_iolist'.
Called from line 350 (address 15737620b) in routine '_FRF'.
Called from line 56 (address 6106040c) in routine 'PRERR2'.
Called from line 132 (address 4610355c) in routine 'IPRSCM'.
Called from line 429 (address 4505303c) in routine 'IFSEQU'.
Called from line 1816 (address 4454631d) in routine 'IFS6P'.
Called from line 1184 (address 4336537b) in routine 'IFPDRV'.
Called from line 198 (address 4237106a) in routine 'IFP'.
Called from line 293 (address 10526562c) in routine 'XSEM02'.
Called from line 311 (address 2064b) in routine 'MAINAL'.
Called from line 306 (address 4617d) in routine '$START$'.
End of Traceback.
 The IBM AIX RS6000 completes but outputs numerous occurances of the
following log file message:
1525-097 A READ statement using decimal base input found the invalid digit
'' in
the input file. The program will recover by assuming a zero in its
place.
Avoidance: Other than shortening the DEQATN entry there is no avoidance.
@
34535; ; Optimization/Design Sensitivity - PCOMP, MAT8, SDRCOMP - UWM 202
0, UFM 5503
In solution 200, if ply failure indice responses are requested for a
composite element whose PCOMP entry references a material entry that does
not define the composite failure allowables for one or more of the plies,
then the following fatal and/or warning messages are issued and the
failure index responses are not used for that element in the optimization
analysis.
** USER FATAL MESSAGE 5503 (DSAD13)
ELEM NO. 42649 PLY NO. 1 REFERENCED ON A DRESP1 ENTRY COULD
NOT BE FOUND.
*** USER WARNING MESSAGE 2020 (SDRCOMP)
ONE OR MORE OF THE ALLOWABLE STRESS CONSTANTS WERE INCORRECT FOR
ELEMENT 1 PLY 4
ELEM NO. 1 PLY NO. 1 REFERENCED ON A DRESP1 ENTRY COULD
NOT BE FOUND.
USER ACTION: CHECK THE ELEMENT IDS REQUESTED ON THE DRESP1 ENTRY.
IF A CFAILURE RESPONSE TYPE IS REQUESTED, CHECK THAT THE
CORRESPONDING PCOMP AND MAT1
BULK DATA ENTRIES HAVE BEEN CORRECTLY SPECIFIED.
USER INFORMATION : USE OF UNREFERENCED ELEMENTS IS NOT RECOMMENDED
AND MAY INDICATE A USER INPUT ERROR.
RESULTS FOR OTHER RESPONSES ARE NOT AFFECTED.
Avoidance: Specify the failure indices. If isotropic ply properties are
needed desired, use an equivalent MAT8 entry.
@
34548; 70.6; Acoustic Analysis - EPOINT, CAABSF - SFM 6719 (MRGIDX)
In frequency response, if EPOINTs exist in a fluid-only analysis with
CAABSF elements with no SPCs, the following fatal message will be
encountered:
*** SYSTEM FATAL MESSAGE 6719 (MRGIDX)
MATRICES HAVE DIFFERING NUMBERS OF ROWS:
[A] = 48 x 48 [B] = 44 x 44 [C] = 48 x
48 [D] = 48 x 48
0FATAL ERROR
Avoidance: Insert the following DMAP alter or add a dummy grid point and
constrain all of its degrees-of-freedom.
compile fdrmgen
alter 'if ( noa',''
 type parm,,i,n,noue $
if ( noa=-1 and noue=-1 ) then $
compile freqrs
alter 'if ( noa'(3),''
IF ( noa>-1 or noue>-1 ) then $
endalter
@
34552; 70.6; Acoustics - EPOINTs, GMA, GKAM - Fails with no message
If EPOINTs are used in a fluid-only model in solution 111, the run will
fail without message after the eigenvalue extraction and before the
frequency response output.
Avoidance: Use the following alter:
compile gma
alter 'gkam'
if ( fonly and noue>-1 ) then $
matgen ,/iee/1/noue $
paraml phfa//'trai'/1/s,n,nohset $
paraml phfa//'trai'/2/s,n,noaset $
type parm,,i,n,nhe,nae $
nhe=nohset+noue $
nae=noaset+noue $
matgen ,/vhe/6/nhe/nohset/noue $
matgen ,/vae/6/nae/noaset/noue $
merge phfa,,,iee,vhe,vae/phdfh/1 $
merge MFHH,,,,vhe,/mfhh1 $
equivx MFHH1/mfhh/-1 $
merge K1FHH,,,,vhe,/k1fhh1 $
equivx K1FHH1/k1fhh/-1 $
endif $ fonly and noue>-1
@
34558; 70.7; P-Elements - BEAM, OUTRCV - Wrong Answers at View Points
The results output at view points for beam p-elements is wrong if a curved
coordinate system is specified as the output coordinate system on the
OUTRCV entry.
@
34581; ; Heat transfer - VIEW3D - FATAL, DIMENSION LENGTH x IS SMALLER T
HAN "VFZTOL"
It is possible running a view factor calculation using the VIEW3D module
(this is the default view factor method selected by PATRAN) to encounter
the message:
*** FATAL *** DIMENSIONAL LENGTH: 1.732432600E-11 IS SMALLER THAN
"VFZTOL": 1.000000013E-10
REASSIGN "PARAM,VFZTOL,X.XE-XX" TO BE SMALLER THAN THE
DIMENSIONAL LENGTH.
Avoidance: Comment out the VIEW3D input so that the older view factor
module is used. In the test problem that encountered the poor message, the
old view factor module issued the messages below which identified the
problem:
*** USER FATAL MESSAGE 4528 (VNASIN)
ERROR IN CHBDY CARD xx AN AREA4 ELEMENT MUST BE A QUADRILATERAL
 CONTAINING NO ANGLE GREATER THAN OR EQUAL
TO 180 DEGREES. THE GRID POINTS MUST BE SPECIFIED SEQUENTIALLY AROUND
THE QUADRILATERAL.
*** USER FATAL MESSAGE 4530 (VNASIN)
ERROR IN CHBDY CARD yy THE FOUR GRID POINTS MUST ALL BE IN ONE
PLANE
@
34585; 2001; Database - SFM 1041 (GALLEX)
A limit in the database directory size may cause large jobs to fail with
the message SYSTEM FATAL MESSAGE 1041 (GALLEX).
Avoidance: This limitation can be overcome by increasing the buffsize
and/or setting system(158)=1000 and system(159)=500 to allow for a larger
expansion of the database.
@
34590; GL; Eigenvalue analysis - EIGB, SINV, CYCLIC4 - SFE 4276, FREAD EC 3
002
In SOL 16, cyclic buckling, if the SINV method is used, and no eigenvalues
are found for the current harmonic, then the run fails with the following
message in CYCLIC4.
*** SYSTEM FATAL ERROR 4276.FREAD ERROR CODE 3002 PID= 0
Avoidance: Use the Lanczos method.
@
34591; 70.6; Cyclic Buckling - SOLs 116 and 16, CYSUP - SFM 3001 and UFM 1126
The use of CYSUP in SOL 116 causes SFM 3001 in UPARTN. In SOL 16, if
eigenvalues are requested at more than one harmonic, then UFM 1126 in
CYCLIC3 is issued.
Avoidance: Insert the DMAP alter:
compile sol16
alter 1
file pvec=save,ovrwrt $
COMPILE MODECYC $
ALTER 'CALL *MODERS'(,-1) $
DELETE /MR1,,,, $
REACT = -1 $
endalter
@
34592; ; Aeroelastic Analysis/Random Analysis - EPOINT - Missing Random O
utput
Random analysis with an aeroelastic model fails to yield the acceleration
random output if any random output requests for EPOINT dofs are made.
Avoidance: Make two separate runs, with the epoint data recovery being
requested in a separate runs. The runs will be more efficient if restart
is used.
@
34602; 70.7; Frequency Response - STRESS(BILIN), STRESS(CORNER) - Wrong Answe
rs
Complex corner stress output for QUAD and TRIA elements is incorrect in
 MSC/NASTRAN Version 70.5.2. The value for the real part of the stress at
the grid point will be overwritten with the imaginary value. The results
at the center of the element are not affected.

Avoidance: Other than not requesting the corner stress option, there is no
avoidance.

@
34604; 70.6; P-elements - Estimate - Over-prediction of Resources
MSC/ESTIMATE may predict significantly more resources than required for
P-elements. This over-prediction is partially a result of the uncertainty
of the P value.
Avoidance: Do not use MSC/ESTIMATE for P elements.
@
34613; 70.6; Cyclic Symmetry - LOADCYN - UFM 4342, Job Aborts
If a cyclic symmetry job encounters the fatal message:
USER FATAL MESSAGE 4342 (CY1BLD)
THE LOAD BULK DATA CARD GIVEN BELOW IS INCONSISTENT WITH SPECIFIED
ANALYSIS HARMONICS OR SEGMENT OR SYMMETRY TYPE.
then the job will not produce an "END OF JOB" and will issue the traceback
below:
Traceback:
Offset 0xffffffff in procedure
Offset 0x0000014c in procedure WriteUnit
Offset 0x00001094 in procedure _xlfWriteFmt
Offset 0x00000884 in procedure cy1bld_
Offset 0x0000063c in procedure cy1ccl_
Offset 0x000008c8 in procedure cycl1_
Offset 0x00000114 in procedure xsem12_
Offset 0x000009a0 in procedure mainal
--- End of call chain ---
Avoidance: Although the job should not issue a traceback, the fatal
message output is correct and the user should correct the LOADCYN input.
@
34652; 2001; Documentation - RFORCE, CTRIAX6 - METHOD 1 and 2 Produce the Sa
me Loads
Remark 9 in the MSC/NASTRAN V70.5 Quick Reference Guides for the RFORCE
bulk data entry states:
9. METHOD=2 should be used for the CTRIAX6. METHOD=1 produces null loads.
This is incorrect. Method 1 produces the same loads as method 2.
@
34665; 70.6; Estimate - Nonlinear, HEXA, PENTA, TETRA - Under-predicts Memory
In a NonLinear analysis (SOL 66, 99, 106, 129, 153,159) the ESTIMATE
 utility erroneously assumes the rotational degrees of freedom are fixed
for grids attached only to 3-dimensional elements. The result is that the
memory and disk requirements may be under-predicted.
@
34666; 70.7; Optimization/Design Sensitivity or Superelement- GENEL - Possib
le Wrong Answers
If the number of GENEL's is an odd number greater than 1, the stiffness
terms produced by the GENEL's will be in error on the second and
subsequent design cycles in SOL 200. Similarly, in other solution
sequences, having GENELs in more than one superelement, or in a
superelement and residual structure can result in some of the GENELS not
generating their respective stiffness matrices which will lead to wrong
answers.
Avoidance: Insert the following dmap alter.
compile semg1 $
alter 1 $
type parm,,i,n,nsize $
alter 'sma3.*nogenl','' $
if ( nokggx<0 ) then
paraml kjjz//'trailer'/2/s,n,nsize $
else $
paraml kjjx//'trailer'/2/s,n,nsize $
endif $
matgen ,/knull/7/nsize/nsize $
sma3 gei,knull/kjjy/lusets/nogenl/-1 $
$
@
34671; 70.7; Beam Library - PBARL, PBEAML - Incorrect Sensitivity for Custom
Cross Sections
If custom beam cross section are added to the beam library and they are
applied in the model without providing analytical sensitivities of the
beam properties with respect to the beam dimensions, the sensitivities
calculated using finite difference techniques will be too large in
magnitude by 50%.
Avoidance: There is no easy avoidance except to manually scale the
gradients. The sensitivities are of the correct sign and are therefore of
some value when optimizing the cross section.
@
34699; ; Heat Transer Analysis - Coupled Convection - SFM 6934
The fatal message below will be encountered if a connection grid is also
specified as an ambient point.
*** SYSTEM FATAL MESSAGE 6934 (EMGPOM)
FOR ELEMENT ID = 100020 AND FILE TYPE = 1
EMGPRO DETERMINES A DIFFERENT COUNT OF ACTIVE CONNECTION SILS THAN
DOES CALLER OF EMGPOM
EMGPRO COUNT = 5 CALLING ROUTINE COUNT = 6
Avoidance: If the PATRAN interface is being used, make sure there are no
overlapping points between application region one and application region
two while applying coupled convection LBCs.
@
34706; 70.6; Restarts - RSSCON - UFM 9086
 Running a nonlinear analysis restart with RSSCON entries results in USER
FATAL MESSAGE 9086. In other solution sequences the restart may be
inefficient. The restart fails because of a logic error in subdmap
PHASE0.
Avoidance:
compile phase0
alter 'if ( noxsort )',''
equivx geom1/ngeom1/-1 $
endalter
@
34714; 70.6; Optimization/Design Sensitivity - Superelements - UFM 7052, Wron
g Answers
RTYPE=FREQ was introduced in Version 70 as a "friendly" alternative to
RTYPE=EIGN. When RTYPE=FREQ is used with superelements,
it can produce an incorrect user fatal message:
*** USER FATAL MESSAGE 7052 (DOMPTC)
ILLEGAL PROPERTY NAME

This error was observed when there was no DESOBJ request. When a DESOBJ
request was included, the job ran, but the sensitivity results were
incorrect.
Avoidance: Use RTYPE=EIGN and not RTYPE=FREQ when there are superelements.

@
34740; 70.7; Postprocessing - Corrupted XDB
Access Objects with internal keys exceeding 214747 in solution sequences
other than linear statics will corrupt the object maps in the xdb. The
internal key id is a 1 to n list of any particular type of output such as
grids, bar elements, quad elements etc. In the Version 70.5 system, logic
has been added to prevent integer overflow before the access key
formation. A FATAL error indicating that translation has stopped is
generated if an overflow condition is encountered.
Avoidance: Each superelement restarts the 1 to n list for each type of
output, thus splitting the model into superelements is one avoidance.
@
34756; 2001; Documentation/Aeroelastic User's Guide - Inconsistent Descriptio
n of REFC Field
The Aeroelastic User's Guide should be consistent in referring to the REFC
field on the AERO and AEROS entries as being the reference chord. The
following places in the Guide incorrectly refer to REFC as the semichord:

Section 8.2, page 350, should refer to chord rather than semichord.
Section 8.13 (example HA145KR ) incorrectly inputs the reference semichord
on the AERO bulk data entry. The REFC should be 40.0 and the same answers
could be produced if the reduced frequency values were doubled.
Section 9.1, page 578 should refer to chord rather than semichord.
@
34766; 70.7; Beam Library - PBARL, NSM, TYPE=HEXA = Incorrect Mass Calculatio
n
For the beam library, PBAR section properties are computed based on the
user's PBARL input. Due to a code error, this computation is incorrect
for the NSM (non-structural mass) property calculation for section
TYPE=HEXA. Therefore, if the mass properties of the BAR are used in the
analysis, the answers will be incorrect.
Avoidance: The PBEAML/PBEAM computation is correct for TYPE=HEXA, so
PBEAML's could be used in place of PBARL's for this section type if the
bar is not being designed. The error is benign in a static analysis
without inertial relief or gravity loads.
@
34767; 70.6; SSSALTERS - checka - Incorrect Degree of Freedom Identifiers
The output for the checka alter was modified for V69.1. Due to an error in
the alter, the degree of freedom identifiers for the matrices REACCOLN
and REACCOLA will be incorrect.
Avoidance: Change the following line in the alter:
MATGPR GPLS,USETB,SILS,REACcoln//'H'/'G'//printol $
to:
MATGPR GPLS,USETB,SILS,REACcoln//'H'/'N'//printol $
and change the following line:
MATGPR GPLS,USETB,SILS,REACcola//'H'/'G'//printol $
to:
MATGPR GPLS,USETB,SILS,REACcola//'H'/'A'//printol $

@
34774; 70.7; Fluid Elements - GRIDB - UFM 315
In Version 69 and later MSC/NASTRAN releases, when fields 3, 4, and 6 of
the GRIDB entry are used the job will fatal with the following error
messages:
*** USER FATAL MESSAGE 315 (IFPDRV)
FORMAT ERROR ON BULK DATA ENTRY GRIDB x
Although the fields are not used, Remark 6 of the Quick Reference Guide's
description of the GRIDB entry clearly states:
"Fields 3, 4, and 6 are ignored, which facilitates the conversion of GRID
entries to GRIDB entries. Note that the fields are the same except for
fields
1 and 9 when a cylindrical coordinate system is used."
Thus, allowing information in these fields is an important design feature.
When blanks are used on fields 3, 4, and 6 of the GRIDB in an attempt to
circumvent the above error, if structure with GRIDs is present, then
arbitrary errors occur specifying that GRIDs are not present, when they
are. Such as:
 *** USER FATAL MESSAGE 2007 (GP2)
ELEMENT 402 REFERENCES UNDEFINED GRID POINT 1381
@
34781; 70.7; Eigenvalue Analysis - READ - Poor Message, SFM 7340 (LNNHERR)
If a Lanczos eigenvalue analysis requests more modes than exist, it is
possible that the job will fail with the message:
*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
process error reported by subroutine LNNP2CD
see LNNP2CD abstract (ier = -729)
*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
warning reported by subroutine LNNDRVD
see LNNDRVD abstract (ier = 729)
This error has been observed only on small problem where some dofs did
not have independent mass terms.
Avoidance: An avoidance is to set f2 to limit the search region within
plausible limits. For small models where all the eigenvalues are desired
a Givens method may also be selected.
@
34783; ; Case Control - MPCFORCE - Incorrect Subcase Labels
If mpcforce output is requested in a multiple subcase run, and not all
subcases have an mpc callout, then the subcase labels for the mpcforce
output will be incorrect. The results, however, are correct. The labels
will be in ascending order of the subcase ids. Using the following case
control setup,

Avoidance: Ignore the subcase labels, the results are correct. Else
place the subcases with MPC requests first.
@
34784; 2001; Documentation - SYSTEM(198), QRG - Incorrect Default Value
The MSC/NASTRAN Quick Reference Guide states that the default for
SYSTEM(198) (a.k.a MINFRONT) is 8. The default is actually machine
dependent is documented in the Configuration and Operations Guide.

@
34795; ; Nonlinear Static Analysis - PARAM,BUCKLE, Thermal Loads - UFM 20
44
In a PARAM,BUCKLE,1 restart with thermal loads, if the PARAM,LOOPID
specifies a loop ID which occurred in a subcase earlier than the previous
subcase in the cold start run, then the restart run will fail with UFM
2044 in the NLCOMB module.
Avoidance: In the restart run remove all subcases after the subcase in
which PARAM,LOOPID occurred. For example, if the cold start contained 4
subcases and PARAM,LOOPID value occurred in the 2nd subcase, then remove
the third and fourth subcases in the restart run.
@
34824; ; Nonlinear Analysis - CQUAD4,CTRIA3 TABLEM1 - Wrong Intermediate
Output
 All output for the QUAD4 and TRIA3 elements at intermediate steps (NLPARM
INTOUT > 1) is wrong when there exists a temperature dependent material
property.
Avoidance: Other than ignoring the intermediate output there is no
avoidance.
@
34833; 70.5; Frequency Response - Corner Force - Missing or Wrong Corner Forc
e
If corner force output ( stress(corner...)=x ) is requested in a frequency
response analysis, the corner force for the QUAD4 will either be missing
or incorrect. This error has been corrected for Version 70.5.
Avoidance: Use the center stress.
@
34834; 70.5; Eigenvalue Analysis - PARAM,RESVEC - Wrong Answers
If a C-set is defined for a superelement, the component modes are passed
thru a filtering process to remove any which might be a linear combination
of the static transformation. In V70, if residual vectors are not
requested, the component modes will not be passed thru the filter. This
may lead to either rank-deficient matrices or wrong answers. If the
filtering is not performed, the reduced stiffness matrix will incorrectly
contain the original eigenvalues on the diagonal and there will be no
off-diagonal terms (as there should be if the filtering is not performed).
Avoidance: If residual vectors are requested with at lease one loading or
U6-set dof defined, then the filtering will be performed. Otherwise, use
Version 70.5 or later version.
@
34850; ; Postprocessing - DBC, PARAM,POST - OPNPFL Message to log Rather
than f06
DBC OPNPFL messages may be written to the log file rather than the f06.

@
34874; 70.7; Eigenvalue Analysis - PARAM,RESVEC, Superelements - Poor Answers
If residual vectors are requested in dynamic analysis (PARAM,RESVEC,YES)
and none are appended, the original modes will be lost without warning.
The user should also see related error 34480.
Avoidance: Either do not include residual vectors, make sure that
residual vectors are calculated, or include the following DMAP alter:
compile resvec
alter 'WERE AVAILABLE AFTER FILTERING'
equivx phx/aphx/-1 $
equivx lama/alama/-1 $
endalter
@
34902; 70.7; Optimization/Design Sensitivity - CBUSH, CVISC, CDAMP - UFM 1126
 Optimization (SOL 200) produces User Fatal Messages 1126 when damping
element properties ( CDAMP, CVISC, CBUSH) are active design variables and
superelements are used in the model.
Avoidance: Apply the following DMAP alters:
compile desopt
alter 'type DB'
type db belmdsf, bdictdsf, belmdsb, bdictdsb $
alter 'dopr5'
type db zuzr01
equivx tabdeq/zuzr01/always $
compile fea list
alter 'type parm'
type db zuzr01
alter 'dspt1'(2),'dspt1'(2)
PARAML zuzr01//'TRAILER'/1/S,N,NDVAR $ no. of des. var.
alter 'dspt1'(3),'dspt1'(3)
PARAML zuzr01//'TRAILER'/1/S,N,NDVAR $ no. of des. var.
matprn zuzr01 // $
alter 'trd2'(2),'trd2'(2)
TRD2 CASERS,TRL,NLFT,DIT,KHH,BHH,MHH,PHT,zuzr01,,,/
UHTDS,,TOLSCR/
SOLTYP/NOUE/NONCUP/0 $
COMPILE PSLGDV
alter 'emg,'(3),'emg,'(3)
IF ( DOFREQ>0 OR DOMTRAN>0 ) EMG,
EST,CSTMS,MPTS,DIT,,,ett,,,,BGPDTS,GPSNTS,,,,ditid/
,,,,BELM,BDICT/0/0/ $
compile super3
alter 'lcgen '(2),''
dbstatus cases//s,n,nocases $
if ( delta and nocases<1 and seid=0 ) then $
LCGEN CASErS,SLT,ETT/CASESX $
else $
LCGEN CASES,SLT,ETT/CASESX $
endif $
@
34925; 2001; Documentation - QRG - DISPLACEMENT
Remark 2b for the DISPLACEMENT Case Control Command should be modified
from

SORT2 is the default in transient response analysis (structural and heat

transfer). SORT2 is not available for real eigenvalue (including
buckling) complex eigenvalue, or flutter analysis. If SORT1 is selected
for one or more of the commands ACCE, DISP, ENTH, FORC, HDOT, MPCF, OLOA,
SPCF, STRA, STRE, and VELO than the remaining commands will also be output
in SORT1 format.
to

SORT2 is the default in transient response analysis (structural and heat

transfer). SORT1 is the default in frequency response analysis. SORT2 is
not available for real eigenvalue (including buckling), complex
eigenvalue, or flutter analysis. If SORT1 is selected in a transient
solution for one or more of the commands ACCE, DISP, ENTH, FORC, HDOT,
MPCF, OLOA, SPCF, STRA, STRE, and VELO then the remaining commands will
 also be output in SORT1 format. If SORT2 is selected in a frequency
response solution for one or more of the commands ACCE, DISP, FORC, MPCF,
OLOA, SPCF, STRA, STRE, and VELO then the remaining commands will also be
output in SORT2 format.
@
34926; 2001; Superelements - SEALL - SEP2DR, Logic error 110
If the SEALL command refers to an undefined superelement, directly or via
the SET command, then the run terminates with "SEP2DR LOGIC ERROR 110".
Avoidance: Correct the undefined superelement references.
@
34940; ; Nonlinear Heat Transfer - RADBC, NLITER - SFM 4276, Traceback
In the radiation dominate problem, the stiffness term is approximately
equal to 4*T**3. Therefore, if the guess at the initial temperature value
is too low for the problem, the temperature may rapidly exceed an infinite
value and the traceback that follows will occur:

Trap encountered:
SIGTRAP - Trace trap FP overflow
Traceback:
Offset 0x00000014 in procedure zerod_
Offset 0x00000564 in procedure nrdbcd_
Offset 0x00001124 in procedure nlemg_
Offset 0x00000780 in procedure nliter_
Offset 0x00000088 in procedure xsem15_
Offset 0x00000990 in procedure mainal
--- End of call chain ---
Avoidance: Make a better estimate of the initial temperature on the TEMPD
bulk data entry. In general it is safer to have a very high initial
temperature than too low.
@
34949; ; Plotting - XYPLOT - Bad Scaling of XYPLOT data
The XYPLOT capability fails to scale the xy-plots properly when small
values of displacement exist in the entire frequency range of interest.
Resolution/Avoidance: Use Microsoft Excel to plot the data or scale th
e
model so the response values are not so small.
@
34956; ; Aeroelastic Analysis - Panels in XY Plane Produce a Singular Mat
rix
The Quick Reference Guide AERO and AEROS descriptions caution against
placing a lifting surface in the plane of symmetry. The problems
associated with this modelling practice in the XZ plane of symmetry have
been removed in Version 70.6. However,
the XY symmetry was not addressed so that vehicles flying in ground effect
(SYMXY=-1) with any panel or body
 in the XY plane still produce a singular matrix. The user is also
cautioned not to attempt to perform a SYMXZ=1 (symmetric) analysis with
ALL panels and bodies on the XZ plane.
@
34967; 70.7; Eigenvalue Analysis - PARAM,RESVEC - PARAM,RESVINER Ignored
If PARAM,RESVINER,YES is specified to calculate residual vectors for the
inertial loads due to the unit acceleration of mass and there are no
LOADSET-LSEQ loadings or U6 set degree of freedom, then PARAM,RESVINER,YES
will be ignored. This may lead to poor answers.
Avoidance: Include the following DMAP alter for V70.5
compile semodes
alter 'call modef',''
CALL MODEFSRS MR,USET,DM,CASEM,DYNAMICS,MAA,MKAA,GPLS,SILS,EED,
EQEXINS,,PG,GM,GOT,KFS,YS,LLL,KLL,
bgpdts,scstm,cstms,medge/
,,,,PHA,LAMA,PHA0,LAMA0/
APP/NORSET/NOASET/FALSE/0/NOASET/METH/METHF/TRUE/
NOQSET/NSUB/0 $
compile modefsrs
alter 'subdmap',''
SUBDMAP MODEFSRS MR,USET,DM,CASES,DYNAMICS,MAA,KAA,GPLS,SILS,EED,
EQEXINS,VAFS,PG,GM,GOT,KFS,YS,LLL,KLL,
bgpdts,scstm,cstms,medge/
PHA,LAMA,PHFA,LAMAF,PHSA,LAMAS,PHSA0,LAMAS0/
APP/NORSET/NOASET/FS/NOFASET/NOSASET/METH/METHF/
NASOUT/NOQSET/NSUB/MODETRAK $
alter 'call.*resvec',''
CALL RESVEC PG,USET,GM,KFS,YS,bgpdts,scstm,cstms,medge,,
PHSA0,LLL,KLL,LAMAS0,MSAA,GOT,KSAA,,VAFS,MIS/
PHSA,LAMAS/
0/NORSET/0/'A '/FALSE $
$
compile resvec list
alter 99
matprn px2,ox2,ox3// $
message //'RESVPGF='/RESVPGF $
alter 'if ( noxset'(3),''
if ( noxset>-1 or append ) then $
alter 'matmod.*ratios.*bigrats'
matprn ratios//$
$
$ alter for CSR 34874
alter 'WERE AVAILABLE AFTER FILTERING'
equivx phx/aphx/-1 $
equivx lama/alama/-1 $
$
alter 'PARTN.*PAB,,VAFS/PX1',''
PARTN PAB,,VAFS/PXaa,,,/1 $
alter 'EQUIVX.*PAB/PX1',''
EQUIVX PAB/PXaa/-1 $
alter 40,55
endif $ replace 'else' with endif
IF ( RESVINER='YES' OR INRLM<0 ) THEN $
VECPLOT ,,BGPDTS,SCSTM,CSTMS,,,MEDGE,/
VGT///4////ALTSHAPE//SEID $
TRNSP VGT/VG $
$
 IF ( NOMSET>-1 OR NOSSET>-1 ) THEN $
UPARTN USET,VG/VF,,,/'G'/'F'/'N'/1 $
ELSE $
EQUIVX VG/VF/-1 $
ENDIF $
if(eigset = 'a')then $
PARAML USET//'USET'//////'F'/S,N,NOFSET
if(nofset<>noaset)then $
upartn uset,vf/va,,,/'f'/'a'/'o'/1 $
else $
equivx vf/va/-1 $
endif $
mpyad mxx,va,/paa $
append pxaa,paa/px1 $
else $
UPARTN USET,MFF/MOF,,,/'F'/'O'/'A'/1 $
MPYAD MOF,VF,/POI $
APPEND PO,POI/PX1 $
endif $
ELSE $
EQUIVX PO/PX1/-1 $
ENDIF $ RESVINER='YES' OR INRLM<0
$
endalter

@
34968; ; Eigenvalue Analysis - PARAM,RESVINER - Missing Residual Vectors
If residual vectors are requested for the residual structure (or in a
model which does not contain superelements), and the model is free-free,
then if PARAM,RESVINER,YES is used, the associated residual vectors will
not be created.
Avoidance: Move everything except one grid point into a superelement and
request the residual vectors for the superelement. Then the vectors will
be included.
@
34969; 70.5; Optimization/Design Sensitivity - EMG - UFM 4659, UIM 4656 from
EHEXD
When superelements are used in a shape optimization task with solids
elements, the job may fail in EMG during the sensitivity analysis phase
with the user fatal messages: UFM 4659 and UIM 4656 from routine EHEXD.
Other similar error message may also be issued depending on the type of
elements being used in the model.
Avoidance: Do not use superelements.
@
35007; 70.6; Optimization/Design Sensitivity - PARAM,DDRMM,0 - UFM 5629, UFM
9058
If PARAM,DDRMM,0 is specified in SOL 200 with ANALYSIS=MFREQ or MTRAN,
then the run will fail with UFM 5269 or UFM 9058, respectively.
Avoidance: Insert the following DMAP alter:
compile fea
alter 'delete \/msf1,bsf1,ksf1,,'(,-1)
if ( drmm ) drmm=false $
 alter 'IF ( DROPT\=2'(2)
delete /msf1,bsf1,ksf1,, $
IF ( DOSPCF
) call CFORCE1 MSF,M2SF,BSF,B2SF,KSF,K4SF,K2SF,,/
msf1,bsf1,ksf1/
APP/APP/false $
delete /MMG1,BMG1,KMG1,, $
IF ( DOMPCF
) CALL CFORCE1 MMG,M2MG,BMG,B2MG,KMG,K4MG,K2MG,,/
MMG1,BMG1,KMG1/
APP/APP/true $
alter 'destrn\=true'(,-1),'destrn\=true'(,2)
alter 'ELSE if ( drmm )'
destrn=true $
drmm=false $
endalter
@
35029; 70.7; Postprocessing - DBC - LAMA Objects have Incorrect Attributes in
the XDB
When subcases in optimization refer to eigenvalue or buckling analysis,
the LAMA objects in the XDB will have incorrect attributes and this may
effect the ability of other products such as MSC/InCheck to process the
results. In addition, if a frequency range is specified rather than an
explicit number of modes, NASTRAN DBC translation may be incorrectly
terminated.
@
35040; ; Static Analysis - SORT2, PUNCH - Punch Output has Zero Results
If SORT2 is requested for static analysis results, zero results are
written to the PUNCH file.
Avoidance: The following DMAP alter uses OFP to write the data in f06
format to an alternate fortran unit thus avoiding the punch unit.
$
$ Output of QUAD4 Element Stress data
$ for post-processing of printed stress in
$ a fortran unit separate from the usual
$ f06 file.
$
COMPILE SEDRCVR,SOUIN=MSCSOU,NOLIST,NOREF
ALTER 'OFP.*OES2S,OSTR2,,,,,','' $
type parm,,i,n,linespp $
linespp=getsys(linespp,9) $
paraml kelm//'trailer'/1/s,n,noelem $ No. of nl elem. in model
noelem=20*noelem $ calc. lines per page to 20 x no. of elem.
putsys(noelem,9) $ set lines per page
putsys(13,2) $ write printed output to unit 13
ofp oes2s,ostr2 // s,n,cardno $
putsys(linespp,9) $ reset lines per page to original value
putsys(6,2) $ restore unit 6 for printed output

Version 68.2 will give correct result if SYSTEM(210)=1 is set on the

NASTRAN entry.
@
35050; 70.6; Miscellaneous - Data Transfer between Window and UNIX
 When an input file for MSC/NASTRAN is generated in a Microsoft Windows
environment and is moved to a window on a unix computer with the drag and
drop feature, it causes input errors. Each line of input has a "|" symbol
at its end, symbolizing an illegal, unprintable character (carriage
return) in the input, leading to many illegal input messages.
Avoidance: An avoidance is to use a file transfer utility such as ftp
instead. It may require some experimentation to find out if the ftp
utility you use avoids inserting these illegal characters. They are
carriage return symbols. MSC has had success using the ftp option in the
EXCEED product.
MSC/NASTRAN itself will delete these spurious carriage returns from its
input files, starting with Version 70.6. If other files are transferred
this way, such as scripts used to run MSC/NASTRAN, the symptoms of this
problem are difficult to detect. When you submit one of these damaged
scripts it will return immediately with no action and no diagnostics,
unless you have included commands in the script that request diagnostic
output. If you list the script on the screen with the "more" command, the
spurious carriage returns do not appear on the screen. They can be seen
with a vi editor, and removed with that editor or by other means.
@
35054; 70.6; Optimization/Design Sensitivity - INREL - UFM 6143
If inertia relief (PARAM,INREL,-1) is requested in SOL 200 and there are
more than 12 static load cases then the run will fail with UFM 6143 in the
MERGE module in subDMAP INERSEN.
Avoidance: Insert the following DMAP alter in V70 or V70.5:
compile inersen
alter 'nld6\=' $
type parm,,i,n,nz $
nz=nld6-5 $
alter 'vn6r'(2),''
matgen ,/vn6r/4/1/nld6r//6/nz/i661/0/1 $
alter 'vn6r'(4),''
matgen ,/vn6r/4/1/nld6r//6/nz/i661/0/1 $
endalter
@
35066; 2001; Optimization/Design Sensitivity - DRESP2 - UFM 315
When a DRESP2 bulk data entry is missing a continuation line an error
message is issued but the message indicates that a subsequent DRESP2 is in
error. The message is similar to the following:
*** USER FATAL MESSAGE 315 (IFPDRV)
FORMAT ERROR ON BULK DATA ENTRY DRESP2 1001
SORTED ENTRY COUNT = 584
USER INFORMATION : THE FORMAT OF DRESP2 ENTRY HAS BEEN MODIFIED IN
MSC/NASTRAN VERSION 68. PLEASE REFER TO THE QUICK
REFERENCE GUIDE FOR MORE INFORMATION.
USER INFORMATION : PLEASE NOTE THAT THE FORMATS OF THE FOLLOWING
DESIGN OPTIMIZATION ENTRIES HAVE BEEN MODIFIED IN
MSC/NASTRAN VERSION 68 :
DCONSTR DOPTPRM DRESP1 DRESP2 DSCREEN DTABLE
DVPREL1 DVPREL2
*** USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA ENTRY DRESP2 1001
SORTED ENTRY COUNT = 584
 ^^^ USER FATAL MESSAGE 9002 (IFPS)
^^^ ERROR(S) ENCOUNTERED IN THE MAIN BULK DATA SECTION
^^^ SEE MESSAGES ABOVE. ERROR ENCOUNTERED IN MODULE IFP,

Avoidance: Look for the DRESP2 with the ID preceding the one listed on
the error message.
@
35067; 70.7; Frequency Response - PBUSH, PELAST - UFM 5423, 6719
In SOL 111, if there are frequency dependent elements, and if the a-set
equals the g-set, the modal reduction of the frequency dependent matrices
is not performed. In a structural analysis, this results in UFM 5423, and
in an acoustic (uncoupled) analysis, the result is SFM 6719.
Avoidance: Add a dummy SPC.
@
35096; 70.7; Iterative Solver - ITER, DEFORM - SFM 6980 (SB85RD)
A job that uses the iterative solver and specifies DEFORM type loads
failed with the message:

*** SYSTEM FATAL MESSAGE 6980 (SB85RD)

LARGEST BLOCK ROW OF THE INITIAL MATRIX ( 85606 WORDS) EXCEEDS THE
SPECIFIED MAXIMUM RECORD LENGTH ( 65536 WORDS).
USER ACTION: PLEASE CONTACT MSC CLIENT SUPPORT.
*** SYSTEM FATAL MESSAGE 3008 (UTMCOR)
MODULE = SOLVIT
INSUFFICIENT CORE AVAILABLE FOR SUBROUTINE
USER ACTION: INCREASE OPEN CORE BY AT LEAST 20070 WORDS
PROGRAMMER INFORMATION: LOCATION KEY = M
FATAL ERROR
Avoidance: Increasing memory may solve the problem but it is likely that
the sparse solver will need to be selected.
@
35117; 70.7; Acoustic Analysis - FORCE(PHASE) - Wrong Answers
When magnitude/phase force output is requested for fluid elements,
real/imaginary outputs are printed in the magnitude/phase format instead.
Avoidance: There is no avoidance.
@
35121; 70.7; Optimization/Design Sensitivity - DRESP1, DRESP2 - UFM 7139 (DOP
R3C)
When an unreferenced DRESP1 entry of RTYPE=FREQ appears in a superelement
optimization job the fatal message below may occur:
*** USER FATAL MESSAGE 7139 (DOPR3C)
ON DRESP2 entry ID = 10, RESP1 ID = 1001 DOES NOT REFERENCE
AN EXISTING DRESP1 ENTRY.
Avoidances:
1) Remove the unreferenced DRESP1 entry using RTYPE=FREQ or
2) Use RTYPE=EIGN instead of RTYPE=FREQ
@
35138; 70.7; Documentation/SSSALTER - prldbckla.v706 - no longer needed
 Buckling with preload has been included in the standard 105 solution
sequence in v70.6. Therefore, the prldbcka alter is no longer needed.
Avoidance: The new capability is invoked through the STATSUB Case Control
command. The documentation for STATSUB in Version 70.6 follows:
Format:
STATSUB[(BUCKLING)]=n
(PRELOAD)
Examples:
STATSUB=23
STAT=4
STATSUB(PRELOAD)=7
Describer Meaning
BUCKLING Subcase id number corresponding to static subcase of buckling or
varying load (default in buckling analysis)
PRELOAD Subcase id number corresponding to static subcase of pre
load or
constant load (default in dynamic analysis)
n Subcase identification number of a prior SUBCASE specified for s
tatic
analysis (Integer>0)
Remarks:
1. Unchanged
2. Unchanged
3. Remove this remark, the limitation on LOADSET has been removed.
4. In dynamic analysis only one STATSUB command may be specified in each
dynamic subcase. In buckling analysis with a preload, both
STATUS(BUCKLING) and STATSUB(PRELOAD) must be specified in each buckling
subcase.
@
35139; 70.7; Inertia Relief - param,inrel,-1 - UFM 9105
If param,inrel,-1 is included in the input file for a solution sequence
that does not use inertia relief (e.g., sol 103), then the job will
terminate with the following error messages:

^^^ USER FATAL MESSAGE 9105 (PHASE0)

^^^ A SUPORTI BULK DATA ENTRY IS NOT PRESENT FOR INERTIA RELIEF
ANALYSIS (PARAM,INREL,-1).
Avoidance: Remove param,inrel,-1 from the input file since it is not used
in a non inertia relief run.
@
35141; 70.7; Frequency Response/Complex Eigenvalues - PARAM,MODACC, CVISCi -
Wrong Answers
If PARAM,MODACC,>-1 is specified in SOLs 110, 111, 146, or 200 with
ANAL=MFREQ and the model contains viscous damping then wrong answers will
result.
 Avoidance: Insert the following DMAP alter:
compile gma
alter 'call xADD5 K2XDD'
equivx b2ydd/b2dd/-1 $
endalter
@
35143; 70.6; Optimization/Design Sensitivity - DSADJ - SFM 1015 (GBKMGR)
The fatal message:
SYSTEM FATAL MESSAGE 1015 (GBKMGR)
may occur in the DSADJ module and has been attributed to an incorrect
assessment of the amount of open core available.
Avoidance: The message has only occurred when running with a small memory
request. A possible avoidance is to increase the memory request.
@
35154; 70.7; Superelements - BEGIN SUPER - SFM 3007
In a superelement analysis using PARTS, if more than 12 points occupy the
SAME location at the boundary in the same superelement, then the job will
fail with the following messages:
*** SYSTEM FATAL MESSAGE 3007 (SEP1X)
ILLEGAL INPUT TO SUBROUTINE SEPBSD
0FATAL ERROR
Avoidance: Do not define more than 12 points occupying the same location
at the boundary for a single part superelement.
@
35158; 70.7; DMAP - LCGEN - SFM 3007, Poor Message
If the first input, usually CASES, to LCGEN is purged then SFM 3007 is
issued. SFM 3001 should be issued instead.
Avoidance: There is no avoidance.
@
35187; 2001; Miscellaneous - UFM 6440 (MODGM2)
If double-field entries are used to enter property ID's longer than 8
digits, then it is possible to obtain USER FATAL MESSAGE 6440 indicating
that a property id is not found even though it exists in the input file.
The cause of this error is the sorting routine which uses only the first 8
digits of the id. The problem will only occur if property ID's which are
both 8-digit and longer than 8-digits exist in the input.
Avoidance: Either keep the property ids less than 9 digits, or use all
property ids greater than 8 digits in length.
@
35188; 70.7; Restarts - SFM 471
The database of large restart runs may be corrupted when data spills from
one file to another. In the restart run the error may lead to the
message:
** ERROR ** UNABLE TO VALIDATE DATABASE FILE FORMAT. DATABASE MAY NOT
 HAVE BEEN BUILT ON A COMPATIBLE PLATFORM.
BIOPNR ERROR ** TIMESTAMP/RECL VALIDATION FAILED. FILX = 30, LOGNAME
= DBALL5 , NSBUF3 = 4096
FILE = /disk/tmp/test.dball
TIMESTAMP OFFSET INVALID.
*** SYSTEM FATAL MESSAGE 471 (DGETTS)
ERROR ATTEMPTING TO OPEN THE FOLLOWING PHYSICAL FILE NAME (PFILEX=
30) :
/disk/tmp/test.dball
USER ACTION: 1) CHECK THE READ/WRITE PRIVILEGES GRANTED THIS FILE.
2) CHECK MAXIMUM NUMBER OF FILES THAT MAY BE OPENED
SIMULTANEOUSLY.
3) CHECK THAT THIS DATABASE FILE IS COMPATIBLE WITH THIS
PLATFORM.
4) ENSURE THAT NO OTHER JOB HAS OPENED THE DATABASE FOR
UPDATING.
0FATAL ERROR
Avoidance: Increase the buffsize. Using fewer but larger database files
may also avoid this error, but if the system is limited to 2gb files, this
avoidance may not be possible..

@
35197; 2001; Documentation/Optimization - QRG, PBUSH, DVPRELi - Unexpected Re
sults
The Quick Reference Guide fails to document that when a property on a
PBUSH entry is designed a negative FID number must be specified on the
DVPREL1 and DVPREL2 entries. If a positive FID number is specified on a
DVPREL1/2 entry, unexpected results may occur.
Avoidance: Use a negative word position number from the element property
table for the FID field on the DVPREL1/2 entry. The following table
provides the word positions for the PBUSH entry in the element property
table.
EPT Section for PTYPE = "PBUSH"
Word Type Item
1 I Property ID
2 through 7 R K1 through K6
8 through 13 R B1 through B6
14 R GE1
15 through 18 R SA,ST,EA,ET
@
35200; 2001; Documentation - QRG - FSLIST Missing from QRG
Version 69 and later Quick Reference Guides do not have a description for
the FSLIST bulk data entry.
Avoidance: See the V68 Quick Reference Guide or the Fluid-Structure
Seminar Notes for the description. An abbreviated description follows:
FSLIST
Defines the fluid points (RINGFL entry) that lie on a free surface
boundary.
Format:
 FSLIST, RHO, IDF1,IDF2,IDF3........
Field Contents
RHO Mass density at the surface. Real>0.0; the default is ta
ken from DRHO
on the AXIF entry).
IDFi Identification number of RINGFL entry. (Integer>0 or
Character="AXIS" in first and/or last field only)
Remarks:
1. This entry is allowed only if an AXIF entry is also present.
2. The order of the points must be sequential with the fluid on the right
with respect to the direction of travel.
3. The word "AXIS" defines an intersection with the polar axis of the
fluid coordinate system.
4. If the fluid density varies along the boundary, there must be one
FSLIST entry for each interval between fluid points.
@
35206; 70.7; External Superelements - PARAM,EXTOUT - Wrong Answers, UFM 5423
If external superelements are used and restarts are performed, the restart
logic will insist on re-processing the external superelements and any
downstream superelements whether or not any changes are made to the
model. If the superelements are created using PARAM,EXTOUT,DMIGDB, this
may result in incorrect answers or UFM 5423 indicating an incompatible
matrix multiply.
Avoidance: When restarting from a run which read in external
superelements which were created using PARAM,EXTOUT,DMIGDB insert the
following DMAP alter:
compile extin
alter 'IF.*NOKAA.*0.*THEN',''
IF ( NOKAA>0 and nomatk<0) THEN $

@
35217; 70.7; External Superelements - PARAM,EXTOUT,DMIGDB - Inefficient Resta
rts
If external superelements are used and restarts are performed, the restart
logic will insist on re-processing the external superelements and any
downstream superelements whether or not any changes are made to the model.

Avoidance: The following alter will disable the deletions for external
superelements. If any changes are made to the external superelements, this
alter should not be used.
compile phase0
alter 'INVOKE ALL RESTART DELETIONS HERE',''
if(setype<>'EXTRNA')RESTART ,,,HISTS/TLIST/TRUE $
@
35234; 70.7; Miscellaneous - DPD - UFM 2079, Poor Message
The text of UFM 2079 is incorrect. The text reads:
*** USER FATAL MESSAGE 2079 (DPD) GRID OR SCALAR POINT <nn> HAS AN
UNDEFINED COORDINATE REFERENCED ON A <xxxx> ENTRY.
Both of the the values "nn" and "xxxx" may be wrong and words "UNDEFINED
COORDINATE" should be "ILLEGAL COMPONENT".
@
35246; 70.7; Matrix Methods - MPYAD - SFM 4276 EC 3002 or 32 on Large Matrice
s
It is possible that MPYAD may fail with SFM 4276 EC 3002 or 32 when very
large, dense matrices are being multiplied. The client model that
encountered this error had a very long duration transient analysis (more
than 100000 time steps). This model failed during mode-acceleration data
recovery in module DDR2 (using MPYAD).

Avoidance: It is possible that providing more memory may allow the run to
finish. For the problem mentioned above, using residual vectors
(PARAM,RESVEC,YES) instead of mode-acceleration allows the run to
complete. If you are encountering this problem in a run using
mode-acceleration, try using residual vectors instead of
mode-acceleration.
@
35259; 70.7; Nonlinear Transient Response - LSEQ - Wrong answers
If LSEQ entries are used in SOL 129 and the DAREA and LID fields have the
same ID, then the dynamic loads which are based on the LSEQ entries will
be doubled and wrong answers will result.
Avoidance: Insert the following DMAP alter:
compile nltran
alter 'trlg ',''
TRLG CASESX2H,USETD,DLT,,BGPDTS,SILS,CSTMS,TRL,DIT,
GMD,GOD,,EST,MPTS,/,,,PDT,,TELH,DLT1/
/S,N,DCOL/IMETHOD/RSTIME/BETA/S,N,FAC1/S,N,FAC2/
S,N,FAC3//TABS $
endalter
@
35261; 70.7; Heat Transfer - CHBDYG, NLCOMB - UFM 4017
If the area of a CHBDY element is less than 1.E-10, than the job may fail
with the message:
0*** USER FATAL MESSAGE 4017, ESTNL AND FILE (103) ARE NOT
PROPERLY ALIGNED
*** SYSTEM FATAL MESSAGE 3007 (NLCOMB)
ILLEGAL INPUT TO SUBROUTINE NLCKDT
0FATAL ERROR
Avoidance: This error may be caused by a poor selection of units. Change
the units to increase the size of the element.
@
35277; 70.7; Static Analysis - LOAD, TEMP, GRAV - Wrong Answers
If a LOAD bulk data entry references a TEMP bulk data entry which has
already been referenced by a previous subcase (TEMP(LOAD)), then the
temperature loads are ignored and wrong answers are produced for all
subcases which contain GRAV or TEMP loads. In addition, then the warning
message below, will not be issued:
USER WARNING MESSAGE 3204 (SSG1)
UNABLE TO FIND SELECTED SET ( 2) IN TABLE (SLT ) IN SUBROUTINE
(SSG1)

Avoidance: Do not reference TEMP bulk data entries from the LOAD bulk
data entry. In order to combine temperature and mechanical loads, use a
LOAD case control entry to specify the static loads, in combination with
the TEMP(LOAD) case control entry to specify the TEMP loads.
@
35280; 70.7; Coupled Fluid/Structure Analysis - PARAM,PANELMP - Wrong answers
In a coupled modal frequency response analysis (SOL 111), if there is no
ASETi or OMITi entries, rigid elements, multipoint constraints nor single
point constraints, all panel participation factors are zero.
Avoidance: Add 1 dummy grid with all degrees of freedom constrained.
@
35289; 70.7; Miscellaneous - DBLOCATE - UFM 453
DBLOCATE fails if the MASTER dbset is assigned twice under different
logical names and the first assignment is not used on a DBLOCATE
statement. The following is an example of the input and fatal message:
ACQUIRE NDDL
ASSIGN ADB_1='test.MASTER.2'
DBLOCATE LOGICAL=ADB_2 WHERE(DBSET='ADB' OR DBSET='AEDB') $
ASSIGN ADB_2='test.MASTER.2'
DBDICT
ENDJOB

*** USER FATAL MESSAGE 453 (LCXLOD)

THE DATABASE NUMBER (IDBSNO)= 3 BEING USED IN THIS RUN DOES NOT
HAVE ENOUGH INFORMATION TO START FROM.
@
35291; 70.7; NAST/API Toolkit - NDDL Qualifiers not reset to their defaults
NDDL qualifiers are not reset to their default values when submitting
multiple MSC/NASTRAN jobs via the API Toolkit. Though this was the
original design intent, this has been shown to cause errors if the second
and subsequent jobs depend on qualifier default values as is the case with
restarts. The application programmer should reset these values manually
via dmap in such cases.
@
35295; ; Documentation/Linear Static User's Guide - Inertia Relief
In SOL 101, inertia relief is calculated differently than in SOL 1. The
description of inertia relief in the Linear Statics User's Guide describes
the method used in SOL 1 to calculate the accelerations. In SOL 101, the
approach is different. In SOL 101, 6 additional loads are created. These
loads are created by applying unit accelerations in each of the 6 dof
about PARAM,GRDPNT. The matrix URA is the accelerations required on these
6 new loads to provide static equilibrium when combined with the original
loadings. In SOL 1, the accelerations about the SUPORT dof are calculated.
@
35298; 70.7.2; DMAP - SDR1, UMERGE1, UMERGE - Poor message
When two matrices are being merged row-wise in modules SDR1, UMERGE1,
SSG4, DDR2, and READ and the number of columns in each matrix do not
match, then the following message is issued. This error also occurs in
UMERGE and UMERGE1 if the the number of zeros and non-zeros in the
partitioning vector do not match the corresponding number of rows and
columns in the sub-matrices.
0*** SYSTEM FATAL ERROR 4276.GETSTR ERROR CODE 31 PID= 0
along with the following traceback
Traceback:
Offset 0x00000020 in procedure raise
Offset 0x00000020 in procedure tracbak_
Offset 0x00000184 in procedure nsexit_
Offset 0x0000016c in procedure pexit_
Offset 0x00004000 in procedure gperr_
Offset 0x00000480 in procedure getstr_
Offset 0x000001ac in procedure mrgcol_
Offset 0x00000ec4 in procedure merge_
Offset 0x00000218 in procedure sdr1b_
System Fatal Message 6143 should have been issued instead.
Avoidance: There is no avoidance.
@
35314; ; Buckling Analysis - Insufficient memory - SFM 3001,1015, Out of
range, Error 924
In a buckling analysis and possibly other eigenvalue solutions, if
insufficient memory is allocated, then depending on the machine and number
of cpus requested, the job may fail with any one of the following unclear
messages:
1. *** SYSTEM FATAL MESSAGE 3001 ()
THE INPUT DATA BLOCK IN POSITION xx DOES NOT EXIST
2. *** SYSTEM FATAL MESSAGE 1015 (GBKMGR)
PHYSICAL FILE NOT FOUND IN FILE ASSIGNMENT TABLE
BIOMSG: ERROR 4276 HAS OCCURRED IN ROUTINE GBKMGR , FILE INDEX
(FILEX) = 0
STATUS = 0
3. BIORD ERROR - FILX = 0(00000000 HEX) OUT OF RANGE
BIOMSG: ERROR 924 HAS OCCURRED IN ROUTINE IONAST , FILE INDEX
(FILEX) = 0
 BIOWRT ERROR - FILX = 0(00000000 HEX) OUT OF RANGE
0*** SYSTEM FATAL MESSAGE 1022 (XIO)
ERROR OCCURED IN BIOWRT
0********* DUMP OF XIO PARAMETERS
0XIO NAME = 407
ATTEMPTED OPERATION = CLOSE
FILEXX = 7
UNIRFX = 25908493
B = 018FDF6A
CLR = 1060
CBP = 1060
FLAGS = 00000400
Avoidance: Increase memory.
@
35315; ; Parallel Processing - Process remains active after job is killed
If a model using parallel on HPUX 10.20 is 'killed' by the system (for
example, due to a full disk or exceeding file size limits), it is possible
that a process may still remain active, possibly with a high CPU usage.
Avoidance:

If this occurs, use 'kill -9 pid' (where pid is the process id obtained
using ps -ef) to 'kill' the remaining process.
@
35324; ; Submittal - No Output
If an MSC/NASTRAN output file is opened by some other program, that other
program may prevent MSC/NASTRAN from using the file. If it does, an error
message such as the following may be displayed:
*** SYSTEM FATAL MESSAGE (pgm: nastran, fn: run_job)
Unable to rename:
./d10101d.out
to
./d10101d.out.1
Avoidance: Close any MSC/NASTRAN output files in use by other programs.
@
35333; ; Nonlinear Analysis - CBUSH, PBUSHT - Nonlinear Affects Ignored
If a nonlinear model contains a CBUSH element, no nonlinear materials, and
PARAM,LGDISP is not set, NASTRAN will solve the problem as a linear
analysis.
Avoidance: Add PARAM,LGDISP to the deck.
@
35370; 70.7.2; Eigenvalue Analysis - READ, EIGRL - Distributed Memory Parallel
Hangs
MSC/NASTRAN Version 70.5.4 for IBM/SP computers introduces Distributed
Memory Parallel Lanczos as a beta capability. There are known problems
with the capability that may cause the system to hang. The problems
include
1. If the master node does not find any roots the job will fail, and if
bat=no is set the job will hang. This problem may be avoided by selecting
a different number of processors. This problem may be also be caused by
 error 2 below.
2. If dmp=n is specified on the submittal line but nastran numseg=x is
not specified in the data deck or rc file, numseg will automatically be
set to n. This action changes the shift logic which often causes the
master node to find any roots. Error number 1 above is then encountered.
An avoidance is to explicitly specify numseg.
3. Distributed Parallel Lanczos hangs the system if multiple subcases
with different boundary conditions exist. The only avoidance is to submit
the job as two separate runs.
4. Modal Frequency response jobs that use Lanczos are likely to hang in
the frrd1 module. There is no avoidance.
5. Restarts are not supported, this is a limitation that may not be
addressed for several releases.

@
35410; ; Nonlinear Transient Response - SPC, LSEQ - SFM 6920 in MATMOD
In SOL 129, if the number of unique, with respect to DAREA ID, LSEQ Bulk
Data entries is less than the number of subcases and static enforced
displacements are specified, then SFM 6920 will be issued from the MATMOD
module.
Insert the following DMAP alter:
compile nltran
alter 'MATMOD YST',''
paraml yst//'trai'/1/s,n,yscol $
if ( yscol>1 ) then $
MATMOD YST,,,,,/YS,/1/1 $
else $
equivx yst/ys/-1 $
endif $ yscol>1
alter 'MATMOD YST'(2),''
paraml yst//'trai'/1/s,n,yscol $
if ( yscol>1 ) then $
MATMOD YST,,,,,/YSt1,/1/1 $
else $
equivx yst/yst1/-1 $
endif $ yscol>1
@
35411; 70.6; Submittal - news=yes - UFM 756 (XCSA)
If an include statement is the first line of an input deck or follows
immediately after a NASTRAN statement and news=yes is requested, then the
job may fail with the message:
USER FATAL MESSAGE 756 (XCSA)
THE BEGIN CARD IS NO LONGER SUPPORTED IN THE EXECUTIVE CONTROL
SECTION.
USER INFORMATION: THE BEGIN STATEMENT IN THE EXECUTIVE CONTROL
SECTION HAS BEEN REPLACED BY THE SUBDMAP DMAP
STATEMENT. SEE USER MANUAL SECTION 5.
ENDDATA
*** USER FATAL MESSAGE 765 (XREAD)
 END-OF-FILE ENCOUNTERED ON UNIT 1. THE INPUT FILE IS INCOMPLETE.
USER INFORMATION: THE FOLLOWING GUIDE LINES SHOULD BE FOLLOWED FOR
PREPARATION OF AN INPUT DECK.
FILE MANAGEMENT SECTION (FMS) CAN BE ENDED BY EITHER AN "ENDJOB"
OR A NON-FMS CARD.
EXECUTIVE CONTROL SECTION CAN ONLY BE TERMINATED BY A "CEND"
CARD.
Avoidance: Specify the Executive control statements in the input file and
include the rest of the data.
@
35415; 70.7; Optimization/Design Sensitivity - REPCASE, SET - SFM 3003 (DOPR3
)
The fatal message shown below may be issued in an optimization job if
there are REPCASE subcases but no SETt case control commands (all the data
recovery requests = ALL)
*** SYSTEM FATAL MESSAGE 3003 (DOPR3) ATTEMPT TO READ PAST RECORD IN DATA
BLOCK
Avoidance: Place a set command in the case control.
@
35460; 2001; Nonlinear Transient Analysis - NLTRD, PARAM,SNORM - SFM 4276
A nonlinear transient run may give a traceback similar to the following
when param,snorm is used.
+80 Offset 0x000001c4 in procedure getnml_
+81 Offset 0x0000003c in procedure gshnod_
+82 Offset 0x00001688 in procedure nqd4d_
+83 Offset 0x00000258 in procedure nembrd_
+84 Offset 0x00001124 in procedure nlemg_
+85 Offset 0x000004c0 in procedure nltrd_
+86 Offset 0x000001a4 in procedure xsem15_
Avoidance: Do not specify param,snorm.
@
35483; 70.7; Performance - SGI Only, Implausible Warning Message with ff_io=n
o
MSC/NASTRAN V70.5(r4) mistakenly omitted support for EAG FFIO in the
template file. On some jobs this may lead to a slight degradation in I/O
performance and hence longer run times.

Avoidance: This can be corrected by updating:

<install-dir>/bin/nast705.lcl
After line 143, which currently reads:
%MSC_ARCH=irix64:%%threads:% print "MP_SUGNUMTHD_VERBOSE:
%thread_verbose%"
the following lines can be added:
%MSC_ARCH=irix64:%%ff_io=^[ya]:%%ff_io_opts:% export
FF_IO_OPTS='%ff_io_opts%'
%MSC_ARCH=irix64:%%ff_io=^[ya]:%%ff_io_defaults:% export
FF_IO_DEFAULTS='%ff_io_defaults%'
 %MSC_ARCH=irix64:%%ff_io=^[ya]:% print "EAG FFIO enabled."
%MSC_ARCH=irix64:%%ff_io=^[ya]:% print "FF_IO_OPTS: $FF_IO_OPTS"
%MSC_ARCH=irix64:%%ff_io=^[ya]:% print "FF_IO_DEFAULTS:
$FF_IO_DEFAULTS"
%MSC_ARCH=irix64:%%ff_io=^n:% print "EAG FFIO disabled."
%MSC_ARCH=irix64:%%ff_io=^n:% unset FF_IO_OPTS
%MSC_ARCH=irix64:%%ff_io=^n:% unset FF_IO_DEFAULTS
@
35508; ; Eigevalue Analysis - READ - UIM 4158, Massless Mechanisms
Eigenvalue analysis of models with massless mechanisms may fail in the
READ module with the following messages:
*** USER INFORMATION MESSAGE 4158 (DFMSA)
---- STATISTICS FOR SPARSE DECOMPOSITION OF DATA BLOCK KFF FOLLOW
NUMBER OF NEGATIVE TERMS ON FACTOR DIAGONAL = 5
MAXIMUM RATIO OF MATRIX DIAGONAL TO FACTOR DIAGONAL = 1.0E+15 AT
ROW NUMBER 4*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
process error reported by subroutine LNNP1ED
see LNNP1ED abstract (ier = -729)
*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
process error reported by subroutine LNNDRVD
see LNNDRVD abstract (ier = -729)
Avoidance: Add a small amount of mass to the mechanism or remove the
mechanism.
@
35547; 70.5.5; ESTIMATE - Will not work if there is a blank space in path
If a directory in the path to the file contains a blank space (for example
('Program Files'), MSC/ESTIMATE will not be able to find the input file
and will fail.
Avoidance: Do not include blank spaces in directory names.
@
35551; 70.7; Superelements - THRU, SPOINT, EXTRN - UFM 6739 (SEP2MP)
If the THRU range is used on the EXTRN entry for a set of SPOINTs, the
program will incorrectly issue UFM 6739 stating that the SPOINTs are
specified with a nonzero degree of freedom
Avoidance: Do not use the THRU option on the EXTRN for SPOINTs.
@
35562; 70.7; Aeroelastic Analysis - TRIM - Wrong Answers
Version 70.5 introduced the capability in SOL 144 to trim an aircraft
that had more free variables than there are free degrees of freedom. The
results from this Generic Control System capability will be in error
unless the aerodynamic extra points (AESTAT and AESURF) specified on the
TRIM bulk data entry are in ascending ID order.
Avoidance: When there are redundant control surfaces, list the known trim
variables on the TRIM entry in ascending ID order. The ID's are those
entered on the AESTAT and AESURF bulk data entries.
@
35603; 70.7; Parallel Processing - Distributed Memory Parallel - Confusing Me
ssages
 At submittal, confusing error messages may be issued to the terminal when
running MSC/NASTRAN V70.5.3 or V70.5.4 on AIX in distributed mode. The
various messages may lead you to believe the job has failed even though it
hasn't. The error messages will be displayed if the IBM Parallel
Environment for AIX V2.4 or later is installed.
Avoidance: The messages may be ignored or they may be avoided by changing
the following line in the <install-dir>/bin/nast705.dmp file from:
%MSC_ARCH=aix:% print -u3 $host %cpu_use% %adapter_use%
to
%MSC_ARCH=aix:% print -u3 $host
@
35615; ; Aeroelastic Analysis - scr=mini, multiple subcases, SOL 145 - Mi
ssing output
If multiple flutter subcases are specified in a SOL 145 job and the job is
run with scr=mini, then the job will only process the first subcase and
ignore subsequent subcases.
Avoidance: Insert the following DMAP alter:

compile super3
alter 'saverstl\=',''
compile sefluttr
alter 'call super3'(,1)
if ( saverst='sedr' ) saverstl=true $
compile sedisp
alter 1
file ol1s=ovrwrt $
endalter
@
35619; 70.7; Complex Eigenvalue Analysis - Number of Desired Roots not Found
Prior to MSC/NASTRAN 70.7 the Complex Lanczos eigenvalue method would
require the user to specify many shifts to find many modes in a lightly
damped system. Beginning in Version 70.7, an the system will
automatically choose new shifts until the desired number of modes is
found.
@
35622; 70.5.2; DMAP - ISHELL may have Difficulties Locating the Called Program
In versions prior to V70.5.2, the ISHELL module may have difficulty
locating the called program file. This may result in the fatal message:
^^^ PRECLOSING FORTIO UNIT 25
^^^ ISHELL PROCESS RETURN CODE: 1
^^^ ISHELL FAILED WITH RETURN CODE 1
Starting with V70.5.2, the search logic has been improved to search for
different file extensions and thru the path correctly.
Avoidance: If you are using ISHELL in a version prior to V70.5.2, use a
program name containing the exact name specified in the ISHELL statement
(upper-case only, maximum of 8 characters including the extension) and
make sure that the program is in the current directory.
@
35630; ; Postprocessing/Eigenvalue Analysis - PARAM,POST- Wrong Eigenvect
or Labels in XDB
If multiple eigensolutions are requested in SOL 103 with PARAM,POST,0,
then the eigenvectors will be labeled incorrectly in the xdb file.
Avoidance: Use PARAM,POST,-1 or insert the following DMAP alter:
compile super3
alter 'EQUIVX CASECC',''
EQUIVX CASErs/CASESEP4/-1 $
endalter
@
35633; ; Acoustic Analysis - Missing Output
In solution 108, if displacements are requested for fluid grids, and
accelerations are requested for structural grids, then no fluid grid
pressures will be output..
Avoidance: Solution l111 works correctly.
@
35646; 70.7; Acoustic Analysis/Superelements - CAABSF - UFM 9101
If a fluid superelement is used, and frequency dependent acoustic
absorbers (CAABSF) are correctly put in the residual with their fluid
grids (in order to keep them frequency dependent), then an error message
similar to the following is erroneously invoked causing the run to fail
prematurely:
^^^ BULK DATA AND CASE CONTROL PARTITIONING, RESTART CHECKING, AND
UNDEFORMED PLOTTING IS
^^^ INITIATED FOR SUPERELEMENT 0
^^^ USER FATAL MESSAGE 9101 (PHASE0)
^^^ FLUID ELEMENTS ARE DETECTED IN THIS SUPERELEMENT.
^^^ USER ACTION: ENSURE THAT ALL FLUID ELEMENTS ARE ASSIGNED TO
SUPERELEMENT 2000
Avoidance: Insert the following DMAP alter:
compile phase0
alter '9101'(,-1),''
paraml ects//'trailer'/5/s,n,trwd5 $ bits 161-176 and 65-80
type parm,,logi,n,fluidel $
fluidel = MOD(trwd5/2**6,2)>0 or
MOD(trwd5/2**5,2)>0 or
MOD(trwd5/2**4,2)>0 $
if ( fluidse>-1 and seid<>fluidse and
((acous>0 and seid>0) or (fluidel and seid=0)) ) then $
endalter $
@
35647; 70.7; Elements - CBEAM, GP2 - UFM 2007, Undefined Grid Point
If a model contains both CBAR and CBEAM elements and the value of BIT is
set to 0.0 on a CBEAM entry and p-elements are not being used, the run
will fail with UFM 2007 stating that the BEAM element references an
undefined GRID point.
 *** USER FATAL MESSAGE 2007 (GP2)
ELEMENT xx REFERENCES UNDEFINED GRID POINT yyy
Avoidance: Do not place a value in the BIT field of the BEAM element
unless you are using p-elements.
@
35657; 2001; Documentation/QRG - THERMAL - Clarification of Remark 1
The following remark should be added to THERMAL case control command
description in the Quick Reference Guide.
3. In a transient heat transfer analysis the SID on the punched TEMP bulk
data entries equal the time step number.
@
35695; ; Restarts - Nastran fails to trap invalid dbset that have been co
pied.
Nastran fails to trap invalid copies of DBSETs used in RESTART (or reuse)
runs. The following sequence demonstrates the problem:
> nastran s31498a.dat # cold start run creates
s31498a.MASTER and s31498a.DBALL
> cp s31498a.MASTER s31498b.MASTER # copy s*a.MASTER to s*b.MASTER;
DBALL is not copied
> nastran s31498b.dat # run using s*b.MASTER; nastran
finds s*a.DBALL using autoassign
> nastran s31498a.dat # run using s*a.MASTER; nastran
fails to recognize s*a.DBALL was modified above and is no
# longer in synch with
s*a.MASTER
@
35726; 2001; Elements - CBUSH - UWM 979 (XYTRAN), THE ID COULD NOT BE FOUND
Trying to use XYPLOT for CBUSH postprocessing results in the following
error message:
*** USER WARNING MESSAGE 979 (XYTRAN)
AN XY-OUTPUT REQUEST FOR POINT OR ELEMENT ID n
-FORC- CURVE IS BEING PASSED OVER. THE ID COULD NOT BE FOUND IN
DATA BLOCK 106
@
35743; 70.7; Aeroelastic Analysis - IFP - UFM, TRIM Entries with no Fxed Para
meters
If you define a TRIM subcase in MSC/NASTRAN in which all trim parameters
(aerodynamic extra points) are FREE, the resultant MSC/NASTRAN execution
will fail due to an illegal TRIM Bulk Data Entry.MSC/NASTRAN considers it
to be illegal to have a TRIM entry with zero PARAM/VALUE pairs. That TRIM
entry would imply that all parameters were FREE, something that is not
necessarily illegal, even for regular TRIM.
Avoidance: Specify at least one parameter that is fixed at some value.
@
35745; 2001; Elements - PBARL/PBEAML, SOLs 1-16 - UFM 1126 (GNFIST) or 2011 (
TA1)
PBARL and/or PBEAML entries do not work in SOL 1. They cause the
following error message:
 in V69.0 - V70.0;
*** USER FATAL MESSAGE 1126 (GNFIST)
DMAP MODULE MODEPT IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME
= IEPT , WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE IEPT =OVRWRT
FATAL ERROR
or UFM 2011 (TA1) in V70.5;
*** USER FATAL MESSAGE 2011 (TA1)
NO PROPERTY ENTRY FOR ELEMENT TYPE BAR
*** USER FATAL MESSAGE 2011 (TA1)
NO PROPERTY ENTRY FOR ELEMENT TYPE BEAM
0FATAL ERROR
Avoidance: Insert the following DMAP alter in SOLs 1-16:
MALTER 'IFP9 EPT',''
IFP9 EPT.7/IEPTX/S,N,NOGOIFP9 $
endalter
@
35754; 70.7; Documentation - Missing Beam Server Documentation in NT C.O.G.
The documentation for the beam server is missing from the "Configuration
and Operations Guide, Windows NT Edition".
Avoidance: The UNIX documentation can be used in its place.
@
35760; GL; Transient Analysis - MPCFORCES - Inconsistent Answers
In a transient analysis, the MPCFORCES may appear incorrect because they
may vary when the independent and dependent degrees of freedom are
reversed. This is due to the inconsistency between the way the
"equilibrium" is calculated using the modified Newmark integration scheme
(which is evaluated at 3 time steps-- Un+1, Un, and Un-1) and the MPCFORCE
data recovery (which is evaluated at Un).
Avoidance: Use a smaller time step if mpcforces are desired. The error
is negligible if (delta t)**2 * acc << disp
@
35762; 2001; DMAP - MATMOD, P1=4 - Possible Wrong Answers
Matmod option 4 (p1=4) does not produce the correct output matrix when
using form 2 (LUSETS).
Avoidance: Use form 1 (SILS) instead of form 2

@
35765; ; Nonlinear Analysis - ITER - UFM 3032 (NLITER), Poor Message
If the ITER entry uses the free field keyword approach but it is entered
incorrectly a misleading fatal message may result. For example the entry
"ITER,100,IPAD=3" lead to the following fatal message:
*** USER FATAL MESSAGE 3032 (NLITER)
UNABLE TO FIND SELECTED SET ( 100) IN TABLE (
0FATAL ERROR
 This entry should have been:
ITER,100
,IPAD=3
@
35778; ; Superelements - SELOC, SEP1X - Poor Message
If the GRID points or POINTs listed on a SELOC entry in the PBi fields do
not have the same relative geometry as the GRID points listed in the PAi
fields, then the following misleading error message may occur:
*** USER FATAL MESSAGE 6790 (SEPO21)
THE FOLLOWING SUPERELEMENTS ARE DETACHED:
Avoidance: Be sure that the relative geometry is correct.
@
35781; 70.7; Nonlinear Analysis - PSHELL, MID2=-1 - Nonlinear Stresses, SWM 6
065 (DBC)
If plane stress elements are mixed with plane strain elements (i.e.,
PSHELL with MID2=-1) then garbage may appear in the nonlinear stress
output and if PARAM,POST,0 is selected then the DBC module will issue the
following warning message:
SYSTEM WARNING MESSAGE 6065 (DBC)
DBCMSG ERROR MESSAGE 22 FROM SUBROUTINE ELMBLK
Avoidance: There is no avoidance.
@
35790; ; RFORCE - SOL 103, Followk - Wrong Answers
RFORCE follower force effects (param,followk) were included in Version
70.6. However, it was not documented that only RFORCE method 2 is
supported. If method 1 is chosen follower force effects are ignored.
Avoidance: Use method 2 for the RFORCE.
@
35819; 70.7; Superelement - RROD, RBAR- UFM 6591 (SEP2RP)
If RROD or RBAR elements are used to connect interior and exterior degrees
of freedom on a PART superelement, the run may incorrectly fail with the
message:

*** USER FATAL MESSAGE 6591 (SEP2RP)

RIGID ELEMENT DEFINES DEPENDENT EXTERIOR POINTS
Avoidance: An avoidance which sometimes works is to interchange the ga
and gb points, and their related data. This behavior has not been found
using other rigid elements, such as the RBE2.

@
35821; 2001; Documentation/QRG - PCOMP - ELFORCE Not Required for Ply Stresse
s
In the MSC/NASTRAN Quick Reference Guide, Remark 7 under the PCOMP entry
which states that an ELFORCE and ELSTRESS request is necessary for
composite ply stresses is no longer true. Only a STRESS request is
required.
@
35845; ; Iterative Solver - PARAM CDITER, NASTRAN ITER=YES
The iterative solver produces incorrect results for the SPCFORCE RESULTANT
when the linear gap (PARAM,CDITER) is used.
Avoidance: Use the direct solver or ignore the spcforce resultant output.
@
35851; 70.6; DMAP - ISHELL - Access Violation on NT
A call to the ISHELL module in Version 70.5 on Windows NT may fail with
an Access Violation.

Avoidance: Two possible avoidance exist. If the problem occurs, use

Version 70.6 or higher. Else, use character parameters whose length is 8
characters with no blank spaces. The workaround for Version 70.5 may not
always work. Whether or not the access violation occurs depends on the
contents of the various data fields and is unpredictable.
@
35857; ; Nonlinear Transient Analysis - PCOMP, STRESS - SFM 1013, Poor Me
ssage
Composite materials run in sol 129 that request stress output may fail
with the message:
*** SYSTEM FATAL MESSAGE 1013 (GBKMGR)
CANNOT FIND CORRECT CLUSTER POINTER RECORD
BIOMSG: ERROR 4276 HAS OCCURRED IN ROUTINE GBKMGR , FILE INDEX (FILEX)
= 0

Avoidance: Do not use the STRESS= case control command since nonlinear
transient analysis does not support composite data recovery. Else
substitute equivalent PSHELL/MAT2 entries for the PCOMP entries.
@
35876; 70.7; SSSALTERs - checka - Effective Weight Matrices Incorrect
Output for the checka.vxx alter has the row and column output for the
modal effective mass and FRACTOT output incorrectly labeled. It should
state that each column represents a mode and each row represents an
associated degree of freedom.
Avoidance: Edit the alter to correct the printout form or request an
updated copy from MSC.
@
35877; 70.7; Rigid Elements - RSPLINE - UFM 5289, Incorrect DOF Identified
If an RSPLINE specifies dependent degrees of freedom that have already
been specified as dependent by another RSPLINE entry, then USER FATAL
MESSAGE 5289 is correctly issued. However, the UFM incorrectly calls out
the independent degrees of freedom on the entry as dependent and lists the
independent degrees of freedom twice. The dependent degrees in error are
also listed but they may be obscured by the incorrect data.
Avoidance: Inspect the RSPLINE and other MPC entries to determine which
are in error.
@
35882; 70.7; SSSALTERs - modevala.vxx - Matrix STRAINED is off by a Factor o
f 2
If PARAM,EVAL,-1 is used in the modevala.vxx alter, the matrix for the
modal strain energy (STRAINED) is off by a factor of 2. If the
recommended value of PARAM,EVAL,-2, is used this problem does not occur.
Avoidance: Modify the alter to divide the MATRIX by 2 as follows:
Replace:
DIAGONAL STRAINE/STRAINED/ $
with
DIAGONAL STRAINE/STRAIND2/ $
add straind2,/strained/(.5,0.) $

@
35886; 70.7; Loads - LOAD(TEMP), LOAD - No UFM, Unexpected Answers
As documented in the Quick Reference Guide, the TEMPERATURE load ID should
be unique with respect to all other loads in the model. However, this rule
is not enforced, and misleading results can occur under the following
situation.
1) The TEMPERATURE load has same set ID as a mechanical (e.g. FORCE,
PLOADi) or body (e.g. GRAV) load.
2) The mechanical or body load is scaled by a non-unity value on a LOAD
bulk data entry.
Although the Quick Reference Guide LOAD card documentation implies that
the TEMPERATURE loads are ignored by the LOAD card, the temperature load
delta-T will be scaled by the load factor on the LOAD card if there is a
mechanical/body load with the same set ID as the TEMPERATURE set.
Avoidance: Ensure that the TEMPERATURE set is unique with respect to all
other loads in the model. MSC/NASTRAN does not issue any warning or
information messages when the TEMPERATURE set is not unique.
@
35892; 70.7; ACCESS - Sample Program Build Errors
The MSC/ACCESS sample programs are built with the command
msc705 access build
Failing to use the msc705 (msc706, msc71, et al) and access scripts will
result in the makefile failing. Alternatively, you could set MSC_BASE and
MSC_ARCH to the correct values and directly invoke
access build
However, under no circumstances can you use
make
or
make build
@
35893; 70.7; Plotting - OUTPUT(POST) - SFM 4276. BIOMSG ERROR 3012
If OUTPUT(POST) and SETS DEFINITION are incorrectly both included in the
case control the following error message is output:
19 OUTPUT(POST)
20 SETS DEFINITIONS
BIOMSG: ERROR 3012 HAS OCCURRED IN ROUTINE QOPEN , FILE INDEX (FILEX)
= 0
The message should say the same thing as Remark 1 of the SETS DEFINITION
description of the MSC/NASTRAN Quick Reference Guide:
"Either SETS DEFINITION or OUTPUT(POST) may be specified but not both."
@
35895; 70.7; Buckling - EMG, TEMP(LOAD) - SFM 3003 EMG ETT
If a buckling job, SOL 105, contains multiple static/buckling subcase
pairs, and if a temperature set (e.g. TEMP(LOAD)) is called out in the
second or later static subcase that does not exist, then the following
error occurs:
SYSTEM FATAL MESSAGE 3003 (EMG) ATTEMPT TO READ PAST THE END OF A LOGICAL
RECORD IN DATA BLOCK ETT (FILE 107) IN SUBROUTINE EMGTAB
Avoidance: Add the referenced temperature set entries to the bulk data
file.
@
35896; ; Optimization/Design Sensitivity - PCOMP, Z0 - Unexpected Answers
When optimizing ply thicknesses in Sol 200, the user must remember to
recalculate Z0 if the midplane is to remain at the nodes. It is common
for users to leave field 3 of the PCOMP entry (Z0) blank, and let
MSC/NASTRAN calculate Z0 as -t/2. But in optimization, this field is not
updated automatically, which can lead to unexpected results.
Avoidance: Either:Create another design variable for Z0 (field 3 of
PCOMP) and use DLINK to calculate updated Z0 or use an additional DVPREL1
to calculate z0 = 1/2 ( c0 + x1 + x2 + x3 + ... xn) where co is the
thickness of invariant plies and the xi's are designed ply thicknesses)
@
35901; 70.7; Optimization/Design Sensitivity - Rigid Element Sensitivity - UF
M 1126 in UMERGE
When a shape optimization job includes grid points which are associated
with rigid elements and are designed (referenced by DVGRID/DVSHAP/DVBSHAP)
and the job is run with the option of SCR=no, it may fail in subdmap
respsen with the following error message:
*** USER FATAL MESSAGE 1126 (GNFIST)
DMAP MODULE UMERGE IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME
= DUG , WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE DUG=OVRWRT
Avoidance:
 1) Use SCR=YES
2) Use the following dmap alter if SCR=NO,
compile respsen
alter '\/ dug ','\/ dug '(,1)
umerge uset, uny, dum / dugz / 'G'/ 'N'/ 'M' $
equivx dugz / ugy / always $
alter '\/ dug '(2),'\/ dug '(2,1)
umerge uset, upnds1, dum / dugz / 'G'/ 'N'/ 'M' $
equivx dugz / upf / always $
compile update
alter 'restart ygp'
DBVIEW RMGP=RMG WHERE (DESITER=DESP) $
RESTART RMGP,RMG,//true $
endalter
@
35914; 70.7; Static Analysis - MPCForce output - Wrong Answers
In linear static analysis (SOLs 101 and 200), if there exist MPC entries,
but no rigid element entries, and some subcases, but not all select those
MPCs, then MPCFORCE output may be labeled with the incorrect SUBCASE ID
listed in the upper right hand corner of the printed output and the forces
will be wrong.
Avoidance: Add a rigid element or insert the following alter:
compile sedisp
alter 'IF ( NOQMG',''
IF ( NOQMG>-1 and not(delta) ) THEN $
if ( nomset<0 ) then $
PARAML USET//'USET'//////'g'/S,N,NOgSET $
matgen ,/qmx/7/nogset/nloads $
else $
alter 'call cforce'
endif $
endalter
@
35918; 70.7; Acoustic Analysis - PARAM,WTMASS - Inefficient Processing
When param, wtmass is used in conjunction with fluid elements it causes
the mass matrix to be of form 1 (unsymmetric). This causes an unsymmetric
solution, which requires more memory and cpu time than a symmetric
solution.

Avoidance: The following alter avoids this problem:

compile semg list noref
alter 'MPYAD.*MJJX1','' $ add sym flag
matprn wtmassg// $
MPYAD MJJX1,WTMASSG,/MJJX/////6 $ force symmetry
@
35921; 70.7; Aeroelastic Analysis - SPLINE3 - UFM 2331 (MKSPL)
Aeroelastic Analysis with SPLINE3 entries may fail with the following
message in Version 70.5:
*** USER FATAL MESSAGE 2331 (MKSPL)
BOXES PICKED ON SPLINE3 ENTRY ID = xxx HAVE NOT BEEN GENERATED BY
CAERO ENTRY ID = yyy
 USER INFORMATION: CAERO ENTRY ID = yyy GENERATES BOXES yyy
to zzz.
Avoidance: Use SPLINE2 entries if possible. If that is not possible,
order GRID components by internal order which can be determined by using
PARAM,USETPRT,1 .
@
35931; 70.7; Restarts/P-Elements - PARAM,PVALINIT - Wrong Answers
If a restart is performed from a p-version run and PARAM,PVALINIT is not
specified, then the restart will default the starting p-value to the first
p-value rather than the last p-value computed in the cold start.
Avoidance: Specify PARAM,PVALINIT or insert the following DMAP alter:
compile ifps1
alter 'type ',''
type parm,,i,n,seid,gnmax,pvalid,pvaltmp,np,adptindx,pval0,
oldseq1 $
type parm,nddl,i,n,pvallast $
endalter
@
35932; ; Nonlinear Analysis - NLRGAP - Radial Gap without Friction, UFM 3
16 (IFPDRV)
Using the NLRGAP does not allow input of the nonlinear forces without
reference to a TABLED1 which specifies O.O for friction. Leaving the
field 8 blank yields the following message:
*** USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA ENTRY NLRGAP 20
SORTED ENTRY COUNT = 36
for the input item:
SID GA GB plane tabk tabg tabu radius
NLRGAP 20 90 101 XY 21 22 0.0

Avoidance: The avoidance is to specify a table for the TABU field with
values of 0.0 for friction.
@
35935; ; Nonlinear Analysis - NLRGAP - UFM 2093 (DPD4)
An NLRGAP element that identifies a non-existent grid id produces a fatal
message that incorrectly identifies the grid. The following fatal message
was produced by an NLRGAP element that referenced a nonexistent GRID id of
90:
*** USER FATAL MESSAGE 2093 (DPD4)
NOLIN OR NLRGAP SET 20 REFERENCES GRID 100 COMPONENT 1
SIL 61 WHICH IS NOT IN UD SET.
*** USER FATAL MESSAGE 2093 (DPD4)
NOLIN OR NLRGAP SET 20 REFERENCES GRID 100 COMPONENT 2
SIL 62 WHICH IS NOT IN UD SET.
*** USER FATAL MESSAGE 2067 (DPD)
UNDEFINED GRID POINT 90 ON NOLIN- ENTRY.
Avoidance: Correct the NLRGAP entry.
@
35937; ; Performance - CPU Time not reported consistently
The CPU time reported in the f04 file is not consistently reported across
all platforms. Thus, comparisons of CPU time between machines that are
reported differently are meaningless. The method of reporting for most
systems is shown below:
Compaq: user + system
Cray: user
HP: user
IBM: user + system
NEC: user
SGI: user
Sun: user
Windows: user + system
Avoidance: To compare CPU times across different platforms, use the time
summary at the end of the LOG file.
@
35939; 70.7; Miscellaneous - INCLUDE - UFM 398, Unexpected Answers or Messag
es
In Version 70.6, if a referenced include file does not exist, the job will
issue the following fatal message
*** USER FATAL MESSAGE 398 (GETLIN)
A REQUESTED INCLUDE FILE WAS NOT FOUND
.
.
THE INCLUDE REQUEST IS SKIPPED
but the job will continue. This may lead to unexpected results as the
information in the INCLUDE file will not be included.
Avoidance: Make sure that the referenced INCLUDE file exists.
@
35993; 70.7; Optimization/Design Sensitivity - DRESP1 - UFM 6482 (DOPR3R)
If the DESOBJ references a DRESP1 that refers to a dynamic response
(e.g., FRDISP or TACCL) and there are constraints on nondynamic responses
(e.g., WEIGHT or DISP) then the message:
USER FATAL MESSAGE 6482 (DOPR3R)
FOR SUBCASE 0, THE ANALYSIS TYPE IS GLOBAL BUT THE DESOBJ
COMMAND REFERENCES AN INCOMPATIBLE
DRESP1 ENTRY WITH RID = n AND RTYPE = FRACCL.
0FATAL ERROR
is produced in error and it incorrectly states that the response type is
incompatible with the analysis type.
Avoidance: Modify the input to have the DESOBJ refer to a DRESP2 and then
create a DRESP2 as shown in the example
Example,
CASE control has:
DESOBJ=40001
 Bulk data has
DRESP1, 40000, ZIEL1, FRACCL , , , 2, 140., 6294
dresp2 40001 kludge 40000
dresp1 40000
deqatn 40000 kludge(fraccl) = fraccl
@
35996; ; SSSALTERs - trncomp1 - SFM 3002 (SDR2)
If the sssalter trncomp1 is used with PARAM,DDRMM,-1 or in SOL 112 T then
the message:
SYSTEM FATAL ERROR 3002 (SDR2 )
EOF ENCOUNTERED WHILE READING DATA BLOCK SILS (FILE 106) IN SUBROUTINE
SDR2.
is issued whenever a STRAIN case control request occurs.
Avoidance: To avoid the error the follwing sdr2 dmap statements must be
changed.

SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,SILS,ETT,OL2,bgpdts,
,,UG,EST,,,,,GPSNTS,,,/,,,,OEF1AX,/
APP1/S,N,NOS/FRCNCMP $
SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,SILS,ETT,OL2,bgpdts,
,,UG,EST,,,,,GPSNTS,,,/,,,OES1AX,,/
APP1/S,N,NOS/STRNCMP $
to:
SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,SILS,ETT,OLt,bgpdts,
,,UGdisp,EST,,,,,GPSNTS,,,/,,,,OEF1AX,/
'reig'/S,N,NOS/FRCNCMP $
SDR2 CASEDR,CSTMS,MPTS,DIT,EQEXINS,SILS,ETT,OLt,bgpdts,
,,UGdisp,EST,,,,,GPSNTS,,,/,,,OES1AX,,/
'reig'/S,N,NOS/STRNCMP $
@
36000; ; Nonlinear Transient Analysis - NLTRD2 - Performance Degradation
In many problems nltrd2 runs slower in Version 70.5 than in previous
versions. The investigation continues; however, it has been found that
increasing smem from the default of 100 to 1000 significantly improved the
performance of nltrd2 to where it exceeded prior versions performance.
Avoidance for V69: Nastran command should include keyword
"sysfield=mapio=no" in all v69 runs to improve the performance.
@
36014; 70.7; Aeroelastic Postprocessing - FLD&S - Unrestrained Unit Solution
not available
The static aeroelastic solution sequence (SOL 144) will store the "unit
solutions" (displacements and aeroelastically corrected forces associated
with the unit perturbation of the aerodynamic extra point) on the XDB if
PARAM, POST,0 is specified in the Bulk Data. Due to a typographical
error, the variant of the unit solution forces associated with the
unrestrained formulation are not correctly passed to the DBC module for
storage. Consequently they are not available to be visualized in
 MSC/PATRAN V8.1+ (which has been enhanced to recognize these data as part
of MSC/FLD V1.0).
No other results are affected since the data are computed. They are simply
not stored for subsequent post processing.
Avoidance: The following DMAP alter against V70.6 will correct the problem
compile aeopu list noref
$
$ Correct the DMAP error for unrestrained unit solutions on
$ the structural mesh (FUM==>FURM; IUM==>IURM)
$
alter 'dbc *aebgpdt.*prxks',''
DBC AEBGPDT, prxks, perxks, peuxks //
'BEPT' /'RM'/'FRM'/'FURM'/////////////////
0/AEDBCP/S,N,CP/'AERO'/0/GEOMU/LOADU/POSTU/
DBCDIAG/DBCCONV/DBCOVWRT/0///// $
alter 'dbc *bgpdts.*perxg',''
DBC BGPDTS, PRXG, PERXG, PEUXG, PIRXG, PIUXG //
'BGPDT' /'RM'/'FRM'/'FURM'/'IRM'/'IURM'///////////////
0/DBCPAE/S,N,CP/'AERO'/0/GEOMU/LOADU/POSTU/
DBCDIAG/DBCCONV/DBCOVWRT/0///// $
@
36042; 70.7; Optimization/Design Sensitivity - BEAM Library, DVPREL1 - Wrong
Answers
The design of the PBARL creates internal constraints on the bar dimensions
to insure, for example, that the inner radius of a tube does not exceed
the outer radius. The calculation of these constraints will be in error
if the DVPREL1 that designs a beam dimension that is included in the
constraint has a non-zero value for C0.
Avoidance: There is no avoidance other than to recast the design task in a
way that eliminates the need for a C0 term.
@
36066; 70.7; P-Elements - Incorrect Answers for p>=10
P-Elements with p>=10 may yield incorrect applied loads, boundary
conditions, and displacement coordinate systems.

@
36073; 2001; Acoustic Analysis - No Component Modes Computed for METHOD(FLUID
)
If an uncoupled acoustic analysis is performed and some of the fluid
grids are placed into the a-set because frequency dependent elements are
attached to them, then in SOL 108 no component modes are computed when the
eigenvalue extraction method is selected using the METHOD(FLUID) Case
Control command. Instead the METHOD Case Control command has to be used.
This is due to a logic error in SEMRM which does not look for
METHOD(FLUID).
Avoidance: A workaround is to use METHOD or to use the following
DMAP-alter (tested with V70.6):
compile semrm
 alter 'PARAML CASES'
if(method=0) PARAML CASES//'DTI'/1/226//S,N,METHOD $
endalter
@
36075; 2001; Eigenvalue Analysis - PARAM,RESVEC - SFM 2198 in FBS
If residual vectors are requested for an unconstrained superelement
which has no physical exterior points (no T-set), the run will fail with
SFM 2198 in FBS which means that the LOO matrix is missing.
Avoidance: Either have at least one exterior point for the superelement,
or insert the following alter:
compile sekr
alter 'IF ( NLKOT>-1 )','IF ( NLKOT>-1 )'
IF ( NLKOT>-1 OR RESVEC='YES') THEN $
endalter
@
36076; ; Miscellaneous - INPUTT2 - SFM 6551 with READ ONLY file
On NT computers, if the permission of a file used for INPUTT2 is
"read-only", the program may fail with SYSTEM FATAL MESSAGE 6551
indicating that the end-of-file was detected incorrectly.
Avoidance: Change the file permission to be full access.
@
36078; ; Complex Eigenvalue Analysis - SCR=NO - UFM 9050 (SEKRRS)
If while running a Complex Eigenvalue Analysis in MSC/NASTRAN Version 70.6
the run encounters
^^^ USER FATAL MESSAGE 9050 (SEKRRS)
^^^ RUN TERMINATED DUE TO EXCESSIVE PIVOT RATIOS IN MATRIX KLL.
and SCR = NO, set SCR = YES this may correct the problem.
@
36096; 2001; Eigenvalue Analysis - GIV, MGIV, HOU, MHOU - UWM 4648,3034
When the tridiagonal eigensolution methods GIV, MGIV, HOU, or MHOU are
selected, and the roots are closely spaced, the messages below may
result:
0***USER WARNING MESSAGE 4648---- THE MODAL MASS MATRIX IS NOT POSITIVE
DEFINITE.
0*** USER WARNING MESSAGE 3034, ORTHOGONALITY TEST FAILED. LARGEST TERM =
1.00E+00 ,NUMBER FAILED = 36
Avoidances:
An avoidance is to select the Lanczos method instead. It tests for
orthogonality as a requirement for stopping the solution process.
Another avoidance when using the tridiagonal methods is to stlightly
perturb the system so that the repeated roots are separated. It was found
by experiment with a 10 by 10 problem with 9 repeated roots in one cluster
that only the first four modes had good orthogonality. If 10 stiffness
terms on the diagonal active in a cluster of 9 roots were perturbed so
that each differed by 1.E-8 independent, orthogonal vectors were produced.
The changes in the eigenvalues were not large enough to appear in the 7
digit format used to print them.
@
36104; 2001; Optimization/Design Sensitivity - DEQATN - Infinite cpu Loop, re
ad error
If a large DEQATN entry is used, then depending on the machine, the job
may terminate without any messages in the f06 or f04 files or the run may
appear to be caught in an infinite loop. Depending on the machine, any of
the following messages may appear in the log file. This situation can
occur even if the number of non-zero characters are below the 32,000
limits.
1. fmt: read unexpected character
apparent state: unit -1 named
last format: (I2,4X,I4)
Unit -1 is a (null) formatted internal file
*** Execution Terminated (115) ***
2. 1525-097 A READ statement using decimal base input found the invalid
digit '' in the input file. The program will recover by
assuming a zero in its place.
Avoidance: Reducing the size the DEQATN entry.
@
36112; ; Transient Response Analsyis - DRMS1 DMRH1 - UWM 5186, Missing Ou
tput
DMAP modules DRMS1 and DRMH1 will issue the message:
*** USER WARNING MESSAGE 5186. (DRMH1C)
A CHANGE IN WORD 2 OF THE OFP-ID RECORDS OF DATA BLOCK 101 HAS BEEN
DETECTED.
PROCESSING OF THIS DATA BLOCK HAS BEEN TERMINATED.
and no output will be produced when the input OUGV1 matrix contains more
than 1 type of grid output request (e.g. displacements AND velocities).

Avoidance: The easiest avoidance is to only include one type of output

request in the case control (e.g. DISP or VELO or ACCE). A possible DMAP
avoidance would involve breaking the UG matrix into 3 separate matrices
(disp, vel, acc) before running through SDR2 to make 3 separate OUGV1s.
@
36133; 70.7; Optimization/Design Sensitivity - Rigid Element Sensitivity - UF
M 1126 in DSVGP5
If rigid element sensitivity is requested in SOL 200 for static and
frequency response analysis, then UFM 1126 is issued from module DSVGP5
in subDMAP PSLGDV for data block ENX3.
Avoidance: Insert the following DMAP alter:
compile pslgdv
alter 1
file enx3=ovrwrt/kmmb=ovrwrt/kmntb=ovrwrt/kmnb=ovrwrt/knnb=ovrwrt/
gtkmn=ovrwrt/ym=ovrwrt/egxx=ovrwrt/egx3=ovrwrt/mmmb=ovrwrt/
mmntb=ovrwrt/mmnb=ovrwrt/mnnb=ovrwrt/gtmmn=ovrwrt/gtmmny=ovrwrt/
gtmyg=ovrwrt/gtkmny=ovrwrt/gtbmny=ovrwrt/gtbyg=ovrwrt/gtkyg=ovrwrt/
gtkyg1=ovrwrt/gtk4yg=ovrwrt/gtk4mny=ovrwrt/egx4=ovrwrt $
endalter
@
36136; 70.7; Optimization/Design Sensitivity - PBEAML, PBARL - Incorrect Sens
itivities
The sensitivities of the stress recovery locations with respect to the
section dimensions are in error for the CHAN and Z section types. This
error will result in incorrect stress response sensitivities.
Avoidance: There is no avoidance. The magnitude of the error has been
under 10% in the examples that have been seen. Thus, the designs produced
using the incorrect sensitivities should be adequate.
@
36155; 70.7; Design Sensitivity - DSAMODES, SENSITY - Poor Performance
The eigenvector sensitivity capability of the dsamodes module requires
the identification of the row of the eigenvector which contains the
largest value. Currently this is done by a do loop in the subdamp which
checks the magnitude of each term in the vector in turn. This can be very
time consuming if the size of the eigenvector is large.
Avoidance: Insert the following DMAP alter:
compile dsamodes
alter 'do while ( mode'(,-1)
PARAML USET//'USET'//////'f'/S,N,NOfSET $
matgen ,/row1/4/nofset/1 $ row matrix of 1 2 3 4 ....
trnsp row1/rowt $ column matrix of 1 2 3 4 ....
alter 'ir\=0','ir\=0'(,5)
matmod phj,,,,,/phjn/2////1. $ filter out values less than one
add phjn,rowt/phjx///1 $ element-by-element multiply
norm phjx/junk///s,n,irr $ find row number
ir=int(irr) $
endalter
@
36160; ; User's Modifiable - blank common block - Possible Wrong Answers
The blank common block name for the User's Modifiable has been changed for
Version 70.5. If the old name common / / is used, the results may be
unpredictable.
Avoidance: Replace the following common block
common / /
with
common /BLNKCM/
@
36215; 2001; Documentation/QRG - PCOMP - Failure Index Output
In the MSC/NASTRAN V70.5 QUICK REFERENCE GUIDE, Remark 7 under the PCOMP
bulk data description needs to be changed:
from:
7. ELFORCE and ELSTRESS requests must be present for all elements for
which ply stress or failure index output is desired.
to:
7. In order to get failure index output the following must be present:
 a) ELSTRESS or ELSTRAIN case control commands,
b) Sb, FT, and SOUTi on the PCOMP bulk data entry,
c) Xt, Xc, Yt, Yc, and S on all referenced MAT8 bulk data entries if
stress allowables are used,
or Xt, Xc, Yt, Yc, S, and STRN=1.0 if strain allowables are used.
@
36230; 2001.0.1; Aeroelastic Analysis - SPLINEi, SET1, SET2 - Unexpected Answer
s
If an input file contains both a SET1 and a SET2 Bulk Data entry with the
same identification number, one or the other willl randomly be used in the
spline(s) that reference that SETi identification number. There is no way
to control which entry will be used and no error message is issued that
indicates that one of the two inputs is being ignored.
Avoidance: There is no real avoidance. You must ensure that your SET1 and
SET2 entries do not have common identification numbers.
@
36245; 2001; Documentation - DEFUSET - UFM 316 When DEFUSET References PA or
PS Sets
If a DEFUSET entry references the PA or PS sets the following fatal
message will be encounterd:
*** USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA ENTRY DEFUSET ********
SORTED ENTRY COUNT = xx
@
36247; 2001; Documentation - LOADSET - Wrong Answers
Multiple loadset entries may not be specified in a dynamics run. If there
are loadsets in each subcase the results for the subsequent subcases use
the initial loadset and are usually zero. However if the user defines the
same assumed DAREA id on each of his LSEQ entries wrong results will occur
for the subsequent subcase. The load is calculated from
using the initial loadset. A warning or fatal message should be issued if
the loadset is ignored. This problem has been fixed in the code in V2001.
@
36250; ; Static Analysis - SUBCOM, SUBSEQ, SDR2 - SFM 3002
If a static analysis has SUBCOMs and SUBSEQ, the job may issue SYSTEM
FATAL MESSAGE 3002 from the SDR2 module.

@
36251; 70.7; Beam Library - PBEAML, PBARL - Incorrect Properties for "T1" Sec
tion
PBEAML and PBARL produce incorrect I2 and J properteis for the "T1" type
section.
Avoidance: If possible use the "T" type section.

@
36257; 70.7; Nonlinear Analysis - UWM 4698 (DCMPD)
Nonlinear analysis may identify degrees of freedom large factor diagonal
 ratios at every iteration with the UWM 4698 shown below. This may result
in a large amount of output. Version 70.7 has been enhanced so that this
output may be controlled by setting SYSTEM(60)=-999.
USER WARNING MESSAGE 4698 (DCMPD)
STATISTICS FOR DECOMPOSITION OF MATRIX KLLRH .
THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN
1.00000E+07 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL.
@
36286; 70.7; Loads - RFORCE - Wrong Answers
In Version 69 and later systems, RFORCE will produce incorrect moment
loads. Moments are produced when CONM2 entries have an offset specified.
Moments are also produced for the mass associated with a PBEAM entry when
an offset is specified and for the PBEAM entry when a channel cross
section is specified. .
Avoidance: The only avoidance is to specify an additional GRID at the
offset location and use a rigid element to attach it to the main GRID.
@
36287; 70.7; Optimization/Design Sensitivity - MODTRAK, scr=no - SFM 3001 in
DOM12
When running SOL 200 with mode tracking (MODTRAK) and scr=no, the job may
terminate with the following diagnostic message in the f06 file:
*** SYSTEM FATAL MESSAGE 3001 (DOM12)
THE EXECUTIVE FILE -112 DOES NOT EXIST.
PROGRAMMER INFORMATION: SUBROUTINE DOM11C
Avoidance: Run with scr=yes or insert the following DMAP alter:
compile fea
alter 'if ( modetrak'(2)
IF ( MODETRAK>0 ) THEN $
delete /mtrak,,,, $
modtrl mtraki//nomod/nomtk $
copy mtraki/mtrak $
endIF $ MODETRAK>0
alter 'delete /mtrak'(,-1),'delete /mtrak'(,3)
endalter
@
36299; 70.7; Miscellaneous - OFP - Inefficient Processing
Efficiency has been improved in the OFP module for MSC/NASTRAN Version
70.7 for jobs that request only plot output.

@
36350; 70.7; FMS - INIT MASTER(NORAM) - SFM 1026
In Version 70.6, if the fms statement INIT MASTER(NORAM) is used on UNIX
machines, then the job may terminate
with the following error messages:
*** SYSTEM FATAL MESSAGE 1026 (XIO)
ATTEMPT TO ACCESS A FILE WHICH IS NOT OPENED
0********* DUMP OF XIO PARAMETERS
0XIO NAME = 423
 ATTEMPTED OPERATION = FILPSX
FILEXX = 23
Avoidance: Remove INIT MASTER(NORAM).
@
36366; ; SSSALTERs - mtacca.vxxx - Wrong Answers
SSSALTER mtacca.vxxx assumes that the modes are mass-normalized. If your
modes are not mass-normalized, wrong answers will result.
Avoidance: Use the default (MASS) eigenvector normalization.
@
36389; 70.7; Miscellaneous - Case Control, MODES - Poor Performance
If a very large value of x is requested with the modes=x case control
command, then the performance will be very inefficient. An example of
this situation is illutrated below:
Subcase 1
modes = 20
disp = all
Subcase 21
modes = 9000
disp = none
Avoidance: Request a reasonable modes value.
@
36391; 2001; Documentation - Superelements/CLOAD - UFM 5423
In the MSC/NASTRAN V70.5 Quick Reference Guide, remark 1 of the CLOAD case
control command should read:
"1. This command may only appear in the residual structure subcases (see
SUPER command) and it must be specified in all of them."
@
36410; ; Static Analysis - TEMP(INIT) - Wrong Answers for Multiple Subcas
es
In Sol 101, if TEMP(INIT) is not defined above the subcase level, it is
ignored and incorrect results occur. Contrary to the Quick Reference
Guide documentation TEMP(INIT) is the recommended way for specifying the
datum (reference) temperature for static analysis with temperature loads.
However, TEMP(INIT) must be made above the subcase level.
@
36428; 70.7; Optimization/Design Sensitivity - DOPR1 - UWM 7602
The following message is correctly issued but is confusing because it
states the analyis value which in this case is not greater than the
maximum value allowed. The message should state the design value.
*** USER WARNING MESSAGE 7602 (DOPR1I)
ON PROPERTY ENTRY PBEAM WITH PID= 7610 AND FID= -9, THE
PROPERTY VALUE 9.93100E+06 IS
GREATER THAN THE MAXIMUM PMAX VALUE 1.00000E+07 SPECIFIED ON THE
DVPREL2 761001 ENTRY.
@
36450; 70.7; P-Elements - GMBC, Multiple Edge Boundary Conditions - SFM 6363
 If there are multiple edge GMBC's for different coordinate directions
using different methodon a p-element, the following error results:
*** SYSTEM FATAL MESSAGE 6363 (EVLEDG)
THE ENFORCED DISPLACEMENT VECTOR COLUMN DOES NOT EXIST
FOR EDGE ID = -21 FOR COMPONENT = 3
USER ACTION: CONTACT MSC CLIENT SUPPORT.
PROGRAMMER ACTION: CHECK ENFORCED DISPLACEMENT VECTOR ROUTINE
(IERROR=2).
Avoidance: A workaround is to use the same method for the multiple
boundary conditions.
@
36451; ; Superelements - MATMOD option 16 - Wrong Answers
When using a determinate interface with an external superelement, using
MATMOD option 16 may lose rigid-body modes due to numeric roundoff.
Avoidance: When creating external superelements, do not use a determinate
interface for the component models.

@
36463; 70.7; Optimization/Design Sensitivity - DVPREL1, PBUSH, PSHELL - UFM 7
190
When an optimization task includes DVPRELi entries refrencing PBUSH as
well as other property entries, it fails in module DOPR1 with the
following error message:
*** USER FATAL MESSAGE 7190 (DOPR1I)
PTYPE PBUSH DEFINED ON A DVPREL1 2 ENTRY CAN NOT BE FOUND
ON A PROPERTY ENTRY.
USER ACTION: DELETE THE DVPREL1 ENTRY OR INCLUDE AN ADDITIONAL
PROPERTY ENTRY.
Avoidance: Remove DVPRELi entries which do not reference PBUSH entries.
@
36465; 70.7; Aeroelastic Analysis - Wrong Answers for QHHL in tpl ha145i
The tpl deck ha145i fails in Version 70.6 due to an interaction between
the incomplete Case Control used in this test case and the changes made to
NASTRAN to support multiple aerodynamic symmetries.

Under those changes, the AERO/AEROS symmetries (which are single valued
for a run) are "replaced" by symmetry in Case Control. However, for upward
compatibility, the AERO/AEROS symmetries are used in the absence of any
Case Control. A subtle requirement of this upward compatiblity logic is
that the Cases must be "aerodynamic" cases for the procedure to function.
Unfortunately, HA145I doesn't have a complete Case Control, so the fact
that the run is aeroelastic cannot be determined from Case. Consequently,
the defaulting rules (whereby the AERO/AEROS would have provided the
symmetry) were not applied to this case. Instead, the value from Case
Control (-2) that represents the fact that the symmetry is NOT supplied in
CASE is mistakenly returned to the DMAP. AMG (which should have failed on
illegal symmetry option) responds in a random way to this illegal value.
Consequently, the deck gives spurious results.
To correct the problem, an FMETHOD entry is needed in the Case Control of
 the HA145I test deck. The corresponding FLUTTER entry is not required in
the Bulk Data because the test case terminates before needing it.
@
36528; 70.7; Optimization/Design Sensitivity - DEQATN - Infinite CPU Loop
If a DEQATN contains a double precision constant (e.g. 10.0D0) in a design
optimization task, an infinite loop will be encountered and the job will
enter the PREDOM subdmap and never complete. Version 70.7 outputs USER
FATAL MESSAGE 224 to correct this problem.
Avoidance: Remove the double precision constant from the DEQATN.
@
36530; 70.7; TRANS - MAT5, MAT9, MAT10 - forrtl: error (65): floating invlaid
The trans program may fail to convert a binary XDB file to a neutral file
when MAT5, MAT9, or MAT10 entries are present in the MSC/NASTRAN input
file. This error has only been duplicated on DEC ALPHAs running UNIX thus
far. The error message from trans is: forrtl: error (65): floating
invlaid.
Avoidance: Set PARAM,GEOMU,0 which eliminates all of the undeformed
geometry or edit the PHASE0 subdmap DBC statement to eliminate the writing
of the MPT table.

@
36546; ; Nonlinear Buckling Analysis - QUAD4 - Spurious Modes
For a rare case of boundary conditions, spurious modes in nonlinear
buckling (Sol 106) may occur for coarse meshes of CQUAD4 elements. The
rare case of boundary conditions is described as follows. All rotational
degrees of freedom in a CQUAD4 are constrained. The transverse
displacements are not constrained or only partially constrained.
Avoidance: Refine the mesh.
@
36547; 70.7; Superelements - Limit of the number of superelements - SFM 4276,
EC 1044
Models that contain more than 100 superelements may fail with SYSTEM FATAL
MESSAGE 4276, ERROR CODE 1044 caused by an integer overflow in routine
GETOFF.

@
36563; 70.7; Miscellaneous - SGI, OS level 6.5 - Receiving signal number 10:
Bus error
On the SGI when running OS level 6.5, if the scratch disk is full, then
the job may fail with the following unclear messages in the log file:
Receiving signal number 10: Bus error
Job aborting..
Avoidance: Allocate more space with the scratch disk..
@
36564; ; DBC - Restart of an Existing XDB
If a cold start run is made but an XDB exists from a prior run. then DBC
 may not correctly correlate the number of modes and duplicate results may
be stored.
@
36569; 70.7; Miscellaneous - SUBCOM,SUBSEQ - Missing Error Messages, Unexpect
ed Answers
The number of coefficients for SUBSEQ are limited to 200, which is
documented in the QRG. However, if more than 200 coefficients are used--in
the case of combining more than 200 subcases--only the first 200
coefficients are used without any warning or fatal messages.
Avoidance: If more than 200 coefficients are needed, create a separate
subcase
@
36589; 70.7; Performance - MCE1, MPC, RBE - Poor Performance
When there are a large number of m-set dofs, the mce1 module may exhibit
poor performance. In cases seen to date, the decomp portion of mce1
finishes in a reasonable time, but the next portion, which happens to be
an fbs with many right-hand sides, ran for hours without completion.
This condition can be detected by looking at the run time for the decomp
operation. If the fbs takes more than 10 times this value, the avoidance
below may be worth considering.
Use the mce1.v706 subdmap from MSC's web site for sssalters. It has been
verified that it works correctly for Version 70.5 as well. On the model
in question, it reduced mce1 times from over 15 hours to minutes.

@
36611; 2001; Optimization/Design Sensitivity - BNDGRID, DVGRID, MODTRK - UFM
1126 in GP4
If mode tracking is requested with BNDGRID and DVGRID entries present,
then the run will fail with the message:
*** USER FATAL MESSAGE 1126 (GNFIST)
DMAP MODULE GP4 IS ATTEMPTING TO OUTPUT DATABLOCK 203 NAME =
USET0 , WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE USET0=OVRWRT
Avoidance: Insert the following alter:
compile desopt
alter 'lcgen '(,-3),''
if ( ambs or gmbs ) then $
endalter
@
36617; 70.7.2; Matrix - SOLVE - Possible Wrong Answers with Low Memory
If insufficient memory is available for the SOLVE module with SPARSE, then
the results may be incorrect.
Avoidance: Increase memory.

@
36628; 70.7; Heat Transfer Analysis - RADSET - Unnecessary Calculation of Vie
w Factors
According to the MSC/NASTRAN Quick Reference Guide, the RADSET entry
"Specifies which radiation cavities are to be included for radiation
enclosure analysis." However, in a multiple radiation enclosure model,
if an enclosure is NOT referenced on the RADSET entry, the view factors
are still calculated.
Avoidance: There is no avoidance.
@
36630; 70.7; DBC - Changing tstep or freq between subcases causes failure in
dbc
If tstep or freq is changed between subcases and param,post,0 is specified
to create a dbc database, then dbc will not complete and the following
messages will be issued to the f06 file:
*** SYSTEM WARNING MESSAGE 6065 (DBC)
*** DBCMSG ERROR MESSAGE 44 FROM SUBROUTINE DBCAOL
OUTPUT LIST DOES NOT MATCH EXISTING IN THE DATABASE, MERGE/RESTART
OPERATION IMPOSSIBLE.
*** SYSTEM WARNING MESSAGE 6066 (DBC)
*** DBCMSG ERROR MESSAGE 34 FROM SUBROUTINE DBC
THE DBC MODULE HAS BEEN TERMINATED BECAUSE OF THE ABOVE FATAL DBC
ERROR.
USER INFORMATION: SUBSEQUENT DBC CALLS WILL NOT BE EXECUTED, AND
THE DATABASE CONVERSION IS INCOMPLETE.
Avoidance: The only avoidance is to execute the subcases in different
runs.
@
36632; 70.7; Grid Point Stress - PARAM,CURV - Floating point overflow
If PARAM,CURV,1 is specified then the run may fail in the CURV module with
the following traceback:
Trap encountered:
SIGTRAP - Trace trap FP overflow
Traceback:
Offset 0x00000240 in procedure pstr2s_
Offset 0x00001bd4 in procedure curv3_
Offset 0x00000198 in procedure curv_
Offset 0x000000e4 in procedure xsem14_
Offset 0x00000984 in procedure mainal
Avoidance: The input that demonstrated the error contained a STRESS=set
id. The error was avoided when this request was changed to STRESS=ALL.
@
36636; 70.7; Heat Transfer - VIEW, RADSET - UWM 6427, UIM 4534, UWM 6137 (UDS
NRD)
If an initial run is used to generate a radiation matrix punch file and a
subsequent run inputs the punch file and defines a RADSET entry that
attempts re-calculate the view factor for some but not all of the
cavities, then the job will fail with the following messages:
*** USER WARNING MESSAGE 6427 (VIEWD)
VIEW FACTORS WILL NOT BE COMPUTED FOR RADIATION CAVITY ID= 85
BECAUSE A RADIATION MATRIX IS ALREADY DEFINED.
USER INFORMATION: THE RADIATION MATRIX IS DEFINED ON RADLST OR
 RADMTX BULK DATA ENTRIES.
*** USER INFORMATION MESSAGE 4534 (RMG2)
2 ELEMENTS HAVE A TOTAL VIEW FACTOR (FA/A) LESS THAN 0.99, ENERGY
MAY BE LOST TO SPACE.
*** USER WARNING MESSAGE 6137 (UDSNRD)
INPUT MATRIX IS RANK DEFICIENT, RANK = 10 IERROR
USER ACTION: CHECK MODEL
*** USER WARNING MESSAGE 5225 (DCMP)
ATTEMPT TO OPERATE ON THE SINGULAR MATRIX KLLRH IN SUBROUTINE DCMP.
0FATAL ERROR
@
36645; 70.7; Rigid Elements - RBE3, GP4 - UFM 6490
When an extremely large number of RBE3 elements is used, the job may fail
indicating an RBE3 input error:
*** USER FATAL MESSAGE 6490 (WRGMTD)
FATAL ERRORS HAVE OCCURRED WHILE PROCESSING MPC AND RIGID ELEMENT
ENTRIES.
PROGRAMMER INFORMATION: SIL NUMBER ******** DOES NOT EXIST IN THE
DEPENDENT SIL TABLE.

Other symptoms are that run times may increase greatly if the number of
RBE3 elements increases. This effect begins to be noticeable at about
10,000 RBE3 elements, and becomes the dominant source of cost in runs
where there are hundreds of thousands of RBE3 elements.
The RBE3 code was revised in V70.7 to provide increased efficiency. The
original code was made with the assumption that no model would have over
100 RBE3 elements. More modern general software tools for reading long
records repeatedly, introduced in V70.7, were used to improve efficiency.
Several other enhancements were made at the same time, as described in TAN
4155. The net effect is that a model that runs for 15 hours on a Cray
computer in V70.5 now runs in several hundred seconds in V70.7. TANs can
be found on the MSC home page by going to the search engine and putting in
the appropriate words. At present the TAN number can not be used in the
search. Two words that will find this TAN are RBE3 and V70.7.
@
36665; 70.7; Aeroelastic Analysis - PFAERO - Infinite Loop PFAERO
If the user attempts to run an aeroelastic solution sequence (SOL 144, 145
or 146 or SOL 200 with aeroelastic analyses) without any aerodynamic data,
an infinite loop will occur in the PFAERO subdmap. This is caused by a
failure to trap the absence of the requisite aerodynamic data.
Subsequently, the PFAERO subdmap (responding to data reuse logic) will
endlessly attempt to create the aerodynamic model.
Avoidance: You must have at least one aeroelasticity-related bulk data
entry to avoid the infinite loop and trigger the "missing data" error
traps.
@
36678; 70.7.2; Eigenvalue Analysis - EIGR, EIGRL - SFM 7340
The Lanczos method of eigenvalue extraction may give nonstandard error
messages. For example a run may terminate with the following messages.
*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
process error reported by subroutine LNNP2CD
see LNNP2CD abstract (ier = -725)
 *** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
warning reported by subroutine LNNDRVD
see LNNDRVD abstract (ier = 725)
Avoidance: The LNNDRVD or LNNDRVS messages meanings are given below:
ier error condition.
-723 insufficient workspace revealed during lanczos iteration
-724 failure of tridiagonal ql iteration to converge
-725 number of eigenvalues computed exceeds internal storage limitations
-726 i/o error with lanczos vector file
-727 failure of bidiagonal ql iteration to converge
-728 convergence failure in gram-schmidt orthogonalization -- probable
cause - indefinite m (vibration) or indefinite k (buckling)
-729 repeated singular matrix -- probably ill-posed problem
-730 number of trust regions too large
-731 factorization restore i/o error
-732 ran out of computer resources.
-733 inconsistent treatment of number of eigenvalues in finite interval
-734 could not compute zero eigenvalues for a buckling case
@
36682; 70.7; Miscellaneous - Requesting Huge mem may yield Wrong Answers
Reguesting more than 1200mb of memory on an RS6000 system may result in
wrong answers or a fatal message.

@
36685; 70.7; Inertia Relief - INREL, SUPORT1 - SFM 3007 in UPARTN
If there are multiple subcases with SUPORT1 command in some, but not all,
subcases and PARAM,INREL,-1 is specified in the Bulk Data then the run
will terminate with:
*** SYSTEM FATAL MESSAGE 3007 (UPARTN)
ILLEGAL INPUT TO SUBROUTINE PARTN
Avoidance: Insert the following DMAP alter:
compile statrs
alter 'if ( norset',''
if ( norset<6 and norset>0 ) then $
endalter
@
36701; 2001; Dynamic Analysis - DAREA - Unexpected Answers
If on the DAREA entry the coordinate system is specified as "0" for a grid
point it will be interpreted as "1". For example:
$ pi ci
DAREA 410 3 0 1
is interpeted by the program as
DAREA 410 3 1 1
This may cause unexpected answers.
Avoidance: Specify the desired degree of freedom.
@
36706; ; Old Design Senitivity - CBEAM - Possible Wrong Answers
If BEAM elements are used in a modal analysis with the old design
sensitivity analysis, then the design sensitivity coefficients may be
incorrect.
Avoidance: 1. Replace the beam elements with bar elements.
2. Use SOL 200 (v707.7) to perform the same task if the user
insists using beam elements.
@
36709; ; Optimization/Design Sensitivity - DSVG1D - UFM 5172 When Materia
l has no Density
If a designed property (DESVAR/DVPREL1) references a material with zero
density, then the job will fail with the following messages:
*** USER FATAL MESSAGE 5172 (DSVG1D)
THE VARIATIONAL ELEMENT ID IN MDICTDSF INCONSISTENT WITH THE
VARIATIONAL ELEMENT ID IN DSPT1, MODULE DSVG1 IS TERMINATED.
*** SYSTEM FATAL MESSAGE 3001 (DSVG1)
THE INPUT DATA BLOCK NAMED MDICTDSF AT POSITION 1 DOES NOT
EXIST.
Avoidance: Specify a non-zero density on the material entry.
@
36727; 70.7; Dynamic Analysis w/Differential Stiffness - SUPORTi, STATSUB, sc
r=yes - UFM 4407
If differential stiffness is requested in dynamic analyis, supported
degrees-of-freedom are specified (see the SUPORT or SUPORT1 entries), and
the job is run with scr=yes, then the mass matrix will not be reduced to
the supported degrees-of-freedom and UFM 4407 will be issued.
Avoidance: Run with scr=no or insert the following DMAP alter:
compile phase1dr
alter 'or kdiff',''
if ( true ) then $
endalter
@
36730; 70.7; Superelements - SEBULK, SECONCT - UFM 6794
When using part superelement, only a single boundary point is allowed at a
particular boundary location (within a specified tolerance), this is
design intent. However, it has been observed in a model that the job
terminates with the following messages:
*** USER FATAL MESSAGE 6794 (SEPSEB)
DUPLICATE BOUNDARY POINTS DETECTED IN UPSTREAM SUPERELEMENT 1
WHILE SEARCHING 0 AT GRID 7001.
LIST FOLLOWS:
13403 49682
even though the points identified do not appear to be at the same location
within the tolerance. The exact cause of this error is not known at this
point.
Avoidacne: Use the SEEXCLD command to remove all but a single boundary
point at each PART.
@
36737; 70.7; Optimization/Design Sensitivity - PARAM SOFTEXIT YES, SDR1 - UFM
4574
If PARAM,SOFTEXIT,YES is specified with static or frequency response
analysis in SOL 200, then the run will fail with UFM 4574 and SFM 3001 in
module SDR1 and subDMAP SEDISP on the final data recovery.
Avoidance: Insert the following DMAP alter:
compile desopt
alter 'dbdelete',''
endalter
@
36741; 70.7; Response Spectrum Analysis - RSPEC - SFE 4276 EC 777
If a response spectrum analysis is performed in Version 70.5 on HP
computers, then the job may fail with the following message and traceback
in the RSPEC module:
Process traceback ...
( 0) 0x00093e90 sysrcv + 0x2a0 [/nast2/msc705t2/hpux/analysis]
( 1) 0x01126d78 _sigreturn [/nast2/msc705t2/hpux/analysis]
( 2) 0x004935e8 rspcod_ + 0x740 [/nast2/msc705t2/hpux/analysis]
( 3) 0x00199024 rspec_ + 0xbc [/nast2/msc705t2/hpux/analysis]
( 4) 0x00098d44 xsem14_ + 0x2fc [/nast2/msc705t2/hpux/analysis]
( 5) 0x0008fbc8 mainal_ + 0x9b0 [/nast2/msc705t2/hpux/analysis]
real 19.5
Avoidance: The only avoidances are to run Version 70 or generate the
curves on another computer.

@
36752; 70.7; Miscellaneous - MSCACT - UFM (mscact) Unknown keyword
The MSCACT utility incorrectly flags the keyword assignment "summary=jid"
as an error and outputs the message shown below. However, "summary=job"
is accepted and generates the requested functionality.
USER FATAL MESSAGE (mscact)
unknown keyword "summary=jid".
Avoidance:
1) Use "summary=job" instead of "summary=jid".

2) Alternatively, the error can be corrected by changing line 1180 of

<ins-dir>/msc705/util/mscact.c from
else if( !strncmp( "job", q, lq ) )
to
else if( !strncmp( "jid", q, lq ) )
Once the change is made, rebuild MSCACT as described in the "Configuration
and Operations Guide".
@
36764; 70.7; Acoustic Analysis - FORCE(PHASE), Velocities and Pressure levels
- Wrong Answers
The acoustic output titled "COMPLEX ACCELERATIONS VELOCITIES AND PRESSURE
LEVELS" prints an incorrect format for the element ID when the
MAGNITUDE/PHASE option is requested. The incorrect format is that the
"PRESSURE(DB)" column is truncated and the phase values are misaligned.
Here is an example of the incorrect MAGNITUDE/PHASE output:
( P E A K )
C O M P L E X A C C E L E R A T I O N S V E L O C I T I
E S A N D P R E S S U R E L E V E L S
(MAGNITUDE/PHASE)
ELE-ID EL-TYPE X-ACCELERATION Y-ACCELERATION Z-ACCELERATION
X-VELOCITY Y-VELOCITY Z-VELOCITY PRESSURE (DB)
1.4994E-43 HEXPR 1.364242E-11 -2.376055E-10 4.083463E+04
.0 .0 .0 5.298811E+0
.0 .0 .0
.0000 .0000 -1624.7582
Here is the output for real/imaginary:
C O M P L E X A C C E L E R A T I O N S V E L O C I T I
E S A N D P R E S S U R E L E V E L S
(REAL/IMAGINARY)
ELE-ID EL-TYPE X-ACCELERATION Y-ACCELERATION Z-ACCELERATION
X-VELOCITY Y-VELOCITY Z-VELOCITY PRESSURE (DB)
107 HEXPR 1.364242E-11 -2.376055E-10 4.083463E+04
.0 .0 .0 5.298811E+01
.0 .0 .0
-5.428147E-13 9.454022E-12 -1.624758E+03
The error only affects the printed output and not the formats of OUTPUT2
and DBC modules.
Avoidance: Request REAL/IMAGINARY output.
@
36776; 70.7; Optimization/Design Sensitivity - PBAL, DLINK, DOM9 - UFM 6499 (
DOM9D)
An optimization job which designs PBARL dimensions with DLINK entries
fails in DOM9 with the following message:
THE GRADIENT OF CONSTRAINT NUMBER NNNNN IS ZERO
CONSTRAINT IS REMOVED FROM THE ACTIVE SET
** THIS IS A VIOLATED CONSTRAINT
** A FATAL ERROR WILL BE GENERATED

**** ONE OR MORE VIOLATED CONSTRAINTS HAS A ZERO GRADIENT ****

NO PROGRESS IS POSSIBLE
OPTIMIZATION TERMINATED
*** USER FATAL MESSAGE 6499 (DOM9D)
THE DOT OPTIMIZATION ALGORITHM HAS ENCOUNTERED AN ERROR THAT MAKES
FURTHER PROGRESS IMPOSSIBLE.
SEE MESSAGES ABOVE.
Avoidance: There is no avoidance.
@
36777; 70.7.2; Optimization/Design Sensitivity - ANALYISIS=MODES and MFREQ - SF
M 3001 in DSAH
If a model has fluid-structure interaction in SOL 200 and both
ANALYSIS=MODES and ANALYSIS=MFREQ are requested, the run may fail with SFM
3001, stating that data block PHG does not exist.
Avoidance: The following alter will avoid the FATAL error, but has not
been confirmed to give correct answers.
compile fea
alter 'IF.*DOSUPER3 AND NOT.*FSCOUP',''
IF ( DOSUPER3 ) THEN $
@
36780; 70.7; Submittal - vers= - Does not get right version
If a different version than the latest installed (or latest known by the
invoked nastran command if under the form nastxxx where
xxx is a version number) is requested, the rc files of the requested
version are picked up, but not the executable and SSS.
@
36790; ; Heat Tranfer Analysis - VIEW, UDSFA - UFM 6136, estimate=mem
MEM=estimate did not yield the correct estimate for a large heat trasnfer
radiation problem. The job subsequenty Failed with the message:
*** USER FATAL MESSAGE 6136 (UDSFA)
INSUFFICIENT CORE FOR NUMERIC PHASE OF SPARSE DECOMPOSITION.
USER ACTION: INCREASE CORE BY 5217107 WORDS
0FATAL ERROR
@
36792; 70.7; Aerodynamic Analysis - PARAM,OPGTKG,1 - Missing output
PARAM,OPGTKG,1 requests a printout of the aerodynamic splining matrices.
The capability was inadvertently deleted from Version 70.6.
Avoidance: Insert the following DMAP alter:
compile phase1a
alter 'np\=seid',''
equivx uset0/aeusets/-1 $
equivx bgpdts/aebgpdts/-1 $
dbdict datablk=(aebgpdts,aeusets) select=(name,size,
modltype,aeconfig,symxy,symxz) qualcurr $
dbview bgpdtf=aebgpdts where (modltype='aerostrc' or
modltype='structur' ) $
dbview usetff=aeusets where (modltype='aerostrc' or
modltype='structur' ) $
IF ( OPGTKG>=0 ) then $
MATGPR bgpdtf,usetff,,GPJK//'K'/'G' $
MATGPR bgpdtf,usetff,,GDJK//'K'/'G' $
endIF $ OPGTKG>=0
endalter
@
36815; 2001; Composites - TEMPP3 - Wrong Answers
The TEMPP3 thermal loading bulk data entry is obsolete and is no longer
supported. The TEMPP3 entry may give wrong answers for composites with
 inplane or normal thermal gradients.
Avoidance: Use the TEMPP1 bulk data entry instead which works correctly
in all cases and includes the same features as the TEMPP3.
@
36821; ; Optimization/Design Sensitivity - RHO - Incorrect Material Sensi
tivities
The design sensitivities with respect to the material densities of
transient response analysis that has a steady state response may be
incorrect.
Avoidance: Do not specify the material density as a response for a steady
state analysis since the response is not affected by the density.
@
36823; 2001; Nonlinear Transient Analysis - OTIME, SDR1 - BIOMSG: ERROR 211
in BLDPK
In nonlinear transient analyis, if the value of an output time on an
OTIME= request exceeds the time value of the final time step, the job
will fail in the SDR1 module with the message:

BIOMSG: ERROR 211 HAS OCCURRED IN ROUTINE BLDPK , FILE INDEX (FILEX)
= 0
STATUS = 0
the same job runs with fewer time step intervals for output (NO) on the
TSTEPNL card or with fewer OTIME requests in the case control.
Avoidance: Do not use the OTIME entry in nonlinear transient analysis.
The Quick Reference Guide does not indicate that OTIME is supported for
nonlinear transient solutions.
@
36836; ; FMS - INIT - UFM 1257
If a large value is assigned to a DBSET member, e.g.,
init scratch logical(scratch(50gb))
then the job may fail with the following unclear messages:
*** USER FATAL MESSAGE 1257 (DBMAP)
THE MAX-SIZE SPECIFIED ON THE INIT FMS STATEMENT OR SMEM VALUE IN THE
RC FILE OR SMEM
KEYWORD IN THE COMMAND PROCEDURE FOR THE MEMFILE OF DBSET SCRATCH
IS 100 BLOCKS
AND IS LESS THAN THE RECOMMENDED MINIMUM OF 110 BLOCKS.
USER ACTION: SPECIFY A MAX-SIZE OR SMEM VALUE AT OR ABOVE THE
RECOMMENDED VALUE.
Avoidance: Either increase the buffsize or decrease the dbset value
@
36842; ; Superelements - PARAM,SESEF - UFM 5423
If the fraction of total strain energy (PARAM,SESEF) is requested and
external superelements exist, the job will fails with USER FATAL MESSAGE
5423.
 Avoidance: Include the following alter:
compile postreig, list
alter 'MPYAD.*KAA,PHILS',''
type parm,,i,n,temp
dbview kaanew=kaa, where(seid=temp and wildcard)
temp=seid $
paraml kaanew//'trailer'/1/s,n,noaset $
MPYAD KAAnew,PHILS,/FA $
@
36870; 2001; Nonlinear Static Analyis - MPCFORCE, SE, INTOUT - SFM 5423 (DMPY
AD)
If mpcforce output is requested in a nonlinear analysis, and the model
contains superelements and intermediate output is requested, then the job
will fail with the following messages:
*** USER INFORMATION MESSAGE 7321 (SEDRDR)
DATA RECOVERY FOR SUPERELEMENT 20 IS NOW INITIATED.
*** SYSTEM FATAL MESSAGE 5423 (DMPYAD)
ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES
COLS ROWS FORM
TYPE NZWD DENS
TRAILER FOR KMG IS 101108522 1310 2
2 12 3
TRAILER FOR UGJ IS 102 1108522 2
2179608 8275
TRAILER FOR PM IS 103 5 1310 2
2 0 0
TRANSPOSE FLAG IS 0
0FATAL ERROR
Avoidance: Insert the following DMAP alter:
compile sedisp
alter 'PARAML UGI'(,-1)
if ( app='nlst' ) then $
equivx ugi/ugj/-1 $
else $
alter 'PARAML UGI'(,2)
endif $
endalter
or do not request intermediate output. If intermediate output is
important, break each subcase into multiple equivalent subcases--one for
each load increment.
@
36930; 2001; Miscellaneous - ECHOON, ECHOOFF - Doesn't Always Work Correctly
ECHOON and ECHOOFF entries may be used in the Executive, Case Control, or
Bulk Data but there are limitations. ECHOOFF must not be the first entry
in the Executive Control section else it will be ignored. ECHOOFF does
not handle continuation entries (for example SET commands) in the Case
Control correctly. If after and ECHOOF command has been placed in the
Bulk Data, PARAM entries are the only entries that follow the ECHOON
command, they will not printed in the unsorted output.

@
36943; 70.7; Aeroelastic Analyisis - GUST - Missing Output for Gust Analysis
 In SOL 146, the gust loads are computed inside the GUST module and then
transformed to generalized coordinates. Only the generalized loads are
output. This makes it impossible for clients to manipulate these loads in
any meaningful way.

Avoidance: There is no simple avoidance.

@
36949; 70.7; DMAP - GPFDR - Extra Output
The GPFDR module with the third parameter xflag=1 (the cross strain
energy output) gives extra output for the first subcase. There should be
no results output for the first subcase.
@
36978; 70.6; Optimization/Aeroelasticity - DMI, W2GJ - Downwash Matrices not
Kept
In SOL 200 with static aeroelasticity, user input downwash matrices (W2GJ)
are not kept after the first cycle.
Avoidance: Apply the following dmap alter in V70 and V70.5:

COMPILE PFAERO
ALTER 'dmiin'
CALL DBSTORE W2GJ,,,,//0/1/'DBALL'/0 $
$
COMPILE FEA
ALTER 'fea' $
TYPE PARM,,I,Y,IFOUND $
ALTER 'pfaero'
IF ( DESCYCLE >1 ) CALL DBFETCH /W2GJ,,,,/0/1/0/0/S,IFOUND $
$
@
36997; 70.7; Optimization/Design Sensitivity - Premature Termination - UIM 64
64
An optimization job starting from a feasible region may terminate for a
hard convergence after the first design cycle without showing any
improvement in either objective or constraint and showing any changes in
the design variables. Here is the termination message:
*** USER INFORMATION MESSAGE 6464 (DOM12E)
RUN TERMINATED DUE TO HARD CONVERGENCE TO AN OPTIMUM AT CYCLE NUMBER
= 1.
Avoidance: None.
@
36998; 70.7; SSSALTER - nlgyroa.v706 - Gyroscopic Effects Ignored
The version 70.6 nlgyroa.v706 sssalter does not include the gyroscopic
effect.
Avoidance: Replace nlgyroa.v706 with the following alter. A copy of the
revised alter is also available on the MSC web site.
$
$ THIS ALTER IS CONFIDENTIAL AND A TRADE SECRET OF THE
$ MACNEAL-SCHWENDLER CORP. THE RECEIPT OR POSSESSION OF
$ THIS ALTER DOES NOT CONVEY ANY RIGHTS TO REPRODUCE OR
$ DISCLOSE ITS CONTENTS, OR TO MANUFACTURE, USE, OR SELL
$ ANYTHING HEREIN, IN WHOLE OR IN PART, WITHOUT THE
$ SPECIFIC WRITTEN CONSENT OF THE MACNEAL-SCHWENDLER CORPORATION.
$
echooff $
$
$ nlgyroa.v706 - updated 8-13-99
$
$ updated July 7, 1998 - tlr
$
$ added PARAM,RFORCE,i - set i to the RFORCE to be used by this alter
$ the default is the first RFORCE in the sorted
$ bulk data
$
$ nlgyroa.v69 - updated for v69 -- 9-25-96
$
$ tested on system built on 3-14-95 VERSION 68.2
$
$************************************************************************
******$
$ ALTER FOR SOL 106 TO INCLUDE CENTRIPETAL TERMS.
$
$ THE ALTER WILL ADD CENTRIPETAL TERMS TO THE TANGENT STIFFNESS MATRIX
$ DURING MODES CALCULATION SPECIFIED BY THE PARAMETER NMLOOP. TO USE
THIS
$ ALTER THE USER MUST SUPPLY A RFORCE FOR THE RESIDUAL STRUCTURE.
$ SUPERELEMENTS SHOULD NOT BE USED, THE CENTRIPETAL EFFECT OF UPSTREAM
$ SUPERELEMENTS WILL NOT BE CALCULATED.
$ THE COORDINATE SYSTEM REFERENCED ON THE RFORCE CARD MUST BE CENTERED
ON
$ THE AXIS OF ROTATION. THE GRID ID ENTRY IS IGNORED FOR THE
CALCULATION
$ OF THE CENTRIPETAL TERMS. THE ROTATION IS ASSUMED TO BE ABOUT THE
$ REFERENCED COORDINATE SYSTEM (CID) ORIGIN, THE MAGNITUDE (A) AND
ROTATION
$ VECTOR COMPONENTS (N1, N2, & N3) DETERMINE THE ORIENTATION.
$
$ THE PARAMETER NMLOOP SHOULD BE SET FOR THE END OF THE FIRST SUBCASE.
$
$ ALTER DEVELOPED BY DAVID BELLA FOR THE MACNEAL-SCHWENDLER CORPORATION,
$ AUGUST, 1992.
$
$************************************************************************
******$
$
compile nlstatic $
alter 'NORMAL MODES WITH NONLINEAR STIFFNESS'(,1)
$message //'testing xxxxxxxxxxxxxxx' $
$
$ CALCULATE CENTRIPETAL TERMS
$
CALL CENTRIP SLT,BGPDTS,EQEXINS,CSTMS,SILS,MGG,
geom3/
KCJJ/ $
$
dbstatus kcjj//s,n,iskcjj $
 if (iskcjj>0)then $
message //' ' $
message //' message from NLGYRO alter - ' $
message //' centripetal stiffness terms calculated' $
message //' adding into stiffness matrix for eigenvalue'/
' solution' $
message //' ' $
$ matprn kcjj// $
else $
message //'dmap warning message' $
message //'no centripetal stiffness terms calculated' $
message //' ' $
endif $
$
CALL SEMR2 kcjj,GOT,USETnl,GM/kcAA,kcFF,kcMG,kcSF/
ERROR/NOMSET/NOSSET/NOOSET/NORC/NOQSET/NOTSET $
matprn kcjj,kcaa// $
$
$ ADD CENTRIPETAL TERMS TO STIFFNESS
$
alter 'call.*gma.*casesx2.*maa'(,-1)
add5 kaa,kcaa,,,/kaanew $
equivx kaanew/kaa/always $
matprn kaa// $
add5 ksf,kcsf,,,/ksfnew $
equivx ksfnew/ksf/always $
add5 kmg,kcmg,,,/kmgnew $
equivx kmgnew/kmg/always $
$
alter 'method=-1'
add5 kllrh,kcaa,,,/klltemp $
equivx klltemp/kllrh/-1 $
$
$
ENDALTER $
$
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
$$$$
$
COMPILE CENTRIP $
$
SUBDMAP CENTRIP SLT,BGPDT,EQEXIN,CSTM,SIL,MJJ,geom3/KCEN/ $
$
$********************************************************************
$ *
$ This DMAP sequence calculates Coriolis and Centripetal effect *
$ matrices in global coordinates. *
$ *
$ For node i: *
$ *
$ [Bcoriolis] = 2[x]m *
$ *
$ [Kcentripetal] = [x][x]m *
$ *
$ | 0 -Omega(z) Omega(y) | *
$ [x] = | Omega(z) 0 -Omega(x) | *
$ | -Omega(y) Omega(x) 0 | *
$ *
$********************************************************************
$
TYPE PARM,,I,N,SLTRECNO,geom3rec,wordno,sid,id $
TYPE PARM,,RS,N,OMEGA $
TYPE PARM,,CS,N,N1C,N2C,N3C $
TYPE PARM,,CS,N,OMEGAC $
type parm,,i,y,rforce=-1
$
$ SEARCH FOR RFORCE ENTRY
$
$tabprt geom3//'geom3'///1 $
$tabprt slt//'slt'///1 $
dbstatus slt//s,n,noslt $
if(noslt<>1)then $
message //'FATAL - FATAL - FATAL' $
message //'slt does not exist in alter - error' $
message //'run terminated by nlgyro alter' $
exit $
endif $
$
if(rforce<0)then $
$
$ take first RFORCE in SLT
$
message //' ' $
message //' PARAM,RFORCE not specified, first RFORCE entry '/
'will be used for nlgyro alter' $
message //' ' $
SLTRECNO= 1
DO WHILE ( SLTRECNO > 0 ) $
$
PARAML SLT//'DTI'/S,N,SLTRECNO/1//S,N,LTYPE/ $
IF ( (SLTRECNO <> -1 ) AND ( LTYPE = 10 ) ) THEN $
$
$ EXTRACT SPIN VECTOR FROM RFORCE CARD.
$
PARAML SLT//'DTI'/SLTRECNO/4//S,N,CID/ $
PARAML SLT//'DTI'/SLTRECNO/5/S,N,RPT/ $
PARAML SLT//'DTI'/SLTRECNO/6/S,N,N1/ $
PARAML SLT//'DTI'/SLTRECNO/7/S,N,N2/ $
PARAML SLT//'DTI'/SLTRECNO/8/S,N,N3/ $
SLTRECNO = -2 $
$
ENDIF $
$
SLTRECNO = SLTRECNO + 1 $
$
ENDDO $
$
else $
$
$ param,rforce specified
$
message //' ' $
message //' param,rforce specified, searching for'/
' rforce number'/rforce $
message //' ' $
$
geom3rec=1 $
do while (geom3rec>0) $
$ message //'top of loop - geom3rec='/geom3rec $
PARAML geom3//'DTI'/s,n,geom3rec/1//S,N,ID/ $
$ message //'id ='/id/' geom3rec ='/geom3rec $
if(id=5509)then $
$ found rforce entries
wordno = 4 $
do while(wordno>0) $ find rforce requested
PARAML geom3//'DTI'/geom3rec/s,n,wordno//s,n,sid $
$ message //'wordno='/wordno/' sid ='/sid $
if(sid=rforce)then $
$ found correct rforce
message //' ' $
message //'found requested rforce entry number'/sid $
message //' ' $
wordno = wordno + 2 $
paraml geom3//'DTI'/geom3rec/s,n,wordno//s,n,cid $
message //'cid ='/cid $
wordno = wordno + 1 $
paraml geom3//'DTI'/geom3rec/s,n,wordno/s,n,rpt $
message //'rpt ='/rpt $
wordno = wordno + 1 $
paraml geom3//'DTI'/geom3rec/s,n,wordno/s,n,n1 $
wordno = wordno + 1 $
paraml geom3//'DTI'/geom3rec/s,n,wordno/s,n,n2 $
wordno = wordno + 1 $
paraml geom3//'DTI'/geom3rec/s,n,wordno/s,n,n3 $
message //'n1,n2,n3 ='/n1/n2/n3 $
wordno = -2 $
else $
if(wordno>0)wordno = wordno + 10 $
endif $
enddo $ wordno>0
endif $ id=5509
if(geom3rec>0)geom3rec=geom3rec + 1 $
enddo $
endif $
IF ( RPT = 0. ) THEN $
$
OMEGA= 0.
RETURN $
$
ENDIF $
$
$ THESE ARE NEEDED LATER ON
$
MATGEN ,/I1X1/1/1/ $
MATGEN ,/I6X6/1/6/ $
$
$ CALCULATE RADIANS PER UNIT TIME
$
OMEGA= 2.0 * PI(1) * RPT $
OMEGAC = CMPLX(OMEGA,0.) $
$
$
$ CREATE A COLUMN VECTOR, MJJTSUM, FROM THE DIAGONAL OF MJJD
$ TO BE USED FOR PARTITIONING PURPOSES LATER
$
DIAGONAL MJJ/MJJC/'COLUMN' $
$
$-------------------------------------------------------------------$
$-------------------------------------------------------------------$
$
$ CALCULATE ROTATION MATRIX IN THE CID SYSTEM
$
IF (N1 <> 0.) THEN
$
$ DETERMINE FACTOR FOR X-ROTATION (in CID system)
$
MATGEN ,/PVC6/6/6/1/1/4 $
MATGEN ,/PVR6/6/6/2/1/3 $
MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $
MPYAD ONE6X6,I6X6,ONE6X6/ROTX/1/-1 $
DELETE /PVC6,PVR6,ONE6X6,,/ $
N1C = CMPLX(N1,0.)
$
ENDIF $
$
$-------------------------------------------------------------------$
$
IF (N2 <> 0.) THEN
$
$ DETERMINE FACTOR FOR Y-ROTATION (in CID system)
$
MATGEN ,/PVC6/6/6/2/1/3 $
MATGEN ,/PVR6/6/6/0/1/5 $
MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $
MPYAD ONE6X6,I6X6,ONE6X6/ROTY/1/-1 $
DELETE /PVC6,PVR6,ONE6X6,,/ $
N2C = CMPLX(N2,0.)
$
ENDIF $
$
$-------------------------------------------------------------------$
$
IF (N3 <> 0.) THEN
$
$ DETERMINE FACTOR FOR Z-ROTATION (in CID system)
$
MATGEN ,/PVC6/6/6/0/1/5 $
MATGEN ,/PVR6/6/6/1/1/4 $
MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $
MPYAD ONE6X6,I6X6,ONE6X6/ROTZ/1/-1 $
DELETE /PVC6,PVR6,ONE6X6,,/ $
N3C = CMPLX(N3,0.)
$
ENDIF $
$
$-------------------------------------------------------------------$
$
$ ADD FOR TOTAL ROTATION (in CID system)
$
ADD5 ROTX,ROTY,ROTZ,,/ROTTOT/N1C/N2C/N3C $
$
$-------------------------------------------------------------------$
$-------------------------------------------------------------------$
$
$ GENERATE G-SIZE CORIOLIS MATRIX IN THE GLOBAL SYSTEM
$
$
$ FIRST, GENERATE G-SIZE MATRIX IN THE CID SYSTEM
$
ADD ROTTOT,/OMEGA/OMEGAC $
MATMOD OMEGA,SIL,,,,/OMEGAGG,/5 $
$
$-------------------------------------------------------------------$
$
$ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS
$ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5)
$
PARTN OMEGAGG,MJJC,/,,,OMEGAXX/ $
$
$-------------------------------------------------------------------$
$
$ TRANSFORM CORIOLIS ACCELERATION MATRIX TO GLOBAL SYSTEM
$ FIRST, TRANSFORM FROM CID TO BASIC
$ THEN , TRANSFORM FROM BASIC TO GLOBAL
$
$ CHECK IF THE CSTM EXISTS. IT WILL NOT EXIT IF THE GLOBAL SYSTEM
$ IS THE SAME AS THE BASIC SYSTEM. ERROR IN MATMOD. THEREFORE
$ GENERATE THE IDENTITY MATRIX
$
$
PARAML CSTM//'PRESENCE'////S,N,NOGTOB=0 $
IF (NOGTOB < 0) THEN
DIAGONAL MJJ/GTOBY/'SQUARE'/0. $
ELSE
MATMOD CSTM,SIL,BGPDT,,,/GTOBY,/5//-1 $
ENDIF $
$
PARTN GTOBY,MJJC,/,,,GTOB/ $
$
$ CHECK IF THE CID IS EQUAL TO ZERO (BASIC)
$
IF (CID = 0) THEN
EQUIVX GTOB/CTOG/ALWAYS $
ELSE
MATMOD CSTM,SIL,BGPDT,,,/CTOBX,/5//CID $
$
$ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS
$ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5)
$
PARTN CTOBX,MJJC,/,,,CTOB/ $
MPYAD CTOB,GTOB,/CTOG/1 $
ENDIF $
$
SMPYAD CTOG,OMEGAXX,CTOG,,,/OMEGAYY/3////1 $
MERGE ,,,,OMEGAYY,MJJC,/OMEGAGGX/ $
$
$-------------------------------------------------------------------$
$-------------------------------------------------------------------$
$
$ CALCULATE CENTRIPETAL ACCELERATION MATRIX
$
$ MULTIPLY MASS BY CENTRIPETAL ACCELERATION
$ TO CALCULATE CETRIPETAL FORCE MATRIX
$
SMPYAD OMEGAGGX,MJJ,OMEGAGGX,,,/KCEN/3////////6 $
$
RETURN $
END $
$
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
 $
echooff $
@
37010; ; Optimization/Design Sensitivity - DRESP1 - SFM 4276, EC 31
When an optimization deck includes a DRESP2 entry which references the
same DRESP1 ID more than once, the following system fatal message is
issued from module DOM9:
BIOMSG: ERROR 31 HAS OCCURRED IN ROUTINE GETSTR , FILE INDEX (FILEX)
= 0
STATUS = 0
followed by:
0*** SYSTEM FATAL ERROR 4276.GETSTR ERROR CODE 31 PID= 0
Avoidance: Do not duplicate the same DRESP1 ID on a DRESP2 entry.
@
37027; 2001; Aeroelastic Analysis - RESTART - SFM 3001 for QHHA or Inefficien
t Restart
A read-only restart in aeroelastic solution 145 will unnecessarily
recompute and decompose the AJJT for a list of existing Mach numbers and
reduced frequency even though the list is a subset of the original list.
Another symptom is SFM 3001 for QHHA.
Avoidance: If the original MKAERO1 entry is re-entered as it was input
on the initial run, then the AMG and DECOMP modules in the PFAERO subDMAP
will not be re-run.

@
37031; 70.7; Aeroelastic Analysis - AESYMXZ, AESYMXY, AECONFIG, IFP - UFM 603
Case control commands are free field input and should allow users to add
comments at the end of the lines so long as a dollar sign precedes the
comment. The new Version 70.6 aeroelastic analysis case control commands
do not allow comments to be specified on the same line as the command. The
simple input file below fails in IFP due to the comments onthe AE* case
control commands.
sol 144
cend
aesymxz=symmetric $ this is a comment on the aesymxy line
aesymxy=asymmetric $ this is a comment on the aesymxy line
aeconfig=csr $ this is a comment on the aeconfig line
spc = 1 $ this is a comment on the SPC line
begin bulk
param,checkout,yes
enddata
@
37032; 2001; Eigenvalue Analysis - PARAM,RESVEC - Possible Wrong Answers
When residual vector is requested, it has been observed on some models
that the additional augmented eigenvalue is lower than the highest
calculated mode. This has been observed on a very small problem and also
on problems using virtual mass.
 Avoidance: Insert the following DMAP alter:
compile resvec
alter 'matmod oxf',''
type parm,,rs,y,resvoxf=.001 $
matmod oxf,,,,,/ox5,/2////resvoxf $ filter again
alter 'append=',''
append=(noox5<0 or ox5r<10000) $ set flag for below
endalter
or inspect the augmented eigenvalue table, if the additional mode is lower
than the highest calculated frequency, use one of the following
avoidances:
1. Specify F2 on the EIGRL entry.
2. Specify a smaller maxset.
3. If it is a small problem, use AGIV or AHOU
@
37042; 70.7; Matrix Methods - MPYAD - SFM 4276 (GINO)
If working memory is small relative to problem size, data is stored past
the end of working memory, destroying executive tables. The result will
ultimately be a confusing error message, such as:
*** SYSTEM FATAL MESSAGE 4276 (GINO)
ERROR CODE 100 PID= 0
*** USER INFORMATION MESSAGE 4276 (GINO)
TO OBTAIN A NASTRAN DUMP RESUBMIT JOB WITH DIAG 44 INSERTED IN THE
EXECUTIVE CONTROL SECTION.
1 * * * END OF JOB * * *
*** SYSTEM FATAL MESSAGE 1123 (GDBSNO)
DATABASE NUMBER: -10855138 IS OUT OF RANGE. DBSPTR = -1085513983
USER ACTION : MAKE SURE ALL ATTACHED DATABASE(S), INCLUDING THE
DELIVERY DATABASE, ARE VERSION (e.g. 69) COMPATIBLE.
0FATAL ERROR
Avoidance: Increase memory.
@
37043; 70.7; Rigid Elements - RBE3, GP4 - Poor Efficiency
When there are more than 50,000 rbe3 elements run times in the GP4 module,
where they are processed, may be very long.

Avoidance: Avoidances are to reduce the number of rbe3 elements, or put

them in several superelements. Each superelement will call GP4 for only
the elements within it.
@
37051; 70.7; Aeroelastic Analysis - TRIM, ASG - SFM 4276, EC 100
The message "SYSTEM FATAL ERROR 4276 ERROR CODE 100" is issued from the
ASG module in an aeroelastic trim analysis when there are no SUPORTed
degrees of freedom but there are aerodynamic extra points (AESTAT and/or
AESURF entries) that are not constrained on the TRIM entry.
Avoidance: Constrain or remove the aerodynamic extra points or add to the
SUPORT degrees of freedom.
@
37055; 70.7; Eigenvalue Analysis - Component Modes - Excessive resource usage
 If component modes are requested and the number of q-set points is not
exactly equal to the number of component modes then the reduced stiffness
and mass matrices and the resulting residual structure matrices will
contain many small off-diagonal terms which should be exactly zero. This
condition contributes to inefficiencies in subsequent matrix operations.
Avoidance: Insert the following DMAP alter:
compile semr3
alter 'dotrip\=',''
alter 'dotrip\='(2),''
alter 'nullkqq'
nullkqq=0 $
alter 'cmmat'
if ( noqset1>noqset ) then $
partn cmmat,,vq1qx/xxx,,,/1 $
equivx xxx/cmmat/-1 $
else if ( noqset1<noqset ) then $
merge cmmat,,,,,vq1qx/xxx/1 $
equivx xxx/cmmat/-1 $
endif $
endalter
@
37059; 70.7; Miscellaneous - NEC SX-5, HPIO - Job Hangs
It has been reported that when HPIO is turned on the NEC SX-5 and the disk
space is full, then nastran will not terminate properly and the job will
hang.
Avoidance: Kill the job manually.
@
37078; 2001; Random Response - CBAR, PLOAD1 - SWM 3002, SFM 4276, missing PSD
output
If PSD element force or stress output for a BAR element is requested in a
random analysis, and the loading involves PLOAD1 entries, then the
requested random output will be missing. Depending on the machine, the
job will issue one of the following messages:
SWM 3002 "EOF ENCOUNTERED WHILE READING DATA BLOCK OES2 (FILE 107) IN
SUBROUTINE RAND2"
SFM 4276 in the SDRX module.

Avoidance: There are two possible avoidances:

!. Replace the BAR elements with BEAM elements.
2. If the BAR element must be used, replace PLOAD1 with equivalent point
loads.
@
37080; 2001; Optimization/Desgin Sensitivity - PBAR - Wrong Answers
When an optimization job includes CBAR elements which are loaded by PLOAD1
entries, the end-B maximum stress, end-B minimum stress and end-B Safety
margin in compression for the static case are incorrectly selected from
the stress data file.
Avoidance: Use item code + 1 for selecting end-B maximum stress and end-B
 minimum stress. However, there is no avoidance for getting the correct
end-B safety margin.
@
37088; 70.7.2; External Superelements - param,extdrout,dmigop2 - Wrong answers
When using the enhanced external superelement capability, if the boundary
points have a different order in the assembly model than in the component
model, then the boundary solution written out for PARAM,EXTDROUT,DMIGOP2
will be in the incorrect order, resulting in incorrect results for data
recovery on the component.
Avoidance: Insert the following dmap alter.
compile super3
alter 'call.*extdrout','call.*extdrout'(3,)
CALL EXTDROUT ULS,OL1M,EQEXINS,MAPS,uset/
DUM/APP1/' '/TRUE $
CALL EXTDROUT UH2,OL1R,EQEXINS,MAPS,uset/
DUM/APP/SOLTYP/TRUE $
ELSE $
CALL EXTDROUT ULS,OL1M,EQEXINS,MAPS,uset/
DUM/APP1/SOLTYP/TRUE $
alter 'call.*extdrout'(4,),'call.*extdrout'(6,)
CALL EXTDROUT UG,OL1M,EQEXINS,MAPS,uset/
DUM/APP1/' '/TRUE $
CALL EXTDROUT UH2,OL1R,EQEXINS,MAPS,uset/
DUM/APP/SOLTYP/TRUE $
ELSE $
CALL EXTDROUT UG,OL1M,EQEXINS,MAPS,uset/
DUM/APP/SOLTYP/TRUE $
compile extdrout, list
alter 1,1
SUBDMAP EXTDROUT UG,OL,EQEXINS,MAPS,uset/
DUM/APP/SOLTYP/EXTER $
alter 'matmod',''
paraml uset//'uset'//////'g'/s,n,lgset/'a'/s,n,noaset $
matgen eqex/intext/9/0/lgset $
$if(lgset<>noaset)then $
$ upartn uset,intext/intexta,,,/'G'/'A'/'O' $
$else $
equivx intext/intexta/-1 $
$endif $
copy eqex/eqexinxxx $
copy x_ul/x_ulc $
call ted intexta,x_ulc,eqexinxxx/dummy $
$
alter 76,76
OUTPUT2 dummy,X_OL//0/EXTDRUNT///'MATPOOL'/XNAME2 $
tabprt dummy//'matpool'///1 $
tabprt x_ol//xname2///1 $
$
compile ted
subdmap ted intexta,x_ul,eqex/dumout
mpyad intexta,x_ul,/extdb/1 $
matmod extdb,eqex,,,,/EXTDBx,/16/0 $ X_MATUL
copy extdbx/dumout $
 return
end
$
@
37089; GL; Aeroelastic Analysis/Superelements - SPLINEi - AECOMP Datablock
Does not Exist
If external superelements are used in an aeroelastic solution and all
splines are contained within the external superelements, the run will fail
claiming that datablock AECOMP does not exist.
@
37102; 2001; Superelements - Buckling, SPCD - Wrong Answers
If buckling (SOL 105) is run with superelements and there are enforced
displacement on the upstream superelements, then the static answers will
be incorrect and hence the eigenvalues will be wrong.
Avoidance: Insert the following DMAP alter:
compile phase0
alter 'ISOLAPP\=NSOLAPP',''
noop() $
alter 'ISOLAPP\=ISOLAPP'(,-1)
IF ( SEID>0 ) ISOLAPP=NSOLAPP $
endalter
or place all enforced displacements on the boundary into the residual
superelement.
@
37106; 70.7; P-Elements - TEMPP1 - QUADp Gives Wrong Answers for TEMPP1 Loads
The QUAD p-element gives wrong results if it is loaded with temperature
gradients (TEMPP1 entry).
Avoidance: There is no avoidance.

@
37126; 2001; Documentation - XYPLOT - LONG Command Default
The MSC/NASTRAN V70.5 Quick Reference Guide case control description in
the XYPLOT section for the LONG command says that the default is YES.
This is not correct. The default is NO.
@
37128; 70.7.2; External Superelements - PARAM,EXTDROUT,DMIGOP2 - SFM 3007 (UPAR
TN)
If the original component model for an external superelement has one or
more superelements defined, and the residual structure is equal in size to
the ASET chosen for the reduced matrices, then the data recovery run on
the component will fail with the following messages if the
PARAM,EXTDROUT,DMIGOP2 option is used.
*** SYSTEM FATAL MESSAGE 3007 (UPARTN)
ILLEGAL INPUT TO SUBROUTINE CALCV
Avoidance: Insert the following alters.
 compile extdrin
alter 'upartn.*uset.*ul',''
paraml uset//'uset'//////'g'/s,n,lgset/'a'/s,n,noaset $
if(lgset<>noaset)then $
UPARTN USET,EXTDB/UL,,,/'G'/'A'/'O'/1 $
else $
equivx extdb/ul/-1 $
endif $
@
37130; 2001; Optimization/Design Sensitivity - ANALYSIS=FLUTTER, SE - SFM 30
01
In SOL 200, if ANALYSIS=FLUTTER and superelements exist, the analysis
results will be wrong if the aerodynamic model is splined to an upstream
superelement. Further,the job may fail with SFM 3001 in DVSG1 indicating
that datablock DUG does not exist. See also CSR 38501
Avoidance: Run the model without using superelements, or use this alter
for V70.7.

compile respsen
alter 'dsflte','dsflte'
DSFLTE KHHx,BHHx,MHHx,QHHL,FLUTAB,r1tabr,casem/
CPHFL,CPHFR,VTQU $
alter 'dug0','vtqu0'(2)
IF ( NODUD>0 ) CALL xSEDRDR,
DUD,EMAP,CASEXX,SLT,ETT,MAPS,bgpdts,USET,
GOAT,GOAQ,GM,CASEs,casem/
DUG/
RSONLY/'DYNAMICS'/false $
IF ( NODVD>0 ) CALL xSEDRDR,
DVD,EMAP,CASEXX,SLT,ETT,MAPS,bgpdts,USET,
GOAT,GOAQ,GM,CASEs,casem/
DVG/
RSONLY/'DYNAMICS'/false $
alter 'dugx'(3),'vtqux'(3)
alter 'doflutd'(6)
IF ( NOT(IMAGE) AND NOPSLG>-1 ) THEN $
alter 'cdelb1'(5)
endif $ not(image) and nopslg>-1
alter 'dsfltf','dsfltf'
if ( seid = 0 ) then $
dsfltf vtqu,cdelk,cdelb,cdelm/delfl $
else $
dsfltf vtqu,cdelk1,cdelb1,cdelm1/delfl $
endif $
compile phase1a
alter 'ISOLKNT=NSOLAPP+1','ISOLKNT=NSOLAPP+1'
if (not desopt) isolknt=nsolapp+1 $

@
37132; 70.7; SSSALTERs - sedampa doesn't supposrt r-set and c-set
If R- or C-set dof exist in a superelement and the sssalter sedampa is
used, the associated superelement will not have modal damping applied. The
alter is intended to terminate the run with a message stating that C- and
 R-set dof are not supported by the alter.
Avoidance: Do not use R- or C-set dof with the alter. The alter has been
replaced by a standard feature, which does allow R- or C-set dof to exist
in Version 70.7.
To correct the alter, replace
paraml uset//'uset'////s,n,nousetxx//'R'/s,n,norset/
'C'/s,n,nocset $
with
paraml uset0//'uset'////s,n,nousetxx//'R'/s,n,norset/
'C'/s,n,nocset $

@
37146; 70.7; Aeroelastic Analysis - RESTART, GUST - SFM 5423, Incompatible Ma
trices
If an aerelastic gust analysis attempts to change the number of retained
modes in a restart, then "SYSTEM FATAL MESSAGE 5423" is issued indicating
an "ATTEMPT TO ADD INCOMPATIBLE MATRICES (SADD5)"
Avoidance: Perform a cold start.
@
37151; 2001; Documentation/QRG - Plot Command AXES is miss typed as AXIS
In case control section of Quick Reference Guide the Plot command AXES
page header and format spell it as AXIS which is incorrect.

@
37152; 70.7; Optimization/Design Sensitivity - MODTRK - SFM 3001
"SYSTEM FATAL MESSAGE 3001" occurs in the MODTRK module when all three of
the following are present:
1. shape design variables,
2. at least one subcase with mode tracking
3. and at least one subcase with ANALYSIS=STATICS
Avoidance: Place the MODTRAK request above the subcase level
@
37169; 2001; Documentation/QRG - ADAPT - Typographical Error
The Quick Reference Guide description of the ADAPT entry contains the
typographical error:
TY=PEUNIP, SIGT=OL22
it should be
TYPE=UNIP, SIGTOL=22

@
37173; 2001; Optimization/Design Sensitivity - PARAM,SOFTEXIT - Incomplete De
sign History
 In SOL 200, if PARAM SOFTEXIT YES is used and soft convergence is
achieved, the DESIGN CYCLE HISTORY does not include the results
from the final exact analysis.
Avoidance: The information that should be in the DESIGN CYCLE HISTORY can
be obtained using the P1 and P2 parameters on the
DOPTPRM entry.
@
37178; 70.7; Miscellaneous - Missing UFM 315 - Unexpected Answers
If a negative sign is the only character in a field of a bulk data entry
it may be ignored. This may lead to unexpected answers. For example if
the negative sign was intended for the value in the following field as in
the following entry:
CBAR 70101 51001 464 1322 0. 0. 1.0
++000363
++000363 - .238 - .238
+000364
Avoidance: Specify "ECHO=PUNCH" in the Case Control and perform a search
(grep) on negative signs followed by blanks ("- ").

@
37190; 70.7; Optimization/Design Sensitivity - PBARL - Wrong Answers
Use of PBARL with "T" section in an optimization task may cause an
infeasible result with Version 70.5 or earlier release of MSC.Nastran.
Avoidance:
1. Use equivalent PBAR with DVPREL2 equations defining the PBAR
properties.
2. Use a new version such as v70.7.
@
37195; 70.7; Optimization/Design Sensitivities - Rigid Elements - Incorrect S
ensitivities
When a shape optimization job has locations of rigid elements being
designed it may result in incorrect sensitivities or may have difficulty
converging. This may lead to excessive time in the DSVGP5 module.

Avoidance: There is no avoidance. Version 70.7 corrects this error and

shows a dramatic improvement in performance of the DSVGP5 module for some
problems.
@
37222; ; Optimization/Design Sensitivity - RESTART - Inefficient Processi
ng
Restarts into SOL 200 will not be efficient if the design model is not
included in the cold start run.
Avoidance: 1. Include the design model in the cold start run
or
2. Insert the following alter in the cold start run:
compile prtsum
alter 1
 DBEQUIV WHERE (DESITER=0 and WILDCARD) ,
CONVERT (DESITER=1), OVRWRT $
endalter
@
37224; 2001; Optimization/Design Sensitivity - PCOMP, DSAD - SFM 4276, Poor M
essage
In solution 200, when using the cfailure criteria force item code 7
(interlaminar stress), NASTRAN will either fail in the DSAD module with
system fatal 4276 or give
wrong answers if the code is specified for the last lamina (ply). This
fatal occurs because there are no interlaminar stresses for the last
lamina since interlaminar stresses are between laminae. Also, using the
cfailure criteria as force item code 8 (max of FP and FB) will fail unless
the lamina (ply) number specified on the DRESP1 is the final one.
@
37227; 70.7.2; Nonlinear Analysis - Creep Law May Cause Overflow - SFM 4276 in
nmcrd
Some forms of creep law may cause an overflow condition in nonlinear creep
analysis. The error causes an overflow in the nmcrd subroutine which
will be identified in the log file.
Avoidance: A different creep law may avoid the error.
@
37247; ; Transient Analysis - TEMP(LOAD) - Wrong Answer
In transient analysis, thermal strains are not subtracted from the total
strains before computing the state of stress. This leads to wrong
answers.
Avoidance: This is a limitation, there is no avoidance.
@
37253; 2001; Cyclic Symmetry - Virtual Mass, MFLUID - Wrong Answers
Virtual mass will cause wrong results when used with cyclic symmetry if a
fluid segment is modeled. This path should be blocked with a fatal error.
Fully enclosed fluids in each segment will give correct answers.
Avoidance: Use acoustic fluid elements instead.

@
37263; ; Virtual Mass - MFLUID - PARAM,VMOPT,1 Does Not Function Properly
When param, vmopt, 1 is present, the virtual mass should be added to the
structural mass at the g-set, that is, before any constraints are applied.
In Version 70.6 and 70.7, setting this parameter has no effect and the
solution method adds virtual mass after all constraints are applied. This
may or may not affect the results
Avoidance: The alter below will make vmopt, 1 take effect.
compile semg list noref $
Alter 'if ( vmopt=1 and','' $ always do what user requests
IF ( VMOPT=1 ) THEN $
Compile semrm list noref $
 Alter 'if ( vmopt=0 or','' $
IF ( VMOPT=0 ) ) THEN $

Note that the original dmap would over-rule the user's request if it did
not seem to "make sense". The alter always honors the vmopt request.
Note that if vmopt=1 is used on medium- and large-size models, run times
and computer resource requirement may be very large. sssalter vma.f707
is scheduled to be available on 15 Oct 99. It allows a more accurate
answer when vmopt=0 (the default value) is used.
@
37264; 70.7.2; Performance - MPYAD - MPYAD Method 1, Storage 2 not selected
Under certain condition, the MPYAD Method 1, Storage 2 method, which is
optimized for dense matrices, will not be correctly selected. This happens
when all of the B and D matrices just fit into working memory.
Avoidance: The work around is to either increase or decrease memory.
@
37272; ; Elements - CHEXA - Stress Output at Gauss Points does not work w
ith IN=2
Element Stresses at the Gauss integration points cannot be produced for
the 8-point CHEXA element when the integration option, IN=2 is used.
Instead, the stresses at the corner points are generated. The description
in the QRG for the PSOLID bulk data entry states that if the stress flag
is "1" or "GAUSS" the integration points will be the location of the
stress output, this is incorrect.
Avoidance: There is no avoidance for this integration option.
@
37276; ; Nonlinear Transient Heat Transfer - TEMPC, TRAN - Fails to Conve
rge
If an nonlinear transient heat transfer analysis has enforced temperature
boundary conditions, that is temperature boundary condition that are a
function of time, MSC/NASTRAN uses the large conductance approach to
enforce the temperature boundary condition. This is achieved by using a
CELAS2 entry with stiffness equal to 1.0E10, and an SLOAD entry on the
node with a temperature of 1.0E10. Thus, the temperature is equal to u =
F/K or SLOAD divide by the K value. In some case this hard coded value
for the spring stiffness may be to large and it may adversely affect
convergence.
Avoidance:
1. Replace the TEMPBC,TRAN with SLOAD and a CELAS2 entry, and set the
stiffness smaller.
2. Use convergence on P only on the iterations where convergence
difficulties are encountered.
@
37284; 70.7; Utilities - MSGCMP.EXE, Windows NT - Missing Utility
The msgcmp.exe executable is missing from the Version 70.5 NT release of
MSC/Nastran.
Avoidance: Download msgcmp.exe from the MSC ftp site or contact your
local support person. The file is located in the /pub/outing directory.
Since this is an executable, use the BIN option for downloading.
@
37290; 2001; Documentation/QRG - PARAM,POST - Patran may use the xdb
The Quick Reference Guide documentation of parameter POST,0 does not
mention the possibility to use the xdb database with MSC.Patran. This
should be added and it should be clarified that the parameters MODEL and
SOLID are not supported by MSC.Patran.
@
37291; 70.7; Optimization/Design Sensitivity - GE - Wrong Sensitivities
In a dynamic optimization job, if the design variables include damping
coefficients and the adjoint method path is invoked, then the dynamic
response sensitivity coefficients with respect to all design variables
will be incorrect. This error has been corrected in Version 70.7.

Avoidance: Turn off the adjoint method using the following dmap alter
compile descon
alter 'dsad'(1,-1)
pexist = true
alter 'dsad'
pexist = false
@
37295; 2001; Documentation/QRG - MATHP - Poor Convergence
The following sentence need to be added to the Quick Reference Guide
description of the TAB1 parameter of the MATHP entry:
Stresses are negative for compression and positive for tension.
If this convention is not followed the job may fail to converge or issue a
traceback.
@
37307; ; Nonlinear Analysis - TEMP(MAT), HEXA - Wrong Equivalent Stress
The "EQUIVALENT STRESS" for the nonlinear HEXA element is wrong if there
are temperature dependent material properties.
Avoidance: There is not avoidance.
@
37309; 70.7; DMAP - APPEND - UFM 4276 EC 1039
The APPEND module may fail and issue the message "USER FATAL MESSAGE
4276.OPENDV EC=1039"
Avoidance: There is no avoidance. The error is fixed in Version 70.7.

@
37319; 70.7.2; Optimization/Design Sensitivity - DESGLB - Unexpected Answers
DESGLB is intended for the selection of design constraints at the global
level (e.g., weight, volume). If DESGLB is used above the subcase level
with the intent to apply the constraints to all the subcases, the
constraints will be ignored for the second and subsequent subcases.
Avoidance: Use of the DESSUB above the first subcase applies it to all
subsequent subcases until a new DESSUB request is entered. DESGLB is
 reserved for subcase independent response only.
@
37335; 70.7.2; Restart - NMLOOP, LOOPID - Inefficient Restart, Modes Recalculat
ed
In a restart from SOL 106 to SOL 111, the stiffened modes are being
recalculated unnecessarily.
Avoidance: Insert the following dmap avoidance.
compile semfreq $
alter 'call modefsrs'(,-1) $
equivx lama/lamas/-1 $
equivx pha/phsa/-1 $
if ( false ) then $
alter 'call modefsrs' $
endif $
endalter
@
37336; ; SSSALTERs - SEGYRO, CYGYR0 Alters - Wrong Results with Parameter
DEPEND=-1
The SEGYRO alter puts the gyroscopic terms in the damping matrix and not
in the mass matrix for param, depend,-1 and this leads to wrong answers.
The assumption eigenvalue lambda = i*Omega is not valid in a rotating
reference frame.
The Corillis forces are equal to the mass times the cross product of the
rotation vector times the velocity vector in the ROTATING SYSTEM. ie.
f=m(omega X V ). The assumption that lambda is proportional to i*omega
is only good for a limited set of response cases where the applied load
has the same frequency as the rotations. An example would be an unbalanced
turbine. The problem would diverge at the critical rotational frequency.
@
37341; 2001; Static Analysis - PARAM,CDITER and INREL - Wrong Answers
If PARAM,INREL,-1 is specified along with PARAM,CDITER,n then no message
is issued stating that PARAM,INREL will be ignored The Version 70.5
Release Guide and QRG indicate that PARAM,CDITER and INREL are mutually
exclusive. A fatal message will be issued in Version 2001.
@
37344; 2001.0.1; Acoustic Analysis - ACSRCE, DAREA - Wrong Answers
When using an ACSRCE power input during an acoustic frequency response
analysis, if the scale factor on the DAREA is increased, then QDOT is
erroneously factored instead of power.
Avoidance: Scale the values on the TABLEDi instead. Also, the Quick
Reference Guide needs to document the equation used: P(f)=A*C(f)
@
37356; 70.7.2; Nonlinear Transient Heat Transfer - DLOAD - SFM3002 (TAH1)
If transient thermal analysis loads (QVOL, QHBDY) are combined with
temperature dependent material properties (MATT4), then the analysis will
fail with SYSTEM FATAL MESSAGE 3002 in the TAHT module.
 Avoidance: There is no avoidance.
@
37382; ; Nonlinear Analyis - CBUSH1D, PARAM,LGDISP - Wrong Answers
Small initial values of stiffness for the CBUSH1D element, i.e. K=1, may
lead to wrong results when PARAM, LGDISP,1 is specified. The problem
occurs because the compression displacement is not limited to the length
of the element.
Avoidance: Specify a larger value for the stiffness.
@
37392; 70.7.2; Acoustic Analysis - CELAS1, CAABSF - UFM 5423
If frequency dependent elements are present in both the fluid (CAABSF) and
the structure (CELAS1) then the run will fail with UFM 5423.
Avoidance: Insert the following DMAP alter:
compile freqrs
alter 'if ( fs )'(2),''
if ( false ) then $
alter 'if ( fs )'(3),''
if ( false ) then $
alter 'if ( fs )'(4),''
if ( false ) then $
alter 'if ( fs )'(5),''
if ( false ) then $
endalter

@
37394; ; Eigenvalue Analysis - EIGRL, EIGR - Incorrect Termination Method
The Lanczos method for Real Eigenvalue analysis may not find all desired
roots when ND is specified on the EIGRL bulk data entry. Lanczos may
report the reason for termination as "REQUIRED NUMBER OF EIGENVALUES FOUND
IN RANGE" even though the number of desired eigenvalues is not output.
Avoidance: Specify a frequency range on the eigrl entry.
@
37399; 70.7.2; DMAP - OUTPTR Module does not Process STRAIN Requests
The OUTPRT module introduced in Version 70.7 does not process STRAIN
requests. As a results the partitioning vector that is generated will not
include these points unless they are requested with other output requests.
Avoidance: An avoidance is to also request STRESS for the same output
set.
At present, this module is used only in sssalter spdra.v707.
@
37400; 70.7.2; P-Elements - RFORCE - Wrong Answers
Wrong answers may be obtained when RFORCE is applied to p-elements where
p>2.
Avoidance: The only avoidance is to use the origin of the basic
coordinate system to locate the rotation vector.
@
37402; 70.7.2; Optimization/Design Sensitivity - BNDGRID - UFM 6499(DOM9), UFM
9050(SEKRRS)
When the C field on a BNDGRID entry in a shape optimization task is left
blank or takes the value of zero, this input error is not detected.
Therefore, the job may fail in one of the following ways:
1. UFM 6499 (DOM9) due to a zero gradient of a violated constraint
2. UFM 9050 (SEKRRS) due to excessive pivot rations in Matrix KLL.
3. Wrong Answers
Avoidance: Specify a non-zero component on the C field.
@
37407; ; Superelements - SEP1X, DTI, SETREE - SFM 4276 EC 3001
If external superelements are used in a multi-level model containing PART
superelements (BEGIN SUPER), and the multi-level tree is defined using a
DTI,SETREE entry, the run may fail in the SEP1X module with SYSTEM FATAL
MESSAGE 4276. The fatal message will indicate that "ERROR 3001 HAS
OCCURRED IN GREAD".
Avoidance: Replace the DTI,SETREE entry with an SETREE entries.
@
37411; 70.7.2; Parallel Processing - DMP - FATAL, Infinite CPU Loop
If MSC.Nastran is run with DMP=n (n > 1) on an NEC system, and there is a
FATAL message, then the processors which do not get the FATAL hang. They
accumulate CPU time, but never finish.
Avoidances: Avoidances are given below; however, this and several other
problems related to DMP have been corrected in Version 70.7.2. DMP users
should install this later version.
1. Modify the $MSC_BASE/bin/nast707.lcl file:
insert after line 294:
%MSC_ARCH=superux:% errorcode=$?
insert at the end of the file:
%MSC_ARCH=superux:% exit $errorcode

2. Or run without DMP being specified.

@
37437; 70.7.2; Frequency Response - dmp>1 and multiple FREQ subcases - Infinite
loop
A job with multiple frequency subcases submitted with Distributed Memory
Parallel, dmp>1, in solution 111 will hang.
Avoidance: Apply the following dmap alter:
compile freqrs
alter 'ckdd,cbdd'(,-1) $
putsys(1,287) $
alter 'ckddf,cbddf'(,-1) $
putsys(1,287) $
compile semfreq
alter 'fs and'(,2),''
jump endloop $
alter 'if ( loopflag'(,-2)
 label endloop $
endalter
@
37454; 70.7.2; Parallel Processing - DMP, ITER=YES, SPCF - SFM
If distributed memory parallel processing is specified with the iterative
solver and spcforces are requested, the the job will fail.
Avoidance: Either do not request spcforces, distributed processing, or
the iterative solver.
@
37457; 70.7; Optimization/Design Sensitivity - PBARL, Smax, Smin - Possible W
rong Answers
In an optimization run, if PBARL is used and only Smax or Smin is
specified instead of the four stress data recovery points, then wrong
answers may occur.
Avoidance: There are two possible avoidances:
1. Specify all four stress data recovery points instead of using Smax or
Smin.

or
2. Use PBAR instead of PBARL.
@
37458; 2001.0.1; DBC - DBC, PARAM,POST - Job Crashes during BLKWRT
A job with thousands of subcases may fail in DBC during a BLKWRT. The
following message was encountered:
MSC.XL DATA BASE CONVERSION STARTED
PROJECT 0 0 0 0 0 0
0 0 0 0
PATHQUAL 0 0 0 0 0 0
0 0 0 0
PATHLINK 0 0 0 0 0 0
0 0 0 0
PATHINT 0 0 0 0 0 0
0 0 0 0
POST-PROCESSING DATA RECOVERY INITIATED
SUBCASES 1 0 1 0 0 1
0 0 0 0
ILLEGAL DATA BASE NUMBER PASSED TO BLKWRT
0FATAL ERROR
Avoidance: There is no avoidance.
@
37462; 70.7.2; Parallel Processing - DMP>1, CBAR, CBEAM, SEQP - SFM 4276
If a statics job is submitted with Distributed Parallel Processing, dmp>1,
and the job contains BAR or BEAM elements that specify disjoint GRID
points for the orientation vector then the job may fail in the SEQP module
or hang.
Avoidance: There are several avoidances:
 1. Do not use distributed parallel processing.
2. Modify the CBAR, CBEAM entries to explicityly define a coordinate
direction.
3. Modify the CBAR, CBEAM entry to reference an attached GRID point.
4. User Version 70.7.2 where this error has been corrected.
@
37483; 70.7; Nonlinear Analysis - SFM 3007 (NLCOMB)/ILLEGAL INPUT TO SUBROUTI
NE NLCXGR
MSC.Nastran may fail with SYSTEM FATAL MESSAGE 3007 (NLCOMB), ILLEGAL
INPUT TO SUBROUTINE NLCXGR on Sun/Solaris systems when run with MEM =>
2Gb.
Avoidance: Either:
1. Reduce memory below 2Gb
2. Run with version 70.7 or higher.
@
37485; 70.7.2; Optimization/Design Sensitivity - Incorrect Sensitivities for St
atic Aero
Incorrect sensitivities and subsequent poor optimization performance
occurs on long word machines, such as the CRAY, for static aeroelastic
sensitivity analysis when there are more design variables than there are
AESTAT and AESURF entries.
Avoidance: Run the job on a short word machine, reduce the number of
design variables, or increase the number of AESTAT/AESURF entries to
satisfy the above inequality.
@
37486; 70.7.2; SSSALTERs - GPSP, PARAM,AUTOSPC - be-set dofs may be autospcd
The be-set is a new set introduced in Version 70.7. It is not used in any
solution sequences at present. Its only use is in the rflaga.v707 alter
package.
The be-set should not be autospcd, even when singular. This is analogous
to the suport entry. It sometimes is autospcd, although the circumstances
when this happens are not well understood yet.
Avoidance: An avoidance is to always used an spcoff entry for every
be-set point. It may not be necessary, but it should not cause problems
either. Example
uset, be, 17, 3
spcoff, 17 3
@
37487; 70.7.2; Miscellaneous - USET - UFM 2104 (GP4), Poor Message
The be-set is new in Version 70.7. When a degree-of-freedom is placed in
a mutually exclusive set, such as the s-set, and the be-set, the following
truncated message appears:
*** USER FATAL MESSAGE 2101 (GP4)
SCALAR POINT 1 ILLEGALLY DEFINED IN SETS US

The phrase "UBE" should also appear.

If you see a message of this type look for be-set entries on the same dof
. These dofs are entered on USET, BE-type inputs.
@
37495; 70,7; Documentation - Hoffman's Theory incorrect in Composites Seminar
Notes (NAS113)
The Equation for Hoffman's Failure Theory in Chapter 5 of the NAS113
(Analysis of Composites Materials) semninar notes is incorrect. There is a
sign error for one term. The term:
(SIG1*SIG2) / (Xt*Xc)
should be preceded by a negative sign in the equation.

@
37496; 2001; Frequency Response - CELAS2 , STRESS - Wrong Answers
In a direct frequency response analysis, the stress output for the CELAS2
element is incorrect.
Avoidance: Use the CBUSH element instead,

@
37503; 2001.0.1; Acoustic Analysis - Misleading UWM 5462 (IFS5P)
The presence of acoustic elements always leads to the following
misleading warning message:
*** USER WARNING MESSAGE 5462 (IFS5P)
OPTIONS IN PSOLID ID = 800000 FOR FLUID ELEMENT ARE NOT ALLOWED.
DEFAULT VALUES WILL BE TAKEN.
Avoidance: Ignore the warning message.
@
37508; ; GPFORCE - PARAM,NOELOF - Incorrect Format on PUNCH Output
If PARAM,NOELOF,1 is specified the punch output format for GPFORCE is
incorrect.
Avoidance: There is no avoidance for this option. The f06 output is
correct.

@
37585; 707; Case Control - SET - Duplicate Output
If a SET command is used to request grid point output, and the set
contains a thru range referencing non-existing grid points, then duplicate
output will be printed for the last existing grid point. For the
following example, no grid point exists between the range of 10001 and
10010, the displacement for grid point 12 is printed twice.
SET 1 = 11,12 , 10001 THRU 10010 $
DISP(PHASE,SORT2) = 1

Avoidance: Remove the reference to the non-existing grid points in the

set.
@
37590; 70.7.2; GPFORCE - GPFDR - Output does not include MPC forces in SOL 144
When Grid Point Forces (GPFORCE) are requested in SOL 144 (Static Aero),
 the forces of multipoint constraint (MPCFORCE) should be included in the
output, but are not.
Avoidance: Add the MPCFORCE=ALL or MPCFORCE(PLOT)=ALL output request to
Case Control to cause the MPC forces to be computed. Once computed, they
will be included in the GPFORCE output.
@
37593; 70.7.2; Superelements - CSUPER, STATSUB, SPC - UFM 1126 (GNFIST)
In a normal modes analysis with image superelements (CSUPER), if the
structure is preloaded (STATSUB) and spcs are applied to the interior of
the primary superelement, then the job will fail with the following
unclear error messages:
*** USER FATAL MESSAGE 1126 (GNFIST)
DMAP MODULE MATREDU IS ATTEMPTING TO OUTPUT DATABLOCK 202 NAME
= KDSF , WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE KDSF=OVRWRT
Avoidance: Replace the following nddl statement
PATH KDSF PEID,MPC,SPC,MTEMP,NLOOP,DESITER,PVALID,
APRCH,HIGHQUAL,K2GG,M2GG,AUXMID,STATSUB,P2G $
with
PATH KDSF seid,PEID,MPC,SPC,MTEMP,NLOOP,DESITER,PVALID,
APRCH,HIGHQUAL,K2GG,M2GG,AUXMID,STATSUB,P2G $
@
37597; 70.7.2; SSSALTER - delmodea.vxx, Part Superelements - Modes Not Deleted
When using the delmodea.vxx alter with PART superelements, and the
DMI,SKPMODE entry is placed in the main bulk data section, then selected
modes will not be deleted. The following warning messages will be issued:
^^^ USER WARNING MESSAGE: MODE PARTITIONING MATRIX NOT FOUND IN BULK DATA
^^^ PLEASE CHECK INPUT DECK FOR DMI, SKPMODE
ENTRIES
^^^
^^^ THE MODE DELETION ROUTINE CANNOT CONTINUE - NORMAL PROBLEM SOLUTION
CONTINUING WITH ALL MODES...
Avoidance: Either request a copy of the updated alter from your local
support engineer or put the DMI, SKPMODE in the last partitioned bulk data
section.
@
37598; 70.7.2; Parallel Processing - DMP, Component Modes - Infinite loop
If distributed parallel processing is selected by setting dmp>1 on a
component modes analysis the job may hang.
Avoidance: Insert the following DMAP alter:
compile cmpmode
alter 'xofp'
if ( getsys(np,265)>1 ) then $
delete /cmlama,cmphixz,cmmi,cmoeigs,eigvmats $
disutil ,,,,/cmlama///3 $
 disutil ,,,,/cmphixz///3 $
disutil ,,,,/cmmi///3 $
disutil ,,,,/cmoeigs///3 $
disutil ,,,,/eigvmats///3 $
paraml cmphixz//'trai'/1/s,n,nozset $
else if ( getsys(np,265)=1 ) then $
disutil cmlama,,,////3 $
disutil cmphixz,,,////3 $
disutil cmmi,,,////3 $
disutil cmoeigs,,,////3 $
disutil eigvmats,,,////3 $
endif $
compile resvec
alter 'read ',''
read kzz,mzz,,,,,,,,,,,/
ALAMA1,APHx1,Mx,ox,Ex,/
'MODES'/S,N,NV////-1/'LAN'//1.e8//999/0////prtsum $
if ( getsys(np,265)>1 ) then $
delete /alama1,aphx1,mx,, $
disutil ,,,,/alama1///3 $
disutil ,,,,/aphx1///3 $
disutil ,,,,/mx///3 $
paraml aphx1//'trai'/1/s,n,nv $
else if ( getsys(np,265)=1 ) then $
disutil alama1,,,////3 $
disutil aphx1,,,////3 $
disutil mx,,,////3 $
endif $
endalter
@
37600; ; Nonlinear Analysis - PSHELL - Incorrect NLSTRESS Format
The NLSTRESS format is wrong if both plane-strain and plane-stress shell
elements exist in the same input file. That is if multiple PSHELL
records exist, some with MID2>-1 and some with MID2=-1.
Avoidance: This output can be turned off by requesting NLSTRESS=NONE.
Another avoidance is not to put both types of shell elements into a single
nonlinear run. The other results (including the stresses using the linear
format) are not affected.
@
37602; 70.7.2; Static Analysis - DMP, dmp>1 and iterative solver - SFM 4276 (UN
PACK)
If a SOL 101 run is made with Distributed Memory Parallel and the
iterative solver is selected, then the job may crash with the following
message:
*** SYSTEM FATAL MESSAGE 4276 (UNPACK)
ERROR CODE 20 PID= 0
Avoidance: Do not use DMP. The error is corrected in Version 70.7.2
@
37606; 70.7.2; Parallel Processing - DMP>1 and mergeresults=yes, OLOAD - Wrong
Answers
In static analysis with domain decomposition (a.ka., Distributed Memory
Parallel, dmp>1) and mergeresults=yes on the nastran submittal line, the
OLOAD output and OLOAD maximums will be wrong for the boundary points if
distributed loads like PLOAD and GRAV are applied.
 Avoidance: Insert the following DMAP alter:
compile sedisp
alter 'pj0s','pj0s'(,1)
upartn usetx,pj/pax,pga,,/'g'/'a'//1 $
merge ,,pax,,,,spcpart/pag0/1 $ Loads at boundary due to local
$ distributed loads
DISUTIL pag0,,,/pg0///8 $ sum of all boundary dist. loads
disutil pg0,,,/pg00///3 $ broadcast to slaves
if ( getsys(np,265)>1 ) equivx pg00/pg0/-1 $
PARTN pg0,,SPCPART/,pay,,/1 $ get local boundary dist. loads
add5 pay,pjxl,,,/pj0s $ add in boundary point loads
umerge1 usetx,pj0s,pga,,/pj1/'g'/'a'//1 $ combine
endalter
@
37615; 2001.0.1; Eigenvalue Analysis - EIGRL - UWM 4698 (DCMPD), SFM 7340
When Lanczos encounters a massless mechanism it will first provide the
following message:
*** USER WARNING MESSAGE 4698 (DCMPD)
STATISTICS FOR DECOMPOSITION OF MATRIX (NONE) .
THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER
THAN
1.00000E+15 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL.
GRID POINT ID DEGREE OF FREEDOM MATRIX/FACTOR DIAGONAL
RATIO MATRIX DIAGONAL

1 R2 1.00000E+15
3.37236E+00

It may then issue a fatal message of the form:

*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
process error reported by subroutine LNNP2CD
see LNNP2CD abstract (ier = -729)
*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
warning reported by subroutine LNNDRVD
see LNNDRVD abstract (ier = 729)
*** SYSTEM FATAL MESSAGE 3003 (READ)
ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK
SCRATCH (FILE 315) IN SUBROUTINE REIGL3
Avoidance: An avoidance is to add an spc to the DOF listed in the first
message.
@
37633; ; Restart - EIGRL, EIGR - Inefficient Restart
In a restart run, if an unreferenced EIGRL (or EIGR) is added to the
restart run input, then the modes will be recalculated even though the
entry is not referenced.
Avoidance: Remove the unreferenced EIGRL/EIGR entry.
@
37635; ; Nonlinear Transient Analysis - RESTART, PARAM,G, GE -Different
Answers
In a nonlinear transient restart analysis, if large displacement is
turned on (param,lgdisp,1), and param W3 damping exists in the model (with
PARAM,G ), the results may be different than the one in the coldstart
run. The difference is due to the possible non-syncronization of the
stiffness update between the coldstart and restart runs.
Avoidance: Recommended damping in nonlinear transient (SOL 129) is param,
W4 damping with GE value in MAT1.
If stiffness update syncronization is desired with param W3 damping, set
KSTEP=1 to force the stiffness update.
Care must be exercise when there is adaptive gap element since the
penalty stiffness can be fictituously
high inducing an unrealistic high damping force.
@
37642; ; Restarts - CBUSH - SWM 2072 for card type 2601
If a model contains CBUSH elements, then upon RESTART SYSTEM WARNING
MESSAGE 2072 will be issued by the RESTART module when it compares the
GEOM2 data block.
Avoidance: No avoidance is necessary, ignore the message.
@
37646; 2001; Documentation/QRG - CSUPER - Improve Remark 3 of CSUPER Entry
Remark 3 of the description of the CSUPER bulk data entry in the QRG
should be improved to read:
3. For identical or mirror image superelements, the grid point IDs, GPi,
may appear in any order. However, if they are not in the same order as the
external GRIDs of the primary superelement, then the SEQSEP entry is also
required. In case of external superelements, the GRID IDs must be in the
order that the terms in the associated matrices occur in.
@
37653; 2001; Documentation - CREEP - Thermal Correction
The Quick Reference documentation for the CREEP entry should be changed to
further explain what turns on the thermal correction for the creep strain
rate. If the T0=0.0 or blank, then the thermal correction is not
utilized. However if PARA,TABS is used, TABS is added to T0 and then the
thermal correction is utilized even if T0 is blank or 0.0. If thermal
correction is utilized then, EXP must be meaningful and have the
consistent units.
@
37661; 70.7.2; Superelement - Repeated SUPER Command - UFM 7323, Possible Wron
g Answers
If the SUPER case control command is repeated incorrectly as shown below
.
subcase 100
super = 0,1
.
subcase 200
super = 0,2
.
subcase 300
super = 0,1
.
 then following unclear fatal message is issued
*** USER FATAL MESSAGE 7323 (SEDRA)
SUBCASES FOR SUPERELEMENT 0 HAS A LOAD SEQUENCE NUMBER = 1
WHICH IS OUT OF ORDER.
The job will continue to execute and the results are unpredictable.
Avoidance: Correct the offending SUPER case control command.
@
37700; 70.7.2; Parallel Processing/Statics - Distributed Memory Processing - SF
M 7492
If DMP is specified on the command line for a SOL 101 run, the following
error message may appear:
*** SYSTEM FATAL MESSAGE 7492 (PRESOL)
INCONSISTENT VALUE FOR NUMBER OF LOCAL GRIDPOINTS ON PROCESSOR 1:
ACCORDING TO SIL TABLE, THE NUMBER OF LOCAL GRIDPOINTS IS 100763
ACCORDING TO GEQMAP, THE NUMBER OF LOCAL GRIDPOINTS IS 101073
USER ACTION: PLEASE CONTACT MSC.SOFTWARE CLIENT SUPPORT.

Avoidance: Do not use DMP. This error is corrected in Version 70.7.2.

@
37719; ; Documentation - V70.7 Release Guide Clarification
Section 2.7 of the Version 70.7 Release Guide documents a new keyword
"DCMPSEQ". This section of the guide does not give examples and it
contains a typo. The Guide states that this keyword is the same as
SYSTBM(206), this is a typo, it should be SYSTEM(206). It states that the
new option is invoked by adding 64 to the value of the keyword; however
the default of SYSTEM(206) is not documented. The default value of
SYSTEM(206) is 4 for all systems except NEC, the default for NEC is 0.
This keyword must be set on the NASTRAN command line in the input deck.
For all systems except NEC the input to activate this option is:
NASTRAN DCMPSEQ=68
or
NASTRAN SYSTEM(206)=68
@
37748; 2001; Miscellaneous - V70.7, New Log File Messages - Wrong DOF Outpu
t
The following new messages are sent to the log file in Version 70.7:
10:32:43 Finite element model generation started.
10:32:43 Finite element model generated 1 degrees of freedom.
10:32:43 Finite element model generation successfully completed.
The second message above always reports 1 DOF which is in almost all
cases wrong.
@
37764; 70.7.2; Parallel Processing/Frequency Response - DMP, PARAM,RESVEC - UFM
Distributed Memory Parallel frequency response analysis will fail when
 PARAM,RESVEC,YES is used to specify residual flexibility.
Avoidance:
Either:
1) Run using a newer version (V70.7.2 or newer)
2) Do not use residual vectors with dmp
3) Insert the following DMAP alter:
compile resvec
alter 'read '(,-1)
$ For dmp>1, deactivate parallel
np=getsys(np,197) $ capture NUMSEG
if( np>1 ) putsys(0,197) $ set NUMSEG=0
alter 'read '
if( np>1 ) putsys(np,197) $ restore NUMSEG
compile cmpmode
alter 'read '
if ( getsys(np,197)>1 and getsys(np,265)>1 ) then $
$ Since the master node has all modes, then
$ then we will generate a dummy mode on the
$ slaves to avoid hang-ups.
delete /cmlama,cmphixz,cmmi,cmoeigs,eigvmats $
nozset=1 $
paraml cmkxx//'trai'/1/s,n,noxset $
matgen ,/cmphixz/4/nozset/noxset $
matgen ,/lmat/7/3/nozset $
matgen ,/cmmi/nozset/nozset $
lamx lmat,/cmlama $
modtrl cmlama// /0/0/0/0/0 $
endif $ getsys(np,197)>1 ) then $
endalter
@
37773; 70.7.2; Matrix Methods - PARTN/MERGE - Poor Performance
The PARTN and MERGE modules perform poorly when the partitioning vectors
are purged.
Avoidance: There is no avoidance but this has been corrected in Version
70.7.2.
@
37775; 70.7.2; Miscellaneous - Q4SKEW,Q4TAPER, TETRAAR, and T3SKEW do not work
in Unix rcfile
On unix workstations, if you attempt to set SYSTEM cells 189, 190, 191, or
218 in the "rcfile" using the cell name (Q4TAPER, Q4SKEW, TETRAAR,
T3SKEW), the default values will be used for these cells with no warning
messages issued,
Avoidance: Use the SYSTEM cell number in the "rcfile" for these cells.
For example:
nastran system(189)=1.
nastran system(190)=1.
nastran system(191)=100.
nastran system(218)=1.
@
37793; 2001; Nonlinear Normal Modes - RESTART, PARAM,FOLLOWK - Wrong Answers
Nonlinear normal mode analysis by a restart in SOL 103 from SOL 106
 database gives erroneous solution in presence of follower forces in
Version 70.7. The follower
stiffness is not included in this restart runs. This error may be avoided
if the normal mode analysis is performed in SOL 106 during the coldstart
run, or
SOL 103 is used with a statsub option for the preload. However, a
restart procedure is required if free-free modes are subject to
a nonlinear deformation under the preload.
Avoidance: For free-free modes, insert the following DMAP alter into the
SOL 103 restart
compile semg
alter 'call semg1'
type parm,,char8,y,followk='yes' $
TYPE PARM,,RS,N,stol=0. $
type db slt $
dbstatus slt//s,n,noslt $
PARAML CASES//'DTI'/COLNO/4//S,N,LOADID $
if ( followk='yes' and noslt>0 and loadid>0 and
seid=0 ) then $
TAFF SLT,BGPDTS/ESTF,GPECTF/
LUSETS/LOADID/0/1. $
EMG ESTF,CSTMS,MPTS,DIT,,UGni,ETT,EDT,
DEQATN,DEQIND,BGPDTS,GPSNTS,,,,/
KFELM,KFDICT,,,,/
1/0/0//1/////////-1/1. $
EMA GPECTF,KFDICT,KFELM,BGPDTS,SILS,CSTMS,,/
KFGG0,/-1 $
dbview cstm0=cstms where (peid=0) $
emakfr bgpdts,cstms,slt,mjj,cstm0,scstm/
kfrfgg/loadid $ rforce stiffness
ADD5 KFGG0,kfrfgg,,,/Kfgg $
PARAML KFGG//'TRAILER'/3/S,N,UNSYM//S,N,NOKFGG $
IF ( UNSYM<>6 AND NOKFGG>-1 ) THEN $
$ Symmetrize stiffness matrix if unsymmetric
TRNSP KFGG/KFGGT $
ADD5 KFGG,KFGGT,,,/KF/1./-1. $
norm kfgg/dummy///s,n,xnorm/1 $
stol = xnorm*1.e-11 $ tolerance recommended
matmod kf,,,,,/kfa,/2////stol $
paraml kfa//'null'////s,n,notnull $
if ( notnull=-1 ) modtrl kfgg////6 $ symmetric
ENDIF $ UNSYM<>6 AND NOKFGG>-1
add kfgg,kjj/kfjj $
equivx kfjj/kjj/-1 $
ENDIF $ FOLLOWK='YES '
endalter
@
37797; 70.7.2; Documentation - diag= keyword does not work for all diags
The "diag" command line keyword does not work for all possible diag flags.
In particular, diag flags 5 and 6 cannot be set from the command line or
an RC file.
Avoidance: Set these diags in the bulk data file using the standard DIAG
statement, DIAG 5,6
@
37799; 2001; Optimization - DRESP2, DRESP1 - SFM 3007 (DOPR3)
 In an optimization run, if a DRESP2 references both DRESP2 and DRESP1, and
the DRESP1 references a property PTYPE (e.g., PBAR) which points to more
than one element, then the job will fail with the following error
messages:
*** SYSTEM FATAL MESSAGE 3007 (DOPR3)
ILLEGAL INPUT TO SUBROUTINE DOPR3J
0FATAL ERROR
*** SYSTEM INFORMATION MESSAGE 1156 (XCLNUP)
GINO FILE 110 WAS LEFT OPEN AT PROGRAM TERMINATION.
Avoidance: There are two possible avoidances:
1. Avoid referencing a DRESP2 with a DRESP2, reference a series of DRESP1
instead.
2. Reference ELEM instead of property on the PTYPE field on the DRESP1
entry.
@
37805; 70.7.2; Optimization/Design Sensitivity - ANALYSIS=MFREQ - UFM 9068
The message "USER FATAL MESSAGE 9068 NO DRESP1 OR DRESP2 ENTRIES EXIST FOR
DESOBJ=n" may be issued in error for a solution 200 analysis that contains
only frequency or transient analysis subcases and the objective is a
WEIGHT or VOLUME response. The fatal will only occur if the DRESP1 ID
that is just below the ID called out by the DESOBJ request has a blank
ATT1 field on the DRESP1 bulk data entry; i.e., the response is RTYPE =
EIGN, FREQ or LAMA.
Avoidance: Provide an ID on the DRESP1 entry for the WEIGHT or VOLUME
objective that is lower than any other DRESP1 ID or ensure that
the ID that is just below the objective ID has a non-blank ATT1 value.
@
37806; ; Documentation - V70.7 Release Guide
Release Notes, p50, "File Size Reduction for Modal Frequency Response"
The efficiencies described in this section require the use of sssalter
spdra.v707, available on the msc internal web site. Contact your local
MSC Technical Representative to get a copy of this dmap alter.
@
37812; 70.7.2; P-elements - InterFace Elements - Incorrect EPSILONs
The EPSILON table produced when interface elements are used is incorrect
in the NEC Version 70.7.0 release of MSC.Nastran The EXTERNAL WORK is
correct however.
Avoidance: There is on known avoidance. The error has been corrected in
Version 70.7.2.
@
37832; 70.7.2; Heat Transfer - TSTRU, TEMP(INIT) - SFM 3001(TA1)
If TEMP(INIT) is set to the same ID as TSTRU then the job will fail with
the message:
*** SYSTEM FATAL MESSAGE 3003 (TA1)
ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK ETT
(FILE 106) IN SUBROUTINE TA1ETD
 Avoidance: Specify a different set ID for TEMP(INIT).
@
37833; 70.7.2; Installation - INSTALL.txt and INSTALL.htm on CD state incorrect
MPT version
The files INSTALL.txt and INSTALL.htm on the V70.7 CD state that for
running distributed parallel MSC.Nastran on SGI R8K/ R10K/R12K, MPT 1.5 is
required.
This is a typo, MPT 1.3 is required.
@
37834; 70.7.2; Estimate - buffsize=estimate keyword broken in V70.7
In Version 70.7, if the keyword buffsize=estimate is used, then the job
will fail with the following error messages in the f06 file.
NASTRAN BUFFSIZE=ESTIMATE $(/USERS/SWAN/.NAST707RC[26])
*** USER FATAL MESSAGE 17 (LOOKUP)
SYMBOL "ESTIMATE" IS UNDEFINED
followed later by:
*** USER WARNING MESSAGE 782 (SYSCHK)
THE NASTRAN BUFFSIZE PARAMETER CANNOT BE SMALLER THAN MACHINE DEFAULT
= 2049
USER INFORMATION: NASTRAN BUFFSIZE PARAMETER IGNORED.
USER ACTION: USE THE INIT FMS STATEMENT TO SET THE DESIRED BUFFSIZE
FOR EACH INDIVIDUAL DBSET.
*** USER INFORMATION MESSAGE 480 (SYSCHK)
JOB TERMINATED DUE TO ABOVE ERROR(S) IN THE NASTRAN STATEMENT
PROCESSOR.
FATAL ERROR
Avoidance: Specify a specific buffsize.
@
37835; 70.7; Installation - HP-UX, Install Script Reports No Disk Space Avail
able
Under some circumstances on HP-UX systems, mscsetup may report "0 MB" disk
space available on NFS or automounted filesystems. The full circumstances
are not yet known.
Note: The error is a direct result of incorrect information being returned
from the df(1) command on HP-UX 10.20 and 11.00. The specific command
used by mscsetup is "df -k", and sample output showing the error is:
$ df -k /net/trinity/scratch
/tmp_mnt/net/trinity/scratch (trinity:/scratch ) : 0 total
allocated Kb
0 free allocated
Kb
0 used allocated
Kb
45 % allocation
used
Another invocation of df, that does not return sufficient information,
appears to have better results:
 $ df /net/trinity/scratch
/tmp_mnt/net/trinity/scratch (trinity:/scratch): 2032506 blocks
-1 i-nodes
Finally, an alternative command yields the correct results:
$ bdf /net/trinity/scratch
Filesystem kbytes used avail %used Mounted on
trinity:/scratch 1981000 862335 1016253 46%
/tmp_mnt/net/trinity/scratch
Avoidance: Ignore the disk space warning messages and force mscsetup to
install the s/w if you know sufficient space is available to install the
selected products.
@
37837; 70.7.2; Documentation/Estimate - Minimum Values
The ESTIMATE utility will apply a minimum value to estimates for the
following:
Parameter Keyword Default Value (MB)
--------- ---------- ------------------
memory memmin 16
disk dskmin 1
sdball dballmin 1
scratch scratchmin 1
scr300 scr300min 1
The minimum value is applied after the conservative value (i.e., "memco",
"dskco", "dballco", "scratchco", and "scr300co" respectively). Thus,
setting "memco=0.5" to reduce all memory estimates by 50% will have *no*
effect if the estimate falls below the minimum value.
@
37843; 70.7.2; Loads - PLOAD4, Large Field Format - UFM 4577 (GP3C)
If a PLOAD4 references an element with large field format, the job may
fail with the message:
USER FATAL MESSAGE 4577 (GP3C)
PLOAD4 LOAD SET= 1 REFERENCES ELEMENT ID =*******************
WHICH WAS NOT FOUND AMONG TRIA3, TRIA6, QUAD4, QUAD8, HEXA, OR PENTA
ELEMENTS

Avoidance: There is no avoidance other than renumbering the elements.

@
37844; 2001; Accounting - XSEMDR - Can not Determine Solution Sequence
In releases prior to Version 70.7 the f04 Execution Summary ended with
the last module specific to the solution. Beginning in Version 70.7 the
f04 Execution Summary ends with the last line XSEMDR END. For example:
13:19:02 0:13 212.0 0.0 12.8 0.0 NLTRAN
527 EXIT BEGN
13:19:02 0:13 212.0 0.0 12.8 0.0 XSEMDR
END
This prevents the accounting program provided my MSC which reads the last
line from determining the solution sequence.
@
37859; 70.7.2; Miscellaneous - Missing TPL file mcso42.dat in V70.7
The file mcs042.dat was missing from the Version 70.7 tpl. It has been
added to the Version 70.7.2 tpl.

@
37862; GL; Miscellaneous - FLEXlm can not be used with system checkpoint ca
pabilities
MSC.Nastran jobs (as well as any other program) that use FLEXlm licensing
or any other socket-based communications cannot be checkpointed by
system-level checkpoint/restart capabilities, such as SGI's cpr utility.
If attempted messages similar to the following will be issued in the log
file:
Checkpointing id 259527 (type PID) to directory /tmp/test.cpr
CPR Error: Uncheckpointable fd, pid 259527 fd 10 type socket
Checkpoint failed. Restarted all processes back to their running states
CPR Error: ckpt_create child failed: errno 47 (Checkpoint/Restart error)
Failed to checkpoint process 259527
---------------------------------
Avoidance: Use a node-lock authorization code for licensing and avoid the
use of any socket connections by MSC.Nastran.
@
37872; 2001; SSSALTERs - checka, PARAM,LANSUP - SFM 3007
If PARAM,LANSUP,1 is used in the checka sssalter in version 70.5, 70.6, or
70.7, the run will fail with SYSTEM FATAL MESSAGE 3007 indicating an
illegal operation while merging matrices.
Avoidance: Replace:
UMERGE USET,DMI,Dar/DM1/'A'/'R'/'L' $ DN changed to DAR
with:
equivx dar/dm1/-1 $

@
37873; 70.7; Old Design Sensitivity - SFM 3001, MPTS Does Not Exist
If the old sensitivity (SOL 103) is used and no material entries are
present, the run will fail with SFM 3001 indicating that data block MPTS
does not exist.
Avoidance: 1) use SOL 200 to perform similar task or
2) or add a dummy material to the input file if the old dsa is
used.
@
37880; 70.7.2; Parallel Processing - DMP - SPCForce Output with Multiple Subcas
es - UFM 5423
Distributed Memory Parallel jobs, dmp>1, with multiple subcases will fail
with the following message if spcforce output is requested.
 *** USER FATAL MESSAGE 5423 (SADD5)
ATTEMPT TO ADD INCOMPATIBLE MATRICES, (SADD5)
COLS ROWS FORM TYPE NZWD DENS
TRAILER FOR QG0S IS 1 450 2 2 0 0
TRAILER FOR QGI IS 2 450 2 2 24 267
0FATAL ERROR
Avoidance: Either do not select dmp or do not select spcforces.
@
37886; 2001; Nonlinear Analysis - RFORCE, PARAM,FOLLOWK - Poor Performance EM
AKFR
Nonlinear analysis of large problems with RFORCE as a follower force may
exhibit poor performance in EMAKFR.
Avoidance: There is no avoidance.
@
37892; 70.7.2; Documentation/COG - V70.7 smaster=estimate fails
The "MSC.Nastran V70.7 Configuration and Operations Guide" incorrectly
states that "smaster=estimate" is valid, it is not.
Avoidance: Set an actual size if the MASTER DBSet default size needs to be
increased.
@
37898; 2001; Superelements - SDAMPING; PARAM,SESDAMP,YES - NORM=MAX or POINT,
Wrong Answers
Superelement modal damping (PARAM,SESDAMP,YES) produces incorrect
answers when
MAX or POINT normalization is selected on the EIGR or EIGRL entries.
Avoidance: Use MASS normalization which is the default.
@
37907; 70.7.2; Optimization/Design Sensitivity - DRESP1, PCOMP - UFM 5503
The message:
UFM 5503 ELEM NO. n PLY NO.y REFERENCED ON A DRESP1 ENTRY COULD NOT BE
FOUND
may be issued in error for composite responses (DRESP1
RTYPE=CSTRESS,CSTRAIN or CFAILURE) if the SOUTi requests on the underlying
PCOMP entries have different values for the different layers (i.e., some
are YES and some are NO or blank).
Avoidance: Make all the SOUTi requests on a given PCOMP entry the same if
design responses are to be imposed on elements associated with the PCOMP.
@
37913; 70.7.2; Submittal - nastran Command does not recognize LSF Submittal
The nastran command does not automatically recognize a submittal via LSF.
For NQS, the nastran command will recognize it is running in a batch job
and set the defaults batch=no and notify=no. In addition, if submitted as
a DMP job by LSF, the nastran will not recognize the list of selected
 hosts.
Avoidance: The problem may be avoided by:
1. set notify=no and batch=no in an RC file or on the command line when
running under LSF.
2. Set hosts=$(print $LSB_HOSTS | sed 's/^ *//;s/ *$//;s/ */:/g') on
the command line when running a DMP job under LSF
@
37918; 2001; Documentation/DMAP - DISUTIL not documented
The MSC.Nastran DMAP Modules and Datablocks CD for Version 70.7 mentions,
but does not document the module DISUTIL The following is the
documentation for this module:
Broadcast data blocks between slave and master processors for parallel
processing.
Additional processing may occur on the master processor.
Format:
Format for DISMETH=1 or 2:
DISUTIL B,X,R,EQMAP/
EPSSE/
NSKIP/S,N,EPSI/1 or 2 $
Format for DISMETH=3 on master processor:
DISUTIL DB,,,////3 $
Format for DISMETH=3 on slave processor:
DISUTIL ,,,,/DB///3 $
Format for DISMETH=4 or 5:
DISUTIL RESMAX,RESMAX0,CASECC,HEADCNTL////4 or 5 $
Format for DISMETH=6:
DISUTIL UG,SPCPART,GEQMAP,/UGG///6 $
Format for DISMETH=7:
DISUTIL PG,SPCPART,GEQMAP,/PGG///7 $
Format for DISMETH=8 or 9 on master processor:
DISUTIL MATS,,,/MATM///8 or 9 $
Format for DISMETH=8 or 9 on slave processor:
DISUTIL MATS,,,////8 or 9 $
Input Data Blocks:
B Rectangular matrix which is the local load vector (with local
values in local a-set).
X Solution of the equation [A][X]=[B] which is the local solution
matrix (with global values in local a-set).
R Residual matrix which is local (with local or global values in local
a-set)
EQMAP Table of degree-of-freedom global-to-local maps for domain
decomposition.
DB Any data block to be broadcast from the master to the slave
processors.
RESMAX Resultant or maxima matrix.
RESMAX0 Resultant or maxima matrix for residual structure.
CASECC Table of Case Control command images.
HEADCNTL List of integer codes for header print control
UG Displacement matrix in g-set for the current processor (local).
SPCPART Partitioning vector for domain decomposition.
GEQMAP Table of grid based local equation map indicating which grid
resides on which processors/partitions for domain
decomposition.
MATS Any matrix on slave processors.
 Output Data Blocks:
EPSSE Table of epsilon and external work.
UGG Displacement matrix in g-set for all processors (global).
PGG Force matrix in g-set for all processors (global).
DB Any data block to be broadcast from the master to the slave
processors.
MATM Any matrix on master processor.
Parameters:
NSKIP Input-integer-default=1. Record number in CASECC
corresponding to the first subcase of the current boundary
condition.
EPSI Output-integer-default=1. Static solution error ratio flag. Set to 1
if the error ratio is greater than 1.E-3.
DISMETH Input-integer-default=1. Method.
1 Compute epsilons and external works assuming a-set
components of RUF are local values (which occurs in a direct
solution)..
2 Compute epsilons and external works assuming a-set
components of RUF are global values (which occurs in a in
iterative solution).
3 Broadcast table or matrix from master to slaves.
4 Broadcast VECPLOT resultant output from slaves to master
and combine on master.
5 Broadcast VECPLOT maxima output from slaves to master
and combine on master.
6 Broadcast displacement matrices from slaves to master and
merge into global displacement matrix on master.
7 Broadcast force matrices from slaves to master and
add/merge into global force matrix on master.
8 Broadcast any matrix from slaves to master and add all
matrices on master.
9 Broadcast any matrix from slaves to master and append
columnwise all matrices on master.
Remarks:
1. All executions of DISUTIL must be synchronized across all processors.
2. EPSSE may be purged. EPSSE contains:
a. Sequential number of subcases
b. Superelement (domain) identification number
c. Epsilon error ratio
d. Strain energy (external work)3. RESMAX0 may be purged, on the slave
processors.
4. UGG may not be purged on the master processors.
5. For DISMETH=8 or 9 then MATS and MATM must be in machine precision.
@
37925; 2001; Optimization/Design Sensitivity - PCOMP - UFM 5629 (DSAD1C)
If RTYPE=STRESS or STRAIN in a composite optimiztion deck with NOCOMPS=1,
the following fatal message is issued.
*** USER FATAL MESSAGE 5629 (DSAD1C)
THERE ARE DRESP1 ENTRIES FOR RESPONSE TYPE = STRESS BUT THERE ARE
NO CORRESPONDING OUTPUT RESPONSES.
Avoidance: There is no avoidance.
@
37932; 2001; Optimization/Design Sensitivity - DVMRELi, DVCRELi, scr=no - Wro
ng Answers
 In an optimization run, if the design variables involve material
properties (DVMRELi) or connectivity (DVCRELi) only, and scr=no is used,
then the run will converge prematurely. A symptom of this problem is the
unexplainable job termination after the first design cycle with no change
in the objective function.
Avoidance: Run the job with scr=yes or insert the following alter:
compile update
alter 'RESTART EPTS,EPTSP'
DBVIEW mPTSP=mPTS WHERE (DESITER=DESP) $
RESTART mPTS,mPTSP,/UHIST $
DBVIEW geom2SP=geom2S WHERE (DESITER=DESP) $
RESTART geom2S,geom2SP,/UHIST $
endalter
@
37944; ; Documentation - VECPLOT - DMAP Module and Datablock Dictionary
The Version 70.7 DMAP Module and Data Blocks Dictionary has incorrectly
indicates that the VECPLOT module has only one output datablock.
Starting in Version 70.7 the VECPLOT module has two output datablocks.
This is correctly documented in the V70.7 Release Notes.

@
37952; 70.7.2; Documentation/COG - Estimate Program, Precedence of Keywords
The precedence of ESTIMATE processed keywords in the MSC V70.7
Configuration and Operations Guide listed in Section 6.2 should be as
follows:
1) ESTIMATE keywords, which are also NASTRAN keywords that are recognized
in the bulk data file (Example: NASTRAN BUFFSIZE=4097)
2) ESTIMATE keywords on the command line, Example: msc707 estimate
myjob.dat memco=2.5
Note: NASTRAN only keywords will cause an error (example:
auth=1700@server)
3) ESTIMATE keywords, which are also NASTRAN keywords, contained in the
nastran INI and/or RC files
(example: buffsize)

4) ESTIMATE keywords contained in data-file-directory/.estimaterc on UNIX,

or
data-file-directory\estimate.rcf on NT, where
data-file-directory is the directory containing the input data file
5) ESTIMATE keywords contained in $HOME/.estimaterc on UNIX, or
%HOMEDRIVE%%HOMEPATH%\estimate.rcf file on NT
6) estimate.ini file in the install-dir/msc707/arch on UNIX
(example: /msc/msc707/hpux/estimate.ini)
There is no equivalent on NT.
@
37969; ; Superelements - Poor Message, UFM 6790 (SEPO21)
 UFM 6790 shown below is output when a list superelement has no points
that can be identified as boundary points. The fatal message is followed
by table print which may be ignored.
*** USER FATAL MESSAGE 6790 (SEPO21)
THE FOLLOWING SUPERELEMENTS ARE DETACHED:
1 3 62000000007 52000000008
212000000002 222000000003 232000000004
242000000005 252000000006 61101529088
71102053376 81102577664 91103101952
101103626240 11 22000000001
42000000007 52000000008 212000000002 22
2000000003 232000000004 242000000005
252000000006 -1 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
@
37978; 2001; Nonlinear Analysis - HEXA - Poor Convergence for Small Loads
Geometric nonlinear analysis (SOL 106 with PARAM LGDISP > 0) with solid
elements and extremely small loads may fail to converge.

@
37981; 2001; Nonlinear Analysis - UWM 5426 for DEFORM Entry Should be a FATAL
Using a DEFORM entry in the case control for SOL 106 (nonlinear statics)
will produce:
*** USER WARNING MESSAGE 5426, DEFORM BULK DATA CARDS REFERENCED BY DEFORM
CASE CONTROL COMMAND ARE NOT DEFINED.
THE REQUEST FOR STRESSES AND/OR FORCES IS IGNORED.
This will cause data recovery for stresses and/or forces to be ignored.
It will also lead to misleading results, because the DEFORM entry is not
applicable to Sol 106.
Recommendations:
1) change Warning 5426 to Fatal for sol 106
2) Add the following remark to Deform Bulk data section:
4. The deform entry, when called by the deform case control command, is
applicable to linear static, inertia relief, differential stiffness, and
buckling (Solutions 101,105,114, and 200) and will produce fatal messages
in other solution sequences. Use SPCD to apply enforced displacements in
solution sequences for which DEFORM does not apply.
3) Modify remark 2 in the Case control FROM:
DEFORM is only applicable in statics, inertia relief, differential
stiffness, and buckling problems.
TO:
DEFORM is only applicable in linear statics, inertial relief, differential
stiffness, and buckling problems (Sols 101, 105, 114, and 200), and will
produce a fatal message in other solution sequences.
@
37990; ; Nonlinear Analysis - Beam with Warping Failed to Converge - SFM
4549
A nonlinear analysis of a BEAM with the warping degree of freedom
specified may results in failed convergence or SFM 4549 followed by a
traceback:
SYSTEM FATAL MESSAGE 4549,
SINGULAR (C+KBRR) MATRIX ENCOUNTERED FOR BEAM ELEMENT 1001
Traceback:
Offset 0x00003e28 in procedure nmepd_
Offset 0x000009d0 in procedure nmatd_
Offset 0x000022dc in procedure nbem2d_
Offset 0x0000017c in procedure nembrd_
Offset 0x00001124 in procedure nlemg_
Offset 0x00000790 in procedure nliter_
Offset 0x00000088 in procedure xsem15_
Offset 0x00000990 in procedure mainal
@
38012; 2001; Documentation/QRG - PARAM,FOLLOWK
The Quick Reference Guide incorrectly states that PARAM,FOLLOWK is
available only in solutions 103, 105, and 107. It is also available in
solutions 108, 109, 111, and 112.
@
38013; ; DMAP - PARTN - TYPE is Ignored
Specification of TYPE is ignored in the PARTN module. This may cause a
problem in user dmaps. The avoidance is dependent on the intent of the
DMAP.

@
38016; ; Superelements - SESET failures in SEP1X using THRU range entry
It has been observed that the SESET entry may cause failures when a THRU
range is specified where the low GRID ID is less than the starting GRID
ID of the model.
Avoidance: An avoidance is to make the starting grid equal to the low the
low grid identification.
@
38021; 70.7.2; Documentation/COG - Fujitsu, XDB, TRANS
The V70.7 Configuration and Operations Guide states that the Fujitsu
VPP,Vx UXP/V systems require that the XDB file be translated. This
information is incorrect. The correct line would be similar to that of HP
HP-UX right underneath the Fujitsu line.
@
38023; 2001; Documentation/QRG - PBEAML - NSM(B) not included on PBEAML entry
The Quick Reference Guide incorrectly omits the field for the NSM
(non-structural mass) on end B of the PBEAML entry. The last field on this
entry is the NSM for end B.
@
38031; ; Nonlinear Transient Thermal Analysis - Missing Results from XDB
A multiple subcase transient thermal analysis will be missing temperature
results in the OP2 and/or XDB files for the 2nd and subsequent subcases.

Avoidance: A DMAP avoidance is under investigation.

@
38041; 70.7.2; Optimization/Design Sensitivity - DESGLB, GRAV, Multiple Subcase
s - UFM 5423
In an optimization run, if the DESGLB global constraint is used, and there
are two or more static subcases with gravity loads, then the job may fail
with the following error messages:
*** USER FATAL MESSAGE 5423 (SADD5)
ATTEMPT TO ADD INCOMPATIBLE MATRICES, (SADD5)
COLS ROWS FORM TYPE
NZWD DENS
TRAILER FOR EGM IS 2 80532 2 2
1830 114
TRAILER FOR EGKL IS 1 80532 2 2
10976 681
0FATAL ERROR
Avoidance: Use the DESSUB constraint callout instead.
@
38046; 2001; Miscellaneous - OUTPUT2, TABPRT - Incorrect Format Field
Output datablocks (OFP tables) contain information in the header record
which describes the information contained in the datablock. Word 9 is used
to indicate if the information is (1) REAL, (2) REAL/IMAGINARY, or (3)
MAGNITUDE/PHASE. If the data is REAL/IMAGINARY, this flag is incorrectly
set to 1.
Avoidance: There is no avoidance.
@
38055; 70.7.2; Parallel Processing - DMP - Incorrect SPCforces for constrained
& loaded points
In a distributed memory parallel execution, dmp>1, SPCforces are
incorrect for constrained and loaded interior DOFs.
Avoidance: Insert the following DMAP alter:
compile disph1b
alter 'pss'
call dbstore pss,,,,//0/0/' '/0 $
compile sedisp
alter 'kfs,uox,pss'(,-1)
call dbfetch /pss,,,,/0/0/0/0/0 $
endalter
@
38056; 2001; Beam Library - PBARL - ILLEGAL INPUT TO SUBROUTINE DOMPFTN
When the generalized PBARL is designed using the beam library server, the
job fails in DOPR1 with the following message:
 "ILLEGAL INPUT TO SUBROUTINE DOMPFTN".
This message is caused by designing the sixth property (FE) in the EPT
table which is currently not used.
Avoidance: In the bsbrid subroutine, change the
Do 110 i = 1,nprop
do 100 j = 1,ndim
sentyp(i,j)=3
100 continue
110 continue
to
Do 110 i = 1,nprop
if ( iprop .eq. 6 ) go to 110
do 100 j = 1,ndim
sentyp(i,j)=3
100 continue
110 continue
@
38061; 2001; SSSALTERS - cygyroa.vxx, TEMP(LOAD) - SFM 3003 (EMG)
The following message will occur in with the SSSALTER CYGYROA if thermal
loads are used.
SYSTEM FATAL MESSAGE 3003 (EMG)
ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK ETT
(FILE 107) IN SUBROUTINE EMGTAB
The error occurs when the differential stiffness matrix is being
generated, and hence the actual output (e.g. displacements, stresses) is
correct. However, since differential stiffness is not generated, restarts
off this database will not include the desired gyroscopic effects. The
error occurs because of the unique way thermal loads are called out in
Cyclic solution sequences. The TEMP(LOAD) case control command points to a
bulk data LOADCYN entry which inturn points to a temperature set. The EMG
module issues the fatal error because it is searching for the temperature
set called out by the temp(load) command rather than the temperature set
called out on the appropriate LOADCYN entry.
Avoidance: Use this modified version of the cygyroa.V705 alter. Note that
this version requires the addition of PARAM,TEMPSET. See instructions at
top of alter.
$ V70.5
$ modified 12/16/99 by Jack Castro to fix error if TEMP(load) is used.
$ Directions: Insert PARAM,TEMPSET in bulk data with the value of the
$ temperature load set which is called out on the appropriate LOADCYN
entry.
$
$ THIS ALTER IS CONFIDENTIAL AND A TRADE SECRET OF THE
$ MACNEAL-SCHWENDLER CORP. THE RECEIPT OR POSSESSION OF
$ THIS ALTER DOES NOT CONVEY ANY RIGHTS TO REPRODUCE OR
$ DISCLOSE ITS CONTENTS, OR TO MANUFACTURE, USE, OR SELL
$ ANYTHING HEREIN, IN WHOLE OR IN PART, WITHOUT THE
$ SPECIFIC WRITTEN CONSENT OF THE MACNEAL-SCHWENDLER CORPORATION.
$
$
$
$ cygyroa.v69 - updated for v69 -- 9-25-96
$
$ tested on system built on 3-14-95 VERSION 68.2
$
$************************************************************************
******$
$ ALTER FOR SOL 114 TO INCLUDE GYROSCOPIC TERMS.
$
$ THE ALTER WILL ADD CENTRIPETAL, CORIOLIS, AND DIFFERENTIAL STIFFNESS
$ TERMS TO THE SUPERELEMENT STIFFNESS AND DAMPING MATRICES. TO USE THIS
$ ALTER THE USER MUST SUPPLY A RFORCE FOR EACH SUPERELEMENT, INCLUDING
$ THE RESIDUAL STRUCTURE ( SEID=0 ). THE COORDINATE
$ SYSTEM REFERENCED ON THE RFORCE CARD MUST BE CENTERED ON THE AXIS OF
$ ROTATION FOR THE SUPERELEMENT. THE GRID ID ENTRY IS IGNORED FOR THE
$ CALCULATION OF THE GYROSCOPIC TERMS. THE ROTATION IS ASSUMED TO BE
$ ABOUT THE REFERENCED COORDINATE SYSTEM (CID) ORIGIN, THE MAGNITUDE (A)
$ AND ROTATION VECTOR COMPONENTS (N1, N2, & N3) DETERMINE THE
ORIENTATION.
$
$ ONLY THE ZERO HARMONIC SHOULD BE USED IN THIS STEP OF THE PROCEDURE.
$ OTHER HARMONICS MAY BE USED IN THE RESTARTS.
$
$ AFTER THIS INITIAL STATIC RUN THE MODEL CAN BE RESTARTED IN ANY
$ CYCLIC SOLUTION SEQUENCE AND THE GYROSCOPIC EFFECTS WILL BE INCLUDED.
$
$ THE ABOVE PROCEDURE RESULTS IN A RESTART INWHICH REFORMULATION OF
$ THE SUPERELEMENT MATRICES IS NOT PERFORMED. REFORMULATION IS NORMALLY
DONE
$ WHEN THE USER CHANGES THE GEOMETRY, ELEMENT OR MATERIAL PROPERTIES, OR
$ ADDS GRIDS OR SCALAR POINTS. IF REFORMULATION IS DONE ( SEMA ) THEN
$ THE GYROSCOPIC TERMS WILL NOT BE INCLUDED IN THE SUPERELEMENT
MATRICES.
$
$ IF SPEED DEPENDENT DIFFERENTIAL AND GYROSCOPIC TERMS ARE REQUIRED IN
THE
$ RESTART, THEN PARAMETER DEPEND SHOULD BE SET TO -1. IF THIS OPTION IS
$ ACTIVATED, THEN THE RESTART MUST NOT HAVE MASS DEPENDENT LOADS (I.E.
NO
$ GRAVITY LOADING) BECAUSE THE MASS MATRIX IS MODIFIED AND THE RESULTING
$ LOADS WILL BE INCORRECT.
$
$ ALTER DEVELOPED BY DAVID BELLA FOR THE MACNEAL-SCHWENDLER CORPORATION,
$ AUGUST, 1992.
$
$************************************************************************
******$
$
echooff $
compile cycmode $
alter 1 $
putsys (0,192) $
$
COMPILE CYCSTATX $
$ALTER 44 $ V67
$ALTER 51 $ V68
$alter 'if.*(.*dbdict\>=1.*).*call.*dbdict1'
alter 'call.*super3'
$
TYPE DB SLIST $
TYPE PARM,,I,N,LPFLG=0,NOUP $
TYPE PARM,,I,N,SCNDRY=+1,EXTRN=+1 $
TYPE PARM,,I,Y,DEPEND= 0 $
TYPE PARM,,RS,N, OMEGA= 0. $
TYPE PARM,,CS,N, SCALE $
$
$ PREPARE DELETE LIST
$
FILE HISTS=APPEND $
MATGEN ,/TMPMGG/1/2 $
MATGEN ,/TMPBGG/1/2 $
MATGEN ,/TMPKGG/1/2 $
$
RESTART TMPMGG,,/HISTS $
RESTART TMPBGG,,/HISTS $
RESTART TMPKGG,,/HISTS $
$
$ LOOP ON SUPERELEMENTS
$
SEID=-1 $ RE-INITIALIZE FOR SEP2DR
$
DO WHILE ( LPFLG >= 0 ) $
$
IF ( RSONLY ) THEN $ CHECK FOR EXISTENCE OF SUPERELEMENTS
$
LPFLG=-1
NOUP=-1
SEID=0
$
ELSE $ DO SEP2DR WHEN SUPERELEMENTS EXIST.
$
SEP2DR SLIST,EMAP//S,N,SEID/S,N,PEID//S,N,LPFLG/////
S,N,NOUP/S,N,SCNDRY/S,N,EXTRN/S,N,NOMR/'ALL' $
$
ENDIF $
$
MESSAGE //' '/ $
MESSAGE //' USER DMAP INFORMATION MESSAGE - DIFFERENTIAL'/ $
MESSAGE //' STIFFNESS AND GYROSCOPIC EFFECT GENERATION'/ $
MESSAGE //' AND ASSEMBLY FOR SUPERELEMENT'/ SEID $
MESSAGE //' '/ $
$
IF ( SCNDRY = -1 ) THEN $
$
MESSAGE //' '/ $
MESSAGE //' USER DMAP WARNING MESSAGE - SECONDARY'/
' SUPERELEMENTS'/ $
MESSAGE //' MUST BE IDENTICAL ELEMENTS (NOT MIRROR) '/
' WITH LOADING'/ $
MESSAGE //' AND DISPLACEMENTS IDENTICAL TO THE PRIMARY'/
' SUPERELEMENT'/ $
MESSAGE //' '/ $
$
ELSE $
$
$ SET QUALIFIERS
$
CALL SETQ CASES//SEID/PEID/S,MTEMP/S,K2GG/S,M2GG/
 S,B2GG/S,MPC/S,SPC/S,LOAD/S,DEFORM/S,TEMPLD/
S,P2G/S,DYRD/S,METH/S,MFLUID $
$
PARAML GPECT//'PRES'////S,N,NOGPECT $
IF ( NOGPECT >= 0 ) THEN $
$
$
$ CALCULATE CENTRIPETAL AND CORIOLIS TERMS
$
CALL CORICEN SLT,BGPDTS,EQEXINS,CSTMS,SILS,MJJ/
KCJJ,BRJJ/S,OMEGA $
$
IF ( OMEGA = 0. ) THEN $
$
MESSAGE //' '/ $
MESSAGE //' USER WARNING: NO RFORCE SPECIFIED,'/ $
MESSAGE //' GYROSCOPIC TERMS WERE NOT'/ $
MESSAGE //' ADDED TO STRUCTURAL MATRICES.'/ $
MESSAGE //' '/ $
$
ENDIF $
$
$
$ CALCULATE DIFFERENTIAL STIFFNESS
$
$message //'before emg...TEMPLD='/templd $
type parm,,i,y,tempset $
$message //'before emg...TEMPSET='/TEMPSET $
EMG EST,CSTMS,MPTS,DIT,,CYUG,ETT,EDT,,,BGPDTS,GPSNTS,,,,/
KDELM,KDDICT,,,,/
1/0/0//1//TEMPSET/DEFORM//////1/ $
EMA GPECT,KDDICT,KDELM,BGPDTS,SILS,CSTMS,,/KTJJ,/ $
PARAML KTJJ//'NULL'////S,N,NOKTJJ $
$
IF ( NOKTJJ < 0 ) THEN $
$
MESSAGE //' '/ $
MESSAGE //' USER WARNING: NO LOADING SPECIFIED,'/ $
MESSAGE //' DIFFERENTIAL STIFFNESS TERMS'/ $
MESSAGE //' WERE NOT ADDED TO STRUCTURAL'/ $
MESSAGE //' MATRICES.'/ $
MESSAGE //' '/ $
$
ENDIF $
$
$ ADD DIFFERENTIAL AND CENTRIPETAL TERMS
$
ADD5 KTJJ,KCJJ,,,/KJJGYRO/ $
PARAML KJJGYRO//'NULL'////S,N,NOGYRO $
$
IF ( NOGYRO < 0 ) JUMP ENDLOOP $
$
IF ( (DEPEND < 0 ) AND (OMEGA <> 0.) ) THEN $
$
SCALE= CMPLX( -1.0/(OMEGA*OMEGA), 0. ) $
ADD MJJ,KJJGYRO/MJJTEMP//SCALE $
EQUIVX MJJTEMP/MJJ/ALWAYS $
 EQUIVX MJJ/MJJTEMP/NEVER $
$
SCALE= CMPLX( 1.0/OMEGA ) $
ADD BJJ,BRJJ/BJJTEMP//SCALE $
EQUIVX BJJTEMP/BJJ/ALWAYS $
EQUIVX BJJ/BJJTEMP/NEVER $
$
MESSAGE //' '/ $
MESSAGE //' USER INFORMATION: SPEED DEPENDENT TERMS'/
' ADDED'/ $
MESSAGE //' TO STRUCTURAL MATRICES'/ $
MESSAGE //' '/ $
$
ELSE IF ( (DEPEND < 0) AND (OMEGA = 0.) ) THEN $
$
MESSAGE //' '/ $
MESSAGE //' USER WARNING: SPEED DEPENDENCE REQUESTED,'/ $
MESSAGE //' BUT NO RFORCE SPECIFIED.'/ $
MESSAGE //' DIFFERENTIAL STIFFNESS OR'/ $
MESSAGE //' GYROSCOPIC TERMS WERE NOT '/ $
MESSAGE //' ADDED TO STRUCTURAL MATRICES.'/ $
MESSAGE //' '/ $
$
ELSE $
$
ADD KJJ,KJJGYRO/KJJTEMP/ $
EQUIVX KJJTEMP/KJJ/ALWAYS $
EQUIVX KJJ/KJJTEMP/NEVER $
$
$ ADD CORIOLIS TERMS TO SUPERELEMENT DAMPING
$
ADD BJJ,BRJJ/BJJTEMP/ $
EQUIVX BJJTEMP/BJJ/ALWAYS $
EQUIVX BJJ/BJJTEMP/NEVER $
$
MESSAGE //' '/ $
MESSAGE //' USER INFORMATION: CONSTANT SPEED TERMS'/
' ADDED'/ $
MESSAGE //' TO STRUCTURAL MATRICES'/ $
MESSAGE //' '/ $
$
ENDIF $
$
ENDIF $ NOGPECT>=0
$
ENDIF $ SCNDRY <> -1
$
$ TRIGGER REASSEMBLEY AND REREDUCTION OF SUPERELEMENT MATRICIES
$ IN RESTART BY DELETING ASSEMBELLED AND REDUCED MATRICES
$
RESTART ,,,HISTS/TLIST/TRUE $
$
LABEL ENDLOOP $
$
ENDDO $ LPFLG>=0
$
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
$$$$
$
COMPILE CORICEN $
$
SUBDMAP CORICEN SLT,BGPDT,EQEXIN,CSTM,SIL,MJJ/KCEN,BCOR/OMEGA $
$
$********************************************************************
$ *
$ This DMAP sequence calculates Coriolis and Centripetal effect *
$ matrices in global coordinates. *
$ *
$ For node i: *
$ *
$ [Bcoriolis] = 2[x]m *
$ *
$ [Kcentripetal] = [x][x]m *
$ *
$ | 0 -Omega(z) Omega(y) | *
$ [x] = | Omega(z) 0 -Omega(x) | *
$ | -Omega(y) Omega(x) 0 | *
$ *
$********************************************************************
$
TYPE PARM,,I,N,SLTRECNO $
TYPE PARM,,RS,N,OMEGA $
TYPE PARM,,CS,N,N1C,N2C,N3C $
TYPE PARM,,CS,N,OMEGAC $
$
$ SEARCH FOR RFORCE ENTRY
$
SLTRECNO= 1
DO WHILE ( SLTRECNO > 0 ) $
$
PARAML SLT//'DTI'/S,N,SLTRECNO/1//S,N,LTYPE/ $
IF ( (SLTRECNO <> -1 ) AND ( LTYPE = 10 ) ) THEN $
$
$ EXTRACT SPIN VECTOR FROM RFORCE CARD.
$
PARAML SLT//'DTI'/SLTRECNO/4//S,N,CID/ $
PARAML SLT//'DTI'/SLTRECNO/5/S,N,RPT/ $
PARAML SLT//'DTI'/SLTRECNO/6/S,N,N1/ $
PARAML SLT//'DTI'/SLTRECNO/7/S,N,N2/ $
PARAML SLT//'DTI'/SLTRECNO/8/S,N,N3/ $
SLTRECNO = -2 $
$
ENDIF $
$
SLTRECNO = SLTRECNO + 1 $
$
ENDDO $
$
IF ( RPT = 0. ) THEN $
$
OMEGA= 0.
RETURN $
$
ENDIF $
$
$ THESE ARE NEEDED LATER ON
$
MATGEN ,/I1X1/1/1/ $
MATGEN ,/I6X6/1/6/ $
$
$ CALCULATE RADIANS PER UNIT TIME
$
OMEGA= 2.0 * PI(1) * RPT $
OMEGAC = CMPLX(OMEGA,0.) $
$
$
$ CREATE A COLUMN VECTOR, MJJTSUM, FROM THE DIAGONAL OF MJJD
$ TO BE USED FOR PARTITIONING PURPOSES LATER
$
DIAGONAL MJJ/MJJC/'COLUMN' $
$
$-------------------------------------------------------------------$
$-------------------------------------------------------------------$
$
$ CALCULATE ROTATION MATRIX IN THE CID SYSTEM
$
IF (N1 <> 0.) THEN
$
$ DETERMINE FACTOR FOR X-ROTATION (in CID system)
$
MATGEN ,/PVC6/6/6/1/1/4 $
MATGEN ,/PVR6/6/6/2/1/3 $
MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $
MPYAD ONE6X6,I6X6,ONE6X6/ROTX/1/-1 $
DELETE /PVC6,PVR6,ONE6X6,,/ $
N1C = CMPLX(N1,0.)
$
ENDIF $
$
$-------------------------------------------------------------------$
$
IF (N2 <> 0.) THEN
$
$ DETERMINE FACTOR FOR Y-ROTATION (in CID system)
$
MATGEN ,/PVC6/6/6/2/1/3 $
MATGEN ,/PVR6/6/6/0/1/5 $
MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $
MPYAD ONE6X6,I6X6,ONE6X6/ROTY/1/-1 $
DELETE /PVC6,PVR6,ONE6X6,,/ $
N2C = CMPLX(N2,0.)
$
ENDIF $
$
$-------------------------------------------------------------------$
$
IF (N3 <> 0.) THEN
$
$ DETERMINE FACTOR FOR Z-ROTATION (in CID system)
$
MATGEN ,/PVC6/6/6/0/1/5 $
MATGEN ,/PVR6/6/6/1/1/4 $
MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $
MPYAD ONE6X6,I6X6,ONE6X6/ROTZ/1/-1 $
DELETE /PVC6,PVR6,ONE6X6,,/ $
N3C = CMPLX(N3,0.)
$
ENDIF $
$
$-------------------------------------------------------------------$
$
$ ADD FOR TOTAL ROTATION (in CID system)
$
ADD5 ROTX,ROTY,ROTZ,,/ROTTOT/N1C/N2C/N3C $
$
$-------------------------------------------------------------------$
$-------------------------------------------------------------------$
$
$ GENERATE G-SIZE CORIOLIS MATRIX IN THE GLOBAL SYSTEM
$
$
$ FIRST, GENERATE G-SIZE MATRIX IN THE CID SYSTEM
$
ADD ROTTOT,/OMEGA/OMEGAC $
MATMOD OMEGA,SIL,,,,/OMEGAGG,/5 $
$
$-------------------------------------------------------------------$
$
$ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS
$ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5)
$
PARTN OMEGAGG,MJJC,/,,,OMEGAXX/ $
$
$-------------------------------------------------------------------$
$
$ TRANSFORM CORIOLIS ACCELERATION MATRIX TO GLOBAL SYSTEM
$ FIRST, TRANSFORM FROM CID TO BASIC
$ THEN , TRANSFORM FROM BASIC TO GLOBAL
$
$ CHECK IF THE CSTM EXISTS. IT WILL NOT EXIT IF THE GLOBAL SYSTEM
$ IS THE SAME AS THE BASIC SYSTEM. ERROR IN MATMOD. THEREFORE
$ GENERATE THE IDENTITY MATRIX
$
$
PARAML CSTM//'PRESENCE'////S,N,NOGTOB=0 $
IF (NOGTOB < 0) THEN
DIAGONAL MJJ/GTOBY/'SQUARE'/0. $
ELSE
MATMOD CSTM,SIL,BGPDT,,,/GTOBY,/5//-1 $
ENDIF $
$
PARTN GTOBY,MJJC,/,,,GTOB/ $
$
$ CHECK IF THE CID IS EQUAL TO ZERO (BASIC)
$
IF (CID = 0) THEN
EQUIVX GTOB/CTOG/ALWAYS $
ELSE
MATMOD CSTM,SIL,BGPDT,,,/CTOBX,/5//CID $
$
$ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS
$ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5)
$
PARTN CTOBX,MJJC,/,,,CTOB/ $
MPYAD CTOB,GTOB,/CTOG/1 $
ENDIF $
$
SMPYAD CTOG,OMEGAXX,CTOG,,,/OMEGAYY/3////1 $
MERGE ,,,,OMEGAYY,MJJC,/OMEGAGGX/ $
$
 $-------------------------------------------------------------------$
$
$ MULTIPLY MASS BY CORIOLIS ACCELERATION
$ TO CALCULATE CORIOLIS FORCE MATRIX
$
MPYAD MJJ,OMEGAGGX,/CJJ1 $
MPYAD OMEGAGGX,MJJ,/CJJ2 $
ADD CJJ1,CJJ2/BCOR $
$
$-------------------------------------------------------------------$
$-------------------------------------------------------------------$
$
$ CALCULATE CENTRIPETAL ACCELERATION MATRIX
$
$ MULTIPLY MASS BY CENTRIPETAL ACCELERATION
$ TO CALCULATE CETRIPETAL FORCE MATRIX
$
SMPYAD OMEGAGGX,MJJ,OMEGAGGX,,,/KCEN/3////////6 $
$
RETURN $
END $
$
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
$
echoon $
@
38062; 2001; Cyclic Symmetry - differential stiffness, modes, tempi - UFM 434
2, SFM 3001
In a cyclic symmetry run that uses the differential stiffness (e.g.,
buckling and normal modes with preload), the applied loads must be
uniform. To enforce this requirement, the loads must be applied to
segment one only. If loads are applied to segments other than one, then
the job will fail with following error messages:
*** USER FATAL MESSAGE 4342 (CY1BLD)
THE LOAD BULK DATA CARD GIVEN BELOW IS INCONSISTENT WITH SPECIFIED
ANALYSIS HARMONICS OR SEGMENTS OR SYMMETRY TYPE.
LOADCYN xxxx yyy zzz
To avoid this message, remove loads from segments greater than one.
Furthermore, if the preload consists of thermal loads, then the job will
fail with the following error messages:

*** SYSTEM FATAL MESSAGE 3001 (CYCLIC4)

THE INPUT DATA BLOCK NAMED PAC AT POSITION 4 DOES NOT
EXIST.
Avoidance: Insert the following dmap alter.
compile phase1dr
alter 'lcgen casers'(,1)
if ( app='bkl0' and cyclic ) equivx casers/casew/-1 $
endalter
@
38070; 70.7.2; Parallel Processing - DMP - System error message at completion o
f job
DMP jobs may report the following error text at job completion:
 ERROR 0031-636 User requested child or EOF termination of pm_command
This message does not affect the MSC.Nastran job.
Avoidance: Delete the following line from <install-dir>/bin/nast707.dmp:
%MSC_ARCH=aix:%export MP_NEWJOB=yes
@
38073; ; Cyclic Symmetry - RESTART, STATSUB - Inefficient Restart
If a restart from SOL 114 (Cyclic static) into SOL 115 (Cyclic modes) is
performed to calculate pre-stressed modes of the model, the static
solution will be re-calculated in SOL 115.
There is no avoidance.
@
38096; ; Miscellaneous - Accounting file, Solution type and CPU time are
not written
The Version 70.7 accounting program does not properly identify the
solution name, CPU, or elapsed times in the account files.
Avoidance: Use the following procedure to modify and install a new
accounting logging program.
1) Change the current working directory to <install-dir>/msc707/util
2) Change lines 1440 through 1443 in mscact.c from
if( (f04 = fopen( f04f, "r" )) == NULL ) {
(void) strcat( f04f, ".f04" );
f04 = fopen( f04f, "r" );
}
to
(void) strcat( f04f, ".f04" );
f04 = fopen( f04f, "r" );
That is, delete the "if" statement and its matching closing brace.
3) Build a new version of mscact with the command
msc707 util mscact
4) Install the new account logger program with the command
cp -p mscact ../<arch>/acct
chmod a+rX ../<arch>/acct
where <arch> is the architecture of the current platform, i.e., aix,
alpha, hpux, irix, irix64, solaris, sppux, superux, unicos, unicosc90,
unicosts, or uxpv.
@
38097; 70.7; Installation - FLEXlm - /etc/inittab file is wrong.
mscsetup for MSC.Nastran Version 70.7 does not set the /etc/inittab line
 with the full path of the flexlm utility. If the system is subsequently
rebooted FLEXlm will not run automatically.
Avoidance: If FLEXlm is already installed, and you want the new version of
FLEXlm, install it but do not have mscsetup set the /etc/inittab file; if
FLEXlm is being installed for the first time, modify the /etc/inittab file
after the installation is complete and add the full path for the
<install-dir>/bin/flexlm utility.
@
38100; 70.7; Installation - No Output
When installing MSC.Nastran on Intel computers, the last step in the
installation asks if you would like to enter the authorization code. If
you choose not to enter the code, the program does not create the
authorize.dat file, but the nastxxx.rcf file will incorrectly point to the
non-existant file. If you submit a Nastran job, no output files will be
created.
Avoidance:
Either:
1) enter the authorization code when installing.
2) edit the nastxxx.rcf file to remove the "AUTH=" line
3) when you try to run MSC.Nastran, enter the following on either the
command line or in the "optional keywords" box
pause=yes
If you do this, the command window trying to run the program will stay
open and you will obtain the error message indicating that the
authorize.dat file does not exist.
@
38115; 70.7; Installation - Missing library libxlsmp.a[smprt.o]
IBM AIX clients that don't have the Fortran compiler installed will be
unable to run MSC.Nastran V70.7.0 and V70.7.2 until the libxlsmp.a library
is installed on their system or in the <install-dir>/msc707/aix directory.
The following message is issued in the log file:
MSC/NASTRAN started Mon Jan 03 09:46:03 EST 2000
Could not load program /tmp/vxlq/n707/msc707/aix/analysis
Could not load library libxlsmp.a[smprt.o]
Error was: No such file or directory
Avoidance: Install the missing library.
@
38116; ; Superelements - CSUPER, PARAM,INREL - SFM 5423 (DMPYAD)
If SOL 101 with inertia relief is used with external superelements, the
external superelement data recovery run will fail with:
*** SYSTEM FATAL MESSAGE 5423 (DMPYAD)
ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES
indicating that UOO and CRX are not correct. This is due to an error which
results in CRX not being written to the EXTDRUNT.
Avoidance: Include the following DMAP alter in the residual structure run:
 compile super3
alter 339
dbview crx0=crx (where seid=0 and wildcard) $
equivx crx0/ol1m/-1 $
@
38119; 2001; Postprocessing - SOL 105, PARAM,POST,0 - Missing Output
Under PARAM,POST,0 in SOL 105, the buckling eigenvalues and eigenvectors
will not be available in MSC.Patran.
Avoidance: Use PARAM,POST,-1 or insert the following DMAP alter:
compile super3
alter 'CASEdr,ol2s'(,-1)
app=app1 $
alter 'CASEdr,ol2s'
app='reig'
endalter
@
38123; 2001; Optimization/Design Sensitivity - DEQATN, DRESP2 - Wrong Answers
If multiple DRESP2's refer to the same DEQATN, only the first (lowest
numbered DRESP2 is checked to see if the number of DRESP2 arguments
agrees with the number DEQATN arguments. If a subsequent DRESP2 has a
mismatch with the number of arguments, the error goes undetected and
unpredictable answers are produced.
Avoidance: There is no avoidance other than not sharing a DEQATN id on
multiple DRESP2's. If this is too tedious, the user must be vigilant in
checking that the number of arguments on the DRESP2 agree with the number
of arguments on the left hand side of the DEQATN in every DRESP2 except
the lowest numbered one.
@
38127; 2001; Optimization/Design Sensitivity - DRESP2, DSAPRT - Incorrect Sen
sitivity Data
When the formatted prints of sensitivity data (case control command
DSAPRT) for DRESP2 responses list a frequency value of the response, the
value has been incorrectly scaled upward by a factor of ten. A similar
error exists for time values. Also, values that are greater than 100.
appear as *******.
Avoidance: Frequency/time values for DRESP1 responses are printed
correctly. The user must be aware that printed DRESP2 frequency/time
values are 10 times greater than they should be. It should also be noted
that when a DRESP2 response spans several frequency, only the lowest
frequency is printed.
@
38137; ; SSSALTERs - errors in alter1ga.vxx and alter9ga.vxx
SSSALTERs alter1ga.vxx an dalter9ga.vxx have errors in them when using OTM
(Output Transformation Matrices) with OUTPUT2 or in frequency response.
alter1ga.vxx does not correctly write the tables identifying the terms in
the OTM if OUTPUT2 is used. alter9ga.vxx does not correctly use the OTM if
it is used in frequency response.
Avoidance: Contact MSC and request an updated version of the alters or
get them from the MSC web page at:
 http://www.mechsolutions.com/support/prod_support/nastran/sssalters/index
html
@
38141; 70.7; Miscellaneous - OUTPUT2, Year 2000 - Incorrect Year Printout
When running nastran jobs in year 2000+, on some unix machines, the
output2 trailer information will print out the date incorrectly. For
example, when using param,post,,-1, the trailer information for OEF1
coming out right after UIM 4114 with a job ran on January 6, 2000 will
appear as follows:
1001 63 1 6 100 0 1
instead of the correct value
1001 63 1 6 0 0 1
In this case field 5 is the year. This problem is fixed in Version 70.7
Avoidance: There is no avoidance. Note that the results are not affected
by the date printout.
@
38143; 2001; Documentation - SYSTEM(173) - Old QUAD4 Formulation
SYSTEM(173) allows the selection of either the pre-V68 QUAD4 formulation
or the V68-V70.5 formulation. This system cell is incorrectly documented
in the Version 70.7 Release Guide. The documentation should state:
1: Requests pre V68 QUAD4 Formulation
2: Requests V68-V70.5 QUAD4 Formulation
Likewise the Version 70.7 Release Guide documentation for the keyword
OLDQ4K incorrectly states that this keyword is equivalent to setting
SYSTEM(173)=2, the keyword is equivalent to setting SYSTEM(173)=1.
@
38167; ; Installation - nastran command: fails with multiple 70.7.X versi
ons
If V70.7.2 of MSC.Nastran is installed in the same directory as V70.7.0,
then both cannot run. The intent was to allow users to use "nast7070" for
V70.7.0 and "nast707" or "nast7072" for V70.7.2. This feature was broken
in the V70.7.x releases.
Avoidances:

1. Install V70.7.2 in a directory with a different MSC_BASE than V70.7.0

2. Use V70.7.2 and not V70.7.0
@
38168; ; Superelements - Incorrect OUTPUT2 units for different SEs w/ PAR
AM OUNIT2
If PART superelments are used, and different values are provided for
PARAM,OUNIT2 in different PARTs, the program will use these values
incorrectly.

1) Setting distinct PARAM, OUNIT2 values in each of the Partitioned Bulk

Data Sections is ignored - only the setting for the residual structure is
used and the OUNIT2 values set under the BEGIN SUPER=xxx are ignored.
 2) Setting distinct PARAM, OUNIT2 values PER SUBCASE using expanded Case
Control causes:
A) All geometry for all SE to be written to the 1st OUNIT2 specified
B) End of Data record on that 1st OUNIT2
C) Results for the residual are then written to that 1st OUNIT2 file
D) The 1st OUNIT2 file is rewound and re-labeled (wiping out the geometry
and SE0 results) and the correct SE results are written.
E) Labeling, results and End-of-Datas are written out for the remaining
SEs.
There is no avoidance.
@
38180; 2001; Frequency Response - DMP>1, SOL 108, fluid-structure decks - Wr
ong answers
If distributed memory parallel, dmp>1, is used with a fluid-structure
in SOL 108, then wrong answers result.
Avoidance: Insert the following DMAP alter:

compile freqrs
alter '197',''
if ( acsyml and adjflg<2 and
not(GETSYS(np,265)>1 and GETSYS(np,197)>0 and
soltyp='modal') ) then $
endalter
@
38195; 2001; Plotting - .plt File Contains Bad Date String
The plot file will contain an invalid date string on any file generated
after Dec 31, 1999. The error is present in all versions of MSC.Nastran,
up to and including V70.7.2. On Version 70.7 the year 2000 was written
simple as "0", for example 1-10-0; however other strings are possible.
This error does not cause any other problems with the plot data.
Avoidance: There is no known avoidance.
@
38207; 2001; Design Sensitivity/Optimization - Static Aero, METHOD - UFM 1126
for M1RR
If the Case Control for a multi-subcase static aeroelastic solution in
SOL 200 contains a METHOD command that varies from subcase to subcase, a
FATAL error can occur in which the M1RR matrix is recomputed. Since the
M1RR matrix is not declared OVRWRT, this causes UFM 1126 shown below. The
error occurs when the METHOD command is the ONLY boundary difference
between the two sets of subcases.
*** USER FATAL MESSAGE 1126 (GNFIST)
DMAP MODULE MPYAD IS ATTEMPTING TO OUTPUT DATABLOCK 201 NAME
= M1RR , WHICH ALREADY EXISTS.
USER ACTION: DELETE THE DATABLOCK, OR USE FILE M1RR=OVRWRT
0FATAL ERROR
Avoidance: Since static aero analysis in SOL 200 is not affected by
METHOD selection, remove the extraneous METHOD selection(s) from the Case
Control or insert the following alter:
 compile desinit
alter 'case '(,-1)
if ( isolapp=3 ) then $
case casers,/casem/'comm'/s,n,nskip///////
'mpc'/'spc'/'supo' $
else $
alter 'case '
endif $
endalter
@
38221; 2001; Documentation - PSHELL, MAT2 - Missing Stress Output for Plate E
lements
When a plate element references a MAT2 entry with an MID larger than
99999999, it is assumed that this MAT2 entry has automatically been
created from a PCOMP entry (see QRG, MAT2 entry, Remark 13). Thus, stress
output will be missing unless PARAM, NOCOMPS, -1 is specified.

Avoidance: Do not use MIDs of MAT2 entries larger than 99999999, or use
PARAM, NOCOMPS, -1.

@
38237; 2001; Documentation - slaveout=yes, mergeresults=no - Clarify Document
ation for DMP
Distributed Memory Parallel executions may specify the keywords slaveout
and mergeresults on the submittal line. These keywords are documented in
the V70.7 Release Guide. The Release Guide should have clarified and
emphasized that the non-default values of these keywords, slaveout=yes and
mergeresults=no are typically only used for debug purposes. The following
should also be noted:
1. If mergeresults=no is specified in a static run the results of the
individual domains will not be sent back to the master and the system
solution will not be obtained.
2. The keyword mergeresults has no affect on a solution 103 or 111 run.
3. In solution 108, if mergeresults=no is specified and slaveout=yes is
not specified, then the results of the slave processors will be lost.
4. In solution 108, it is possible to get a through-put advantage by
saving communication between the master and slaves when mergeresult=no and
slaveout=yes is specified.
@
38238; 2001; Cyclic Symmetry - SFM 3001 (READ), Poor Message
In SOL 115, if the stiffness and mass matrix have common null-columns the
following message will be issued:
*** SYSTEM FATAL MESSAGE 3001 (READ)
THE INPUT DATA BLOCK IN POSITION 13 DOES NOT EXIST.
Avoidance: Make sure degrees of freedom with no stiffness and no mass are
constrained. In Version 70.7 parameter AUTOSPC will do this, but in
Version 70.5 AUTOSPC would not correctly constrain grids on CYJOIN
entries. A better error message is implemented in Version 2001..
@
38254; 2001; Optimization/Design Sensitivity - PBAR - UFM 5626 (DSAD17)
Optimization analysis does not support item codes 18 and 19 for PBAR for
complex stress. If specified the following fatal message is issued:
 *** USER FATAL MESSAGE 5626 (DSAD17)
ITEM CODE 18 REFERENCED ON DRESP1 ENTRY 16 WITH RTYPE =
FRSTRE
IS INVALID FOR ELEMENT TYPE BAR 34.
USER ACTION: CHECK THE SPECIFIED ITEM CODE WITH THOSE LISTED IN
APPENDIX A OF THE MSC/NASTRAN QUICK REFERENCE GUIDE.
Avoidance: There is no avoidance.
@
38257; 2001; DMAP - FBS - Wrong Answers, BIOMSG: ERROR 31
The FBS module may return poor fatal diagnostics (BIOMSG: ERROR 31) or
wrong answers when a left hand fbs is requested incorrectly. An example of
a left hand fbs follows.
The conventional (right hand) equation solved by the FBS module is:
Given the square matrix A, and a collection of vectors X, find B, where
A*X< = B $ FBS module
A is first decomposed into its factors L and U. $ DECOMP OR DCMP modules.
Typically, X and B have many fewer columns than A.
There are applications where X has many more rows than A. That type of
problem is most naturally posed as
X*A<= B
The unknown vectors X have been moved to the left hand side, hence the
name LH FBS. One method to solve this problem is to transpose it.
Assume, for the sake of simplification, that A is symmetric, that is,
A=A'. Then
A*X' = B'
This form of solution requires transposing the input before solution, and
the output after solution. The present lh fbs option still requires
transposing the input X, but the output B need not be transposed. It also
requires use of system cell 72. Tests to date prove that this option
works correctly. An investigation is underway to determine what other
solvers can provide LH FBS operations. Another complication results when
a forward-only or backward-only solution is requested. See the DMAP
Dictionary for a definition of these terms. The definition of forward and
backward is somewhat ambiguous in the context of a transposed equation.
The definition used in MSC.Nastran is given by example below. It is a
good practice when using these advanced forms to verify that you are
obtaining correct answers, at least while debugging your alter. The
revised equation, and a dmap verified by testing to provide correct
answers, is:
$ A*X = B', or X'*A = B

Decomp A/L,,//1 $ 1= CHOLESKY, A=L*L', A must be symmetric psd

$ demonstrate solution with RH FBS for comparison
message //'rh solution, requires two extra transposes' $
trnsp b/bt
 Fbs L,,Bt/X1t///1 $ 1=fwd only
Trnsp X1t/X1 $
Message //'lh solution follows.' $
Type parm,,I,n,c72 $ used to store current value of lh fbs flag
C72=GETSYS(C72,72) $ capture current value
PUTSYS(1,72) $ required for LH
Fbs L,,Bt/X2/2//1 $ 2=LH, 1=fwd only
PUTSYS(C72,72) $ reset default for lh fbs selection. Always reset,
$ or wrong answers may occur in downstream FBS module or subroutine
$ calls
Add X1,X2/error//-1.0 $ error = X1 - X2
Norm error/screrr///s,n,maxerror $
Message //'maximum error in lh fbs. Should be binary zero'/
Maxerror $
There have been reports of errors when using system cell 74, a function
described in the Numerical User's Guide. It is recommended that dmaps
that use this feature be checked for correct answers. There are alternate
procedures where one can avoid the use of this system cell by tranposing
the input and/or ouput set of vectors seen by the FBS module.

@
38263; 2001.0.1; Superelements - Multi-Master, Data Recovery - Missing Output
In a dynamic response analysis using the mulit-master technique, no
output will be generated for the superelements. This error only occurs
in Version 70.5, 70.6, and 70.7.
Avoidance: Insert the following dmap alter.
compile super3
alter 'equivx cases \/casedr1',''
equivx casedr/casedr1/-1 $
endalter

@
38268; 70.7.3; Transient Response - TRLG - Loss of Accuracy
Transient loads are computed in single precision in module TRLG. One
effect that has been observed is the accumulated time after many delta-T
steps is in error due to numeric underflow. This round-off error
typically only occurs if the duration, T is greater than 100000 times
delta-T. Loads such as COS(W*T) may appear to have a slight frequency
shift.
This problem is more likely to occur on 32-bit computers than on 64-bit
computers.
@
38276; 2001; Documentation/QRG - XYPLOT - XMIN Incorrectly Described
In the Quick Reference Guide, remark 2 on the XMIN (XYPLOT) command
incorrectly states that if XMIN is not specified, it is set to the lowest
value of "Y". This should state that it is set to the lowest value of "X".
@
38282; ; Elements - TETRA, SYSTEM(213), DISTORT - UFM 4659, 5271
Using NASTRAN system 213=1 to override the element distortion fatal
 message, USER FATAL MESSAGE 4659, for TETRA elements does not work.
Avoidance: Set a high value for the maximum acceptable aspect ratio on
system(191).

@
38283; ; Parallel Processing - DMP - SGI Only, arrayservices not availab
le
When running distributed parallel (DMP) on the SGI, the job may fail with
the following error messages in the log file:
'MPI: asgetnetinfo_array('(null)') failed : array services not available
S L A V E 1
Avoidance: Change the authentication method in the
/usr/lib/array/arrayd.auth from
AUTHENTICATION NOREMOTE
to
AUTHENTICATION NONE
and make sure you start the array services with '/usr/etc/arrayd start' as
root. Furthermore, make sure that the user account can do a 'rsh
shu01irx date' on the shu01irx machine without running into a 'permission
denied' problem.

@
38293; 2001; Aeroelastic Analysis - Superelements, SPLINE, AMP- SFM 4276 EC
1044 QOPEN
In Version 70.7, a flutter analysis (SOL 145) with superelements may fail
with SFM 4276, EC 1044 (QOPEN) in the AMP module if the superelement
reduction is applied using separate subcases AND the separate subcases do
not include the FMETHOD bulk data entry .The problem occurs because the
splining to upstream superelements, combined with work in Version 70.6 to
support multiple aerodynamic symmetries has coupled the spline matrix data
to both the aerodynamic configuration (AECONFIG) and to the
superelement/boundary condition qualifier paths. To ensure that the
correct linkage among configurations and boundary conditions is performed,
the Case Control is examined for aerodynamic cases to determine the
AECONFIG name of each subcase. In the case of superelement reduction
subcases, these fictitious cases are examined and, since the FMETHOD entry
is not present, they are ignored as aerodynamic cases. As a result, the
spline matrix reduction to the superelement was not done. Thus, in AMP,
these data are missing which results in this error.
Avoidance: Insert the following DMAP alter:
compile phase1a
alter 'AEMODEL.*CASES,EDT',''
dbview case0=cases where (seid=0 and wildcard) $
AEMODEL CASE0,EDT//
S,N,NSKIP/S,N,LPFLG/
S,N,AECONFIG/S,N,SYMXY/S,N,SYMXZ $
compile phase1dr
alter 'call phase1b'(,-1)
dbview case0=cases where (seid=0 and wildcard) $
equivx case0/case00/-1 $
 call dbstore case00,,,,//0/0/' '/0 $
compile phase1b
alter 'AEMODEL.*CASES,EDT',''
call dbfetch /case00,,,,/0/0/0/0/0 $
AEMODEL CASE00,EDT//
S,N,NSKPAE/S,N,LPFLG/S,N,AECONFIG/S,N,SYMXY/S,N,SYMXZ $
endalter
or make sure that all subcases in the SOL 145 superelement run contain the
FMETHOD selector or that the FMETHOD appears above the first subcase.
The splining to upstream superelements, combined with work in V70.6 to
support multiple aerodynamic symmetries has coupled the spline matrix data
to both the aerodynamic configuration (AECONFIG) and to the
superelement/boundary condition qualifier paths. To ensure that the
correct linkage among configurations and boundary conditions is performed,
the Case Control is examined for aerodynamic cases to determine the
AECONFIG name of each subcase. In the case of superelement reduction
subcases, these fictitious cases are examined and, since the FMETHOD entry
is not present, they are ignored as aerodynamic cases. As a result, the
spline matrix reduction to the superelement was not done. Thus, in AMP,
these data are missing which results in this
error
@
38295; 2001; Superelements - RESTART, SEBULK, CSUPER - Wrong answers or fatal
If external superelements are used in a solution, then a change is made to
the external superelement (new matrices generated) for use in a RESTART
run, the updated matrices will not be used.
Avoidance: If you are using external superelements created using
PARAM,EXTOUT,DMIGOP2, the following alter (for V70.5) will provide support
for restart changes to the matrices:
compile extin, list
alter 'FORTIO.*EXISTS.*EXTUNIT.*IOSTAT '
if(iostat=0)then $
dbdelete datablk=(extdb), where(xname=*) $
endif $

@
38304; 70.7; Frequency Response - STRAIN(SORT2) - Missing Output
If STRAIN(SORT2) output is requested in a frequency response run and there
are no other output requests, there will be no strain output produced.
Avoidance: Add any other (SORT2) output request - for example
DISP(SORT2,PLOT)=ALL - to the case control.
@
38306; 2001; Dynamic Analyis - DAREA, LSEQ - Unpredictable Answers
Non-unique Set-ID/DAREA-ID pairs in LSEQ data lead to unpredictable
results. This is because the user cannot be sure as to which of the
static loads specified as part of the non-unique data are actually applied
and which ones are ignored. Accordingly, this scenario must be regarded
as a fatal error. The user can avoid this situation by using and
referencing the combination LOAD bulk data instead of non-unique
Set-ID/DAREA-ID pairs.
@
38308; ; Heat Transfer Analysis - CHBDYE, QUAD4 - UFM 6285
If a heat transfer analysis attempts to calculate view factor from the
edge of plate elements using CHBDYE with sides from 2 to 5, then you will
get UFM 6285
Avoidance: Define the line segment with CHBDYP, LINE boundary elements.
@
38317; 2001; ACCESS - Utilites error forttl:severe(25) record number outside
range
After a successful build, the following error may occur during execution
of any of the sample ACCESS utilities.. The error may occur for both v70.5
and v70.7.
"forttl:severe (25) record number outside range, unit 40, file
c:\filename.xdb"
Avoidance: Download new dbio object libraries, libdbio_vc.lib and
libdbio_irc.lib, from MSC's ftp site.
@
38323; 2001; Utilities - BUILDSSS - UFM 701 PRCMPL
In Version 70.7, when using the buildsss utility to create a new delivery
database with more than one added user dmap, the following error occurs:
*** USER FATAL MESSAGE 701 (PRCMPL)
THE NUMBER OF SUBDMAPS BEING COMPILED EXCEEDS THE MAXIMUM ALLOWED
NUMBER OF 200 FOR THIS VERSION
Avoidance: Compile the additional dmaps in a separate run and add them to
the delivery. Contact MSC Support for details on how to perform this.
@
38337; ; Submittal - Large physical memory (>8GB) may cause failure of na
stran command
Machines with physical memory in excess of 8GB may cause the nastran
command to fail with a message similar to:
MSC/NASTRAN V70.7 (SUN SPARC SunOS 5.7) Wed Jan 26 14:23:19 2000
*** SYSTEM INFORMATION MESSAGE (pgm: nast707, fn:
validate_buffsize_memory)
Assuming "memory=-2147483648" as specified by
"memorymaximum=0.8*physical".
In some instances, an error will not occur, instead the nastran command
may generate an incorrect memory request.
Avoidance: Set the maximum memory size in a system RC file to the value
listed in the "MSC.Nastran V70.7 Configuration and Operations Guide",
Table 4-2, pg 90. For example, on Solaris:
memorymaximum=4GB
@
38349; 2001; P-Elements - UFM 9050 DCMP (70.5) or UFM 3001 VECRMA (70.7)
It is [possible that in a model containing both p- and h-elements that the
 run may fail with either UFM 9050 or UFM 3001.
Avoidance: This problem is under investigation. The model demonstrating
this problem executed successfully when more memory was allocated or the
BUFFSIZE was increased.
@
38365; 2001; Elements - PARAM,SNORM - Wrong Answers
The maximum allowed value of the SNORM parameter is 45 degrees. Higer
values are automatically set lower. For shells meeting at a large angle,
such as a right angle, with a
different number of elements on each side, the calculated normal may be
greater than 45 degrees from one side. This is legitimate, but will fail
the error check.
@
38373; 2001; Rigid Elements and Multipoint Constraint - RBEi, MPC - UFM 5289
If a dof is listed as depednt more than once on an MPC or RBEi entry,
Nastran will issue UFM 5289, indicating an error. Unfortunately, the
message may point to the wrong dof and may not point to the entry which
has the error.

Avoidance: There is no avoidance other than inspecting the MPC and rigid
element entries.
@
38376; 2001; Static Analysis - DMP>1, GPSTRESS and PARAM,POST,0 - Missing out
put in xdb
If grid point stresses are requested with dmp>1 and param,post,0 then the
xdb file will be missing the grid point stresses for the slave domains.
Avoidance: Insert the following DMAP alter:
compile sedrcvr
alter 'GPLS,EGPSTR'(2),''
DBC OPG1,OUGV1,OEF1X,OES1X,OQG1,GPL,EGPSTR,EGPSF,GPDCT,
ELDCT,error1,,,,,,,,,//
'OPG'/'OUG'/'OEF'/'OES'/'OQG'/'GPL'/'GPS'/'SVF'/
'GPDCT'/'ELDCT'/'perror'//////////
-1/DBCPATH/S,N,CP/APP1/ICYCLIC/GEOMU/LOADU/POSTU/
DBCDIAG/DBCCONV/DBCOVWRT/DESCYCLE/////adptindx/lusets $
endalter

@
38378; 2001; DMAP - MATMOD option 12 - Wrong answers, Missing Fatal Message
If the first input, I1, to the MATMOD module, under option 12, is purged
then the module will incorrectly report that NONULL=0 and the O1 output is
purged without warning or fatal even if the second and third inputs (I2
and I3) are not purged and contain null columns. The documentation for
MATMOD option 12 contained in the "DMAP Module and Data Blocks" guide
states that only I2 and I3 may be purged, which implies that a fatal
message should be issued if I1 is purged.
Avoidance: Rearrange the inputs so that the first input is not purged.
@
38381; 2001; Optimization/Design Sensitivity - DRESP1 - UFM 6482 from DOPR3R
USER FATAL MESSAGE 6482 indicates that an invoked DRESP1 entry in a SOL
 200 subcase is not compatible with the ANALYSIS type for that subcase.
Although the message points to a real user input error, it is possible
that the identified DRESP1 type is not the type that caused the error.
Avoidance: The UFM has the correct RID for the DRESP1. Use the RID and
not the response type to identify the incorrect input.
@
38396; ; Superelements - NASTRAN LOCBULK=2 - SFM 4192, SFM 3062 in SEMA
If the multi-master method is used on a Cray and LOCBULK is set to 2, the
run may fail with SFM 4192, indicating undefined grid points.
*** SYSTEM FATAL MESSAGE 4192 (SEMA)
GRID POINT .... IN SUPERELEMENT 0 IS NOT DEFINED IN
BGPDT FOR DOWNSTREAM SUPERELEMENT
and
0*** SYSTEM FATAL MESSAGE 3062 MODULE = SEMA
FAILURE OCCURRED IN UTILITY MAPBTI CALLED IN SUBROUTINE SEMA .
REQUESTED EXTERNAL ID(S) NOT FOUND.

Avioidance: If possible, do not use the multi-master technique, run

using a different computer type, or run using LOCBULK=1
@
38398; 2001; Optimization/Design Sensitivity - DSAPRT, DRESP1 - SFM 4276
If the fomatted sensitivity option is used and there are no retained
responses associated with DRESP1 entries, but there are retained DRESP2
responses, SYSTEM FATAL MESSAGE 4276 and a traceback in dsaprt is
produced.
Avoidance: Use the unformatted sensitivity option or force the retention
of a DRESP1 response.
@
38401; 2001; Restarts - Superelements - Changes in SPCs may yield Wrong answ
ers
If in a superelement restart run, changes are made to the constraints
(spcs) of only one superelment, then wrong answers may result.
Avoidance: Insert the following DMAP alter:
compile phase0
alter 'IF ( LRESTART )'(2)'
IF (KCHNG) CALL CHNGFLG //KCHNG/SEDWN/SEID/PEID $
IF (MCHNG) CALL CHNGFLG //MCHNG/SEDWN/SEID/PEID $
IF (BCHNG) CALL CHNGFLG //BCHNG/SEDWN/SEID/PEID $
IF (PCHNG) CALL CHNGFLG //PCHNG/SEDWN/SEID/PEID $
IF (K4CHNG) CALL CHNGFLG //K4CHNG/SEDWN/SEID/PEID $
endalter
@
38403; ; Postprocessing/Aeroelastic Analysis - TRANS utility does not wor
k with Aero
The trans utility prorgram delivered with Version 70.6 and 70.7 will not
convert a solution 145 xdb file to a ndb file. If executed the utility
yields a message similar to the following:
 TRANS - V70.7 (23-NOV-1998)
XDB FILE: mlgo_flutter2_test.xdb
NEUTRAL FILE: mlgo_flutter2_test.ndb
MSC DATA BASE TRANSMIT PROGRAM - VERSION 70.7COMPILE DATE:
30-AUG-1999

due: [22] Invalid argument

logical unit 42, named 'fort.42'
lately: reading direct unformatted external IO
Abort (core dumped)
Avoidance: The only avoidance is to execute the run on an ieee compatible
system where the post-processing will be performed so that trans is not
necessary.
@
38411; 2001; Optimization/Design Sensitivity - DVPREL2, APRCOD, DOM9 - SFM 42
76, EC1404
If an input file for optimization has DVPREL2 entries and APRCOD is set to
3 on the DOPTPRM entry, the run may fail and issue SYSTEM FATAL MESSAGE
4276 ERROR CODE 1404 in the DOM9 module. .
Avoidance: 1) If the file contains DVPREL2 entries, do not set APRCOD to 3
or
2) If the deck contains DVPREL1 entries, force them to be
retained using a DSCREEN entry.
@
38412; GL; Submittal - "pwd: cannot access parent directories" error
Under not-fully-understood circumstances a small subset of C shell users
may encounter the following error message:
$ /nast/deliv/msc707/unix/bin/msc707
/nast/deliv/msc707/unix/bin/msc707[42]: pwd: cannot access parent
directories
/nast/deliv/msc707/unix/bin/msc707[97]: cd: bad argument count
when submitting a nastran job.
Avoidance: One avoidance is to use the Korn shell (ksh) as the login
shell.
@
38428; ; Aeroelastic Analysis - PAERO4, AEFACT - UFM 6171 (APD4)
If a PAERO4 entry has fewer strips that requested, the program will issue
USER FATAL MESSAGE 6171 (APD4) and terminate. Unfortunately, the fatal
message will incorrectly state that the error is on an AEFACT entry with
the same id as the PAERO4.
Avoidance: Provide the correct number of strips on the PAERO4 entry.
@
38443; 2001; Documentation - DEQATN - Unexpected Results
The documentation for the DEQATN entry does not mention that the
arithmetic is carried out using the type of the input data. For example,
if a term is raised to a fractional power:
 X=A**(1/2)
since both values in the exponent are integers, the value returned for the
exponent is calculated using integer arithmetic or 1/2=0.
Avoidance: Excersize caution when writing equations. If possible, avoid
integer terms.
@
38449; 2001; GPFORCE - MPCFORCE, TEMP(LOAD) - Missing APP-LOAD Output
If there are thermal loads and mpcforces, GPFORCE output may be missing
the applied load (APP-LOAD) output.
Avoidance: There is no avoidance.
@
38455; ; Documentation - Configuration & Operations Guide - Maximum Memor
y on HP-UX
Tables 4-2 and C-7 of the "Configuration and Operations Guide" incorrectly
state the maximum memory size on HP-UX systems is 2 GB. The maximum memory
size on HP-UX systems is:

1 GB (256 MW) on HP-UX 10.20

2 GB (512 MW) on HP-UX 11.00, PA-RISC 2.0
The actual value usable by the "memory" keyword is reduced from this by
approximately 35 MB.
@
38458; ; Nonlinear Analysis - QUAD4 - Plastic In-Plane Bending, Wrong Ans
wers
If a nonlinear analysis is performed with QUAD4 elements with elastic,
perfectly plastic material and the elements are loaded in pure shear and
twise (no bending, no axial load), then the results may incorrectly
indicate that the model can sustain increasing loads even after the
elements in the critical load path have reached the inplane shear plastic
limit. This problem only occurs for coarse meshes subjected to pure shear
loads. It is not likely to affect real world models.

@
38468; 70.7.3; Miscellaneous - GP4, Possibly Other Modules - Infinite Runtime
Sort routines in nastran have an upper limit that may be exceeded if
models have many of one type of element. The deck that encountered this
error had several hundred thousand rigid elements.
Avoidance: There is no avoidance. The limit has been removed in
MSC.Nastran V2001.
@
38478; 2001; Documentation/QRG - CONNECT, PBARL, PBEAML - Non-Existant Develo
per's Guide
The MSC/NASTRAN Quick Reference Guide documentation for the FMS CONNECT
statement, remarks :
 5. Custom beam cross-section evaluator libraries developed by users
should
comply with the MSC/NASTRAN Beam Cross-section Evaluator Developer's
Guide.
There is no such document.
Avoidance: The remark should instead, refer to:
MSC/NASTRAN V69 Release Guide,
Section 3.1 Beam Cross-Section Library
Appendix C: Adding Your Own Beam Cross Section Library
@
38501; 2001; Optimization/Design Sensitivity - Flutter, Superelements - UFM 1
126 for FPHP
USER FATAL MESSAGE 1126 occurs if a flutter analysis is attempted with
superelements in SOL 200 and complex eigenvector output is requested. See
also CSR 37130.
Avoidance: Remove the eigenvector request or make the following dmap
changes:

compile fea
alter 1
type db fphp
endalter
@
38502; ; Optimiztion/Design Sensitivity - DEQATN - SFM 233 (PRSPOP)
In SOL 200, if the DEQATN bulk data entry uses certain keywords such as
"OR" as variable names in the equation argument list, an error message
similar to the following will occur:
*** SYSTEM FATAL MESSAGE 233 (PRSPOP)
UNDERFLOW OCCURRED DURING PROCESSING OF OPERAND STACK
DURING DEQATN EQUATION PROCESSING.
PROGRAMMER INFORMATION: STACK POINTER = 0 WITH MINIMUM = 1
*** COMPILATION DIAGNOSTIC INFORMATION ***
1. STATEMENT TYPE: DEQATN EQUATION
2. EQUATION ID NUMBER: 14
3. EQUATION:
TWALL(OR,IR)=OR-IR
*** SYSTEM FATAL MESSAGE 233 (PRSPOP)
The error message needs to be improved to add the following:
Possible cause: variable name is a keyword such as "OR".
Avoidance: Change the variable name to something different such as "ORR".
@
38519; 2001; Documentation/QRG - PACABS - Units
The Quick Reference Guide description for the PACABS entry should state
that the units of the B, K, and M fields are per unit area.

@
38539; 2001; Distributed Memory Parallel - DMP, DISDCMP - Infinite CPU Loop
MSC.Nastran may get caught in an infinite cpu loop when distributed memory
parallel is used with inertia relief.
Avoidance: Do not use ineria relief with distributed memory parallel in
V70.7.

@
38542; 2001; Miscellaneous - VECPLOT - SWM 3062, Insufficient Memory
Module VECPLOT does not check for insufficient memory. This may lead to
the following unclear message:
0*** SYSTEM WARNING MESSAGE 3062 MODULE = VECPLOT
FAILURE OCCURRED IN UTILITY GTXYZD CALLED IN SUBROUTINE DMTRXD .
PREXYZ FLAG ARGUMENT WAS SET WITH AN INCORRECT VALUE TO GET THE
REQUESTED DATA RETRIEVAL.
*** USER WARNING MESSAGE 3199 (VECPLOT)
NON-FATAL MESSAGES MAY HAVE BEEN LOST BY ATTEMPTING TO QUEUE MORE
THAN 100 MESSAGES.
@
38547; 2001; Aeroelastic Anaysis - Multiple Double Lattice Groups yield Wrong
Answers
If a doublet lattice model (CAERO1 and/or CAERO2) is created in SOL 144
for use in static aeroelasticity, and the model has more than one
interference group, the downwashes representing unit perturbations of the
control devices (the DJX matrix) are incorrect under most circumstances.
The problem is caused due to the erroneous introduction in Version 70.5 of
an assumption that the coordinate system ID and the CAEROi ID are in the
same order. This is true for a single interference group, but not for
multiple groups. As a result, the wrong coordinate system may be used in
the computation of the downwash.
Avoidance: There is no avoidance other than to convert the model to have a
single interference group. Note, also, that unsteady aerodynamics (SOL 145
and SOL 146) are NOT affected.
@
38555; 70.7.2; Installation - NFS Mounted - HP-UX 10.2 and 11.0
The original MSC.Nastran V70.7 (r1) CD may fail to install correctly on
some HP-UX 10.20 and 11.00 systems if the installation base directory is
NFS mounted from another system. The problem is due to incorrect
information being returned from the df(1) utility, this should be visible
from the output of the command
df -k <install-dir>
Avoidance: Choose one of the following avoidances:
1) Use the MSC.Nastran V70.7.2 (r2) release, that works around the HP-UX
utility error.
2) Install on the computer exporting the filesystem. If this is not an
HP-UX system, you can use the "Custom" installation to deselect the local
 computer's binaries and select the HP-UX binaries.
@
38559; 2001; Optimization/Design Sensitivity - DSAD, DRESP2 - SFM 3007
An optimization job may fail in module DSAD with SFM 3007 when the
following conditions are met:
1. The DRESP2 entry only references the DNODE term,
2. The job has more than one subcases,
3. The DESGLB entry references the DRESP2 entry with DNODE terms.

The avoidance is to replace DESGLB with the DESSUB entry and place it in
the last subcase.
@
38564; 70.7.2; Documentation - Linux: how to enable large files (> 2 GB)
The "Configuration and Operations Guide" incorrectly states that large
files are available on Linux. They were not in the version of either the
Linux kernel (2.0.36) or libc (2.0) that MSC.Nastran was built with.
Note: The support required for this was not built into MSC.Nastran, so
using a system with support for large files will not help.
Large file support will be available in a new release of MSC.Nastran
Version 70.7.2 on Linux that is to have a build date and release date of
June 2000.
Avoidance: Use INIT and ASSIGN statements to specify multiple physical
files for any DBSet that exceeds 2GB.
@
38574; 2001; Optimization/Design Sensitivity - DISPRS - Poor Performance
It is possible, when performing modal frequency analysis in SOL 200, to
recover results to the g-set in the DISPRS subDMAP twice with each call
performing identical operations. For large problems, this can entail
significant CPU resources with no benefit.
Avoidance: The following dmap alter removes the redundant call:
compile fea
alter 'if ( psldf'(2)
if (delta) then $
alter 'if (psldf'(2,4)
endif $
endalter
@
38595; 2001; Composites - PCOMP, PARAM,S1 PARAM,POST,0 - Wrong Answers in xdb
If stresses are to be sorted (PARAM,S1) and composite stresses are
requested, then the xdb file will contain incorrect composite stresses.
Avoidance: Insert the following DMAP alter:
compile sedrcvr
alter 'DBC OES1C',''
DBC ES1C,,,,,,,,,,,,,,,,,,,//
'OESC'////////////////////
-1/DBCPATH/S,N,CP/APP1/ICYCLIC/GEOMU/LOADU/POSTU/
 DBCDIAG/DBCCONV/DBCOVWRT/DESCYCLE/////adptindx $
endalter
@
38608; 2001; Documentation/QRG - FLFACT - Remark 4 contains incorrect equatio
n
The formula of Remark 4 on the FLFACT bulk data entry of the Quick
Reference Guide V70.5 has two omissions and a typographical error.
It should read as follows:
F1 (FNF-FMID) (NF-i) + FNF (FNF-F1) (i-1)
Fi =
-----------------------------------------------------------------------
(FNF-FMD) (NF-i) + (FMID-F1) (i-1)
The above formual was taken from the Aeroelastic Supplement and was
compared to output generated by Version 70.5.
@
38610; 2001; Documentation/QRG - EIGRL - System Cell 229 Documentation Incorr
ect
The documentation for SYSTEM cell 229 incorrectly indicates that if it is
given a value of 1, then the scratch database usage for Lanczos will be
decreased. This system cell applied only to the old Lanczos solver and
will not have an impact on the current Lanczos solver in MSC.Nastran
Version 70.7 (The old Lanczos solver is available by setting
SYSTEM(273)=1).
@
38614; 2001; Nonlinear Analysis - PARAM NMLOOP may give poor answers
Nonlinear SOL 106 with PARAM NMLOOP may give poor answers. The
eigenvalues, displacements, and stresses for modes including prestresses
and nonlinear geometry may be based upon matrices that do not reflect a
converged solution. This error started in version 70.7 when a different
version of the history file was used to construct the required matrices.
The use of NLPARM method ITER-1 gives fairly good results.
Avoidance: An avoidance ALTER is available from MSC to improve results.
When using HEXA elements in nonlinear, include NASTRAN SYSTEM(309)=1 , see
CSR 38899.
@
38620; 2001; Superelements - DMIG - UFM 327
DMIG matrices in part superelements result in the following fatal message
*** USER FATAL MESSAGE 327 (IFS2P)
BAD DATA OR FORMAT OR NON-UNIQUE NAME. DMIG also for DMI
The message above occurs when two part superelements adjacent in the
processing order have dmigs or dmis with the same name.
Avoidance: Add a second dmig or dmi matrix. It does not need to be used
anywhere, it is only necessary to have more than one dmig or dmi name.
@
38627; ; Subcases - SPC, DEFORM - UFM 9058
In static analysis, if DEFORM and SPC entries are changed between
subcases the job will fail with the message:
 UFM 9058 (SUPER3) - THE SOLUTION FOR THE RESIDUAL STRUCTURE DOES NOT
EXIST
Avoidance: Execute the subcases as two separate runs.
@
38629; ; Nonlinear Analysis - TEMP(LOAD), PARAM,LGDISP - Incorrect OLOAD
Output
MSC.Nastran considers the fictitious forces for thermal expansion to be
loads. These are output by a case control command OLOAD. The output is
for the initial geometry, even when geometric nonlinear effects are
requested. Since only the OLOAD output is wrong, other items like
displacements, element forces and SPCFORCEs are not affected. Other load
types (like PLOAD and FORCE1) are correctly updated to the new geometry.
@
38632; 2001; Acoustics Analysis - KJJZ of fluid is wrong for prestressed stru
cture
An acoustic analysis of a prestressed structure yields an incorrect fluid
part of KJJZ. The values are off by a factor of 2.
Avoidance: Apply the following dmap alter to Version 70.5 or 70.7:

compile semg
alter 'add5','add5'
add5 kjjl,kjjnl,kdjjnl,,/kjjz $
endalter
@
38634; 2001; SSSALTERs - trncompa.v707, SORT1 - Missing Output
If sort1 output is requested while using the trncompa sssalter, no ply
output will be obtained.
Avoidance: Only request sort2 output.

@
38645; 70.7; Plotting - XYPLOT - Missing Data Points
An XYPLOT of grid points versus temperature results in a heat transfer
analysis results in missing data points.
Avoidance: There is no avoidance.

@
38653; 2001; Mass Matrix - SYSTEM(303), CUTOFF - UFM 17(LOOKUP)
The MSC.Nastran Version 70.7 Release Guide states that the name for
SYSTEM(303) is CUTOFF. If SYSTEM CUTOFF=1 is specified, the program will
issue:
USER FATAL MESSAGE 17 (LOOKUP)
SYMBOL "CUTOFF" IS UNDEFINED
Avoidance: Use the system cell number rather than the name.
@
38654; 2001; Superelements/Nonlinear Analysis - PARAM,POST,0 - SWM 6065 (DBC)
 If superelements are used in SOL 106 and PARAM,POST,0 is used, it is
possible that SWM 6065 may be issued, followed by SWM 6066, indicating
that the DBC module has been terminated. If this occurs, no additional
information will be written to the .xdb file for post-processing.
Avoidance: Insert the following DMAP alter into the file:
compile super3
$ remove ol2s from DBC list
alter 'dbc.*casedr.*ol2s.*casecc.*oltype',''
DBC CASEDR,,,,,,,,,,,,,,,,,,,//
'CASECC'////////////////////
-1/DBCPATH/S,N,CP/APP/ICYCLIC/GEOMU/LOADU/POSTU/
DBCDIAG/DBCCONV/DBCOVWRT/DESCYCLE/////ADPTINDX $
@
38676; 70.7.2; Acoustic Analysis - Missing Panel Participation Factor Output
Prior to Version 70.7.2 fluid/structure problems that have no s-set (no
constraints) will not allow panel participation factors to be printed with
PARAM,FLUIDMP or STRUCTMP.

Avoidance: Use AUTOSPC and an unattached GRID to force a degree of freedom

in the s-set.

@
38682; 2001; Nonlinear Analysis - NMLOOP - Possible Wrong Answers
The use of NMLOOP yields incorrect prestressed modes in Version 70.7 if
the user does not enforce a stiffness update. The prestressed modes are
correct in Version 70.5 because the stiffness is updated automatically at
the end of the subcase.
Avoidance: There are three possible avoidances in V70.7, we recommend the
first option.
1. Insert the following dmap alter:
ECHOOFF $ SWITCH ESTNL, REPLACE MODULE EMAKFR WITH SUBDMAP CENTRIP
COMPILE NLSTATIC $ VER 70.7 BY R.L.H. MARCH 29, 2000
ALTER 'NMLOOP.*CNMLOOP=2','' $
IF ( NEWP>-1 AND NMLOOP<>0 ) THEN
CNMLOOP=2 $
MESSAGE //' DELETE/COPY ESTNL $ CSR 38614'
DELETE /MESTNL,,,, $
COPY ESTNLH/MESTNL $
ENDIF
COMPILE NLKR $
ALTER 'EMAKFR','' $
TYPE DB,EQEXINS $
CALL CENTRIP SLT,BGPDTS,EQEXINS,CSTMS,SILS,MJJ/KFRFGG $
COMPILE SEKDR $
ALTER 'EMAKFR','' $
CALL CENTRIP SLT,BGPDTS,EQEXINS,CSTMS,SILS,MJJ/KFRFGG $
COMPILE CENTRIP NOREF NOLIST
$
SUBDMAP CENTRIP SLT,BGPDT,EQEXIN,CSTM,SIL,MJJ/KCEN/ $
$
$********************************************************************
$ *
$ This DMAP sequence calculates Coriolis and Centripetal effect *
$ matrices in global coordinates. *
$ *
$ For node i: *
$ *
$ [Bcoriolis] = 2[x]m *
$ *
$ [Kcentripetal] = [x][x]m *
$ *
$ | 0 -Omega(z) Omega(y) | *
$ [x] = | Omega(z) 0 -Omega(x) | *
$ | -Omega(y) Omega(x) 0 | *
$ *
$********************************************************************
$
TYPE PARM,,I,N,SLTRECNO $
TYPE PARM,,RS,N,OMEGA $
TYPE PARM,,CS,N,N1C,N2C,N3C $
TYPE PARM,,CS,N,OMEGAC $
$
$ SEARCH FOR RFORCE ENTRY
$
SLTRECNO= 1
DO WHILE ( SLTRECNO > 0 ) $
$
PARAML SLT//'DTI'/S,N,SLTRECNO/1//S,N,LTYPE/ $
IF ( (SLTRECNO <> -1 ) AND ( LTYPE = 10 ) ) THEN $
$
$ EXTRACT SPIN VECTOR FROM RFORCE CARD.
$
PARAML SLT//'DTI'/SLTRECNO/4//S,N,CID/ $
PARAML SLT//'DTI'/SLTRECNO/5/S,N,RPT/ $
PARAML SLT//'DTI'/SLTRECNO/6/S,N,N1/ $
PARAML SLT//'DTI'/SLTRECNO/7/S,N,N2/ $
PARAML SLT//'DTI'/SLTRECNO/8/S,N,N3/ $
SLTRECNO = -2 $
$
ENDIF $
$
SLTRECNO = SLTRECNO + 1 $
$
ENDDO $
$
IF ( RPT = 0. ) THEN $
$
OMEGA= 0.
RETURN $
$
ENDIF $
$
$ THESE ARE NEEDED LATER ON
$
MATGEN ,/I1X1/1/1/ $
MATGEN ,/I6X6/1/6/ $
$
$ CALCULATE RADIANS PER UNIT TIME
$
OMEGA= 2.0 * PI(1) * RPT $
OMEGAC = CMPLX(OMEGA,0.) $
$
$
$ CREATE A COLUMN VECTOR, MJJTSUM, FROM THE DIAGONAL OF MJJD
$ TO BE USED FOR PARTITIONING PURPOSES LATER
$
DIAGONAL MJJ/MJJC/'COLUMN' $
$
$-------------------------------------------------------------------$
$-------------------------------------------------------------------$
$
$ CALCULATE ROTATION MATRIX IN THE CID SYSTEM
$
IF (N1 <> 0.) THEN
$
$ DETERMINE FACTOR FOR X-ROTATION (in CID system)
$
MATGEN ,/PVC6/6/6/1/1/4 $
MATGEN ,/PVR6/6/6/2/1/3 $
MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $
MPYAD ONE6X6,I6X6,ONE6X6/ROTX/1/-1 $
DELETE /PVC6,PVR6,ONE6X6,,/ $
N1C = CMPLX(N1,0.)
$
ENDIF $
$
$-------------------------------------------------------------------$
$
IF (N2 <> 0.) THEN
$
$ DETERMINE FACTOR FOR Y-ROTATION (in CID system)
$
MATGEN ,/PVC6/6/6/2/1/3 $
MATGEN ,/PVR6/6/6/0/1/5 $
MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $
MPYAD ONE6X6,I6X6,ONE6X6/ROTY/1/-1 $
DELETE /PVC6,PVR6,ONE6X6,,/ $
N2C = CMPLX(N2,0.)
$
ENDIF $
$
$-------------------------------------------------------------------$
$
IF (N3 <> 0.) THEN
$
$ DETERMINE FACTOR FOR Z-ROTATION (in CID system)
$
MATGEN ,/PVC6/6/6/0/1/5 $
MATGEN ,/PVR6/6/6/1/1/4 $
MERGE ,,,,I1X1,PVC6,PVR6/ONE6X6/1/0/1 $
MPYAD ONE6X6,I6X6,ONE6X6/ROTZ/1/-1 $
DELETE /PVC6,PVR6,ONE6X6,,/ $
N3C = CMPLX(N3,0.)
$
ENDIF $
$
$-------------------------------------------------------------------$
$
$ ADD FOR TOTAL ROTATION (in CID system)
$
ADD5 ROTX,ROTY,ROTZ,,/ROTTOT/N1C/N2C/N3C $
$
$-------------------------------------------------------------------$
$-------------------------------------------------------------------$
$
$ GENERATE G-SIZE CORIOLIS MATRIX IN THE GLOBAL SYSTEM
$
$
$ FIRST, GENERATE G-SIZE MATRIX IN THE CID SYSTEM
$
ADD ROTTOT,/OMEGA/OMEGAC $
MATMOD OMEGA,SIL,,,,/OMEGAGG,/5 $
$
$-------------------------------------------------------------------$
$
$ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS
$ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5)
$
PARTN OMEGAGG,MJJC,/,,,OMEGAXX/ $
$
$-------------------------------------------------------------------$
$
$ TRANSFORM CORIOLIS ACCELERATION MATRIX TO GLOBAL SYSTEM
$ FIRST, TRANSFORM FROM CID TO BASIC
$ THEN , TRANSFORM FROM BASIC TO GLOBAL
$
$ CHECK IF THE CSTM EXISTS. IT WILL NOT EXIT IF THE GLOBAL SYSTEM
$ IS THE SAME AS THE BASIC SYSTEM. ERROR IN MATMOD. THEREFORE
$ GENERATE THE IDENTITY MATRIX
$
$
PARAML CSTM//'PRESENCE'////S,N,NOGTOB=0 $
IF (NOGTOB < 0) THEN
DIAGONAL MJJ/GTOBY/'SQUARE'/0. $
ELSE
MATMOD CSTM,SIL,BGPDT,,,/GTOBY,/5//-1 $
ENDIF $
$
PARTN GTOBY,MJJC,/,,,GTOB/ $
$
$ CHECK IF THE CID IS EQUAL TO ZERO (BASIC)
$
IF (CID = 0) THEN
EQUIVX GTOB/CTOG/ALWAYS $
ELSE
MATMOD CSTM,SIL,BGPDT,,,/CTOBX,/5//CID $
$
$ USE THE MJJC VECTOR TO PARTITION OUT THE ZERO MASS TERMS
$ ALONG WITH THE SPOINTS (WHICH CAUSE PROBLEMS FOR MATMOD OPTION 5)
$
PARTN CTOBX,MJJC,/,,,CTOB/ $
MPYAD CTOB,GTOB,/CTOG/1 $
ENDIF $
$
SMPYAD CTOG,OMEGAXX,CTOG,,,/OMEGAYY/3////1 $
MERGE ,,,,OMEGAYY,MJJC,/OMEGAGGX/ $
$
$-------------------------------------------------------------------$
$-------------------------------------------------------------------$
$
$ CALCULATE CENTRIPETAL ACCELERATION MATRIX
 $
$ MULTIPLY MASS BY CENTRIPETAL ACCELERATION
$ TO CALCULATE CETRIPETAL FORCE MATRIX
$
SMPYAD OMEGAGGX,MJJ,OMEGAGGX,,,/KCEN/3////////6 $
$
RETURN $
END $
$
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
$
2. Change the iteration strategy on NLPARM to AUTO,1 to force a stiffness
upon convergence. If that is too costly,
apply the preload in two subcases, the first subcase with default
iteration strategy, the second subcase with METHOD
case control command and NLPARM with with AUTO, 1.
3. Apply the preload in a SOL 106 coldstart run and restart in SOL 103 to
calculate the prestressed modes using PARAM,NMLOOP.
CSR 38614 is related to this CSR.
@
38684; 2001; Documentation - PSHELL - Remark 14, (ts)10e8(ts) incorrect.
Remark 14 under the PSHELL description in the MSC/NASTRAN Verson 70 and
70.5 Quick Reference Guide description is incorrect.
It says:
14. If MIDi is greater than or equal to (Ts)10^8(Ts), then parameter
NOCOMPS
is set to +1 indicating that composite stress data recovery is desired.
(MIDi greater than (Ts)10^8(Ts) are generated by PCOMP entries.)
It should say:
14. If MIDi is greater than or equal to 10^8, then parameter NOCOMPS
is set to +1 indicating that composite stress data recovery is desired.
(MIDi greater than 10^8 are generated by PCOMP entries.)

The v70.5 online encyclopedia has the correct remark as does pre v70.0
Quick Reference Guides.
@
38722; 2001; Optimization/Design Sensitivity - PCOMP - UFM 315, UFM 316
In Version 70.7 it is possible to output updated propertied in an
optimization run. If a model contains PCOMP entries and one or more of
the PCOMP entries has an odd number of layers, then any subsequent run
using the updated PCOMP entries generated from the previous puch file
will fail with the following error messages:
*** USER FATAL MESSAGE 315 (IFPDRV)
FORMAT ERROR ON BULK DATA ENTRY PCOMP X
*** USER FATAL MESSAGE 316 (IFPDRV)
ILLEGAL DATA ON BULK DATA ENTRY PCOMP X
 Avoidance: Punch out the PCOMP entries using the small field format
(e.g., param,despch1,x) where x can be -1, -2, -3, -5 -,6, or -7. The
default
is param,despch1,6.
@
38740; 2001; DMAP - DRHM3 corrupts OSE1 table
It has been observed that on Version 70.5 SGI systems that the OES1 table
produced bythe DRHM3 module is corrupted and can not be printed.

@
38743; 2001; Optimization/Design Sensitivity - SPCFORCE, DSCREEN - UFM 315
Specifying the response type (RTYPE) of SPCFORCE on the DSCREEN entry in
SOL 200 will result in UFM 315 ILLEGAL DATA ON BULK DATA ENTRY DSCREEN.
Note that the SPCFORCE response type was first added in V70.7.
Avoidance: Create an equivalent DRESP2 for each DRESP1 of type SPCFORCE
that will be constrained. Then specify the response type (RTYPE) of EQUA
on the DSCREEN entry.
@
38750; 2001; Aeroelastic Analysis - RESTARTS, MKAERO - Inefficient Processing
The dynamic aeroelastic restarts are dependent on the sorted order of the
MKAEROi (i=1 or 2) bulk data entries whenever multiple entries are
present. This creates the situation of regenerating the aerodynamic
matrices when the /,1,99999 method of deleting and reinputting the bulk
data is used. The MKAEROi entries on the restart MUST be in the same
order as the coldstart run otherwise SOL 145 or 146 will rerun AMG/DECOMP
even though the MKAEROi entries are identical to the coldstart but in a
different sorted order.
Aviodance: To avoid this problem the restart should only delete the
specific bulkdata entries required NOT the entire bulkdata input.
@
38769; ; Materials - CQUAD4, MAT8, G12 - UFM 4297
In Version 70.7, if the G12 field on the MAT8 entry is blank, then the job
will fail with the following misleading error messages:
*** USER FATAL MESSAGE 4297 (EQD4D)
QUAD4 ELEMENT WITH ID = 1
HAS AN INTERIOR ANGLE POSSIBLY GREATER THAN OR EQUAL 180 DEGREES.
Avoidance: There are two possible avoidances:
1. Add the G12 field to the MAT8 entry if appropriate.
or
2. Set SYSTEM(173)=2
@
38780; 2001; SSSALTERs - mmecha.v707 - possible errors
It has been observed that the mmecha alter may fail due to a singularity
problem. This problem is due to an incorrect maxratio test in the
presence of 2x2 pivots. This type of pivot can be caused by the Lagrange
Multiplier Method.
 Avoidance: The revised alter for Version 70.7 is available on the
MSC.Software website. If you have trouble getting the updated alter,
contact your local support representative.
@
38788; 70.5; Static Analysis - STRESS, CORNER, PUNCH - Missing Grid ID
When requesting punch output for element forces or stress at the corners
, the punch file does not print the grid ID. It does not know wich fields
are integer and which fields are floating point
Avoidance: There is no avoidance.
@
38815; ; Distributed Memory Parallel - DMP - Sigbus Error on SGI
SGI distributed memory parallel jobs that request 8 or more processors may
fail with a "SIGBUS: Bus error" as follows:
Core from signal SIGBUS: Bus error
(dbx) where
> 0 _realloc(0x100193e8, 0x190, 0x19450152, 0x0, 0x110, 0x9436c5a,
0x94369bd, 0x94369d2)
["/xlv43/6.5.7m/work/irix/lib/libc/libc_n32_M4/gen/malloc.c":367,0xfb2a1f
]
1 parse_file(0x111, 0x1, 0x19450152, 0xfb51ef8, 0x110, 0x9436c5a,
0x45,0x94369d2)
["/home/poppy08/kak1/mpt/1.3/mpi/lib/libxmpi/libxmpi_n32_M3/all/all_parse
c":264,0x9406214]
2 parse_global_opts(0x9425388, 0x1, 0x19450152, 0xfb51ef8, 0x110,
0x9436c5a,
Avoidance: The following changes need to be made to the nast707.dmp file:
From:
%MSC_ARCH=irix64:% print -nu3 -- "${hosts[$i]} -np 1 ${rcmds[$i]}
-c$i -d%r.debug% jid=$jid $args"
To:
%MSC_ARCH=irix64:% print -u3 -- "#"
%MSC_ARCH=irix64:% print -nu3 -- "mpirun ${hosts[$i]} -np 1
${rcmds[$i]} -c$i -d%r.debug% jid=$jid $args"
From:
%MSC_ARCH=irix64:%mpirun -f $cmdfile 2>%out%.%j.unique%.mpirun
To:
%MSC_ARCH=irix64:%chmod a+x $cmdfile
%MSC_ARCH=irix64:%$cmdfile 2>%out%.%j.unique%.mpirun
@
38839; 2001; Superelements - SEBULK, Collectors - SWM 3001 (READ)
If component modes are requested on a collector superelement then the run
will fail with SWM 3001 in READ.
Avoidance: Insert the following DMAP alter:
compile phase0
alter 'call ifps1'
if ( setype='colltr' ) equivx dynamic/dynamics/-1 $
endalter
@
38841; 70.7; Installation - Poor Error Message
 If MSC.Nastran is NOT installed on a remote node properly, then the
following error may occur:
*** USER WARNING MESSAGE (pgm: nastran, fn: validate_dmp_hosts)
beowulf21(LAM n1): No response from MSC.Nastran remote query:
/scratch/msc/bin/nast707 -d0 -rN70.7 frs=/scratch/DMP/dmp_108.dat
drwx=/tmp fw=/scratch/DMP/dmp_108.T6084_15 drwx=/scratch/DMP
drwx=dmp_108 drwx=.
Segmentation fault (core dumped)
Avoidance: Install MSC.Nastran correctly on the remote node. If there is
not an authorization file, then the following error may occur:
*** USER FATAL MESSAGE (pgm: nastran, fn: validate_authorize)
authorize="???????" (LM_LICENSE_FILE)
This file does not exist.
Avoidance: Install the an authorizatoin file.
@
38869; ; Documentation -- DMAP Module Dictionary - APPEND 1st param canno
t be 0
Both the Version 68 DMAP Module Dictionary and the current Version 70.7
online pdf documentation indicate that the 1st parameter for the APPEND
module may be either a "0" or a "1" in order to create an output datablock
[IN1:IN2]. However, if "0" is used then "UFM 5005 INVALID VALUE OF OPTION
PARAMETER DETECTED 0" is issued, followed by "SFM 3007 MODULE = APPEND
ILLEGAL INPUT TO SUBROUTINE APPEND".
Avoidance: Set the 1st parameter to "1" instead of "0".
@
38889; 2001; GPFORCE - TEMP(LOAD) - Wrong Answers
The applied load contribution of the GPFORCE output is incorrect in
TEMP(LOAD) is applied to only a portion of the model.

Avoidance: The only avoidance is to put a small thermal load on ever

element.
.
@
38892; 2001; Optimization/Design Sensitivity - SYSTEM(2), DOPTPRM - Undocumen
ted Feature
If the IPRINT parameter on the DOPTPRM entry is used to obtain
intermediate prints from the DOT optimizer, the output is always written
to unit 6 regardless of which unit other "f06" output is directed to by
NASTRAN SYSTEM(2).
Avoidance: Undocumented paramter IWRITE can be used on the DOPTPRM entry
to direct the intermediate prints to the unit of the users choice. That
is IWRITE=4 places this output in the .f04 file.
@
38895; 70.7.2; DMAP - PARAML - U1 Set not Recognized
Degree of freedom Set U1 is not recognized by DMAP modules dealing with
 degree of freedom sets, e.g. PARAML, VEC, etc.
Avoidance: The only avoidance is not to use the U1 set.
@
38898; 70.7.3; Performance - SYSTEM(107) - Poor Performance
If Parallel processing is selected with SYSTEM(107) (a.k.a. keyword
PARALLEL) on MSC.Nastran Version 70.7 on the Cray J-90 executable cache
memory will be disabled and poor performance results.
Avoidance: Other than not selecting parallel there is no avoidance.
@
38899; ; Nonlinear Analysis - HEXA - Poor Convergence
Poor convergence may be experienced for Nonlinear jobs with HEXA elements
in MSC.Nastran Version 70.7. The problem occurs because the new HEXA
formulation introduced for linear analysis in Version 70.7 is used on the
first iteration and it shouldn't be.
Avoidance: Set SYSTEM(309)=1 to revert to the pre-Version 70.7
formulation for all iterations.
@
38907; ; Complex Eigenvalue Analysis - CEAD - UWM 5411 (REIGL8)
The message below may appear many times when using the complex eigensolver
module CEAD.
*** USER WARNING MESSAGE 5411 (REIGL8)
NEGATIVE TERM ON DIAGONAL OF MASS(VIBRATION) OR STIFFNESS(BUCKLING)
MATRIX DETECTED.
ROW NUMBER = 80 VALUE = -1.20000D+02
Avoidance: This message may be safely ignored when encountered in the
CEAD module.

@
38925; 2001; Nonlinear Static Analysis - slideline contact, PELAST - SFM 3001
in EMG
Job will fail in SOL 106 when line contact problem includes nonlinear
springs in the model. For example, any SLIDE-LINE model with PELAST
and/or PBUSHT, which refer to TABLEDi bulk data cards, will be terminated
during contact nonlinear stiffness generation in SOL 106.
Avoidance: Please insert the following DMAP alter into SOL 106 decks to
avoid this difficulty
COMPILE NLSTATIC $
ALTER 'CALL NLKR'(3),'' $
CALL NLKR MESTNL ,CSTMS ,BTOPSTF , , ,
BGPECT ,BGPDTH ,SILS ,USETNL ,GMNL ,
, ,GPSNTS ,USET1 , ,
DEQATN ,DEQIND ,ETT , , ,
/
BKELM ,BKDICT ,BKGG ,BKFS ,BKAA ,
, , ,BKMG /
S,NOBKGG/-1 /K6ROT /LANGLES /MPCF1 /
NOSSET /NOOSET /'CONTACT'/0 /0 /
 NOOSETX /LOADFAC /LSETID /PLSETID /FOLLOWK /
UNSYMF $
ALTER 'CALL NLKR'(4),'' $
CALL NLKR MESTNL ,CSTMS ,BTOPSTF , , ,
BGPECT ,BGPDTH ,SILS ,USETNL ,GMNL ,
, ,GPSNTS ,USET1 , ,
DEQATN ,DEQIND ,ETT , , ,
/
BKELM ,BKDICT ,BKGG ,BKFS ,BKAA ,
, , ,BKMG /
S,NOBKGG/-1 /K6ROT /LANGLES /MPCF1 /
NOSSET /NOOSET /'CONTACT'/0 /0 /
NOOSETX /LOADFAC /LSETID /PLSETID /FOLLOWK /
UNSYMF $

@
38927; 2001; Elements - CTETRA, PSOLID - Poor Answers, UWM 4683 (READ) MASS M
ATRIX
The mass matrices for mid-side-noded CTETRAs are not calculated correctly
if the defaults values for IN and ISOP fields on the PSOLID are not used.
This may lead to UWM 4683 from the READ module and poor answers.
Avoidance: Leave the IN and ISOP fields blank on the fluid PSOLIDs in
order to use the defaults.
@
38939; ; Licensing - FlexLm - Problem with FlexLm on Windows 2000
In some instances, the authorization process on NT and Windows 2000 may
take almost a minute to return with the authorization. If this is
happening, add the following line to your .rcf file
SYMBOL=LM_NO_SNMP=1
@
38949; 2001; Submittal - hosts keyword on nast707 command is ignored
If the hosts keyword is defined in any of the RC files, then you will be
unable to override it's value on the command line.
Avoidance: A work around is to put the keyword in a local RC file in the
directory where the input file resides. This value will override any
values set in the other RC files.
@
38951; 2001; Submittal - Execution on remote node may cause input and output
to be deleted.
If a Nastran analysis is run on a remote system using the the node
keyword, and the output file directory is visable to the remote node, then
all the output files and the input files will be deleted after the job has
run.
Avoidance: There are two avodiances.
1) Run the job in a directory that is not visable to the remote node
2) Make the following change near the end of the
install_dir/bin/nast707.rmt file.
127,128c127
 < %s.rsh% %node% -l %username% /bin/rm -f "%r.out%*"
< %r.outvis=y:%/bin/rm -f "%r.out%*"
---
> %r.outvis=n:%%s.rsh% %node% -l %username% /bin/rm -f "%r.out%*"
@
38955; 2001; DMAP - PARAML - Overflow Exit, Floating Point Exception
When the PARAML module is used to extract an element from a matrix with
the DMI option, and the element is larger than the range allowed by single
precision parameters, a machine-dependent overflow error exit occurs. On
NT MSC/NASTRAN, the output from the f06 file is:
0THE NUMBER OF NON-ZERO TERMS IN THE DENSEST COLUMN = 1
0THE DENSITY OF THIS MATRIX IS 100.00 PERCENT.
MAIN: The FPU has been reset after floating point exception.
MAIN: For reference, A(MAINAL) = 405290, A(/SYSTEM/) = 1777f60
[similar data is presented in the .log file]
There is no mention in any of the diagnostic output that the PARAML module
was being executed when the overflow occurs. One diagnostic method is to
temporarily put a message statement in front of every suspected module
call where the overflow may have occured, and isolate the offending dmap
statement in that manner.
Avoidance: An avoidance in Version 70.7, once the statement is isolated,
is to divide the offending matrix by 1.e30, then multiply the output from
paraml by the same number, but with all parameters at this point converted
to double precision.
In Version 2001 the scalar module will provide a double precision output
parameter that avoids this problem

@
38960; ; Documentation - QRG, PELAST, PBUSHT - Unexpected Answers
If the initial slope of the nonlinear force-displacement relationship
defined on a PELAST or PBUSHT entry in a nonlinear solution does not match
the corresponding Ki value on the PELAS or PBUSH entry, then the results
may be unpredictable.
Avoidance: Make sure that the initial slope as defined on the TABLEDi
entry matches the corresponding Ki value on the PELAS or PBUSH entry.
@
39031; ; Documentation - DMAP Module Dictionary - MATMOD Option 5, Form 3
Example 3 under the DMAP Module Dictionary's description of MATMOD option
5 should be modified from
Transform KGGBASIC coordinate system 10.
MATMOD CSTM,SIL,,,,/TRANS10,/5//10 $
TRNSP TRANS10/TRANST10T $
SMPYAD TRANS10,KGGBASIC,TRANS10T,,,/KGG10/3 $
to
Transform KGGBASIC to coordinate system 10.
 MATMOD CSTM,SIL,,,,/TRANS10,/5//10 $
TRNSP TRANS10/TRANST10T $
SMPYAD TRANST10T,KGGBASIC,TRANS10,,,/KGG10/3 $
@
39041; 2001; Optimization/Design Sensitivity - PARAM,NOCOMPS -1 not supported
on DRESP1
PARAM NOCOMPS -1 is used to recover shell element stresses for an
equivalent homogeneous element. The use of this capability has been
blocked in SOL 200 so that if the user invokes the equivalent PSHELL on
the DRESP1 entry to obtain stress/strain responses of the element, the
corresponding PSHELL will not be found.
Avoidance: There is no avoidance. Version 2001 will remove this
restriction.

@
39046; ; File Management Section - DBDICT - Missing Dictionary
If the DBDICT FMS command is used in a run as follows, the dictionary will
be blank.
assign master='oldrun.MASTER'
DBDICT
endjob
Avoidance: Use a read-only restart similar to the following to list the
database dictionary.
assign run1='oldrun.MASTER'
restart, logical=run1
dbdict
sol 101
cend
begin bulk
enddata
@
39056; 2001.0.1; Grid Point Stress - GPSTRESS- 4276 (FREAD)
If GPSTRESS output is requested, then following error may occur.

SYSTEM FATAL MESSAGE 4276 (FREAD).

Avoidance: Ensure that case control requests for stress and force specify
the same type of output. If center and grid point output is specified for
one, it should also be specified for the other.
@
39067; 2001; Distributed Memory Parallel - DMP, READ - Poor Performance
Distributed Memory Parallel eigenvalue jobs may experience poor
performance in READ on Compaq Alpha clusters.
Avoidance: There is no avoidance.
@
39088; ; Cyclic Modes - SFM 3001 (DCMP)
 The following message may occur in a cyclic symmetry modal run (SOL 115)
when requesting a tridiagonal solution method such as AHOU or AGIV. The
message is output from the DCMP module if the USET and SIL tables are
mismatched with the input matrices.
*** SYSTEM FATAL MESSAGE 3001 (DCMP)
THE INPUT DATA BLOCK IN POSITION 3 DOES NOT EXIST.
Avoidance: An avoidance is to use the Lanczos method which does not
perform this operation. Another avoidance is to use the DECOMP module
instead:
Compile moders list noref $
Alter 'dcmp.*kww1'(,-1) $
If (cyclic) then $
Decomp kww1/lww1,, $
Else $
Alter 'dcmp.*kww1' $
Endif $

@
39098; 2001; Documentation/QRG - PARAM,USETPRT and PARAM,SETSTR1
In Version 70.7 param,usetsel has been changed such that values greater
than 0. are not acceptable. If USETSEL is set greater than 0 the
following message occurs:
^^^ USER WARNING MESSAGE 9109 (SEKR0)
^^^ USETSEL IS GREATER THAN ZERO WHICH IS OBSOLETE. USETSEL WILL BE
RESET TO -1 TO PRINT ALL SETS.
^^^ USER ACTION: SPECIFY PARAM,USETSEL,0 AND SET NAMES ON
PARAM,USETSTR1 THRU USETSTR4.
Avoidance: The syntax for the r USETSTRi paramerter is:
PARAM,USETSTR1,M:R for more info see the description under TABPRT in the
Release Guide for Version 70.7.
@
39109; ; Eigenvalue Analysis - GIV, MGIV, HOU, MHOU - Poor Answers
When using methods GIV, HOU, or their variants, if there are repeated
roots their eigenvectors may not be orthogonal. A small test problem with
9 repeated roots produced only three orthognal vectors.
Avoidance: An avoidance is to use the LAN method.
@
39125; 2001; Elements - CRAC3D, MAT9 - UFM 3005
If a model has CRAC3D elements that referene MAT9 entries, , then the job
may fail with the following error messages:
**** USER FATAL MESSAGE 3005 (SNGERR)
Attemp to operate on singular matrix AI in subroutine SD92ND.
Avoidance: There is no known avoidance.
@
39129; 70.7; DMAP - DIAGONAL - UFM 5250 DIAGNL Negative non-integer exponent
If the diagonal terms of a matrix are raised to a negative power using
the DIAGONAL command with the 'SQUARE' option, if there are negative
off-diagonal terms, the program will fail with:
UFM 5250 DIAGNL Negative non-integer exponent
Avoidance:
First extract the diagonal terms using a POWER of 1.0, then raise the
resulting matrix to the desired negative power.
@
39150; 2001; Documentation - DSAPRT, TRIM, STABDER, - Formatted Output not Su
pported
The formatted sensitivity prints produced using the DSAPRT case control
command are not supported for aeroelastic responses (DRESP1 RTYPE = TRIM,
STABDER or FLUTTER
Avoidance: Use the unformatted sensitivity prints.
@
39168; ; DMAP - Access Violation Error
A dmap had a purged input data block of the form call xyz a,,c . . . It
in turn called another dmap, with the format call wyz b// $, where b is
the purged input. If an attempt was made to test for the presence of b
with paraml or to print it with matprn, an access violation occurs. If
all inputs to a subdmap are purged, an executive system error occurs.
Avoidance: An avoidance is to add another non-purged datablock to the
calling argument. This purged argument does not need to be used in the
subdmap. Its presence avoids the error that results when all inputs to
the subdmap are purged

MAIN: "Access violation" (C0000005) exception encountered.

@
39172; 2001; DMAP - paraml /dbname/'name' - Datablock is discarded
If the paraml option 'name' is used on a data block in the calling
argument of a subdmap, and the datablock used as input to paraml is an
NDDL-scratch data block that has previously been equivalenced, the data
block is discarded at the end of the subdmap.
Avoidance: There is no known avoidance.

@
39191; 2001; Aeroelastic Analysis - SDR1 - UFM 4574, SWM 3001 UXAX Does not e
xist
Aeroelastic analyses that result in no a-set elastic deformations for unit
perturbations of the controllers will result in the following error
 message:
*** USER FATAL MESSAGE 4574 (SDR1)
DATA RECOVERY STARTED BUT NO DISPLACEMENT VECTORS AVAILABLE
*** SYSTEM WARNING MESSAGE 3001 (SDR1)
THE INPUT DATA BLOCK NAMED UXAX AT POSITION 3 DOES NOT
EXIST.
The problem is that UXAX is purged because there are no elastic
deformations. This is a somewhat rare event, that can happen if you
attempt to model a "rigid" airplane. There are a number of ways in which a
rigid airplane can be better modelled. In particular, in the TRIM
specification, you can set the AEQR field to 0.0. That will result in a
flexible vehicle with no elastic feedback to modify the aero loads. This
is the typical "rigid" behavior that is desired. However, there is a DMAP
avoidance that can be used:
compile aeopu
alter 'mpyad *gdka.*uxax'(,-1)
type parm,,i,n,nou $
DBSTATUS UXAX//S,N,NOU $ CHECK IF THERE ARE RESTRAINED SOL'NS
if ( nou > 0 ) then $
alter 'sdr1 *uset.*uxax.*uergx'
endif $ nou
This alter will avoid the problem and still supply the rigid and elastic
increment forces/moments onto the aerodynamic and structural meshes for
unit solutions and trim. These data can be placed on an XDB file (PARAM,
POST, 0) and visualized in the "Results" tool of MSC.Patran 8.1..
@
39198; 2001; DMAP - MATGPR - UWM 3059
If the 2nd parameter is unspecified on the MATGPR DMAP module, then UWM
3059 is issued. This warning can be safely ignored.
Avoidance: Specify the 2nd parameter.
@
39203; 2001; Nonlinear Analysis - PARAM,POST - Missing Postprocessing Output
PARAM,POST does not work correctly in the Case Control of a Version
70.7 Nonlinear run.
Avoidance: Insert the following DMAP alter;

COMPILE NLSTATIC
alter 'PVT PVT,CASESX2',''
dbstatus casesx2//s,n,nocase2 $
if ( nocase2>0 ) PVT PVT,CASESX2/ $
endalter
@
39208; 2001; Plotting - XYPLOT, XYTRAN - Missing Plot Output
If a non-existant grid point is requested by the XYPLOT command, the
XYTRAN module makes a premature exit. i.e. when a displacement plot
command cannot find the grid point, all of the acceleration and element
stress/force plots could be skipped.
@
39213; 2001; Dynamic Anaysis - FRLG - Wrong Answers.
 In MSC.Nastran it has been found that FRLG can produce incorrect loads
when insufficient memory is given.
Avoidance: If the answers are suspect, run the analysis at only the
suspected problem frequencies and see if the answers change. If the
answers change increase memory and rerun the job.
@
39215; 2001; Mass Matrix - UFM 4646, Grids and Components not Identified
The indefinite mass matrix check may not correctly identify the grid
point and degree of freedom.
AN INDEFINITE MASS MATRIX DOF DETECTED AT GRID: 0 COMPONENT: 0
Avoidance: The run may be allowed to continue by setting system(303).
MSC's elastic elements do not produce an indefinite mass matrix. Massless
mechanisms may cause this error, dmig input, or improperly coded rigid
elements.
@
39219; ; Complex Eigenvalue Analysis - EIGC - Missing Eigenvalues
In MSC.Nastran Version 70.7 the new Block Complex Lanczos Eigenvalue
analysis method automatically reverts to real operations when no damping
matrix exists. This may cause roots to be missed in acoustic/fluid
problems.
Avoidance: Two possible avoidances are:
1. Introduce a false damping matrix by applying the following alter:
compile gma
alter 'acsyml\=',''
acsyml=(acsym='yes' and fs) $
endalter
2. Force the method to always use complex arithmetic by setting
system(108)=16384.
@
39224; 2001; Optimiztion/Design Sensitivity - PBEAM - UFM 7055
When an optimization task tries to design K1, K2 and other quantities
specified on the last two continuations of a PBEAM entry, then USER FATAL
MESSAGE 7055 (DOPR1C/DOPR1Z) will be printed.

Avoidance: Do not design K1, K2 and others on the last two continuations
of the PBEAM entry.
@
39235; 2001; Optimization/Design Sensitivity - UFM 3001 (DSAL)
In a SOL 200 senstivity analysis with multiple frequency subcases, an
error can result if an initial subcase has no responses that require a
direct (as opposed to adjoint) sensitivity analysis and a subsequent
subcase does.
Avoidance: A possible avoidance is to arrange the subcases so that any
that the initial subcase requires a direct analysis. Another avoidance
is to force the use of the direct senstivity calculation for all subcsases
with PARAM AUTOADJ NO.
@
39248; 2001; Optimization/Design Sensitivity - DVPREL1, PBEAML, PBARL - UFM 7
193
The above UFM message can be issued in error when there are both PBARL
and PBEAML entries being designed. The error does not occur when a single
PBARL is designed.
Avoidance: The only avoidance is to replace either the designed PBARL's or
the PBEAML's with PBAR's of PBEAM's and use the original dvprel1/dvprel2
techniques to design these properties.
@
39269; 2001; Optimization/Design Sensitivity - DRESP2 - SFM 4027 (WRITE), EC
3021
An optimization task may fail with the SYSTEM FATAL MESSAGE 4027 when the
following two conditions are met:
1) a DRESP1 response is referenced in the first subcase and
2) a DRESP2 response is retained in the second subcase where the DRESP2
response references to another DRESP2 entry.
Avoidance: Do not use a DRESP2 referencing a DRESP2.
@
39272; 2001; Optimization/Design Sensitivity - DRESP2 - Wrong Answers
An optimization task may generate incorrect type 2 response values when
the nested referenced DRESP2 entries are used in multiple subcases.
Avoidance: There is no avoidance.

@
39273; 2001.0.1; Elements - HEXA, Corner Stresses, SDR2 - Possible Wrong Answer
s
In Version 70.7, if the HEXA element is highly distorted, then the corner
stresses may be incorrect. The element center stresses are correct.
Avoidances: There are two possible avoidances.
1. Use the version 70.5 HEXA formulation by adding
nastran system(309)=1
to the top of the input file
2. Do not use the corner stress values.
@
39275; 2001; Optimization/Design Sensitivity - MODTRK, FREQ3, 4, 5- SFM4276 (
GREAD)
In SOL 200, SYSTEM FATAL MESSAGE 4276 may be issued if mode tracking is
requested and there is a modal frequency subcase that uses the FREQ3,
FREQ4, or FREQ5 entries. in the specification of the excitation
frequencies.
Avoidance: The use of FREQ3, 4 or 5 is not recommended with sensitivity
and optimization analysis since the excitation frequencies are not updated
when the structural model is changed. Further, the sensitivity
 calculations ignore the effect of the change on the excitation frequency
on the design response and are therefore inaccurate, at best, and may
cause the design to go in the wrong direction.
@
39280; 2001; Documentation - QRG - SET description under OUTPUT(POST) Missing
Keywords
The SET description on page 244 of the V70.5 QRG for OUTPUT(POST) is the
missing keywords [ELEMENTS, GRID POINTS].

@
39285; ; Optimization/Design Sensitivity - PARAM EIGD, DSAPRT(EXPORT), IU
NIT
SUBDMAP EIGVECDS does not contain the parameter iunit. Thus, when the
DSAPRT(EXPORT) Case Control command is used, OUTPUT4 fails because IUNIT
is undefined.
Avoidance: Apply the following alter:
INIT MASTER(S)
assign output4='a200.op4' unit=21 formatted delete
$
ID MSC, SWRO
TIME 100
DIAG 5,6,8,56
SOL 200
$need to rewind iunit before anything else
malter 'file.*append'
output4 //-1/iunit $
$ need to modify output2 from rewind to do nothing
malter 'output2.*dscmcol','output2.*dscmcol'
OUTPUT2 DSCMCOL,,,,//0/IUNIT/'DSCMCOL'/OMAXR $
compile eigvecds,nolist
alter 'nddl'
TYPE PARM,NDDL,I,Y,iunit
$need iunit id subdmap eigvecds
alter 'output4'
message // 'iunit='/iunit $
alter 'output4'(2,0),'output4'(2,0) $ remove rewind of iunit
$exit
@
39288; 2001; Elements - CTETRA, PSOLID - UFM 9050 (SEKRRS)
If a CTETRA with midside nodes references a PSOLID that has the IN field
equal to "TWO", the following error message occurs:
^^^ USER FATAL MESSAGE 9050 (SEKRRS)
^^^ RUN TERMINATED DUE TO EXCESSIVE PIVOT RATIOS IN MATRIX KLL.
^^^ USER ACTION: CONSTRAIN MECHANISMS WITH SPCI OR SUPORTI ENTRIES OR
SPECIFY PARAM,BAILOUT,-1 TO
CONTINUE THE RUN WITH MECHANISMS.
NASTRAN should give a warning message saying that IN = "TWO" is not
allowed and that the suggested IN = "THREE" is used instead, and the run
should continue.
 Avoidance: Use IN equal to "THREE".
@
39294; 2001; Optimization/Design Sensitivity - DRESP2, DNODE - Incorrect Sens
itivities
When an optimzation task includes a DRESP2 entry which also references a
DNODE and another DRESP2 entry, incorrect sensitivity coeffcients for the
DRESP2 entry may be produced.
Avoidance: Do not reference both DNODE and DRESP2 on a DRESP2 entry.
@
39304; ; Cyclic Normal Mode/Plotting - Incorrect Results in XDB
Cyclic Symmetry Modal Analysis produces incorrect XDB outputs. Several
calculated modes are truncated due to an apparent error in counting
logic.The printed f06 outputs are correct.
@
39309; 2001; P-elements - GP0 - GMLOAD applied to GMCURV for CBEAMs
For p-elements, if a GMLOAD is applied to a GMCURV in order to load
CBEAMs, a fatal error with no additional information occurs in GP0.
Avoidance: The avoidance is to apply the GMLOAD to the individual FEEDGEs
for the CBEAMs.
@
39314; ; Nonlinear Analysis - UIM 9005, Saved LOOPID is not correct
Even when an increment in nonlinear statics (SOL 106) cannot converge, it
still prints the following message, where "n" is the loopid of the
unconverged increment:
^^^ USER INFORMATION MESSAGE 9005 (NLSTATIC)
^^^ THE SOLUTION FOR LOOPID= n IS SAVED FOR RESTART
This message is not true since increment "n" has not been saved for
restart.
Avoidance: Ignore the message since only increments up to n-1 have been
saved.
@
39326; ; Database - SFM 1041 (GALLEX)
A limit in the database directory size may cause large jobs to fail with
the message SYSTEM FATAL MESSAGE 1041 (GALLEX).
Avoidance: This limitation can be overcome by increasing the buffsize
and/or setting system(158)=1000 and system(159)=500 to allow for a larger
expansion of the database.
@
39342; 2001; DMAP - DEFUSET - Entry Ignored without Warning
The DEFUSET function was removed in Version 70.7. Any such bulk data
entries are now ignored, without warning.
Avoidance: There is no known avoidance.
@
39375; ; FMS - ASSIGN, scr=yes - Database Not Deleted
In Version 70.7, if an ASSIGN statement is specified for the database, the
database will not be deleted at the end of the job even if scr=yes is
specified on the command line.
Avoidances: There are two possible avoidances.
1. Add the following FMS statement to the input file
init master(s)
2. Add a ",temp" to the end of the assign statement.
@
39378; ; Beam Library - CONNECT - Incorrect Properties for Multiple Beam
Servers
When more than one beam server is used, it may produce incorrect the
properties. This error occurs whe the Group names or evaluator class
names show certain pattern in which the first portion of the n-th group
name is identical to the name of the (n-1)-th group where n is => 2. In
addition the CONNECT entry including (n-1)-th group must precede the
CONNECT entry containing n-th group name. For example, the CONNECT entries
given here will produce the incorrect properties for the beamsrv1 group:
CONNECT BEAMEVAL beamsrv locbml and CONNECT BEAMEVAL beamsrv1 locbmls.
Avoidance: Do not use the problematic pattern as the group name for
multiple beam servers.
@
39386; 2001; Documentation/QRG - PARAM,S1 - Incorrect Documentation for SOL 1
09 and 112
Stress sorting was implemented in SOL's 109 and 112 in Version 69. The QRG
incorrectly shows that the associated parameters are not supported for
these solutions. .
@
39387; 2001; Optimization/Design Sensitivity - DRESP2 - SFM 3007 DOM9
An optimization task may fail with the 'SFM 3007 DOM9, SUBROUTINE DOM91'
message when the DRESP2 response references another DRESP2 entry.
Avoidance: 1) Convert the offending DRESP2 entries into a single DRESP2
entry or
2) Use the following dmap alter:
For v70.7
COMPILE DESCON
alter ' DSAD '
DBSTATUS RR2IDR//S,N,NORR2ID $
IF ( NORR2ID>0 ) THEN
PARAML RR2IDR//'TRAILER'/1/S,N,NRR2 $
TYPE PARM,,I,N,NRR2I
NRR2I = NRR2
PARAML RR2IDR//'DTI'/1/ S,N,NRR2I//S,N,R2ID $
IF (R2ID = 0) THEN
MODTRL RR2IDR //0 $
ENDIF
ENDIF
@
39391; 2001; Optimization/Design Sensitivity - EIGN, FREQ, FLUTTER - Wrong Se
nsitivities
Incorrect sensitivities can be produced in SOL 200 when all of the
following pertain:
a. There are superelements
b. A global response is selected as the objective
c. The response is not WEIGHT or VOLUME.
An example of responses are EIGN, LAMA, FREQ and FLUTTER.
Correct answers are provided when the response is constrained. It is only
when it is invoked by the DESOBJ case control command that incomplete
sensitivity information is provided.
@
39421; ; Nonlinear Analysis - GPFORCE, ESE - Loadstep Not Identified
The GPFORCE and ESE output in SOL 106 does not identify the currrent
LOADSTEP.
Avoidance: The loadstep can only be determined by the order of output..

@
39422; 2001; Distributed Memory Parallel - Execution fails with large value o
f "dmp"
Nastran runs with large values of the "dmp" parameter will fail in the
setup of the run. The exact value that is allowed is a function of
various factors like the size of the current path name.
@
39426; 2001.0.1; Optimization/Design Sensitivity - Multiple Eigenvalue Subcases
- Missing Output
If two eigenvalue subcases with different boundary conditions are run in
SOL 200, the second subcases eigenvalues are printed twice but the first
is not printed. The eigenvector results are printed correctly and the
optimzation is performed correctly.

Avoidance: Setting NASPRT=1 will produce correct eigenvalue results at

each design iteration, including the last. Version 70.5 can also be used
to get the correct prints.
@
39433; 2001; Inertia Relief - PARAM,INREL,-2 - UWM 4698
When a matrix to be decomposed requires 2 by 2 pivots, its rows may be
identified as having negative factor terms by maxratio and USER WARNING
MESSAGE 4698 may be issued. 2 by 2 pivots most commonly occur when using
Lagrange Multiplier techniques (LMT). The max ratio test is not valid for
this circumstance, so this warning message may be ignored. LMT methods
are used in sol 101 for param, inrel, -2.
@
39436; 2001; Aeroelastic Analysis - PARAM,POST,0 - Complex Eigenvectors Missi
ng from xdb
 If complex eigenvalues/vectors are extracted using the pk method, they
cannot be written to the xdb file (PARAM POST 0) due to an error in the
datablock labelling
Avoidance: The op2 file (PARAM POST -1) can be used to display complex
eigenvectors in MSC.Patran
@
39439; 2001; Inertia Relief - PARAM,INREL,-2 - Poor Performance
When the automated suport capability for inertia relief is selected
(param, inrel, -2) the problem may run many times slower than when other
methods of inertia relief are used.
Avoidance: A dmap avoidance is given in the sssalter library with the
name ir2a.v707.
@
39440; 2001; Optimization/Design Sensitivity - PCOMP - Wrong Answers
If any thickness of the various plies on a PCOMP entry is defined as a
design variable, Z0 is blank, and a DVPREL1 or DVPREL2 has not been used
to define ZO, then wrong answers may occur. This error occurs because the
the initial Z0 is used for the entire design optimization task even
though Z0 changes as the ply thicknesses change.. Thus, sensitivities and
optimization results are wrong.
Avoidance: A DVPREL1 or DVPREL2 entry must be used to define Z0 ( Z0
=-0.5 times the element thickness). Otherwise, the initial Z0 will be used
for the entire design optimization task even though Z0 changes as the ply
thicknesses change.. Thus, sensitivities and optimization results are
wrong for Z0=blank without DVPREL1 or DVPREL2 definition.
@
39449; 2001; Documentation/Plotting - CTETRA iem codes for nonlinear are wron
g
The Stress-Strain Item Codes, in Appendix A, of nonlinear HEXA, PENTA and
TETRA are wrong in the current Quick Reference Guide, The errors and
corrections are listed as follows:
1) The last row of Nonlinear CHEXA (Code=93), "19-146 Items 3 through 17
Repeated for 8 corners", should be changed to
"19-146 Items 3 through 18 Repeated for 8 corners ".
2) The last row of Nonlinear CPENTA(Code=91), "19-146 Items 3 through 17
Repeated for 8 corners", should be changed to
"19-114 Items 3 through 18 Repeated for6 corners".
3) The last row of Nonlinear CTETRA(Code=85), "19-146 Items 3 through 17
Repeated for 8 corners", should be changed to
"19- 82 Items 3 through 18 Repeated for4 corners".
@
39452; 2001; External Superelements - ASET, USET, Residual Vectors - Possible
Wrong Answers
If an explicit ASET and a user USET (i.e. U1, U2, U3, U4, U5, or U6) are
specified, then the USET table is corrupted. The ASET will be reduced by
the dof which are common with the user set (u1-u6). and wrong answers may
occur.
@
39473; 2001; Nonlinear Analysis/Hyperelastic Elements - PLSOLID, MATHP, PCOMP
- SFM 3003
When hyperelastic material properties (PLSOLID and MATHP) are used
together with composites (PCOMP) or the alternate form of beam property
(PBCOMP) in a nonlinear analysis, Nastran fails in IFP6 if MATHP refers
to any TABLES1 entry.
Avoidances: This error can be avoided if user specifies the material
constants instead of using tables.
@
39482; ; Distributed Memory Parallel - DMP - Misleading f04 mesages
When running jobs using DMP (distributed memory parallel) with more than
2 cpus, the f04 file may only show the execution time of 2 cpus. The log
file does have the correct information. The results are correct.
Avoidance: Refer to the log file.
@
39488; ; Performance - DECOMP - Solver sequencer causes poor performance
The automatic sequencing options in the DCMP and READ solver modules go
from a grid point based sequencing to a DOF sequencing under some
circumstances.
Avoidance: TAN 4593 describes the reasons for this transition, and
methods to avoid the transition. It has been observed that DOF sequencing
can be more than 10 times slower than grid point sequencing.
@
39490; ; Nonlinear Transient Analysis - NLTRD2 - SFM 4276 UNPACK
It is possible that a nonlinear transient analysis may fail with BIOMSG
33, followed by SYTEM FATAL MESSAGE 4276, indicating an UNPACK error.
Avoidance: The exact cause of this problem is under investigation. One
potential cause is epoints. Replacing epoints with spoints may solve the
problem.
@
39502; ; Nonlinear Heat Transfer - RADBC - UFM 4551 (NCONVG)
If a nonlinear heat transfer analysis models radiation across a gap using
the RADBC entry the job is likely to have trouble converging and will
issue the message:
*** USER FATAL MESSAGE 4551 (NCONVG)
*** STOPPED PROBLEM DUE TO FAILED CONVERGENCE
Avoidance: Use a CONV entry to model the radiation across the gap.

Avoidance: Model radiation across gap using convection with exponent equal
to 4.0
Or run the model in transient analysis until it reached steadty state
solution
@
39508; 2001; Nonlinear Analysis/Hyperelastic Elements - PLSOLID, MATHP
Nastran job may fail in nonlinear analysis when the specified number(s) of
NA (order of polynomial function of the distortional strain energy) and/or
 ND (order of polynomial function of thethe volumetic strain energy) in
MATHP card are higher than the real rank(s). The real ranks are
computered from user specified material constants (Aij and Di) or tables
(TABLES1) in MATHP bulk data card by NASTRAN automatically.
Avoidances: This error can be avoided if user controls the number of
order of NA and ND - don't make them higher than the possible ranks based
on the user's input of Material constants (Aij) or Tables (TABLES1) in the
MATHP bulk data card.
@
39509; ; Transient Analysis - PARAM,ADSTAT, LFREQ, STATSUB - UFM 6143
If a modal transient analysis is run using PARAMs ADSTAT and LFREQ and
the case control command STATSUB, then the run will fail with UFM 6143
indicating incorrect input to the MERGE module.
Avoidance: Insert the following DMAP your file:
compile super3
alter 'matmod.*lm1.*mass'
type parm,,rs,y,lfreq
if(lfreq>0.)then $
$
$ correct error if ADSTAT, LFREQ, STATSUB are all used
$ this is a simple correction and will not work for the general case
$
paraml freq//'trailer'/2/s,n,nmodes $
type parm,,i,n,rmmodes
rmmodes=nmodes-NRUH1 $ number of modes removed by LFREQ - NOTE this alter
does
$ not check for HFREQ or NMODES
matgen ,/lfmodes/6/nmodes/rmmodes/NRUH1 $ partitioning vector based on
removed modes
partn freq,,lfmodes/,nfreq,,/1 $
partn mass,,lfmodes/,nmass,,/1 $
equivx nfreq/freq/-1 $
equivx nmass/mass/-1 $
endif $
@
39510; ; Nonlinear Analysis - Incorrect Loadstep ID after UIM 6189
If a Nonlinear analysis issues the message:
*** USER INFORMATION MESSAGE 6189 (NCONVG)
*** BEST ATTAINABLE SOLUTION HAS BEEN FOUND ***
Then for the next load increment the 3 permitted bisections are going to
take place without convergence. The displacement output for the "BEST
ATTAINABLE SOLUTION" will then be labeled with a load factor that includes
the increment from the final step with 3 bisections.
Avoidance: Set the intermediate output (INTOUT) field 9 on the NLPARM
entry to YES.
@
39556; 2001; Eigenvalue Analysis - EIGRL, READ - UFM 5401
USER FATAL MESSAGE 5401 is issued when a Lanczos job does not find any
roots. Possible causes are given. One cause that is not given in the
 message is a large maximum ratio indicating a modeling error.
@
39560; 2001; Optimization/Design Sensitivity - ANALYSIS=FLUTTER SFM 4276 (tbk
) at DSAPRT,DSAD
A design optimization analysis may fail in the DSAD or DSAPRT modules if
ANALYSIS=FLUTTER
*** SYSTEM FATAL MESSAGE 4276 (tbk)
ERROR CODE 7777 PID= 0
*** USER INFORMATION MESSAGE 4276 (tbk)
@
39572; 2001; Normal Modes - SOL 103 - Bad Eigenvalues
Bad eigenvalues may occur on the Version 70.7.2 NEC system with a build
date of December 15, 1999. This error has been corrected in Version
70.7.2 on the R3 CD. The error was fixed without changing the Version.
The build date of the corrected executable is June 28, 2000.
Avoidance: There is no avoidance.
@
39609; ; Miscellaneous - SGI ffio may result in a traceback
The FFIO feature may cause problems on some SGI operating systems. The
jobs may stop wih a traceback in ffopen.
Receiving signal number 5: Trace/BPT trap
Avoidance; Setting ff_io_opts = 'casecon.DB* casecon.db* casecon*SCR*
casecon*scr* may solve the problem. Else turn off FFIO.
@
39612; ; DMAP - ORTHOG Module Incorrectly Declares a Matrix to a Null Mat
rix
The ORTHOG module may incorrectly declare a matrix with small terms to be
null and the following message is then issued.
*** SYSTEM WARNING MESSAGE 6467 (ORTHRD)
A NULL COLUMN WAS ENCOUNTERED AT COLUMN NUMBER = 1 DURING
ORTHOGONALIZATION
@
39624; 2001; P-Elements - MATTi - Wrong Answers
If temperature-dependent material properties are being used with p-element
shells,and the temperature field is not constant, the answers may be
incorrect.
Avoidance: There is no avoidance.
@
39630; ; Modal Frequency Response - multiple subcases, element output - M
issing Output
If multiple subcases exist in a modal frequency response analysis, and
element-type output is requested (e.g., force) in the second or subsequent
subcases without requesting it in the first subcase, then the element
output will be missing.
 Avoidance: There are two possible avoidances. The first one is
recommended.
1. Add the same request in the first subcase. If the printed output is
not desired, use the (plot) option.
2. Use param,ddrmm,-1. This method is not recommended as it may cause
substantial additional runtime.
@
39633; 2001; Nonlinear Buckling Analysis - PARAM,BUCKLE,2 - UWM 8592 (REIG)
Nonlinear buckling analysis in a cold start run was made possible in
Version 70.7 by specifying param, buckle, 2. A buckling analysis should
be able to be selectively perfromed for each subcase that has a method =
n specified. However, if a method request is not made in each subcase
the job will fail with the following message:
*** USER WARNING MESSAGE 8592 (REIG)
THE METHOD RECORD IN CASE CONTROL IS MISSING OR INCORRECT.
THIS WILL CAUSE THE FOLLOWING FATAL ERRORS.
USER ACTION: PUT A METHOD ENTRY IN CASE CONTROL WHICH POINTS TO AN
EIGR, EIGRL, OR EIGB RECORD IN BULK DATA.
*** USER FATAL MESSAGE 3032 (READ)
UNABLE TO FIND SELECTED SET ( 0) IN TABLE (EED ) IN SUBROUTINE
(REIG).
0FATAL ERROR
Avoidance: Insert the following DMAP alter or make the method request
above the subcase level.
compile nlstatic
alter 'cbuckle=-1'(2,1), ''(2,2)
else if ( newp>-1 and buckle=2 ) then $
PARAML casesx2//'DTI'/-1/5//S,N,METHod $
if ( method=-1 ) cbuckle=2 $
endalter
@
39636; 2001; Aeroelastic Analysis - SPLINE - UFM 7358
In an aeroelastic analysis, USER FATAL MESSAGE 7358 may be issued in
error. This has been observed to occur when the the structural grid
points used by the spline are separated by a length that is much less than
one.
Avoidance: Recasting the dimensions of the problem so that the distances
are greater than one should provide a workaround. Another alternative is
to select additional gridpoints that are outboard or inboard of the
selection. This error does not occur in Version 70.5 so this version
could be used as an avoidance.
.
@
39643; 2001; Optimization/Design Sensitivity - PCOMP, DSAD13 - Read Past End
of Record
In a Nastran optimization job (sol 200) the message
ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD (DSAD13)
 is produced with composite responses (DRESP1 RTYPE = CSTRESS,CSTRAIN or
CFAILURE) if the lamina number requested in the ATTA field is greater than
the number of lamina on the associated pcomp entry.
Avoidance: Correct the incorrect lamina number request
@
39663; 2001; Beam Library - DOM9 - Segmentation Error
A large optimization job that utilized the external beam library
capability gave a segmentation error in the DOM9 module.
Avoidance: Use Node Locked instead of Flexm
@
39682; 70.7; Aeroelastic Analysis - GUST, RMS - Wrong Answers
In a symmetric aeroelastic analysis, if an aerodynamic panel lies on the
plane of symmetry wrong GUST RMS values or a rank deficient matrix
result.
Avoidable: For symmetric analysis do not define panels that lie on the
plane of symmetry. Version 70.7 allows aerodynamic panels to lie on the
plane of symmetry by implementation of the AESYMXZ and AESYMXY commands.
@
39695; GL; Eigenvalue Analysis - PARAM,RESVEC,YES - Traceback in READ
PARAM,RESVEC,YES may cause a floating point error in READ. This occurs
when the number of modes plus residual vectors exceeds the rank of the
mass matrix. This problem typically only occurs in very small test models
where all modes may be unexpectedely calculated.
Avoidance: Modify the EIGR(L) record so not all modes are computed. If
there are disjoint submodels, do not get all of the modes for any
submodel for which there will be residual vectors.
@
39697; 2001; SSSALTER library - pstmac2a.v707, missing eigenvalue table - SF
M 3001
When using the pstmac2a.v707 alter, if an eigenvalue table is not
provided, then the job will fail with SFM 3001
Avoidance: There are two possible avoidances:
1. Provide an eigenvalue table

or
2. Modify the following statement from
lamx ,,lamdum1/lamat $
to
lamx lamdum1,/lamat/nummodes $
@
39705; 2001; Virtual Mass - MFLUIED, ZFS - UWM 5411
When using virtual mass, if any grid point of an element listed on an
ELIST record lies very near the free surface, a nonpositive definite mass
matrix may occur. This may result in the following message:
 *** USER WARNING MESSAGE 5411 (REIGL8)
NEGATIVE TERM ON DIAGONAL OF MASS(VIBRATION) OR STIFFNESS(BUCKLING)
MATRIX DETECTED. ROW NUMBER = 56 VALUE = -3.60212D+01
The number of modes calculated may not match the Sturm sequence output.
Avoidance: Change the free surface level slightly so that no grid point
lies near the free surface. Near means a distance less than one percent
of the square root of the area.
@
39713; 2001; Miscellaneous - DMIG - UFM 2070 (MTRXIN)
DMIG matrices may be selected by the (K2gg, M2gg, and B2gg) commands, and
the (K2pp, M2pp, and B2pp) commands. If the same matrix is requested
twice, for example
K2gg = AA $
M2gg = AA $
or

K2pp = AAX $
B2pp = AAX $
Then the following misleading fatal message is output:
*** USER FATAL MESSAGE 2070 (MTRXIN)
REQUESTED DMIG MATRIX AA IS UNDEFINED.
Avoidance: Duplicate the DMIG matrix and rename.
@
39714; 2001; Case Control - TFL - Poor Message, BIOMSG, ERROR 3001 MTRXIN
If the Case Control contains the command TFL = [sid] but there is no TFL
entry in bulk data, a non-standard exit is taken. For Version 70.7 on the
NT, the following exit was taken in the MTRXIN module:
BIOMSG: ERROR 3001 HAS OCCURRED IN ROUTINE FREAD , FILE INDEX =
0.
STATUS = 0

Avoidance: Correct the input data.

@
39718; 2001; Case Control - MPRES - Missing Output Labels
If SORT2 format output is requested with the MPRESS Case Control command,
the associated "PRESSURES FOR ELEMENTS ON FLUID STRUCTURE BOUNDARY"
output will have no element labeling on it.
Avoidance: Request SORT1 output
@
39724; 2001; SSSALTERs - modevala.v707 -MAX Eigenvector Normalization causes
Wrong Ans.
SSSALTER modevala.vxx assumes that the eigenvectors are normalized to unit
generalized mass. If they are not, incorrect results will occur.
 Avoidance: Always normalize your eigenvectors to MASS (the default).
@
39742; 2001; Frequency Response - STATSUB - Missing Output
If there is a static subcase in SOL 111 to apply a prestress followed by a
dynamic subcase, and if output requests are only in the dynamic subcase no
output is obtained.
Avoidance: Put the output request above all subcases or insert the
following DMAP alter:
compile super3
alter 'if ( desopt ) then',''
if ( true ) then $
endalter $
@
39771; 2001; Optimization/Design Sensitivity - PBEAML requires SOUT=YES
If a tapered PBEAML is being designed and the user has a value of NO for
one of the SO stress requests, incorrect sensitivity values will be
created. The optimizer will then be unable to make progress. A fatal
message may be issued or the job may indicate a hard convergence at design
cycle 1.
Avoidance: Do not use SO(j) or SO(B) = NO.
@
39773; 2001; Optimization/Design Sensitivity - PBEAM, PBEAML - UFM 7052, 7197
Solution 200 jobs may fail with USER FATAL MESSAGE 7052 or USER FATAL
MESSAGE 7197 when the field name of an intermediate section of a PBEAML is
referenced on a DVPREL1 entry.
Avoidance: Split a PBEAM with multiple intermediate sections into
multiple CBEAM/PBEAML. For the new PBEAMLs, provide properties for end A
and B only.
@
39786; 2001; Documentation/QRG - Continuation Entry Replication produces UFM
6721
The "=(C)," immediate continuation format for card replication as defined
on page 428 of the v70.5 Quick Reference Guide no longer works.

Avoidance:
Use "," instead. For instance, instead of:
=(C), , *7, *7, *7, *7, *7, *7, *7
use:
, , *7, *7, *7, *7, *7, *7, *7
@
39791; 2001; DMAP - SDR2 - SFM 3001, Wrong Datablock Position Number
The DMAP message:
*** SYSTEM FATAL MESSAGE 3001 (SDR2)
THE INPUT DATA BLOCK IN POSITION 10 DOES NOT EXIST.
 may incorrectly identify the missing data block position.
Avoidance: Use DBSTATUS before the offending line to verify the status of
the input datablocks.
@
39806; 2001; Miscellaneous - Missing "BEGIN BULK" causes part of input to be
ignored
Starting with Version 70.7, if an input deck contains a "BEGINBULK" entry
(no space between "BEGIN" and "BULK") and the bulk data section contains
INCLUDE statements, then bulk data cards that appear BEFORE the "INCLUDE"
statement will be ignored.
@
39808; ; SSSALTERs - alter1ga - QGDYN is Output Twice
If MPCFORCE output is requested in a run using the alter1ga.v707 alter the
run will fail with a message warning that datablock QGDYN is being output
twice.

Avoidance: Either remove the MPCFORCE request or request an updated

version of the alter from MSC.
@
39847; 2001; Documentation - QRG - DCONADD and DCONSTR
When the same DCID specified by a DESSUB command is referenced by DCONADD
and DCONSTR entries, the constraints defined with the DCONSTR entry are
ignored.
Avoidance: Two avoidances exist:
1. Do not reference the same DCID with the DCONADD and DCONSTR entries.
2. If one insists on referencing the same DCID with both entries, include
that DCID in the DCONADD entry as the DCONSTR identification number.
@
39871; ; Distributed Memory Parallel - DMP, MPCs - signal caught, mpi_ab
ort mpi bug
When using Distributed Memory Parallel (DMP) and there are many MPCs, then
a large amount of memory is needed. If insufficient memory is allocated,
then the job will fail with the following messages in the log file
----------
SIGNAL CAUGHT:
Signal number = 15
Signal code = -1
Signal pid = 3426
Signal uid = 9450
Signal tid = 1
Fault addr = 00000d62
FP inexact = 0x0
FP divide0 = 0x40000000
FP underflow = 0x0
FP overflow = 0x20000000
FP invalid = 0x0
----------
TRACEBACK:
Avoidance: If there are large number of MPCs, use a single processor
 instead of DMP.
@
39885; 2001; Materials - MAT2 - UFM 2042 (EMG)
Starting with Version 70.5, if a bulk data file includes PSHELL/MAT2s with
9-digit MIDs where any of the MIDs are greater than 400,000,000 then the
deck fails with the message:
USER FATAL MESSAGE 2042 (EMG) MISSING MATERIAL TABLE ****** FOR ELEMENT
Such PSHELLs/MAT2s are routinly constructed from the PUNCHed echo of a
PCOMP entry but they must not be mixed with other PCOMP entries. The
message has been improved in Version 2001.

@
39889; 2001; Distributed Memory Parallel - DMP, EMG - Job Hangs
A distributed memory parallel Statics, SOL 101, job will hang if one of
the tasks gets a fatal error in EMG.

Avoidance: Other than correcting the input there is no avoidance.

@
39892; 2001; DMP - Missing Library - libmpi.so not found
The distributed version of MSC.Nastran Version 70.7 on linux requires a
file libmpi.so.1. This file was not placed on the CDROM.
There are two possible avoidances:
1. The file may be created by using the lam-6.3.1.tar.gz file which is in
$MSC_BASE/msc707/linux with the following commands:
configure --with-fflags=-fno-second-underscore --with-shared
make install
These commands create /usr/local/lam-6.3.1/lib/libmpi.so.1. Please
copy this file into $MSC_BASE/msc707/linux.
2. The file may be downloaded from the MSC ftp site from:
ftp://ftp.macsch.com/pub/msc-products/NASTRAN/V70.7/linux/

or:
ftp.mscsoftware.com/pub/msc-products/NASTRAN/70.7/linux/
@
39896; 2001; Eigenvalue Analysis - READ, EIGRL, RBE3 - SFM 7340 (LNNHERR)
Lanczos eigenvalue problems that ran in Version 70.5 may fail in 70.7 with
the message:
*** SYSTEM FATAL MESSAGE 7340 (LNNHERR)
process error reported by subroutine LNNP2CD
Avoidance: This error may be related to the re-formulated RBE3 elements,
Setting rotational dof set as independent, appears to be the cause of the
problem.
If this is the case, either setting:
 NASTRAN SYSTEM(310)=1
or removing all rotational dof from the independent set may fix the
problem.
@
39901; ; TRANS/RECEIVE - Missing libF77.so.4
It is possible that the fortran libraries cannot be found for the
NASTRAN utilities 'trans' and 'receive'. This is evidenced by an error
message like:
ld.so.1: receive: fatal: libF77.so.4: open failed no such file or
directory
Avoidance: A possible solution is as follows (example for SUN SOLARIS):
Ensure the following three libraries( libF77.so.4; libM77.so.2;
libsunmath.so.1) exist in directory
<installation-dir>/msc/msc707/solaris.
Then set the environmental variable LD_LIBRARY_PATH to that directory.
@
39904; 2001; Superelement - heat transfer - UFM 6599
Superelement heat transfer works in SOL 101 by specifying APP HEAT.
However, it only works when there is no convection in the problem.
NASTRAN heat transfer applies convection by first defining a boundary
element (CHBDYG or CHBDYE) and then applying CONV with the same element ID
on that boundary element. For example:
chbdyg,101,,area4
,5,6,17,16
conv,101,1,,,99
However, there is a limitation in superelements that all element IDs must
be unique else the following fatal message is encountered:
*** USER FATAL MESSAGE 6599 (SEPUEC)
DUPLICATE ELEMENT IDS CONV 101 AND CHBDYG 101 IN
SUPERELEMENT 0
0FATAL ERROR
Avoidance: Do not use superelements.
@
39931; 2001; DMAP - FIles Required to Create Delivery Database Missing on NT
If you are using Windows NT, the files necessary to create a delivery
database are not in the "del" directory.
Avoidance: Contact MSC and request a copy of the necessary files

@
39942; ; Nonlinear Normal Modes - PARAM,FOLLOWK - Wrong Answers
Using SOL 106 to perform a nonlinear static analysis of a rotating beam
with followk effects leads to wrong results when the modes are calculated
by restarting with SOL 103 using the NMLOOP parameter. Restarting SOL
106 from a converged solution with just the NMLOOP parameter and a
 METHOD case control request yields onty the results of the nonlinear
static analysis, no normal modes are computed.
Avoidance: To get the modes from a restart with SOL 106, add a static
subcase with the appropriate values of SUBID and LOOPID set with the
NMLOOP parameter and METHOD case control command, then the nonlinear
normal modes could be obtained.
@
39962; 2001; Superelements - PARAM,RESVEC - Wrong Answers
When using superelements or residual vectors, there is a hard-coded limit
of 999 eigenvectors in the program for any single component. If a
component has more than 999 modes calculated, the program will incorrectly
truncate it to 999 modes with no warning.
Avoidance: Insert the following DMAP alter to remove the limit:
compile resvec nolist noref
alter 'read.*kzz,mzz',''
paraml kzz//'trailer'/1/s,n,zzsize $
READ KZZ,MZZ,,,,,,,,,,,/
ALAMA1,APHX1,MX,OX,EX,/
'MODES'/S,N,NV////-1/'LAN'////zzsize/0////PRTSUM $
@
39963; 2001; Nonlinear Analysis - CELAS1 - Incorrect Sign
The element force output for CELAS1 in nonlinear solutions has an
incorrect sign in MSC.Nastran Version 70.7.

@
39994; ; Documentation - QRG, YTAXIS - UFM 676
If the XYPLOT command YTAXIS is used, the program will incorrectly issue
UFM 676 indicating that it is not recognized.
*** USER FATAL MESSAGE 676 (IFP1XY)
YTAXIS IS NOT RECOGNIZED ON ABOVE CARD.
Avoidance: Remove the command
@
39995; ; Documentation - DLINK - UFM 7011 (DOPR1B) with Concatenated DLI
NKs
The following remark show be added to the DLINK bulk data description in
the QRG. i
Remark 6.
If a design variable is specified as dependent on a DLINK entry, then it
cannot be specified as independent on another DLINK entry.
@
39997; ; Case Control - PARAM,INREL, REPCASE - UFM 6143
If conventional static solutions are mixed with inertia relief in the
same run and REPCASEs are used, the run may fail with UFM 6143 indicating
 incompatible matrices.
Avoidance: Do not use REPCASEs when conventional statics and inertia
relief are both being executed in the same run..
@
40001; ; Optimization/Design Sensitivity - SDAMPING - Wrong Answers
The BHH value for modal damping is computed as Bi = 2 * Mi * (omega) *
(zeta) . Mi and omega can vary with the design variable and therefore
affect the modal damping sensitivity. However, these perturbations are
ignored in the current sensivity analysis with ANALYSIS = MFREQ. The
discrepancy associated with neglecting this term is most pronounced at
resonant peaks.
Avoidance: Alternative forms of damping, such as PARAM G, cvisc and cdamp
elements and material damping are supported and could be substituted for
the modal damping input.
@
40007; 2001; Distributed Memory Parallel - SUPER - Fails in DISFBS
A distributed memory parallel job with SUPER Case Control commands will
fail in DISFBS. Version 2001 traps this invalid input earlier in SEQP and
issues the following fatal message
*** USER FATAL MESSAGE 7567 (SEQP)
THE SUPER CASE CONTROL COMMAND HAS BEEN SPECIFIED.
USER INFORMATION: THE SUPER COMMAND IS NOT PERMITTED WITH THE
AUTOMATIC MODEL PARTITIONER.
USER ACTION: CONSOLIDATE AND REWRITE THE SUBCASES WITHOUT THE SUPER
COMMAND.
@
40008; 2001; Case Control - K2GG, K2PP, M2GG, M2PP, P2G - Unexpected Answers
DMIG entries are selected by the K2GG, M2GG, P2G, and K2PP, M2PP case
control commands. If a case control command selects a bulk data DMIG
entry, and there are no DMIG or TF entries in the bulk data file, the
module that reads this matrix returns without any error messages, and the
run may go to completion although it is missing the selected DMIG
information.
Avoidance: An avoidance is to put in at least one legal DMIG matrix. The
program will provide good diagnostics and a fatal error exit if any
selected matrices are missing from bulk data.
@
40037; ; Heat Transfer Analysis - VIEW3D - Wrong Answers
The View factor calculation may have trouble detecting elements when the
elements are very close. This may lead to wrong answers.
Avoidance: Decrease the ETOL error estimate from default 0.1 to a small
number such as 0.0001 on the VIEW3D bulk data entry , so that the VIEW3D
program will go and calculate the view factor using the contour
integration method
@
40040; 2001; P-Elements - TEMP(LOAD), ADAPT - SFM 3003
When the ADAPT module is retrieving the thermal loads to calculate the
error estimators, the following fatal error may occur:
 *** SYSTEM FATAL MESSAGE 3003 (ADAPT)
ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK ETT
(FILE 109) IN SUBROUTINE GETEMP
Avoidance: The error appears to occur with multiple element types.
Currently the only workaround is to specify temperatures using TEMPD.
@
40077; ; Documentation - DMAP Module Dictionary - SUBDMAP NOT EXECUTED
If a SUBDMAP is called with no output datablocks or parameters, then it
may not execute. Symptoms will be missing output and a statement in the
.f04 file "SUBNAME NOT EXECUTED." where SUBNAME is the name of the
subdmap.
Avoidance: Add an output block to the calling sequence, or add a
parameter.
@
40088; ; SSSALTERs - mtacca - Wrong Answres
SSSALTER mtacca.vxxx correctly calculates the rigid-body acceleration
terms based on the initial conditions, however it does not include the
effect of the initial rigid-body acceleration on the rigid-body velocity
and displacement terms.
Avoidance: The flexible results are correct. If the relative solution is
obtained (positive value of PARAM,RBMODES), the answers are correct..

@
40094; 2001; Transient Response Analysis - K2PP - SFM 3007 (TRD1)
If a complex matrix is input using the K2PP command, a transient response
run will fail with the message:
*** SYSTEM FATAL MESSAGE 3007 (TRD1)
ILLEGAL INPUT TO SUBROUTINE TRD1H
Avoidance: Use only real matrices in transient analysis.
@
40127; 2001; Heat Transfer - RADBND - Wrong Answers
In a steady state heat transfer analysis if RADBND is used to model the
radiation across a gap then wrong answers may result.

Avoidance: Use RADBC to model radiation across gap instead of view

factor.
@
40133; 2001.0.1; Frequency Response - TLOAD1, TLOAD2 - Wrong Answers
TLOAD1 and TLOAD2 entries define a time dependent load. In frequency
response a Fourier analysis is used to transform the time-dependent load
to the frequency domain. The TYPE field on the TLOAD1 and TLOAD2 entry
which specifies whether the dynamic excitation is and enforced
displacement, velocity, or acceleration, is ignored in frequency response.

Avoidance: There is no avoidance.

@
40134; 2001; Miscellaneous - ESTIMATE crashes when PROJ statement is present
The estimate program fails when a PROJ file management statement is
present.
Avoidance: Remove the PROJ statement.

@
40139; 2001.0.1; DMAP - MATGPR, USET - Bad DOF Labels
When USET entries are present in the bulk data section that place degrees
of freedom in user defined displacement sets, Form 1 (MATGPR GPL,USET,SIL,
...) usage of the MATGPR module may cause the degree of freedom output
identifier code to change from the expected T1/T2/T3/R1/R2/R3 to simply
the component direction number (1 thru 6). This will occur when the
degree of freedom internal number associated with the grid/components
defined on the USET entry is less than the number of grid points in the
bulk data. External grid identification numbers and printed matrix values
are correct. Only the component code identifiers may be displayed
differently. The identifier (whichever one is displayed) is correct.
Avoidance: A possible avoidance is to use the second form of the MATGPR
module (MATGPR BGPDT,USET,,.....).
@
40141; 2001.0.1; Optimization/Design Sensitivity - Optimizer has difficulty mak
ing progress
It has been observed that the Method of Modified Feasible Directions
(Method=1 on the DOPTPRM entry) can sometimes fail to provide a better
design even though it seems that it should. This occurs in situations
when there are many more active constraints than there are design
variables
Avoidance: The SQP algorithm (Method=3 on the DOPTPRM entry) may make
progress in these situations.
@
40156; 2001; Acoustic Analysis - fluid-only models - Missing Acoustic Pressur
e Output
If a model contains only fluid elements then the "COMPLEX ACOUSTIC
RESULTS" output will not be printed.
Avoidance: Insert a dummy structural element or the following DMAP alter:
compile sedrcvr
alter 'fscoup'
fscoup=true $
endalter
@
40157; 2001; Nonlinear Analysis - PARAM,BUCKLE,1, RESTART - SFM 6036
Nonlinear buckling analysis by a restart in SOL 106 with param, buckle, 1
will encounter SFM 6036 (divide by zero) if the number of increments for
the restart subcase is different from 2 or the bisection is triggered.
Avoidance: Specify INTOUT=YES on the NLPARM Bulk Data entry in the last
subcase or insert the following DMAP alter:
compile nlstatic
 alter 'IF ( BUCKLE=2 ) POUTF=1',''
if ( buckle>0 ) poutf=1
endalter
@
40158; 2001; Documentation/Rigid Elements - RBE3 - Different Answers
The reformulated RBE3 element in V70.7 can result in large differences in
answers between V70.5 and V70.7. This has only been observed on models
where rotations have been selected on the connected grid points. The
changes are most noticeable when the distances between connected grid
points are much greater than 1.0. The changes in answer between versions
do not occur when only translations are used for the independent dof.
Avoidance: To obtain same answers as in Versions prior to 70.7 set:
NASTRAN SYSTEM(310)=1
to use the old RBE3 formulation or do not use rotations on the "connected"
grid points on RBE3 elements. The answers will then be identical between
v70.5 and later versions.
@
40167; 2001; Optimizaton/Design Sensitivity - CBUSH - Wrong Answers on NEC
In MSC.Nastran Version 70.7 the NEC system has special code to enhance the
performance of design sensitivity tasks that utilize the adjoint method.
This code yields wrong answers when there are no coordinate systems in the
bulk data file and a CBUSH element is being designed.
Avoidance: Add a dummy coordinate system. For small problems where
performance is not a large concern, setting PARAM AUTOADJ NO will avoid
the problem by avoiding the adjoint method.
@
40171; 2001.0.1; Eigenvalue Analysis - Maxratio Control in READ module is Indir
ect.
Parameter MAXRATIO is used by the DCMP module in statics analysis to
control the degree of near-singularity that is allowed when computing
solutions. It is input to the DCMP module with a user parameter. This
parameter is not an input to the eigensolver, the READ module, although
READ makes tests based on its value. If the DCMP module has not run first
the value READ uses is set to 1.E7. If DCMP has run prior to READ the
value input by the user for DCMP is also used for READ.
Avoidance: DCMP is called in modal analysis when SUPORTs exist, when
superelement reduction is used, when residual vectors are computed, and
for several other reasons. However, it is not always called before the
READ module, that behavior is problem dependent. Look in the .f04 file
for the word DCMP before READ to determine what value has been used.
If you want to change the value that the READ module uses you could
schedule a small matrix to run in DCMP with dmap instructions, and set the
value there.
@
40184; 2001; Optimization/Design Sensitivity Analysis - PLOADi, CBAR - Incorr
ect Sensitivitie
In solution 200, sensitivities are incorrect if a distributed load is
applied on CBAR and/or CBEAM elements and there are more than one design
variable.
 Avoidance: There is no avoidance

@
40228; 2001; Superelements - CSUPER - Floating Point Exception in SEP1X
A model bringing in an external superelement may fail with a floating
point exception in module SEP1X if the GRID points defined on the EXTRN
entry do not exist.
Avoidance: Always make sure that the GRID points on the EXTRN entry exist
in the model.
@
40257; 2001; Dynamic Analysis - PARAM,RSPECTRA, STATSUB - UWM 4568 (INTERR)
Response spectra analysis does not work with prestressing (STATSUB Case
Control command). If attempted the following message is issued:
.*** USER WARNING MESSAGE 4568 (INTERR)
THE NUMBER OF SPECTRUMS SELECTED ON THE, DLOAD CARD DOES NOT MATCH
THE NUMBER OF SUPPORT POINTS.,
CALCULATIONS FOR THIS SUBCASE TERMINATED.
Avoidance: Insert the following DMAP alter:
compile semodes
alter 'call postreig'(,-1)
call dbstore casem,,,,//0/0/' '/0 $
compile postreig list
alter 'equivx '
call dbfetch /casem,,,,/0/0/0/0/0 $
PARAML CASEm//'TRAILER'/1/S,N,ICASE1//S,N,NOCASem $
PARAML CASECC//'DTI'/1/176//S,N,SUPER $ super=-1 means super=all
$ no need to append.
IF ( NOCASem>-1 AND NOT(RSONLY) AND SUPER>-1 ) THEN $
APPEND CASECC,CASEM/CASESEP4/1 $
PARAML CASECC//'TRAILER'/1/S,N,ICASE $
PARAML CASECC//'TRAILER'/3/S,N,MAXL $
PARAML CASECC//'TRAILER'/4/S,N,IPLT $
PARAML CASEM//'TRAILER'/3/S,N,MAXL1 $
PARAML CASEM//'TRAILER'/4/S,N,IPLT1 $
ICASE=ICASE+ICASE1 $
MAXL=MAX(MAXL,MAXL1) $
IPLT=MAX(IPLT,IPLT1) $
MODTRL CASESEP4//ICASE/ICASE/MAXL/IPLT $
ELSE $
EQUIVX CASECC/CASESEP4/-1 $
ENDIF $ NOCASRS1>-1
alter 'sep4',''
SEP4 CASEsep4,PCDB,EMAP,XYCDB,,,/DRLIST/
'REIG'/'SEID'/S,N,NODR/S,N,SEID $
alter ' sedr ',''
IF ( (NOCASem>-1 AND SEID=0) ) THEN $
EQUIVX CASEm/CASEX/-1 $
ELSE $
EQUIVX CASECC/CASEX/-1 $
ENDIF $
SEDR EMAP,CASEx,PCDB,DRLIST,XYCDB,,,
 MAPSF,PHGD,BGPDTSD,GDNTAB/
PHILS,CASEDR,PCDBDR,XYCDBDR/
'REIG'/SEID/S,N,NOUP/S,N,NOSRT1//S,N,NOOUT/
S,N,NOPLOT/S,N,NOXYPLOT/'SEID'/0 $
endalte
@
40261; ; Part Superelements - PARAM,OUNIT2 is ignored
The Nastran Quick Reference Guide indicates that PARAM,OUNIT2 may be
specified in either the Case Control or the Bulk Data. If it is specified
in the Bulk Data after a BEGIN SUPER entry it will be ignored.
Avoidance: Specify the parameter in the case control..
@
40263; ; Transient Response Analysis - EPOINTS, PARAM,ADSTAT - UFM 6143
If a model has EPOINTs and PARAM,ADSTAT is specified on a pre-stressed
structure in a modal transient response analysis, the run will fail with
UFM 6143 indicating that incompatible matrices were attempted to be
merged.

Avoidance: Insert the following alter:

compile super3
alter 'append.*freq0,mass'
if(n1<>nruh1)then $
type parm,,i,n,ted1 $
ted1 = n1 + 1 $
matgen ,/vone2/6/ted1/1/n1 $
merge ,,f0m,,,,vone2/f0m1/1 $
$
else $ n1=nruh1
$
alter 'merge.*f0m,,,,vone'
endif $
@
40275; 2001; DMAP - MATMOD Option 20 - Wrong Answers
If the first input to MATMOD Option 20 is purged, then wrong answers
result.
Avoidance: Rearrange the inputs so that the first input is never purged.

@
40276; 2001; Restarts - SCR=MINI - UFM 2071 in DPD
In any dynamic solution with static loads used to define dynamic loads, if
the initial run is made using SCR=MINI, the restart run may fail with the
following message:
*** USER FATAL MESSAGE 2071 (DPD2) will be issued indicating that the
DAREA set is not defined.
Avoidance: Insert a dummy DAREA or the following DMAP alter in the
restart run only:
compile (insert main subdmap name; e.g., sedtran for SOL 109, sedfreq for
 SOL 108, semfreq for SOL 111, semtran for SOL 112)
alter 'dpd ',''
DPD DYNAMICS,GPLS,SILS,USET,,/GPLD,SILD,USETD,
TFPOOL,,PSDL,,NLFT,TRL,EED,EQDYN/
-1/S,N,LUSETD/0/0/0//0/0/1/123/S,N,NOUE $
endalter
@
40280; ; Nonlinear Analysis - CBUSH, PBUSHT ,CBUSH1D, PBUSH1D - SFM 3003
In SOL 106, the use of CBUSH or CBUSH1D elements with a nonlinear
force-deflection with a lot of zero slope around the origin may lead to
the following unclear fatal message
SYSTEM FATAL MESSAGE 3003 (NLITER)
ATTEMPT TO READ PAST THE END OF A LOGICAL RECORD IN DATA BLOCK MESTNL
(FILE 208) IN SUBROUTINE NLINIT
This is due to the fact that all the nonlinear stiffness is zero for this
iteration. This should rarely occur-usually with single element decks. If
it does add a dummy nonlinear element that will never have a zero slope as
for example:

$ put in a mnl dummy element

$ 2 3 4 5 6 7 8 9 0
CBUSH 6666 333333 777777 888888
PBUSH 333333 K .1
PBUSHT 333333 KN 1212
TABLED1 1212
-2.3 -9.-9 0. 0. 2.3 9.-9 ENDT
GRID 777777 0. 0. 0. 123456
GRID 888888 1. 0. 0. 123456

@
40284; ; Postprocessing - Shock Response results not written to XDB File
Running shock response analysis using Sol 103 with PARAM,POST,0 will
generate an XDB file which contains only the model but no results.
Avoidance: Set PARAM,POST,-1 and the results will be written to the .op2
file.
@
40296; 2001; Optimization/Design Sensitivity - DESPCH - Incorrect Punch File
When parameter DESPCH is set to a non-zero value in an optimization task,
the updated property entries written in the punch file do not correspond
to the current design cycle. In fact, the updated property entries for the
previous design cycle will be written.
Avoidance: use the following dmap alter where the original inputs of ept,
geom2 and mpt are replaced by eptn, geom2n and mptn.
compile desopt
alter 'dom12 '(2),'dom12 '(2)
DOM12 XINIT,XO,CVALO,PROPIF,PROPOF,OPTPRMG,HIS,DESTAB,
geom1n,coordo,,,eptn,geom2n,mptn,epttabf,dvptabf,,,,,,,/
HISADD,NEWPRM,,NEWDES/
DESCYCLE/OBJIN/OBJOUT/S,N,CNVFLG/cvtyp/OPTEXIT//
MDTRKFLG/DESPCH/DESPCH1/MODETRAK/eignfreq//
 PROTYP///fsdcyc $
@
40307; 2001; DMP - UFM 4276 EC 923, Infinite loop on master+slave(s)
In a DMP (distributed memory parallel) analysis if a slave runs out of
disk space then the job may hang and become caught in an infinite cpu
loop.

@
40310; 2001; Optimization/Design Sensitivity - DSAPRT
If a design variable has more than 6 digits in its id, the formatted
sensitivity output will replace the design variable id with ******.
Avoidance: Do not use design variables with more than 6 digits in the id.
@
40315; ; Documentation - DMAP Module Dictionary - DIAGONAL
The DMAP Module Dictionary description of the DIAGONAL Module's WHOLE
option is incorrect for complex matrices.
In case of a complex matrix, the option WHOLE gives the following result:
POWER = 1.: Magnitude of the complex matrix elements
POWER != 1.: The real part of the complex matrix element is raised to the
power
@
40358; 2001; Acoustic Analysis - PACABS - Possible Unexpected Results
SYNTH=-YES requests the calculation of B, K, and M from the specified
tables. However if a value is given for B, K, or M on the PACABS entry,
then any table input is ignored and the data in the B, K, and M fields
is used. If SYNTH is specified as NO and no data is specified for B, K,
or M then TIDi is used.
@
40369; ; Aeroelastic Analysis - HESS, KE-Flutter - SFM, BIOMSG ERROR 3002
The Hessenberg eigenvalue extraction method used in the ke flutter
analysis may fail to converge. When this happens, a misleading message
such as
BIOMSG: ERROR 3002 HAS OCCURRED IN ROUTINE FREAD , FILE INDEX = 0
is produced.
Avoidance: This problem has only been observed on decks where the flutter
input data is questionable. If this message is encountered it is
suggested that the k-method of flutter analysis be chosen combined with
the complex lanczos method.
@
40394; 2001; GPFORCE - GPFDR, DEFORM, REPCASE, TEMP - Wrong Answers
Grid Point Force gives incorrect answers when thermal (TEMP) loads are
combined with DEFORM loads in the same subcase.
@
40395; ; GPFORCE - GPFDR, CQUAD8 - Wrong Answers
The applied load may be incorrect for Grid Point Force output when
thermal loads are applied to CQUAD8 elements.
Avoidance: There is currently no avoidance.
@
40399; 2001; Optimization/Design Sensitivity - DRESP1,RTYPE=ESE - UFM 1126 (G
NFIST)
If RTYPE=ESE is specified on the DRESP1 entry then the job may fail with
USER FATAL MESSAGE 1126.
Avoidance:
compile respsen
alter 'IF ( DOWTVOLD\<=0 ) ELTPRT', ''
@
40408; ; Damping - GE, MAT8 - Possible Wrong Answer
The value specified for the GE, material damping, field on the MAT8
entry is being ignored if it is referenced by a PCOMP entry.
Avoidance: If the GE field is important, replace the PCOMP/MAT8 with
equivalent PSHELL/MAT2/MAT1 entries.
@
40438; ; User Modifiable Nastran - Not Enough Space
The executable built by User Modifiable Nastran can not use memory larger
than 115mb else the following error may be encountered:
Error (errno=12) in opencore memory allocation of xxxx words
MSC.NASTRAN (ocalloc.c): Not enough space
This error may be caused by:
Insufficient swap space.
The maximum possible size was exceeded.
Avoidance: This problem is caused by a typo in the link script (umnld).
In this file, there is an enviroment variable MAXDATA. For non-dmp it
should be set as MAXDATA=0x80000000 and for dmp should be set as
MAXDATA=0x70000000 (it should have 7 zeros).
@
40444; 2001; Optimization/Design Sensitivity - RBEi, RBAR - Wrong Answers
When an optimization job includes both sizing variables, and shape
variables that affects the grid locations associated with the r-type
elements, the response sensitivities with resepcet to a sizing variable
may be null.
Avoidance: use the following dmap alter:
compile desopt
alter 'rgsens='
rgsens = false
alter 'dopr5 '
rgsens = false
compile respsen
alter 'rgsens=('
 rgsens = false
compile pslgdv
alter 'dsvgp4'
rgsens = false
@
40465; ; Complex Eigenvalue - EIGC, Massless DOF - SFM 3002
In MSC.Nastran Version 70.7, if the model contains massless degrees of
freedom and the old lanczos single vector method (system(108)=2) is used,
then the job may fail with the following error messages:
*** SYSTEM FATAL MESSAGE 3002 (CEAD)
EOF ENCOUNTERED WHILE READING DATA BLOCK SCRATCH (FILE 309) IN
SUBROUTINE CLASP
0FATAL ERROR
Avoidance: Use the new default block lanczos method.
@
40475; 2001; Nonlinear Elements - CBUSH1D, SDR2RD - SFM 6985
A FORCE=ALL Case Control request will cause an analysis with CBUSH1D
elements to fail with the message:
*** SYSTEM FATAL MESSAGE 6985 (SDR2RD)
PREVIOUS ELEMENT-ID = 9010
DID NOT REQUEST CORRECT NUMBER OF PHASE-1 OUTPUT WORDS.
Avoidance: Request STRESS=ALL instead or both FORCE and STRESS
@
40479; 2001; DMAP - CEND, ENDALTER - UFM 601 (IFP1D)
The automatic generation of a CEND entry introduced in MSC.Nastran V70.7
causes jobs with user DMAP that does not end with an ENDALTER to fail
with the following message:
*** USER FATAL MESSAGE 601 (IFP1D)
THE KEYWORD ON THE ABOVE CARD TYPE IS ILLEGAL OR MISSPELLED.

Avoidance: Input an ENDALTER statement

@
40497; ; Nonlinear Buckling Analysis - EIGRL - Traceback in lnnrigl, buck
sd
In a nonlinear buckling analysis, if the differential stiffnes matrix is
null then a traceback may occur in the READ module.
Traceback:
Offset 0x000003c0 in procedure chkjacd_
Offset 0x000005fc in procedure bucksd_
Offset 0x0000149c in procedure lnnrigl_
Offset 0x000028b8 in procedure reig_
Offset 0x000000f4 in procedure xsem06_
Offset 0x00000134 in procedure xsemdr_
Offset 0x0000005c in procedure nastmon_
Offset 0x00000984 in procedure mainal
--- End of call chain ---
@
40503; 2001; Transient Response - TRLG - Poor Performance
MPYAD picks inefficient methods. It tends to pick method 1 nt, when that
is almost always the slowest method, usually requiring multiple passes
when there are many time steps.
An alter that avoids this error is:
Compile semtran list
Alter 'trlg.*ppt'(,-1) $ before
type parm,,I,n,oldsys66 $
oldsys66=getsys(oldsys66,66) $
message //'oldsys66'/oldsys66 $
putsys (1,66) $ deselect 1 nt
Alter 'trlg.*ppt' $ after
putsys (oldsys66,66) $ restore original value
@
40514; 2001; Nonlinear Analysis - PARAM,BUCKLE,2 and scr=yes - Wrong Answers
Nonlinear buckling analysis in SOL 106 with PARAM,BUCKLE,2 may give poor
answers if a scratch data base is used (scr=yes) and there are linear
elements in the model.
Avoidance: Insert the following DMAP alter:
COMPILE NLSTATIC
ALTER 1
FILE KDJJ=SAVE,OVRWRT/KDGG=SAVE,OVRWRT/KDSF=SAVE,OVRWRT/
KDSS=SAVE,OVRWRT/KDAA=SAVE,OVRWRT/KDMG=SAVE,OVRWRT/
KDLL=SAVE,OVRWRT $
alter 'call sekdr '(,-1)
dbdelete datablk=(kdaa,kdmg,kdsf,kdjj,kdgg,kdss,kdelm,kddict) ,
where(wildcard) $
endalter
@
40545; 2001; Optimization/Design Sensitivity - DEQATN - Traceback in IFS6P
It has been observed on Cray computers that DEQATN input that exceeds
10000 characters can cause NASTRAN to fail during the processing of the
input bulk data deck. This is well below the 32000 character limit that
is documented. Symptons are erroneous echos of the deqatn input and a
traceback in the log file which includes subroutines PRERR2 and IFS6P.
Avoidance: Run on a computer other than the CRAY. Reduce the number of
characters in a DEQATN to less than 10000. In Version 70.7 a DRESP2 can
reference a DRESP2 so that a single DEQATN could be broken up into a
string of DEQATN's.
@
40547; 2001; P-Elements - Missing Output
If there are both h- and p-elements in the model, and the output requests
differ, for example stress=print for h-elements and stress=print,plot for
p-elements, the p-element output may be missing.
@
40565; 2001; Optimization/Sensitivity - PARAM,EIGD,YES - Wrong Sensitivity Co
efficients
If eigenvector sensitivity is requested (param,eigd,yes) and there is no
 constraint in the model (e.g., spc, rbe, etc.), then the eigenvector
sensitivitiy coefficients may be incorrect.
Avoidance: Add a dummy grid point and constrain the dummy grid point or
apply the following dmap alter:
compile eigvecds
alter 'ssg2'(,-1)
if (omit = -1 and nomset = -1 and single = -1 ) then
equivx fg/ff/always $
else $
alter 'ssg2'(,6)
endif $ no spc, mpc or omit
@
40579; ; Documentation/Optimization - DVPREL1 - PSHELL Property Name for
12I/T**3
The DVPREL1 entry has the ability to reference property field id names;
however, the PSHELL id field 12I/T**3 and other fields that begin with
integers can not be referenced in this manner.

Avoidance: Refer to the field by the field position.

@
40582; 2001; Design Sensitivity/Optimization - PARAM,SOFTEXIT,YES - UFM 2109
in DPD
The message:
USER FATAL MESSAGE 2109
will be issued from the DPD module with an accompanying "NO GRID, SCALAR,
OR EXTRA POINTS DEFINED." when PARAM SOFTEXIT set to yes.

Avoidance: Do not specify PARAM SOFTEXIT YES

@
40620; 2001; Optimization/Design Sensitivity - DSAPRT, PARAM,SOFTEXIT - UFM 4
276 EC 3001
When an optimization task includes the DSAPRT formatted request and
PARAM,SOFTEXIT,YES it may fail in subroutine dr2dx with the message

SYSTEM FATAL MESSAGE 4276 ERROR CODE 3001.

Avodiance:
1) Use either dsaprt(unformatted,...) option or use optexit=4.
2) Use the following dmap alter (for Version 70.7)
compile desopt
alter 'l exitopt'(8),'l exitopt'(8)
CALL EXITOPT XINIT,DESTAB,HIS,,,R1TABRG,RSP2RG,DRSTBLG,
FOL1,TOL2,FRQRPRG,DPLDXI,CONSBL,R1VALRG,R2VALRG,
DSCMG,DEQATN,PLIST2,DEQIND,DXDXIT,DCLDXT,DVPTAB,TABDEQ,
R1MAPR,R2MAPR,DPLDXT,EPTTAB,DBMLIB,,,PROPI,PROPO,,,,,
CASEXX,EDOM,MTRAK,HISADD,WMIDG,dnodel,rr2idrg,EPT,MPT,GEOM2/
DOM11SCR/
CNVFLG/DESCYCLE/DESMAX/OBJSID/
DESOBJ/OUTOPT/0.0/DSPRINT/DSUNFORM/DSEXPORT/TRUE/
 MODETRAK/MDTRKFLG/EIGNFREQ/DSAPRT/PROTYP $
$

@
40628; ; Nonlinear Analysis - CROD, TEMP(LOAD) - Wrong Stress Results for
Linear Elements
If a thermal load is applied in one subcase in a nonlinear analysis and
then is removed in a subsequent subcase, the total thermal load is
removed in one step from linear elements. This may lead to incorrect
stresses for intermediate output. The output at the end of each subcase
is correct.

Avoidance: Insert MATT1 for the linear element using a constant property,
then the stress will be correct for the linear elements at the immediate
output.
@
40644; 2001; SSSALTERs - alter1ha and alter2ha - Incorrect Results
It is possible that the results (printed output) from alter2ha may be
incorrect. This results from the internal sort of the exterior
(connecting) dof being different between the alter1ha run and the alter2ha
run.
Avoidance: Verify that the exterior dof are in the same order between the
runs or request an ujpdated version of the alters for V70.7 (also
available on the MSC web page).
@
40647; 2001; Loads - DMIG UACCEL, SPCFORCE - UWM 3059 SDR1
In Version 70.7, when using DMIG UACCEL to apply inertia loads in SOL 101
and SPCFORCE = ALL is specified, the following error may occur:
*** USER WARNING MESSAGE 3059 (SDR1)
SET IDENTIFIER DOES NOT EXIST. ERROR DETECTED IN SUBROUTINE GETBOL
0FATAL ERROR
Avoidance: Either use GRAV/RFORCE to apply the loads, or remove the
SPCFORCE=ALL case control.
@
40667; 2001.0.1; Random Analysis - SFM 3008/SWM 3002 (RANDOM) Insufficient Memo
ry for RAND5A
If SORT2 output is requested, and PSDF, ATOC, or RALL output are
requested, and the SET being used has a THRU range, the RANDOM module will
fail with SFM 3002, indicating insufficient memory.
Avoidance: This only occurs if all 3 of the above are done simultaneously.
Removing the SORT2 request or using a SET which does not contain a THRU
will allow the problem to run.
@
40670; 2001; P-elements - Missing Output
If the p-element output requests are varied by subcase, for example stress
only for the first subcase and stress and displacement for the second, the
output for the earlier subcases may be missing and/or mislabeled.
 Avoidance: An avoidance is to request the same output for all subcases.
@
40685; 2001; Plotting - IFP1 - Access Violation on NT Systems
When requesting XYPLOT output, several commands require that an equal (=)
sign be used. On NT systems, if any of these commands (XTGRID for example)
do not have the equal sign, the program will fail with an access violation
in IFP1. On other systems, it will terminate with "USER FATAL MESSAGE
(IFP1X) 676 IS NOT RECOGNIZED ON ABOVE CARD"
Avoidance: Be sure to include the equal sign on the appropriate commands.
@
40686; 2001; Acoustic Analysis - RLOADi, ACSRCE - Wrong Answers
If solution 111 is run with multiple subcases loading the fluid structure,
all loading on the fluid structure is lost after the first subcase.
Avoidance: Apply the following dmap alter:
$ begin lost loads fix
compile freqrs,list $
alter 'MATMOD FOL' $
if (ncol > nfreq) then $
$
$ ensure that omega contains nfreq for all subcases
$
file omegax=append $
type parm,,i,n,nsols $
nsols = ncol/nfreq $
trnsp omega/omegat $
do while (nsols > 0) $
append omegat,/omegax/2 $
nsols = nsols - 1 $
enddo $ (nsols > 0)
delete /omega,,,, $
trnsp omegax/omega $
endif $ (ncol > nfreq)
$ end lost loads fix

@
40700; ; Superelements - SWM 6065 (DBC), UWM 4698 (DCMPD)
When using PART superelements (BEGIN SUPER), error messages 6065 and 4698
may use internal numbers, rather than the original ids, to identify GRID
points.
*** SYSTEM WARNING MESSAGE 6065 (DBC)
*** DBCMSG ERROR MESSAGE 19 FROM SUBROUTINE DBCEQG
EXTERNAL GRID ID 2000000334 DOES NOT MEET UPDATE ASSUMPTIONS, CHECK
MODEL BEFORE PROCEEDING.
Avoidance: To guarnatee a specific order boundary grids are biased with
the value of system cell 233. The default value for this system cell is
2,000,000,000. Subtracting this value from the value identified in the
message will obtain the external grid id.

*** USER WARNING MESSAGE 4698 (DCMPD)

 STATISTICS FOR DECOMPOSITION OF MATRIX (NONE) .
THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER
THAN
1.00000E+15 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL.
GRID POINT ID DEGREE OF FREEDOM MATRIX/FACTOR DIAGONAL
RATIO MATRIX DIAGONAL
2000000892 6 1.00000E+15
1.00000E-10
@
40725; 2001; P-elements/Postprocessing - PARAM,POST - Strains overwrite Stres
s in XDB
When stress and strain output is selected in an analysis with P-Elements,
the P-Element strains will over-write the P-Element stresses on the xdb
file.
Avoidance: When saving P-element stresses on the xdb, make stress
requests only. Strain can be saved when no stresses are requested but
will be identified incorrectly on the xdb.
@
40732; 2001.0.1; Acoustic Analysis - CHACAB, CHACBR - Missing Fatal Message, Wr
ong Answers
The CHACAB and CHACBR elements are structural elements and all faces must
connect to structural points. The face on the fluid side should connect to
structural grids that have all degrees of freedom constrained (spc'd)
except the dof that is normal to the fluid. These grids should coincide
with the fluid grids but be separate from them. If the fluid side of the
element attaches directly to fluid points a bad solution will result with
no warning. This can happen when both the CHACAB or CHACBR element and the
fluid points are in the residual.
@
40733; ; Postprocessing/Buckling - PARAM,POST,0 - Missing ESE Results
In a solution 105 buckling analysis the Element Strain Energy results are
written to the xdb file for the static subcase but not the buckling
subcase.

@
40734; 2001; Superelements - SEBNDRY and SECONCT with THRU - UFM 315, Wrong a
nswers, UFM 6782
The THRU format on the SEBNDRY and SECONCT Bulk Data entries causes
various failures or wrong answers:
1. THRU ranges on SEBNDRY continuation entries cause UFM 315.
2. The second THRU range on the SEBNDRY parent entry is ignored
with no warning or fatal message issued and produces wrong answers.
3. Multiple SEBNDRY entries causes UFM 311.
4. THRU ranges with gaps in numbering on the SECONCT entry causes UFM
6782.
Avoidance: For SEBNDRY use only one THRU range. For SECONCT, specify
 separate THRU ranges without any gaps.
@
40773; 2001.0.1; Miscellaneous - ASSIGN, OUTPUT2, OUTPUT4 - SIGNAL 11 ERROR, ac
cess violation
If there are many output2/output4 assign statements, then depending on
the machine, the job may fail with any one of the following messages:
SIGNAL 11 ERROR
MAIN: "Access violation" (C0000005) exception encountered.
MAIN: For reference, A(MAINAL) = 428ce0, A(/SYSTEM/) = 1781f80
MAIN: Exception occurred at address 005A04DD.
MAIN: Program attempted to store data at location CB31AF28.
MAIN: Context Flags 0001003F
Avoidance: Removing one or two of output2/4 statements avoided the error
in this case.
@
40805; 2001; Superelements - PARAM,FIXEDB,-1 - Incorrect Eigenvectors for Com
ponent Modes
If output is requested for component modes using PARAM,FIXEDB,-1 and C-
or R-set degrees of freedom exist, the output eigenvectors and eigenvalues
printed will be incorrect. Note that all other output will be correct.
Avoidance: Insert the following alter into SOL 103 to correct the
problem.
compile semodes
alter 1
putsys(0,192)
$
compile super3
alter 'call.*sedisp.*cmpho',''
type db,cmpho0,cmlama0
CALL SEDISP,
GOT ,GOQ ,LOO ,LAO ,UOO ,
UOX ,USET ,PJ ,UL ,CRJ ,
CASESX ,CASEDR ,ULS ,EDTS ,YS ,
GM ,PSS ,KFS ,KSS ,QR ,
CMPHO0 ,OL1M ,CMLAMA0 ,CMPHA ,PP1 ,
PS1 ,CRX ,CASE0 ,BACK ,FB ,
QRG ,PG ,DYNAMICS,GPLS ,SILS ,
SLT ,DIT ,KSF ,KMG ,MMG ,
MSF ,BMG ,BSF ,K4MG ,K4SF ,
K2MG ,M2MG ,B2MG ,K2SF ,M2SF ,
B2SF ,UH1 ,OL1R ,BGPDTS ,CSTMS ,
QSGF ,QMGF ,FRLR ,CASES ,PVTBC ,
LCDVEC ,SPCPART ,QAG0 /
UG ,QG ,QMG ,OL1S ,PJ1 /
FIXEDB /APP /APP1 /SEID /DLOOP /
ALTRED /NOQG /NLHEAT /DROPT /NONLNR /
RSONLY /GW3 /RW4 /C1G /NOW3W4 /
NOQMG /NOPJ1 /NCUL /PEXIST /FREQDEP /
PARDCMP $
@
40821; 2001; Superelements - BEGIN SUPER, RESTART - Wrong answers
 When using part (BEGIN SUPER) superelements with any type of RESTART then
the data recovery may be incorrect.
Avoidance: Insert a dummy (unreferenced) FORCE Bulk Data entry in the
main Bulk Data section of the restart run.
@
40826; ; DMAP -- OUTPUT2 with MATRIX neutral format - Wrong answers
On short-word computers (all except Cray and Nec), if OUTPUT2 is used to
output a single precision matrix in neutral format then the INPUTT2 of the
neutrally formatted matrix will produce wrong answers.
Avoidance: Use the ADD or ADD5 module to convert the matrix to double
precision before passing it to OUTPUT2.
@
40831; ; Nonlinear Transient Heat Transfer - UWM 6137(UDSNRD), UFM 3007 (
NLTRD2)
If a nonlinear transient heat transfer analysis, solution 159, model has
extra grid points that are not attached to any elements and are not
constrained, then the job may fail with the message:

*** USER WARNING MESSAGE 6137 (UDSNRD),

INPUT MATRIX IS RANK DEFICIENT, RANK = 7816 IERROR
USER ACTION: CHECK MODEL
0*** UNSUCCESSFUL MATRIX DECOMPOSITION.
*** SYSTEM FATAL MESSAGE 3007 (NLTRD2)
ILLEGAL INPUT TO SUBROUTINE NLTRD2
Avoidance:
1) Rerun the model using SOL 101 with APP HEAT and insert PARAM,SPCGEN,1
in the bulk data section. The PARAM,SPCGEN,1 entry will cause constraint
entries for the unattached extra grid points to be written to the punch
file and these constraint entries can be subsequently used in solution
159.
2) Rerun the transient model by using the the older and significantly
slower unsymmetric solver by setting the NASTRAN SYSTEM(209)=0.
@
40839; ; Documentation/QRG - RESTART, DMIG - Inefficient Restart or SFM
4252
Restarts with DMIG entries may unneccessarily re-process the entire model
even though no changes to the input have occurred. The way to verify that
this is happening is to run with DIAG 56 and to look in the .f04 file for
"changed" and "DMIG" on the same line. If these words occur on a line
output by the RESTART module, the program will re-run the entire model.

Avoidance: This is currently a limitation of the sort routines used in

Nastran. Other potentially problem entries include BNDFL, DMIAX, ELIST,
DMIAX, ELIST, MFLUID, RADLIST, and RADMTX. If there are no changes to
any BNDFL, DMIAX, DMIG, ELIST, MFLUID, RADLIST, and RADMTX entries, then
insert the following alter to disable the checking for changes in these
entries.
compile ifpl
alter 'restart.*matpool',''
message //'restart checking on MATPOOL disabled' $
@
40843; ; Beam Library - PBARL, PBEAML - SFM 6498 on Large Models
A very large model with PBARL or PBEAML entries may fail with the
following error messages:
*** SYSTEM FATAL MESSAGE 6498 (Beam Lib, allocate_Master_Buffer)
API MESSAGE FOLLOWS.
USER INFORMATION: Memory Limit exceeded.
Could not allocate Memory for the Main Buffer Pool!
MSC.Software has not been able to duplicate this error and the model that
causes this error is confidential. The client model contains over 1.3
million entries, but only a few PBARLl entries.
Avoidance: There are three possible avoidances.
1. Replace the pbarl/pbeaml entries with equivalent pbar and pbeam
entries.
2. Allocate more memory.
3. Run the job on a non-Cray computer since Cray computers require more
memory.
@
40874; ; Heat Transfer Analysis - NLCOMB - Poor Message, UFM 4016 (GETEMP
)
In Version 70.7 Heat Transfer analysis the following message indicating
that there is missing TEMP entry data will be output for any point with an
SPC constraint that does not have a corresponding TEMP entry.
USER FATAL MESSAGE 4016 (GETEMP)
THERE IS NO TEMPERATURE DATA FOR ELEMENT 8127187 IN SET 1
FATAL ERROR
Avoidance: Check all the grids with SPC constraints and ensure that there
is a corresponding TEMP entry.
@
40892; 2001; Documentation/QRG - PBEAM
The following remark should be added to the Quick Reference Guide
description of the PBEAM entry:
Remark 13:
In nonlinear analysis the location of the 8 plastic rods is the same on
the PBEAM as on the PBCOMP SYMOPT=0.
@
40909; 2001; Documentation/QRG - SEUPPLOT, SEPLOT, PLOT - Missing Plot Output
Starting with Parts superelements in Version 69 each PLOT command requires
an SEUPPLOT or a SEPLOT command. This requirement and an example should be
noted under the Nastran Quick Reference Guide PLOT command description.
 An example follows:
some general plot commands such as VIEW etc.
SEPLOT 10
FIND SCALE ORIGIN 1 SET 1
PLOT SET 1 ORIGIN 1 $ SYMBOL 1 LABEL GRID
SEUPPLOT 0
FIND SCALE ORIGIN 1 SET 1
PLOT SET 1 ORIGIN 1 $ SYMBOL 1 LABEL ELEMENT
@
40910; ; Nonlinear Transient Analysis - RESTART/CGAP - Missing CGAP Outpu
t
If a nonlinear transient analysis restart run adds a request for CGAP
element force/stress output the output may be missing.
Avoidances: Request the output in a cold start run.
@
40934; 2001; Optimization/Design Sensitivity - PBARL, PBEAML - Illegal input
to DPNCH6
When both PBARL and PBEAML entries are included in an optimization task,
the job may fail in routine dpnch6 with a SYSTEM FATAL MESSAGE indicating
illegal input to subroutine dpnch6.
Avoidance: Set PARAM,DESPCH,-1
@
40936; 2001.0.1; Documentation - Missing Output, UFM 316
The Nastran Quick Reference Guide does not specify a limit for element
identification numbers. With the large field format 16 digits may be
specified but id's greater than 8 digits may result in missing output or
USER FATAL MESSAGE 316.
Avoidance: Do not specify element ids that are greater than 8 digits.

@
40937; 2001; Miscellaneous - SPOINTs, SEP1 - SWM 4203, UWM 4203
If a model has superelements and contains more than 32,000 SPOINTs the run
may fail after issuing SYSTEM WARNING MESSAGE 4203 and USER WARNING
MESSAGE 4203, indicating that the SEMAP datablock has been marked as
non-executable in SEP1.
Avoidance: If possible use GRID points in place of SPOINTs.

@
40953; 2001; Optimization/Design Sensitivtiy - DRESPi - No Message, Traceback
in dopr3m
If a DRESP2 entry references a DESVAR, DRESP1, DNODE, or DVPRELx with ID =
11000. The job will fail in the DOPR3 module. The traceback from the
log file follows:
 Signal received: SIGSEGV - Segmentation violation
Traceback:
Offset 0x000002e4 in procedure dopr3m_
Offset 0x000004e8 in procedure dopr3_
Offset 0x00000284 in procedure xsem03_
Offset 0x00000110 in procedure xsemdr_
Offset 0x0000005c in procedure nastmon_
Offset 0x000009ac in procedure mainal
--- End of call chain ---
Avoidance: Specify another DRESP id.
@
40974; 2001.0.1; Superelements - part superelements, SNORM - Wrong answers
If PARAM,SNORM or the SNORM Bulk Data entry is specified with part
superelements or BEGIN SUPER and main Bulk Data superelements, then wrong
answers will result.
Avoidance: Insert PARAM,SNORMPRT,2 or 3.

@
40990; 2001.0.1; Superelements - PARAM,SESDAMP - Incorrect Damping
If param,sesdamp,yes is used, then superelement component modes will have
modal damping applied based on the SDAMP command in each superelement's
SUBCASE. The residual structure will use the SDAMP command in the first
residual structure SUBCASE, unless there is none, in which case it will
use the last SDAMP called in a superelement subcase, which may result in
unexpected damping in the solution..
Avoidance: To be safe, if using PARAM,SESDAMP,YES, always explicitly
specify SDAMP in the residual subcase. If zero damping for the residual
is desired, SDAMP should refer to a TABDMP1 with 0.0 damping.
@
41009; 2001.0.1; P-Elements - PBEAM - Non-structural mass is ignored
For beam p-elements, the non-structural mass (NSM) on the PBEAM bulk data
entry is ignored.
Avoidance: An avoidance is to adjust the density to account for the
non-structural mass.

@
41020; ; Optimization/Design Sensitivity - CSUPER - SFM 3007 (MTRXIN)
If external superelements are used in optimization, and DESMAX has a value
greater than the number of cycles required to converge, and the run is not
saving the database (i.e. scr=yes), the run may fail with the following
message:.
SYSTEM FATAL MESSAGE 3007 (MTRXIN)
ILLEGAL INPUT TO SUBROUTINE MTRXIN
FATAL ERROR
Avoidance: Either set DESMAX to the number of cycles required to converge
or run with SCR=NO on the submittal.
@
41021; 2001.0.1; Aeroelastic Analysis - APD - SFM 4276
 If an AEFACT entry has a duplicate station defined, then the APD module
fails with a floating point exception.
Avoidance: The avoidance is to remove or correct the duplicate station
specification.
@
41034; 2001.0.1; P-Elements - OUTPUT - Wrong Answers
If the output requests specified by the DISPLACEMENT, VELOCITY, and
ACCELERATION case control commands are different from the corresponding
output requests on the OUTPUT bulk data entry, then the output may be
incorrect.
Avoidance: The only avoidance is to make sure the output requests are
consistent.
@
41036; 2001.0.1; Miscellaneous - CORDi - UFM 6915 (GP1C)
When defining a coordinate system in MSC.Nastran, three non-colinear
points are required. If the dimensions of the three points are such that
there are more than 15 orders of magnitude difference between them, the
program may incorrectly issue the message:
*** USER FATAL MESSAGE 6915 (GP1C)
indicating that the points are colinear.
Avoidance: Use 3 points which have locations within 10 orders of
magnitude of each other when defining coordinate systems.
@
41039; ; DMAP - DRMH1 - SFM 4276 EC 7777 on DEC ALPHA
Use of the DRMH1module to generate transfer matrices for certain element
types may cause an underflow to occur. Subsequent use of the matrices
will generate the following error on Alpha machines:
*** SYSTEM FATAL MESSAGE 4276 (tbk)
ERROR CODE 7777 PID= 0
Avoidance: This error appears to occur only on Compaq Alpha Unix systems.
If possible, try running on a different type of computer.
@
41050; 2001; Distributed Memory Parallel - RBEi, DMP, SEQP - Traceback in SEQ
P
If a Distributed Memory Parallel, DMP, run contains rigid elements or mpc
equations with many dependent degrees of freedom, then the job may fail in
the SEQP module with the following traceback:
Signal received: SIGSEGV - Segmentation violation
Traceback:
Offset 0x000006d4 in procedure seqcon_
Offset 0x000013b0 in procedure seqmpc_
Offset 0x00000f0c in procedure seqp_
Offset 0x000002c8 in procedure xsem02_
Offset 0x000001b8 in procedure xsemdr_
Offset 0x0000005c in procedure nastmon_
 Offset 0x00000c30 in procedure mainal
--- End of call chain ---
Avoidance: Increasing memory may allow the job to execute but there may
still be a performance penalty that makes Distributed Memory Parallel
undesirable for models with a large number of dependent degrees of
freedom. Version 2001 has been enhanced to provide an estimate of the
memory required.
@
41057; ; Nonlinear Analysis - PCOMP, TEMP(LOAD) - Wrong Answers
A nonlinear analysis with temperature loading applied to composite
elements where the PCOMP specifies an unsymmetric thermal coefficient
expansion in the 1 and 2 direction results in wrong answers.
@
41059; 2001.0.1; Aeroelastic Analysis - AMG - Floating Point Exceptions
The above error can result when using Strip Theory aerodynamics with a
Prandtl-Glauert correction to the lift curve slope (CLA < 0) and a
supersonic Mach number is chosen on a MKAEROi entry. This is because the
correction involves computing a term of the form sqrt(1.0 - M**2); i.e.,
taking the square root of a negative number
Avoidance: If strip theory is to be used with lift curve slope
corrections, use CLA > 0
@
41061; ; Nonlinear Analysis - PLOADX1, PARAM,FOLLOWK - UFM
The follower stiffness for PLOADX1 is incorrect. This could cause some
nonlinear static problems in SOL 106 to diverge.
Avoidance: Turn off the follower stiffness by specifying param, followk,
no.
@
41065; ; Heat Transfer Analysis - VIEW Factor using CQUAD8 yields poor an
swers
The calculation of the VIEW factors for CQUAD8 elements for curve surfaces
such as cylinder and sphere in heat transfer analysis yields poor results
when compared to the CQUAD4 element.
Avoidance: Use the CQUAD4 element.
@
41066; 2001; Documentation - PARAM FOLLOWK - Missing Solution Sequences
The Version 70.5 Quick Reference Guide description of PARAM,FOLLOWK does
not include all applicable solution sequences. The description should
read:
In Sols 103, 105, 106, 107, 108, 109, 110, 111, 112, 115, 116,
FOLLOWK=YES(default) requests......
and on the fourth line:
For FOLLOWK=YES in Sols 103, 105, 107, 108, 109, 110, 111, 112, 115, and
116 a separate static subcase is required .......
@
41069; 2001; Distributed Memory Parallel - DMP, PARAM,RESVINER - SFM 5423 (DM
PYAD)
MSC.Nastran Version 70.7 with DMP with PARAM,RESVINER,YES fails in SOL
111 with:
*** SYSTEM FATAL MESSAGE 5423 (DMPYAD)
ATTEMPT TO MULTIPLY INCOMPATIBLE MATRICES
Avoidance: Insert the following DMAP alter:
compile resvec
alter '\/paa',''
dbstatus vafs//s,n,novafs $
if ( novafs>0 ) then $
merge mxx,,,,vafs,/mxx1 $
else $
equivx mxx/mxx1/-1 $
endif $
mpyad mxx1,dga,/paa $
endalter
@
41071; 2001.0.1; Inertia Relief - Enhancement to PARAM,INREL,-2
An inertia relief option based on measuring deflections from a body-based
coordinate system was introduced in Version 70.5. It is requested with
param, inrel, -2. This method uses constraint equations that cause
deflections to be measured from a coordinate system tied to the model's
center of gravity. In Version 70.7 the rigid body modes used to generate
the constraint equations are based on shapes made by moving a point at the
basic origin through 6 unit motions. Revision 1 of the ir2a.v707 sssalter
package for improved efficiency (dated 2 Jan 2001) now has the modes
calculated around the grid point specified on the param, grdpnt entry.
The default action of this parameter is to use the basic origin, as
before. This change is in the revised sssalter package on the MSC web
site, and is being considered for Version 2001.
The alter to achieve this action is:
compile phase0 list $ use grdpnt as reference for rb modes
alter 'IF ( GRDPNT.*0 OR INREL=-2 ) REFPNT=0','' $ V70.7
IF ( GRDPNT<0 ) REFPNT=0 $

@
41109; ; 6 by 6 gets 7 roots with LAN
Some small eigensolutions compute extra roots by repeating one of the
roots and its eigenvector. This tends to cause fatal exits in the
dynamics solvers such as frrd1 and trd1. this error is caused by making
two or more rows of the modal matrices linearly dependent.
An avoidance is to use the AHOU metod instead.
@
41133; ; Nonlinear Analysis - PCOMP, PARAM.LGDISP - Missing Composite Str
ain Output
Homogenous composite strain output will be missing in nonlinear analysis
when param,lgdisp,1 is specified.
 Avoidance: Request STRAIN output in case control. Also, a request for
STRESS for at least one element and appy the following alter:
compile nlstatic
alter 'sdr2.*oesxc'
SDR2 CASESX2,CSTMS,MPTS,DIT,EQEXINS,,
ETT,EDTS,BGPDTH,,,UGNI,EST,,,,,GPSNTS,
DEQATN,DEQIND,/
,,,OESXd,,/'GNST'//3 $
ofp oesxd/ $

@
41137; ; Eigenvalue Analysis - EIGRL - Lanczos unreliable for diagonal in
put matrices.
The Lanczos method gets eigenvectors with poor orthogonality when there
are repeated roots, and both the mass and stiffness matrix are diagonal.
On some occasions these matrices can also cause overflow exits. Diagonal
matrices occur most commonly on research efforts when using dmap.
Avoidance: Use the AHOU method instead.
@
41142; ; Parts Superelements - SELOC, Exterior SPOINTs - UFM 4156 (SEMA)
If SELOC is used to locate a parts superelement that has exterior degrees
of freedom that include generalized coordinates that are defined by
spoints and qset1 entries, then the run may fail in the SEMA1 module with
the following message when the superelement is attempted to be attached to
the residual:
*** USER FATAL MESSAGE 4156 (SEMA)
DIMENSIONS OR TYPE OF DATA BLOCK SCRATCH ,
SUPERELEMENT 10
TRAILER=( 72 72 1 2 18 662) ARE
INCONSISTENT WITH SUPERELEMENT MAP.
FATAL ERROR

Avoidance: The error occurs because the program assumes every scalar point
has six degrees of freedom. The error may be avoided by using the
internally generated q-set DOFs with the SENQSET entry. Remove the
SECONCT bulk data entry.
@
41146; ; Eigenvalue Analysis - EIGRL - maxset=1 or =2 on eigrl causes fat
al exits
The number of vectors in a block for the LAN method is controlled by the
maxset input (field 7) on the eigrl entry. When set to any explicit
value, several classes of small problems have been observed to fail, with
diagnostics including floating point overflow, lnherr subroutine errors.,
and other obscure diagnostics. Examples include those with Lagrange
Multiplier variables and those with diagonal input matrices.
Avoidance: Use the default value for maxset (leave it blank), or use the
AHOU method.
@
41164; 2001.0.1; File Management Section - EXPAND, INIT - SFM 1012
Allocating too much space to scratch files may lead to poor messages such
 as;
"SPECIFIED RECORD IS BEYOND MAXIMUM FOR THIS FILE SYSTEM" or
"DBSET SCRATCH IS FULL AND NEEDS TO BE EXPANDED.
USER ACTION: USE FILE MANAGEMENT STATEMENT "EXPAND" FOR THIS DBSET."
Avoidance: The diagnostic stating to use EXPAND FMS statement is
applicable only to restarts. The user is referred to the Version 2001
Installation and Operations Guide or tan 4893 on the msc web site for
methods to make larger databases.
@
41165; ; Eigenvalue Analysis - EIGRL - Possible Wrong Answers
If f1 is specified as 0.0 on the EIGRL entry and the model contains a
"large mass" for enforced motion the run may produce poor answers.
Avoidance: Leave f1 blank.
@
41178; ; Parts Superelements - SEP1X - WM 6795, SFM 6790
If a PART superelement does not have any attachment points between it and
the rest of the model, the program will correctly issue SWM 6795, but may
incorrectly issue SYSTEM FATAL MESSAGE 6790 (SEP021) followed by an
incorrect list of superelements.
Avoidance: Be sure that PART superelements have points which connect to
the rest of the model.
@
41192; 2001.0.1; Nonlinear Analysis - PARAM,FOLLOWK, PLOADi - Poor Efficiency
In most cases the formulation used for follower stiffness results in an
unsymmetrical matrix. Since the default value for parameter FOLLOWK is
yes in SOL106 most problems with pressure loads will have an
unsymmetrical stiffness unless param,followk,no is specified. This may
lead to an inefficient solution.
Avoidance: If the follower force effects are not great, a more efficient
solution may be achieved by setting PARAM,FOLLOWK,NO.
@
41198; ; Modal Transient Response - SUPORT1, PARAM,RESVEC - SFM 6424 (DFN
SB)
If a modal transient response analysis has a static preload, a SUPORT1
entry, and PARAM,RESVEC,YES is specified the run may fail with :
SYSTEM FATAL MESSAGE 6424 (DFNSB)
THE NUMBER OF ROWS IN THE LOAD VECTOR AND THE FACTOR MATRIX ARE
INCOMPATIBLE
followed by:
*** USER FATAL MESSAGE 5423 (SMPYAD)
ATTEMPT TO MULTIPLY OR ADD INCOMPATIBLE MATRICES
Avoidance: Insert the following alter into your file:
compile resvec, list
alter 'SSG3.*LXX,,KXX,PX3,,,,,/UX1'(,1) $ 144
PARAML USET//'USET'//////'r'/S,N,NOxSET $
@
41215; 2001.0.1; DMAP - TRNSP, APPEND - Segmentation Violation in TRNSP
The APPEND module will append incompatible matrices and this may cause
problems in subsequent modules including a segmentation violation in TRNSP
module. Two steps have been taken to address this problem in
MSC.Nastran 2001. First, the following warning message is issued:
*** USER WARNING MESSAGE 5423 (APPEND)
ATTEMPT TO APPEND INCOMPATIBLE MATRICES
ID COLS ROWS FORM TYPE
NZWD DENS
TRAILER FOR X IS 101 20 30 2 2
60 10000
TRAILER FOR Y IS 102 40 10 2 2
20 10000
Second, the program will select the greater number of rows for the output
matrix. For example if APPEND is requested to append a 10X40 matrix with a
30X20 matrix instead of creating a 10X60 matrix a 30x60 matrix will be
output and the unspecified terms will be set to 0.
@
41221; ; Distributed Memory Parallel - DOMAIN, dmp - Job Hangs if Insuffi
cient Space
Distributed Memory Parallel jobs, dmp=n, executed on Sun systems may hang
if insufficient space exists. Currently, this is a limitation of the Sun
MPI implementation.
Avoidance: The only avoidance is to provide more space.
@
41237; 2001; Aeroelastic Analysis - GI, SPLINE1, CAERO1 - Floating Point Exce
ption
If a SPLINE1 with the FPS option is used to connect a CAERO1 that only has
one spanwise box, then a floating point exception may occur in the GI
module. This error will also occurs if the spline connects to a chordwise
strip of elements on a CAERO1 that is 1 spanwise box wide.
Avoidance: Split the CAERO1 entry so that it has two span wise boxes, or
use the IPS option on the SPLINE1.
@
41241; ; Miscellaneous - SMPYAD - Segmentation Fault, Signal 11, Alpha On
ly
On Alpha platforms, a run may fail during a SMPYAD operation with a
message in the log file:
Receiving signal number 11: Segmentation fault
Job aborting..
Avoidance: There are 2 possible avoidances either replace the SMPYAD in
the dmap with 2 equivalent MPYAD's, or set NASTRAN SYSTEM(129)=1.
@
41257; 2001.0.1; Complex Eigenvalue Analysis - OFREQ - Extra Output
OFREQ feature does not work for solution output requests in complex
 eigenvalue analysis

@
41260; 2001.0.1; Frequency Response - FRRD1 - SFM 3001
If the model in frequency response does not have any stiffness, the
program will fail in FRRD1 with SFM 3001 indicating that executive file 0
does not exist.
@
41269; 2001.0.1; Elements - SYSTEM(218) - SFM 4276 (tbk) , Arithmetic Trap in g
mtstd
Setting SYSTEM(218) to a real number may cause an arithmetic trap in
routine gmtstd on some machines.
Avoidance: Use the keyword "T3SKEW" instead.

@
41276; ; Lagrange Multiplier Technique - Wrong Answers
When using enforced velocity in transient analysis with the Lagrange
Multiplier approch, the LMT variables may show an instability. This
appears as a high frequency signal at the forcing variable only. If this
is encountered, contact MSC, and ask for details on residual mass.
Mention this Client Service Request Number (41276).
@
41280; ; DMAP - DIAGONAL - Module DIAGONAL may give a bad matrix trailer
If a rectangular matrix (form=2) is processed by the DIAGONAL module using
the 'SQUARE' option, the output matrix will have an incorrect trailer
(form=6) indicating that it is symmetric, when it should be
rectangular(form=2).
Avoidance: If the input matrix to DIAGONAL is rectangular and the
'SQUARE' option is specified, use MODTRL to update the trailer.

@
41301; 2001; Transient Analysis - GPSTRESS- Missing Output
If GPSTRESS output is requested in SOL 112 in Version 70.7, there will
be no corresponding output created.
Avoidance: Insert the following DMAP alter:
compile sedrcvr
alter 'call gpstress'(2),''
call gpstress postcdb,BGPDTN,ECTS,est,CSTMS,estnl,casedr,
eqexins,oes1x,oesnlx,ostr1/
egpsf,ogs1,egpstr,oeds1,ogds1,eldct,gpdct,
ogstr1,egpstra/
app/otape2/statics/app $
endalter
@
41308; 2001.0.1; P-Elements - DDRMM - Incorrect Principal Strains
 For solid p-element dynamics in SOL 112 (modal transient), the principal
strains calculated with DDRMM are incorrect.
Avoidance: A workaround is to turn DDRMM off by setting param,ddrmm,-1.
@
41335; 2001; Nonlinear Analysis - Complex Eigenvalue - Material Damping Ignor
ed
The calculation of complex eigenvalues in SOL 106 does not include the
effect of structural damping defined by Ge on the material entries.
Avoidance: Insert the alter:
compile nlstatic
alter 'call semr2'
equivx kllrh/kaa/-1 $
type parm,,i,n,nobggnl=-1,nok4ggnl=-1 $
EMG MESTNL,CSTMS,MPTS,DIT,,MUGNI,ett,,DEQATN,DEQIND,BGPDTS,
GPSNTS,,,,ditid/
x1,x2,,,belmnl,bdictnl/
-1/0/s,n,nobggnl/s,n,nok4ggnl/NSKIP/
/////////tabs/sigma/K6ROT/LANGLES/NOC//////unsymf $
if ( nobggnl>-1 ) EMA GPECT,bDICTNL,bELMNL,BGPDTS,SILS,CSTMS,,/
bggNL, $
call xadd5 bgg,bggnl,,,/bggt/cu/cu/cz/cz/cz $
dbstatus bggt//s,n,nobggt $
if ( nobggt>0
) CALL SEMRB USET,GM,GOA,BGGt,casesx2,cmlama,bk4zz,mphvz,goq,dit/
BAA,BMG,BSF/
NOA/NOMSET/norc/norset/nocset/nobset/noqset/
notset/false/0 $
if ( nok4ggnl>-1 ) EMA GPECT,KDICTNL,KELMNL,BGPDTS,SILS,CSTMS,,/
K4ggNL,/1 $
call xadd5 k4gg,k4ggnl,,,/k4ggt/cu/cu/cz/cz/cz $
dbstatus k4ggt//s,n,nok4ggt $
if ( nok4ggt>0 ) CALL SEMRB,
USET,GM,GOA,K4GGt,casesx2,cmlama,bk4zz,mphvz,goq,dit/
K4AA,K4MG,K4SF/
NOA/NOMSET/norc/norset/nocset/nobset/noqset/
notset/false/0 $

endalter
or, if possible, use PARAM,G to define the structural
dampin
@
41359; 2001; Documentation/Aeroelastic Analysis - SPLINEi - Unexepected Answe
rs
The USAGE field of the SPLINEi entries is not clearly documented that the
"BOTH" option is the ONLY one in which a single SPLINEi entry will
transform both loads and displacements (wherein the virtual work transpose
of forces to get displacements is used). Since the set of structural
degrees of freedom appropriate for load application may not be the same
set that is appropriate to represent the important deflections for the
aeroelastic correction, the FORCE and the DISP options allow the the
splines to be defined separately. However, FORCE or BOTH spline must
 be defined for each aerodynamic box to yield a "complete" mapping of
forces.
It is unusual (but not unheard of) to have aerodynamics for which forces
are not desired (e.g., to impose 2D-like boundary conditions or "infinite
aspect ratio" boundary conditions one typically uses extraneous
aerodynamics). Because there are these legitimate cases in which the
forces are not desired, no error messages or warnings are issued. Please
be careful that you have correctly and completely modelled the desired
usage of the splines. BOTH, the default USAGE, is a good default unless
you are particularly modeling around or for a particular purpose for which
a separation of force and displacement mapping is desired.

@
41365; GL; Acoustic Analysis - STATSUB, PARAM, RESVEC, RESVINER, INRLM - UF
M 9050
If the STATSUB Case Control or PARAMs, RESVEC, RESVINER, or INRLM is
requested in a fluid-structure model then the job may fail with UFM 9050.
Avoidance: Specify a SUPORTentry for any grid point in the fluid model.

@
41379; ; Nonlinear Transient Analysis - TSTEPNL - Poor Message, SFM 4276,
EC 7777
If NDT on the TSTEPNL entry is incorrectly set to a value less than 3, the
program will fail with a Floating Point Exception.
Avoidance: Specify a value greater than 3 for NDT.
@
41387; ; Parallel Processing - DMP, GPWG - Wrong Answers for GPWG
The GRID POINT WEIGHT GENERATOR output is incorrect in Distributed Memory
Parallel solutions.
Avoidance: The only avoidance is to execute a checkout run without dmp.

@
41389; 2001; Optimization/Design Sensitivity - DRESP2 - Incorrect DRESP2 Resp
onse Output
The print of the DRESP2 response values that are controlled by the P2
parameter on the DOPTPRM entry may be in error. The error occurs when
there are more DRESP2 values than
the length of the maximum record in the RESP12 datablock (this value is
the second word of the trailer of the RESP12 datablock). The incorrect
value that is printed is not used for subsequent
calculations with the exception that it may get used in calculating the
sensitivities of DRESP2's that reference other DRESP2's. Optimization is
not affected.
@
41408; ; DMAP - MATGPR - Illegal Input to MATGPR
The MATGPR module fails with "Illegal input" error messages when
attempting to print matrices that contain the k or ks sets in aeroelastic
analysis.