Giant Metrewave Radio Telescope

Tata Institute of Fundamental Research

GMRT Project, P. B. No. 6, Narayangoan,Tal - Junnar, Dist - Pune. State - Maharashtra, INDIA, PIN - 410 504. Phone : +91-2132-252112/3/6/7, Fax: +91-2132-252120, email: www@gmrt.ncra.tifr.res.in

Introducing GMRT

An illuminated GMRT antenna at twilight NCRA has set up a unique facility for radio astronomical research using the metrewavelengths range of the radio spectrum, known as the Giant Metrewave Radio Telescope GMRT!, it is located at a site a"out #$ km north of %une& GMRT consists of '$ fully steera"le gigantic para"olic dishes of ()m diameter each spread over distances of upto *) km& GMRT is one of the most challenging e+perimental programmes in "asic sciences undertaken "y ,ndian scientists and engineers& Why metre wavelengths : The metre wavelength part of the radio spectrum has "een particularly chosen for study with GMRT "ecause man-made radio interference is considera"ly lower in this part of the spectrum in ,ndia& Although there are many outstanding astrophysics pro"lems which are "est studied at metre wavelengths, there has, so far, "een no large facility anywhere in the world to e+ploit this part of the spectrum for astrophysical research& The site : The site for GMRT, a"out .$ km east of Narayangaon town on the %une-Nasik highway, was selected after an e+tensive search in many parts of ,ndia, considering several important criteria such as low man-made radio noise, availa"ility of good communication, vicinity of industrial, educational and other infrastructure and,a geographical latitude sufficiently north of the geomagnetic equator in order to have a reasona"ly quiet ionosphere and yet "e a"le to o"serve a good part of the southern sky as well&

Antenna configuration : The num"er and configuration of the dishes was optimi/ed to meet the principal astrophysical o"0ectives which require sensitivity at high angular resolution as well as a"ility to image radio emission from diffuse e+tended regions& 1ourteen of the thirty dishes are located more or less randomly in a compact central array in a region of a"out . sq km& The remaining si+teen dishes are spread out along the ' arms of an appro+imately 234-shaped configuration over a much larger region, with the longest interferometric "aseline of a"out *) km& The multiplication or correlation of radio signals from all the (') possi"le pairs of antennas or interferometers over several hours will thus ena"le radio images of celestial o"0ects to "e synthesi/ed with a resolution equivalent to that o"taina"le with a single gigantic dish *) kilometre in diameter5 The array will operate in si+ frequency "ands centred around )$, .)', *'', '*), 6.$ and .(*$ M7/& All these feeds provide dual polari/ation outputs& ,n some configurations, dualfrequency o"servations are also possi"le& The highest angular resolution achieva"le will range from a"out 6$ arcsec at the lowest frequencies to a"out * arcsec at .&( G7/& The design breakthrough : GMRT is an indigenous pro0ect& The construction of '$ large dishes at a relatively small cost has "een possi"le due to an important technological "reakthrough achieved "y ,ndian 8cientists and 9ngineers in the design of light-weight, low-cost dishes& The design is "ased on what is "eing called the `SMART' concept - for Stretch Mesh Attached to Rope Trusses& The dish has "een made light-weight and of low solidity "y replacing the conventional "ack-up structure "y a series of rope trusses made of thin stainless steel wire ropes! stretched "etween .6 para"olic frames made of tu"ular steel& The wire ropes are tensioned suita"ly to make a mosaic of plane facets appro+imating a para"olic surface& A light-weight thin wire mesh made of $&)) mm diameter stainless steel wire! with a grid si/e varying from .$ : .$ mm in the central part of the dish to *$ : *$ mm in the outer parts, stretched over the rope truss facets forms the reflecting surface of the dish& The low-solidity design cuts down the wind forces "y a large factor and is particularly suited to ,ndian conditions where there is no snowfall in the plains& The overall windforces and the resulting torques for a ()-m GMRT dish are similar to those for only a **-m dish of conventional design, thus resulting in su"stantial savings in cost& The dish is connected to a 2cradle4 which is supported "y two elevation "earings on a yoke placed on a '&6 m diameter slewing-ring "earing secured on the top of a .) metre high concrete tower& The weight of the disk is a"out #$ tonnes and the counter-weight is a"out ($ tonnes& The dishes have alt-a/imuth mount& The salient parameters and specifications of each dish are summari/ed in the Ta"le&

GMRT Antenna Specifications

The large si/e of the para"olic dishes implies that GMRT will have over three times the collecting area of the ;ery <arge Array ;<A! in New Me+ico, =8A which consists of *> antennas of *) m diameter and is presently the world4s largest aperture synthesis telescope operating at centimetre wavelengths& At '*> M7/, GMRT will "e a"out # times more sensitive than ;<A "ecause of the larger collecting area, higher efficiency of the antennas and a su"stantially wider usa"le "andwidth "ecause of the low level of man-made radio interference in ,ndia& lectronic !rontends and "ackends : Apart from the novel low-cost design of the para"olic dishes, the instrument has state-of-the-art electronics systems developed indigenously and consisting of the following main su" units&

Antenna feeds at si+ different frequency "ands "etween )$ M7/ and .)$$ M7/, having good polari/ation characteristics as well as simultaneous multi"and operation& <ow-noise amplifiers, local oscillator synthesi/ers, mi+ers, ,1 amplifiers& ?ptical fi"res linking the entire array with the C9@& These are used "oth for the telemetry signals and local oscillator phase reference communication "etween the C9@ and each antenna "ase& A digital *,'$,$$$-channel 1:-type correlator providing upto .*# spectral channels and covering a ma+imum "andwidth of '* M7/

The status report of GMRT as of Aune .BBB& The status report of GMRT as of Aanuary .BB6&

Iluminated GMRT antennas at twilight

Giant Metrewave Radio Telescope

Giant Metrewave Radio Telescope

Giant Metrewave Radio Telescope at sunset

Organization: Location: avelength: !uilt: Telescope st"le: #iameter: $ollecting area: Mounting: e%site:

NCRA 10 km east of Narayangaon, ndia radio !0 to 1!00 M"# $irst lig%t 1&&! array of '0 para(olic reflectors )!m *0,+!0m, alt-a#imut% fully steera(le primary %ttp.//www0gmrt0ncra0tifr0res0in

Giant Metrewave Radio Telescope &GMRT', located near 1une in ndia, is t%e world2s largest radio telescope at metre wavelengt%s0 t is operated (y t%e National Centre for Radio Astrop%ysics, a part of t%e Tata nstitute of $undamental Researc%, 3om(ay0

$ontents
4%ide5

1 6ocation , Tec%nical nformation ' Activities ) 78ternal links

(edit) Location
T%e GMRT is located around 90 km nort% of 1une at :%odad ;6ocated 1&< !2)+0)*=N +)< ,2!&00+=7>0 A near(y town is Narayangoan w%ic% is around 1! km from t%e telescope0 T%e office of NCRA is located in t%e 1une ?niversity campus rig%t ne8t to ?CAA0

(edit) Technical Information

T%ere are fourteen telescopes randomly arranged in t%e central s@uare, wit% a furt%er si8teen arranged in t%ree arms of a nearly =A=-s%aped array ;similar to t%e B6A> giving an interferometric (aseline of a(out ,! km0 T%e GMRT is an interferometer w%ic% uses a tec%ni@ue known as aperture synt%esis to make images of radio sources0 7ac% antenna is )! metres in diameter and, instead of a solid surface like many radio telescopes, t%e reflector is made of wire rope stretc%ed (etween metal struts in a para(olic configuration0 T%is works (ecause of t%e long wavelengt%s ;,1 cm and longer> at w%ic% t%e telescope operates0 7ac% antenna %as four different receivers mounted at t%e focus0 7ac% individual receiver assem(ly can rotate so t%at t%e user can select t%e fre@uency at w%ic% to o(serve0 T%e ma8imum (aseline in t%e array gives t%e telescope an angular resolution ;t%e smallest angular scale t%at can (e distinguis%ed> of a(out 1 arcsecond at t%e fre@uency of neutral %ydrogen ;1),0 M"#>0 Astronomers from all over t%e world regularly use t%is telescope to o(serve many different astronomical o(Cects suc% as gala8ies, pulsars and supernovae0

(edit) Activities
7ac% year on National Dcience Eay t%e o(servatory invites t%e pu(lic and pupils from sc%ools and colleges in t%e surrounding area to visit t%e site w%ere t%ey can listen to e8planations of radio astronomy, receiver tec%nology and astronomy from t%e engineers and astronomers w%o work t%ere0 Near(y sc%ools/colleges are also invited to put t%eir individual science e8periments in e8%i(ition and t%e (est one in eac% level ;primary, secondary sc%ool and Fr0 college> are awarded0

(edit) *+ternal lin,s

GMRT "omepage A-s%aped array Article on Dcience Eay

Coordinates. 1&<0!G)+0)*HN +)<0,G!&00+H7

Retrieved from =%ttp.//en0wikipedia0org/wiki/GiantIMetrewaveIRadioITelescope= Categories. Radio telescopes J nterferometric telescopes J Tata nstitute of $undamental Researc%
-iews

Article Eiscussion *dit this page

"istory

.ersonal tools

Log in / create account

0avigation

Main 1age Contents $eatured content Current events Random article

Interaction 1earch
Go 8earch

A(out Kikipedia Community portal Recent c%anges Contact Kikipedia Eonate to Kikipedia "elp

Tool%o+

K%at links %ere Related c%anges ?pload file Dpecial pages 1rinta(le version 1ermanent link Cite t%is page

Languages

Eeutsc% "rvatski

T%is page was last modified on ,' Fanuary ,009, at 1'.,+0 All te8t is availa(le under t%e terms of t%e GN? $ree Eocumentation 6icense0 ;Dee $op"rights for details0> KikipediaL is a registered trademark of t%e Kikimedia $oundation, nc0, a ?0D0 registered !01;c>;'> ta8-deducti(le nonprofit c%arity0 1rivacy policy A(out Kikipedia Eisclaimers

$ontents 1u%sections GMRT M(servations 6ow fre@uency specific pro(lems o Mapping wit% non co-planar arrays 'E imaging 1oly%edron imaging o 3andwidt% Dmearing

M(servational procedure o Mn-line monitoring Dystem monitoring Eata monitoring

Eata Analysis o Eata 7diting o Eata 7diting and Cali(ration in AIPS $lu8 density cali(ration 1%ase cali(ration 3andpass cali(ration

nversion and deconvolution of GMRT data T%e mages

The GMRT #ata

Recent Galactic plane o(servations at 9)' M"#, using t%e Molonglo Dynt%esis Telescope ;MMDT> ;Gray1&&)a> and at ,0) G"#, using t%e 1arkes *)-m single dis% ;Euncan et al01&&+(>, %ave revealed a num(er of new candidate DNRs0 "owever many of t%ese o(Cects were in complicated regions of t%e Gala8y and t%ese o(servations also suffered from contamination due to t%ermal emission at t%ese %ig% fre@uencies0 Alt%oug% t%e MMDT o(servations were at a relatively low fre@uency, t%e side-lo(es due to near(y strong sources severely limited t%e fidelity of images0 Kit% only single fre@uency data availa(le for most of t%ese fields ;t%ese o(servations did not %ave any overlapping fields>, t%e identification of t%e candidate DNRs in t%ese fields remained inconclusive0 A sample of candidate DNRs, accessi(le to t%e GMRT, were selected from t%ese o(servations0 Continuum o(servations of t%is sample were done wit% t%e GMRT at ',! M"# during t%e period of 1&&9 to ,0000 T%e telescope %ardware, on-line and off-line software were all in a state of (eing de(ugged during t%is

period and o(servations were invaria(ly intermingled wit% de(ugging activity and related o(servations0 Conse@uently, t%e data ac@uired was often unusa(le, sometimes re@uiring several o(serving sessions involving t%e same field to get good data0 T%is c%apter descri(es t%e procedure used for on-line data monitoring and identification of a potential source of data corruption, off-line data flagging and finally imaging using t%e AIPS package0 Dection )01 gives t%e general description of t%e o(servations done at ',! M"#0 maging at t%ese low fre@uencies re@uires t%e use of algorit%ms and tec%ni@ues not usually used for mapping at %ig%er fre@uencies0 T%ese low fre@uency specific pro(lems for inversion of visi(ilities and image deconvolution, and t%e tec%ni@ues used to overcome t%ese pro(lems are discussed in Dection )0,0 Dection )0,0, discusses t%e pro(lem of (andwidt% smearing of sources away from t%e p%ase centre, w%ic% necessitates t%e multi-c%annel mode of o(servations0 T%e procedures used for on-line system and data monitoring are descri(ed in Dection )0'0 T%e detailed procedure used for off-line data editing and cali(ration is descri(ed in Dections )0)0 Dection )0! descri(es t%e procedure used to finally map t%e data and w%ile Dection )0* presents t%e full primary (eam images after correcting for t%e primary (eam attenuation0

GMRT O%servations
',! M"# o(servations for continuum imaging of a sample of nine candidate DNRs were done during t%e period of 1&&9-,0000 T%e parameters of t%ese o(servations are summari#ed in Ta(le )010 Most of t%e o(Cects of interest in t%ese fields were e8tended, wit% emission at angular scales arcmin re@uiring relia(le o(servations at t%e smallest availa(le (aselines0 M(servations at ',! M"# are often affected (y intermittent radio fre@uency interference ;R$ > and, since R$ from near(y sources of emission remains partially correlated for t%e smallest (aselines, data from t%e smallest (aselines are also most severely affected (y R$ 0 $ortunately, R$ is often narrow (and and can (e identified if t%e o(serving (and is split into a num(er of narrow (and fre@uency c%annels0 T%e o(servations were t%erefore done wit% t%e full 1* M"# (and split into 1,9 fre@uency c%annels of eac% %aving a widt% of k"#0 T%e current single side-(and GMRT correlator measures only t%e co-polar visi(ilities, i0e0 only signals of t%e same polari#ation are correlated using t%e

Indian mode of t%e B63A Multiplier and Accumulator c%ips ;MACs>0 All t%e 1,9 fre@uency c%annels of t%e co-polar visi(ilities corresponding to t%e rig%t and left circular polari#ed signals were recorded for all availa(le (aselines0 T%e data was kept in t%e multi-c%annel format t%roug%out t%e imaging process to minimi#e t%e (and-widt% smearing of sources away from t%e p%ase centre ;see Dection )0,0,>0 After editing t%e R$ -affected or ot%erwise (ad data, a typical (andwidt% of a(out M"# was finally used for imaging giving a typical RMD noise of mFy/(eam0 T%e (ackground sky temperature can vary (y a factor of wit%in t%e Galactic plane, resulting in a c%ange in t%e total power output (y similar factors0 T%e correlator samplers are, %owever, optimi#ed to work wit% an input signal of 0 d3m0 To keep t%e sampler inputs at t%is level, automatic level controllers ;A6Cs> are introduced at t%e output of t%e (ase(and ;33> w%ic% effectively c%anges t%e gain to keep t%e output at 0 d3m0 To keep t%e A6C operating point wit%in its linear operating range, an attenuation of 1* d( was typically used for t%e $ and 33 signals0

Ta%le 234: 1arameters of o(servations wit% GMRT $re@uency of M(servations ;M"#> R$ 3andwidt% ;M"#> 3andwidt% used ;M"#> ntegration time ;sec> Average time spent on t%e source ;%r> No0 of Antennas used Ma80 (aseline ;k > Min0 (aseline ; > Ma80 spatial scale;arcmin> Average Antenna Densitivity ;:/Fy> 1rimary (eam ;degree> Dynt%esi#ed resolution ;arcsec> 00', 10) 1*0& ! ',! 1*

RMD noise ;mFy/(eam>

As mentioned earlier, t%e GMRT was in a state of (eing de(ugged during t%e period of t%e o(servations0 Conse@uently, t%e num(er of antennas and t%e longest (aseline availa(le c%anged from o(servation to o(servation0 As a result, t%e resolution c%anges from o(servation to o(servation0 "owever, most of t%e Central Square antennas, plus some of t%e arm antennas were availa(le for all o(servations giving angular resolution in t%e range arcsec0 Dome of t%e fields, w%ic% needed %ig%er resolution were o(served w%en t%e long (aselines antennas were availa(le0 Dince most of t%e target o(Cects were e8tended, it was essential t%at t%e s%ort antenna spacings were well sampled0 "ence t%e t%ree antennas, C0!, C0* and C0& w%ic% provide t%e s%ortest spacings, were used for all o(servations giving a s%ortest (aseline of ;after flagging (ad data> for w%ic% relia(le data was measured0 T%is corresponds to a largest angular scale of arcmin0 Most of t%e sources of interest were well wit%in t%is limit in angular si#e and we (elieve t%at most of t%e emission up to angular scales of arcmin is well represented in t%e GMRT ',!-M"# images0

Low fre5uenc" specific pro%lems

T%e visi(ility measured (y a properly cali(rated interferometer is given (y
;+01>

w%ere ; and

> are t%e antenna co-ordinates in t%e are t%e direction cosines in t%is system,

co-ordinate system, is t%e source

(rig%tness distri(ution ;t%e image> and is t%e far field antenna reception pattern ;t%e primary (eam>0 $or furt%er analysis we will assume , and drop it from all e@uations ;for typing convenienceN>, remem(ering all along t%at

t%e effect of receive radiation to

is to limit t%e part of sky from w%ic% t%e antenna can radians, w%ere is t%e diameter of t%e antennas0

$or a small field of view ; >, t%e a(ove e@uation can (e appro8imated well (y a ,E $ourier transform relation0 T%e ot%er case in w%ic% t%is is an e8act ,E relation is w%en t%e antennas are perfect aligned along t%e 7ast-Kest direction0 "ere, we discuss t%e pro(lem of mapping wit% non-7astKest arrays0 T%e derivation presented %ere of results for devising algorit%ms used for imaging large fields of view presented %ere follow t%e treatment of Cornwell O 1erley ;1&&,> and Cornwell O 1erley ;1&&&>0

Mapping with non co6planar arra"s

7@uation )01 also reduces to a ,E relation for a non-7ast-Kest array, if t%e integration time is sufficiently small ;snaps%ot o(servations>0 "owever modern arrays are designed to ma8imi#e t%e uv-coverage wit% t%e antennas arranged in a 2A2 s%aped configuration ;non 7ast-Kest arrays> ;Mat%ur1&*&>0 $ields, suc% as t%e ones o(served for t%is dissertation, wit% emission at all angular scales, re@uire ma8imal uv-coverage0 Telescopes suc% as t%e GMRT use t%e rotation of eart% to improve t%e uv-coverage and o(servations of comple8 fields typically last for several %ours0 "ence, 7@uation )01 needs to (e used to map t%e full primary (eam of t%e antennas, particularly at low fre@uencies0 6et (e treated as an independent varia(le0 A 'E $ourier > and ; > as a set of

transform of can (e written using ; conCugate varia(les as

;+0,>

Du(stituting for

from 7@uation )01 we get

;+0'>

T%e

in t%e coefficient of

in 7@uation )01 comes from fringe rotation0 n

t%e a(ove e@uation, it corresponds to a s%ift (y one unit along t%e a8is0 Dince t%is is only a c%ange of t%e origin of t%e co-ordinates system, it can (e a(sor(ed in t%e wit%out loss of generality0 ?sing t%e general result

;+0)>

we get
;+0!>

T%is e@uation t%en provides t%e connection (etween t%e ,E sky (rig%tness distri(ution given (y given (y use and t%e result of a 'E $ourier inversion of referred to as t%e Image volume0 "ere after,

to refer to t%is Image volume0 > would (e a s%ift of t%e

T%e effect of including t%e fringe rotation term ;

Image volume (y one unit in t%e conCugate a8is ; a8is> ;s%ift t%eorem of $ourier transformsP 3racewell1&9*and later eds>0 "ence, t%e effect of fringe

stopping is to make t%e plane coincide wit% t%e tangent plane at t%e p%ase center on t%e Celestial sp%ere ;t%e point w%ere t%e tangent plane touc%es t%e Celestial sp%ere> wit% t%e rest of t%e sp%ere completely contained inside t%e Image volume ;$ig0 )01>0

Figure 234: Grap%ical representation of t%e geometry of t%e Image volume and t%e Celestial sp%ere0 T%e point at w%ic% t%e Celestial sp%ere touc%es t%e first plane of t%e Image volume is t%e point around w%ic% t%e ,E image inversion appro8imation is valid0 $or wider fields, emission at points along t%e intersection of Celestial sp%ere and t%e various planes ;la(eled %ere as t%e Celestial Dp%ere> needs to (e proCected to t%e tangent plane to recover t%e undistorted ,E image0 T%is is s%own for ' points on t%e Celestial sp%ere, proCected on t%e tangent plane0

Figure 237: Grap%ical illustration to compute t%e distance (etween t%e tangent plane and a point in t%e sky at an angle of 0

Noting t%at t%e t%ird varia(le

of t%e Image volume is not an independent , 7@uation )0! gives t%e 0 magine t%e Celestial sp%ere defined (y

varia(le and is constrained to (e p%ysical interpretation of

enclosed (y t%e Image volume , wit% t%e top most plane (eing tangent to t%e Celestial sp%ere as s%own in $ig0 )010 7@uation )0! t%en tells t%at only t%ose parts of t%e Image volume correspond to t%e p%ysical emission w%ic% lie on t%e surface of t%e Celestial sp%ere0 "owever, t%e Image volume will (e convolved (y t%e telescope transfer function0 T%e telescope transfer function is t%e $ourier transform of t%e sampling function in t%e frame ;see C%apter ,, page >0 T%e telescope transfer function, referred to as t%e dirty beam and defined as , also defines a volume in t%e image domain0 T%e dirty image volume defined (y t%e relation is a convolution of wit%

0 Dince t%e dirty beam is not constrained to (e finite only on t%e Celestial sp%ere, will (e finite away from t%e surface of t%e Celestial sp%ere corresponding to non-p%ysical emission due to t%e side lo(es of 'E deconvolution using t%e dirty image and t%e dirty beam volumes will 0A

produce a Clean image volume0 An e8tra operation of proCecting all points in t%e CLEAN image-volume along t%e Celestial sp%ere onto t%e ,E tangent plane to recover t%e ,E sky (rig%tness distri(ution is t%erefore re@uired0 Grap%ical representation of t%e geometry for t%is is s%own in $ig0 )010 8# imaging T%e most straig%t forward met%od suggested (y 7@uation )0! for recovering t%e sky (rig%tness distri(ution, is to perform a 'E $ourier transform of 0 T%is re@uires t%at t%e a8is (e also sampled at t%e Ny@uist rate ;3racewell1&9*and later edsP 3rig%am1&99and later eds>>0 $or most o(servations, it turns out t%at t%is is rarely satisfied and doing a $$T along t%e t%ird a8is would result into severe aliasing0 T%erefore, in practice, t%e $ourier transform on t%e t%ird a8is is usually performed using t%e direct $ourier transform ;E$T> on t%e un-gridded data0 To perform t%e 'E $T ;$$T along t%e and a8is and E$T along t%e

a8is> one still needs to know t%e num(er of planes needed along t%e a8is0 T%is can (e found using t%e geometry s%own in $ig0 )0,0 T%e si#e of t%e synt%esi#ed (eam along t%e a8is is compara(le to t%at along t%e ot%er two

directions and is given (y w%ere, is t%e longest proCected (aseline lengt%0 T%e separation (etween t%e planes along s%ould (e 0 T%e distance (etween t%e tangent plane and a point separated (y from t%e p%ase center, for small values of , is given (y 0 $or a field of view of angular si#e , critical sampling would (e ensured if t%e num(er of planes along t%e a8is, , is

;+0*>

w%ere is t%e diameter of t%e antenna ; km for t%e GMRT at ',! M"#>0 T%erefore, for mapping a field of view wit%out distortions, one would re@uire 9 planes along t%e -a8is0 "owever, for mapping wit% Central

Square alone ; km>, one plane is sufficient0 At t%ese fre@uencies, it (ecomes important to map most of t%e primary (eam since t%e num(er and often t%e intensity of t%e sources in t%e field increase and t%e side-lo(es due to t%ese sources limit t%e dynamic range in t%e maps0 "ence, even if t%e source of interest is small, it re@uires a full 'E inversion ;and deconvolution>0 Anot%er reason w%y more t%an one plane would (e re@uired for very %ig% dynamic range imaging is as follows0 Dtrictly speaking, t%e only point w%ic% lies in t%e tangent plane is t%e point at w%ic% t%e tangent plane touc%es t%e Celestial sp%ere0 All ot%er points in t%e image, even close to t%e p%ase center, lie slig%tly (elow t%e tangent plane0 Eeconvolution of t%e tangent plane t%en results into distortions for t%e same reason as t%e distortions due to t%e deconvolution of a point source w%ic% lies (etween two pi8els in t%e ,E case ;3riggs1&&!>0 As in t%e ,E case, t%is pro(lem can (e minimi#ed (y over sampling t%e image w%ic%, in t%is case, implies %aving more t%an one plane along t%e -a8is, even if 7@uation )0* implies t%at one plane is sufficient0

Figure 238: Appro8imation of t%e Celestial sp%ere (y multiple tangent planes ;poly%edron imaging>0

.ol"hedron imaging As mentioned a(ove, emission from t%e p%ase center and from points close to it, lie appro8imately in t%e tangent plane0 1oly%edron imaging relies on e8ploiting t%is (y appro8imating t%e Celestial sp%ere (y a num(er of tangent

planes, referred to as facets, as s%own in $ig0 )0'0 T%e visi(ilities are recomputed to s%ift t%e p%ase center to t%e tangent points of eac% facet and a small region around eac% of t%e tangent points is t%en mapped using t%e ,E appro8imation0 T%e num(er of planes re@uired to map an o(Cect of si#e can (e found simply (y re@uiring t%at t%e ma8imum separation (etween t%e tangent plane and t%e Celestial sp%ere (e less t%an , t%e si#e of t%e synt%esi#ed (eam0 As s%own earlier, t%is separation for a point degrees away from t%e tangent point is 0 "ence, for critical sampling, t%e num(er of planes re@uired is e@ual > divided (y ;

to t%e solid angle su(tended (y t%e sky (eing mapped ; >

;+0+>

Notice t%at t%e num(er of planes re@uired is twice as many as re@uired for 'E inversion0 "owever, since a small portion around t%e tangent point of eac% plane is used, t%e si#e of eac% of t%ese planes can (e small, off-setting t%e increase in computations due to t%e increase in t%e num(er of planes re@uired0 Anot%er approac% w%ic% is often taken for very %ig% dynamic range imaging, is to do a full 'E imaging on eac% of t%e planes0 T%is would effectively increase t%e si#e of t%e field t%at can (e imaged on eac% tangent plane, t%ere(y reducing t%e num(er of planes re@uired0 T%e poly%edron imaging sc%eme is implemented in t%e current version of t%e AIPS data reduction package and t%e 'E inversion ;and deconvolution> is implemented in t%e ;no longer supported> SDE package developed (y T0F0 Cornwell et al0 3ot% t%ese sc%emes, in t%eir full glory, are availa(le in t%e ;recently released> AIPS++ package0 T%e GMRT ',!-M"# data was imaged using t%e IMAGR task in AIPS0 T%is program implements t%e poly%edron algorit%m and re@uires t%e user to supply t%e num(er of facets to (e used and a list of t%e locations of t%e centre of eac% facet wit% respect to t%e image centre and t%e si#e of eac% facet0 T%is list of

facets and t%eir parameters were computed using t%e task SETFC in AIPS, typically resulting in a grid of facets, eac% of si#e pi8els0

!andwidth 1mearing
T%e effect of a finite (andwidt% of o(servation as seen (y t%e multiplier in t%e correlator, is to reduce t%e amplitude of t%e visi(ility (y a factor given (y w%ere is t%e angular si#e of t%e synt%esi#ed (eam, is t%e center of t%e o(serving (and, is location of t%e point source relative to t%e field center and is t%e (andwidt% of t%e signal (eing correlated0 T%e distortion in t%e map due to t%e finite (andwidt% of o(servation can (e understood as follows0 $or continuum o(servations, t%e visi(ility data integrated over t%e (andwidt% is treated as if t%e t%e o(servations were made at a single fre@uency - t%e central fre@uency of t%e (and0 As a result

t%e and co-ordinates and t%e value of visi(ilities are correct only for 0 T%e true co-ordinates at ot%er fre@uencies in t%e (and are related to t%e recorded coordinates as
;+09>

Dince t%e total weig%ts , used w%ile mapping, does not depend on t%e fre@uency, t%e relation (etween t%e (rig%tness distri(ution and visi(ility at a fre@uency (ecomes
;+0&>

"ence t%e contri(ution of

to t%e (rig%tness distri(ution get scaled (y

and t%e co-ordinates get scaled (y 0 T%e effect of t%e scaling of t%e co-ordinates, assuming a delta function for t%e Dirty Beam, is to smear a point source at position into a line of lengt% in t%e radial direction0 T%is will get convolved wit% t%e Dirty Beam and t%e total effect can (e found (y integrating t%e (rig%tness distri(ution over t%e (andwidt% as given in 7@uation )0&
;+010>

w%ere

is t%e (and-s%ape function of t%e R$ (and and

is t%e Dirty

Beam at fre@uency

0 f t%e synt%esi#ed (eam is represented (y a Gaussian of

standard deviation and t%e (andpass (y a rectangular function of widt% , t%e fractional reduction in t%e strengt% of a source located at a radial distance of is given (y
;+011>

7@uation )010 is e@uivalent to averaging a large num(er of maps made from @uasi-monoc%romatic visi(ilities at 0 Dince eac% suc% map scales (y a different factor, a source away from t%e center would move along t%e radial line from one map to anot%er, producing t%e radial smearing mentioned a(ove, convolved wit% t%e Dirty Beam0 T%e effect of (andwidt% smearing can (e reduced if t%e (and is split into fre@uency c%annels wit% smaller c%annel widt%s0 T%is effectively reduces t%e (andwidt% as seen (y t%e mapping procedure and w%ile gridding t%e

visi(ilities, and can (e computed separately for eac% c%annel and assigned to t%e appropriate uv-cell0 T%e $Q correlator used in GMRT provides up to 1,9 fre@uency c%annels over t%e entire (andwidt% of o(servation and t%e visi(ilities can (e retained as multi-c%annel in t%e mapping process to reduce (andwidt% smearing0 Alt%oug% purely from t%e point of view of (andwidt% smearing, averaging c%annels at ',! M"# would (e accepta(le, keeping t%e visi(ility data(ase wit% all t%e 1,9 c%annels is usually recommended to allow identification and flagging of narrow (ands R$ 0

O%servational procedure
$or , t%e amplitude of t%e visi(ility function is proportional to t%e flu8 density of t%e source and t%e normali#ed visi(ility amplitude is converted to flu8 density units using t%e measurement of for a source of known (ut non varia(le flu8 density ;referred to as t%e flu8 density cali(rator>0 Also, t%e comple8 antenna (ased gain can potentially vary as a function of time0 T%ese slow time variations can (e corrected using periodic o(servation of a source of known structure, usually an unresolved source, referred to as t%e p%ase cali(rator0 T%e flu8 density of t%e p%ase cali(rators, t%e flu8 density of w%ic% is potentially varia(le over time scales of days, is also cali(rated using t%e flu8 density cali(rator0 7ac% o(servation t%erefore re@uires at least one o(servation of a flu8 density cali(rator and periodic o(servation of p%ase cali(rators to properly cali(rate t%e data0 T%e B6A flu8 density cali(rators, 3C48 and 3C286 was used for all o(servations0 Mne of t%ree cali(rators close to t%e Galactic plane namely, 1830-36, 1709-299 and 1822-096, were used as t%e p%ase cali(rator0 A flu8 cali(rator was o(served for min at t%e (eginning and at t%e end of eac% o(serving session and t%e p%ase cali(rator was o(served for min at an intervals of min0 T%e planned periodic inCection of noise from a cali(rated noise source to measure t%e system temperature %as not yet (een implemented for t%e GMRT0 As a result, t%e system temperatures for t%e flu8 density cali(rator fields ; > and t%e field (eing mapped ; to > must (e measured and a correction e@ual was

(e applied as part of t%e flu8 density cali(ration0

measured to (e : w%ile was estimated from t%e all-sky maps at )09 M"# ;"aslam et al01&9,P "aslam et al01&&!P "aslam et al01&91>0 $or a few fields, was also measured at few points around t%e source of interest and t%e measured system temperature was consistent wit% t%at estimated from t%e )09 M"# data to wit%in 0 None of t%e p%ase cali(rators used for t%ese o(servations are known to (e varia(le over few %ours0 T%ese cali(rators were t%erefore also used as secondary amplitude cali(rators to effectively correct for any slow variation in t%e receiver temperature0 T%e o(serving sc%edule if t%e GMRT on-line array control system can (e supplied via a computer reada(le file0 T%is file, apart from a few ot%er systemrelated commands, contain instructions a(out t%e source to (e tracked and t%e integration time on eac% source0 "owever since t%e feed(ack of t%e antenna pointing status is not used, t%is file needs to (e tweaked (y inserting delays (etween various commands0 $or e8ample, t%e amount of time taken for antennas to move from one tracking direction to anot%er varies from antenna to antenna, and appropriate time delays must (e inserted (etween t%e various commands in t%is file to make sure t%at t%e data recording (egins only after all antennas %ave reac%ed a given tracking position0 T%e commands for alternately tracking t%e p%ase cali(rator and t%e target source were put in an infinite loop, w%ic% allowed t%e a(ove mentioned periodic o(servations of t%e p%ase cali(rator0 "owever, since only an infinite loop is possi(le, t%e o(servations %ad to (e manually terminated at t%e end of o(serving session0 nformation a(out t%e status of t%e antennas as well as a mec%anism to derive time-of-t%eday information in t%e command synta8 used in t%is file is %ig%ly desira(le and will allow a (etter and more automated o(serving session0

On6line monitoring
T%e data could (e corrupted due to a num(er of reasons including ;1> R$ , ;,> catastrop%ic %ardware failures ;irrecovera(le in a s%ort time>, ;'> intermittent %ardware failures from w%ic% a @uick recovery is possi(le ;e0g0 (reakdown of communication (etween correlator control %ardware and software>, ;)> antenna-(ased (reakdown ;for e0g0, failure of t%e servo system resulting in stoppage of source tracking>, ;!> failure of t%e communication link (etween t%e antenna (ased computer ;A3C> and t%e on-line control computer at t%e C73, ;*> power supply failure to some of t%e antennas, ;+> loss of p%ase-lock for t%e

four local oscillators ;6Ms> or a p%ase Cump in t%e 6Ms, ;9> onset of ionosp%eric scintillations, ;&> pro(lems in t%e online array control, ;10> pro(lems in t%e correlator control software/%ardware, etc0 Eata affected (y any or many of t%ese sources manifests itself in various forms in t%e recorded data0 Duc% data needs to (e flagged from t%e data(ase (efore it is used for mapping0 T%e flagging information for t%e data(ase was generated (y on-line monitoring of t%e critical array control parameters as well (y off-line e8amination of t%e data itself0 T%e procedure used for t%is is descri(ed (elow0 1"stem monitoring t was noticed t%at pro(lems related to items ;'>, ;)>, ;!> and ;10> listed a(ove, occurred fre@uently enoug% to re@uire careful monitoring of various related telescope parameters as well as on-line monitoring of t%e visi(ility data0 T%e GMRT on-line array control system maintains a large amount of information a(out t%e status of t%e various su(-systems0 T%is information is updated once every few seconds and is availa(le in t%e s%ared memory resource of t%e control computer0 Any ar(itrary information from t%is resource can (e e8tracted using t%e table+01 program0 T%is program was used to e8tract t%e following information as a function of time.

t%e packet se@uence num(er t%e a#imut% and elevation angles of t%e antennas t%e error (etween t%e target and current antenna a#imut% and elevation angles t%e antenna servo system related flags

T%e table program produces t%e output in t%e form of a ta(le w%ic% was supplied to a s%ell script w%ic% generated an alarm if any of t%e following conditions occurred.

t%e error on a#imut% or elevation angles of any of t%e antennas e8ceeded a t%res%old value0 T%is usually was an indication of t%e antenna not receiving t%e t%e periodic tracking instructions from t%e GMRT on-line array control system computer t%e packet se@uence num(er did not c%ange for a t%res%old lengt% of time0 T%is usually indicated loss of communication (etween t%e antennas and t%e GMRT online array control system computer t%e flag for servo (rakes was raised for any of t%e antennas0 T%is usually indicated eit%er a malfunctioning antenna servo system and/or t%e detection of e8tra antenna load ;due to wind loading or due to some spurious signal indicating e8tra load>

n case of any of t%e a(ove pro(lems, manual intervention was re@uired0 "owever, t%is rat%er primitive RRautomation22 did %elp enormously in long o(serving sessions0 T%is procedure, w%ile already @uite useful, s%ould ultimately (e made part of t%e link (etween t%e GMRT on-line array control system and t%e correlator control software to ;1> record on-line flagging information, and ;,> control t%e recording of t%e data0

#ata monitoring Mn-line monitoring of t%e visi(ility data was done in two ways0 $irst, t%e +0, matmon program was used to monitor t%e normali#ed correlation coefficient for t%e p%ase cali(rators0 T%is program displays single integration cycle snaps%ots of t%e amplitude ;or t%e p%ase> of t%e visi(ility data for all (aselines in t%e form of a matri80 T%is display was used to determine t%e general %ealt% of t%e system (efore starting t%e o(servations and was useful in @uickly locating catastrop%ic pro(lems0 Mnce t%e o(servations were started, t%e data from t%e correlator was monitored using t%e programs t!a"t, !ant#ol and "lo#$!e ;see C%apter '>0 T%e amplitude and p%ase from all (aselines were continuously displayed as a set of stacked scrolling line plots using t%e program o%%& ;w%ic% uses t!a"t and t%e plotting package of t%e GMRT off-line software>0 T%is display provided a %our long snaps%ot of t%e data covering two or more o(servations of t%e p%ase cali(rator0 Dince most pro(lems in t%e system can (e detected using t%e data from t%e p%ase cali(rators, t%e onset of a pro(lem in t%e data (etween two cali(rator scans was easily detected using t%ese plots0 T%e output of !ant#ol ;namely, t%e antenna (ased comple8 gains> for t%e cali(rator scans was also similarly plotted0 T%ese plots provided information a(out t%e %ealt% of t%e system containing all t%e data on t%e cali(rator scans0 1ro(lems ranging from a significant loss in antenna sensitivity ;e0g0 due to c%ange in t%e antenna pointing offset across % seen for some of t%e antennas> to t%e onset of closure errors due to a malfunctioning correlator or t%e onset of ionosp%eric scintillations were @uickly detected from t%ese plots0 T%e output of "lo#$!e ;t%e closure p%ase for all possi(le triangles> was supplied to anot%er program w%ic% raised an alarm if t%e closure p%ases e8ceeded a t%res%old value for a t%res%old lengt% of time0 T%is procedure @uickly identified time varying correlator related pro(lems @uite effectively0

T%e output of t%e table and "lo#$!e programs as a function of time were also saved in a file0 T%is data was later e8amined and used to generate a flagging ta(le reada(le (y AIPS0 T%ese procedures, effectively generated first order online data flagging information, w%ic% is t%e crucial first step in improving t%e final data @uality0 6ack of time did not permit implementing t%ese procedures as part of t%e on-line array control and correlator control software, (ut must (e done in t%e near future to improve t%e relia(ility and @uality of t%e final data0

#ata Anal"sis
Two types of data analysis were re@uired for mapping. ;1> off-line data analysis ;(efore and after importing t%e data into AIPS> for furt%er, more careful identification of (ad data, and ;,> data analysis involving 'E inversion, deconvolution and p%ase cali(ration for t%e purpose of imaging0 T%is section descri(es t%e procedure used for t%e data analysis re@uired for data editing and cali(ration0 Dection )0! descri(es t%e procedure used for mapping0

#ata *diting
n addition to t%e data flagging information derived from on-line system and data monitoring, t%e data was also e8amined off-line using t%e programs t!a"t, !ant#ol and ba%ba#e ;see C%apter '>0 n long o(serving sessions, t%ere were invaria(ly antennas w%ic% did not produce usa(le signals for some fraction of time0 Mften, t%is was eit%er due to t% malfunctioning of a su(system or t%e antenna (eing taken for some kind of maintenance work over time scale muc% longer t%an t%e t%res%old time intervals set in t%e on-line monitoring programs ;see Dection )0'01>0 Duc% situations were easily identified from t%e plots of amplitude from all (aselines ;or for all (aselines wit% a single antenna> as a function of time for t%e cali(rator scans0 T%e "al&bb' program was also used for (andpass cali(ration of t%e data, using t%e p%ase cali(rator scans0 T%e antenna-(ased average (andpass solutions and t%e (aseline (ased (and-passes were e8amined to identify fre@uency c%annels affected (y intermittent narrow (and R$ or (y any ot%er source of data corruption0 T%e program ba%ba#e uses t%e output of !ant#ol to generate a summary of t%e antenna (ased solutions ;see section '0!0,, C%apter '>0 T%is was found to (e very useful in identifying intermittent (aseline (ased pro(lems in t%e data0 T%e

tasks used for cali(ration are rat%er sensitive to t%e presence of (ad (aselines in t%e data0 $lagging data corrupted (y (aseline (ased errors was t%erefore essential0 T%e algorit%m for t%e computation of antenna (ased comple8 gains implemented in !ant#ol is ro(ust in t%e presence of suc% data and was %ence crucial for t%e identification and flagging of suc% data0
AIPS

T%e a(ove mentioned procedure was followed for data from (ot% t%e polari#ation c%annels separately0 All t%e flagging information generated from on-line monitoring and off-line data analysis, was converted to flagging ta(les to (e used later for flagging data in AIPS0 Eata recorded in t%e native 6TA format was imported to AIPS via t%e $ TD format0 T%e program (l2)&t was used to convert t%e 6TA data(ase into $ TD format0 t was noticed t%at data for some of t%e (aselines was in t%e illegal num(er representation identified as *a* ;not-a-num(er> or In) ;infinite> num(ers0 Most of t%e data analysis programs ;including t%e ro(ust algorit%m of !ant#ol> (e%ave erratically in t%e presence of suc% num(ers0 "owever, computer representation of suc% num(ers is well documented in t%e 777 num(er representation format and can (e easily and relia(ly identified in t%e software0 T%is form of (ad data was t%erefore removed using an on-t%e-fly filter used in all off-line data analysis programs, including (l2)&t0 Anot%er form of (ad data manifests itself in t%e form of t%e normali#ed visi(ility amplitude (eing greater t%an unity0 T%is is again easily identified in t%e software, and t%e (l2)&t program flags suc% data w%ile converting it to t%e $ TD format0 Eue to sync%roni#ation pro(lems (etween t%e various network programs of t%e GMRT data ac@uisition software and t%e online array control software, t%e time stamp of t%e data records is sometimes corrupted and t%e value of time stamp of t%e successive data records does not increase monotonically0 Eata wit% suc% time stamp corruption is unusa(le in AIPS0 Also, a few data records at t%e (eginning of eac% scan were found to (e regularly (ad0 All suc% data records were filtered out using t%e program tma"0 tma" was used as a data filter in a data pipe-line, (efore t%e program (l2)&t0 f t%e o(servation was split into a num(er of 6TA files, t%e program lta"at was also used (efore tma" in t%e data pipeline to concatenate t%e various 6TA files into one file0 6TA files for some o(servations also re@uired c%anging t%e values of some of t%e keywords in t%e Glo(al an Dcan %eaders of t%e 6TA data (ase0 T%is was done using t%e program )& &t0

A typical data pipe-line set up to convert t%e and 6TA file to $ TD format was t%erefore, 4lta"at5 4)& &t5 tma" (l2)&t, w%ere 2 2 indicates t%e flow of data and t%e parent%esis are used to indicate t%at t%e program were used only if necessary0 ?se of ?N Q pipes eliminated t%e necessity of saving t%e intermediate 6TA files, w%ic% would ot%erwise %ave re@uired several giga (ytes of disk space0 $urt%er data editing was done after importing t%e data into AIPS0 T%e (ad data/ antennas/ (aselines identified earlier were translated to AIPS reada(le flagging files and applied inside AIPS0 Barious steps used for furt%er data editing and cali(ration inside AIPS are descri(ed (elow0

#ata *diting and $ali%ration in AIPS

T%e first step in t%e se@uence of data analysis is t%e amplitude and p%ase cali(ration of t%e visi(ilities0 T%is section descri(es, in a stepwise fas%ion, t%e se@uence of various AIPS tasks t%at were used for data cali(ration0 A typical AIPS task depends on a large num(er of parameters, w%ic% control t%e (e%avior of t%e tasks0 T%e settings of t%e relevant parameters of t%ese tasks are also discussed (elow0
10 T%e $ TD file generated using t%e procedure descri(ed a(ove, was imported into AIPS using t%e task FIT+D0 T%is results into a multi-source visi(ility data(ase written in AIPS, (y default in t%e compressed format0 ,0 Normally, multi-source file contains a num(er of ta(les, including t%e C+ ta(le, version 1 ;C+1> At present, t%e C+1 ta(le is not generated w%en t%e GMRT data is converted into $ TD format0 T%is ta(le is also used (y AIPS cali(ration programs as a template to determine t%e time resolution for t%e su(se@uent C+ and antenna gain ta(les ;t%e S* ta(les>0 "ence, t%e task I*D,R was always run to generate t%e ta(le C+10 T%is also generates t%e *, ta(le, w%ic% is used (y A 1D to navigate in t%e data(ase0 T%is, (y default, produces a C+ ta(le wit% a time resolution of ! min0 Dince t%e p%ase cali(rators used for all t%e o(servations were strong enoug% to provide sufficient signal to noise ratio, t%e antenna (ased gain solutions could (e computed for every integration cycle of sec0 T%e C+ and S* ta(les were later used to not only cali(rate t%e data (ut also identify (ad data0 T%e minimum time resolution for gain solutions was t%erefore set to sec (y setting APARM-0.33/0 (efore running I*D,R0 '0 T%e flu8 density for t%e flu8 cali(rators must (e accurately set (efore proceeding for cali(ration0 T%e flu8 density scale of t%e standard B6A flu8 cali(rators 3C48

and 3C286 is encoded in t%e task SET010 T%is task was run on t%e multi-source visi(ility data(ase to set t%e flu8 densities for t%ese flu8 density cali(rators using t%e following settings. 2PC2DE-3"al"345S26RCE-33C2863/33C483/330 )0 T%e task 67F+G reads t%e data flagging information from a disk file supplied via t%e I*FI+E keyword0 T%is task was used to apply all t%e flagging information generated from t%e on-line monitoring and off-line data analysis descri(ed earlier0 !0 T%e visi(ility amplitude of an unresolved source remains constant as a function of (aseline lengt%0 1lot of visi(ility amplitudes as a function of for t%e flu8 density cali(rator scans was e8amined using t%e task 67P+T0 T%e S26RCE keyword was appropriately set for t%e purpose0 M(viously discrepant data were identified, after allowing for some spread due to differences in antenna sensitivities0 Duc% data, w%ic% eit%er %ad very low or very %ig% amplitude can %ave severe repercussions for cali(ration0
67F+G was used to flag t%ese discrepant points wit% I*FI+E-330 T%e keywords A*TE**A/58ASE+I*E, and TIMERA*GE were used to specify t%e offending

antennas/(aselines and any time range for w%ic% t%e data was re@uired to (e flagged0 T%e primary task for cali(ration in AIPS is CA+I80 T%is task is, %owever, rat%er sensitive to t%e presence of (ad/dead antennas0 ;e0g0, in one case ;G'!*0'-10!> t%e presence of a(out 10 (ad (aselines out of a total of a(out '*0 good (aselines ;for antennas C0* and C0!> gave severely under estimated antenna gains>0 T%e initial effort outside AIPS to identify and flag (ad data/(aselines paid good dividends at t%is stage of processing0 *0 n some cases, t%e grap%ical flagging task T7F+G was also used to interactively flag (ad data0 +0 CA+I8 was t%en used to compute t%e antenna (ased comple8 gains for t%e flu8 density cali(rator scans wit% S*7ER-0/5GAI*6SE-0/5S2+I*T-0.330 T%is generates t%e first antenna gains ta(le ;S*1> containing comple8 gain solutions wit% a time resolution of sec0 90 To furt%er identify (ad data, t%e data was e8amined using 67P+T and 7P+2T wit% D2CA+I8-1 and S*7ER-10 Kit% t%ese settings, t%e S*1 ta(le is applied to t%e data on-t%e-fly0 T%is essentially takes care of any s%ort term variations in t%e antenna gains as well as variations in t%e sensitivity of t%e antennas0 T%is procedure %elps in identifying mildly discrepant data, w%ic% cannot (e corrected (y any antenna (ased correction0 As (efore, 67F+G was used to flag suc% data0 S*1 was t%en deleted using t%e task E,TDEST wit% I*E,T-3S*3 and I*7ERS-00 T%is deletes t%e latest S* ta(le generated in any previous run of CA+I80 Dteps ! to 9 were repeated till a satisfactory S*1 ta(le was generated for t%e flu8 density cali(rator0

&0 T%e task C+CA+ was t%en used wit% I*TERP2+5-53M9F345I*TPARM5-50.:/0.:45 CA+S26RCE5-533C2863/33C483/3345S26RCE5-533 to generate t%e second version of C+ ta(le ;C+2> containing t%e median window filtered solutions derived from S*10 T%e resulting ta(le C+2 contains flu8 cali(ration for all sources in t%e data(ase and constitutes t%e (asic flu8 cali(ration ta(le0 100 Dteps ! to & were repeated for eac% p%ase cali(rator to identify and flag (ad data0 T%is resulted into t%e t%ird C+ ta(le ;C+3> containing t%e median window filter solutions for t%e antenna gains, derived from t%e p%ase cali(rator scansP t%is was used later to correct for slow variations in t%e antenna (ased comple8 gains0 A fourt% C+ ta(le ;C+4> was also generated containing t%e t%e gain solutions for t%e p%ase cali(rator wit% a time resolution of sec ;I*ERP2+-32PT3>0 T%is was later used for (and pass cali(ration0 110 GET01 was t%en used wit% CA+S26RCE-33C2863/33C483 and S26RCE-33 to derive t%e flu8 density of t%e p%ase cali(rator0 f all was well in t%e a(ove procedure, t%e derived flu8 density s%ould (e close to t%e value listed in t%e B6A cali(rator manual and t%e error (ars of less t%an 10S ;after multiplication (y t%e ratio of t%e system temperatures for t%e flu8 and p%ase cali(rator fields, since t%e GMRT visi(ility data(ase as yet does not contain t%e ta(le>0 T%is corrects for t%e difference in t%e system temperatures (etween t%e flu8 and p%ase cali(rators0 1,0 n all t%e a(ove, a single, clean fre@uency c%annel, free from R$ , was used0 All t%e cali(rators used were strong enoug% to warrant t%e use of a single fre@uency c%annel0 "owever, to use ot%er c%annels for t%e purpose of imaging, a (andpass cali(ration also needs to (e applied to cali(rate any c%annel dependent gain variations0 n t%e cali(ration procedure adopted %ere, it is assumed t%at t%e time cali(ration ;determined in t%e a(ove procedure> and (andpass cali(ration can (e separated and determined independently0 T%e time cali(ration ta(le ;C+4> was t%erefore applied to all t%e c%annels (efore deriving t%e (and pass cali(ration ta(le, t%e 8P ta(le0 AIPS offers interpolation of t%e 8P ta(le in time to apply (and pass cali(ration to data of t%e target source0 "ence, t%e time cali(rated data wit%in t%e cali(rator scans needs to (e averaged, to generate a scan averaged (and pass solution0 T%e resulting (and pass solutions ;one per cali(rator scan> can t%en (e interpolated in time to take care of any slow variations in t%e (and s%ape0 "owever, data affected (y intermittent R$ needs to (e flagged (efore t%e data is averaged in time0 R$ on cali(rator scans was identified using t%e task F+GIT on t%ese cali(rator scans0 T%is task e8amines t%e data after su(tracting a linear fit to t%e (and s%apes from individual (aselines0 A user specified set of c%annels is used to determine t%e linear fit0 All data wit% residuals outside t%e user specified limits are t%en flagged0 All c%annels were flagged for a given integration time containing (ad data0 T%is was ac%ieved wit% t%e following settings for F+GIT. 8C;A*-C045EC;A*-C145D2CA+I8-145GAI*6SE-4 w%ere, C0 and C1 are t%e first and t%e last fre@uency c%annels to (e used0 Deveral sets of ;C0/5C1> for range of

clean fre@uency c%annels can (e specified, w%ic% alone will (e used for t%e linear fits, via t%e *82,ES and 82, keywords0 T%e flagging criterion can (e specified via t%e APARM keyword0 was also used later on t%e cali(rated data on t%e target sources0 "owever, since t%e signal to noise ratio on individual (aselines for e8tended sources can vary a lot, suc% automated procedures are of limited use0 dentification of (ad data on t%e target source was t%erefore usually done manually using tasks like 67P+T, 67F+G, T7F+G and SPF+G0
F+GIT

1'0 T%e grap%ical data editing task SPF+G was sometimes used at t%is stage to identify and flag (ad data0 T%is task displays t%e data in t%e time-fre@uency plane from one (aseline at a time and provides t%e same interface as t%at of T7F+G to grap%ically flag data0 1)0 Ne8t, t%e 8P ta(le was derived (y time averaging t%e data wit%in t%e p%ase cali(rator scans ;after application of t%e time cali(ration>0 T%is was done using t%e task 8PASS wit% D2CA+I8-145GAI*6SE-4458PASSPRM<:=-1458C;A*-C045 EC;A*-C1 w%ere C0 and C1 refer to t%e first and t%e last fre@uency c%annel to (e used0 1!0 T%e 8P ta(le and (and pass corrected (and s%apes were e8amined using t%e task P2SSM0 6arge oscillations across t%e (and were sometimes found in a few antennas0 T%ese antennas were usually flagged from t%e entire data (ase0 1*0 $inally, t%e task SP+IT was used to apply t%e time and (and pass cali(ration ;t%e C+ and 8P ta(les respectively> to t%e data on t%e p%ase cali(rators as well as t%e data on t%e target source and single source multi c%annel cali(rated data(ases generated0 T%is was done using t%e following settings. D2CA+I8-145GAI*6SE-345 D28A*D-3458P7ER-10

Flu+ densit" cali%ration $lu8 density cali(ration was done using o(servations of one of t%e two B6A flu8 density cali(rators, 3C286 or 3C480 Time varia(ility of t%ese sources %as (een found to (e small from t%e B6A monitoring of t%e flu8 densities of t%ese sources0 T%e a(solute flu8 densities of t%ese sources was derived (y careful o(servations (y 1erley O Crane ;1&9*> using t%e B6A in E-array configuration and t%ey found t%at t%e 3aars scale ;3aars et al01&++> was slig%tly in error0 T%ey adCusted t%e flu8 density of 3C29: to t%at of 3aars value and derived corrections for t%e flu8 densities of 3C286 and 3C480 T%ese corrected flu8 densities are encoded in t%e AIPS task SET01 w%ic% was used to set t%e flu8 densities for t%ese sources used for flu8 density cali(ration0 T%e adopted flu8 densities of 3C286 and 3C48 were and Fy respectively ;at ',! M"#>0 M(servations of 3C48 wit% t%e B6A %as s%own t%at t%e flu8 density derived using SET01 gives t%e ',!-M"# flu8 density wit% an accuracy of S0

T%e flu8 density cali(rators were typically o(served at t%e (eginning and at t%e end of eac% o(servation0 T%e p%ase cali(rators used for t%ese o(servations are also listed as good secondary B6A cali(rators0 T%e flu8 density cali(rator scans were used to derive t%e flu8 densities of t%ese secondary cali(rators as a consistency c%eck0 T%e p%ase cali(rators were also used as secondary cali(rators to correct for any slow variations in t%e antenna gains0 All fields o(served for t%is dissertation also %ad many ot%er sources in t%e field0 $or some of t%ese sources, t%e ',!-M"# flu8 densities were availa(le from ot%er independent o(servations as well ;B6A cali(rators, targeted B6A o(servations or t%e Te8as survey ;Eouglas et al01&&*> w%ic% gives t%e spectral inde8 and point source flu8 densities at '*! M"#>0 T%ese flu8 densities were also used for a consistency c%eck on t%e flu8 cali(ration and to eliminate t%e possi(ility of any systematic flu8 cali(ration error0 T%e (ackground temperature in t%e Galactic plane can c%ange @uite su(stantially for separate pointings0 $or accurate flu8 density cali(ration, one must measure t%e system temperature for t%e flu8 density cali(rator field as well as for t%e target field0 To also account for small time dependent variations in t%e system temperature, it s%ould (e monitored regularly during t%e lengt% of t%e o(servations0 T%e planned periodic inCection of cali(rated noise at t%e frontend of eac% antenna to measure t%e system temperature %as not yet (een implemented at t%e GMRT0 n its a(sence, t%e system temperature was measured at a few positions around t%e target source in t%e Galactic plane and t%e measured system temperature used to correct for t%e differences in t%e (ackground temperature (etween t%e field of interest and t%e flu8 density cali(rator0 T%e (ackground temperature from t%e )09-M"# all sky survey ;"aslam et al01&&!> was also estimated as a consistency c%eck0 Kit% t%is sc%eme, we estimate t%at t%e ',!-M"# flu8 densities from GMRT are accurate to 0

.hase cali%ration Dlow variations of t%e antenna (ased comple8 gains occur on time scales of a few tens of minutes0 T%e relative p%ase variations (etween t%e antennas due to t%is needs to (e corrected so as to p%ase t%e array over several %ours0 T%ese slow variations are measured using periodic o(servations of a p%ase cali(rator0 Dince t%e system temperature at ',! M"# in t%e Galactic plane is a factor of '! %ig%er t%an away from t%e plane, t%e p%ase cali(rators must also (e strong ;typically Fy> to provide enoug% signal to noise ratio for t%e computation of antenna (ased comple8 gains0 Temporal as well spatial variations in t%e

ionosp%eric total electron content at ',! M"# is e8pected to (e t%e maCor source of p%ase corruption0 T%is can produce p%ase variations over t%e scale of t%e array ;and sometimes even across t%e primary (eam of eac% antenna>0 t is t%erefore not advisa(le to use a p%ase cali(rator too far from t%e target field since t%e antenna (ased comple8 gains o(tained from t%e p%ase cali(rator may not reflect t%e p%ase variations in t%e direction of t%e target field0 ?sing t%e antenna (ased p%ase variation derived from continuous o(servations of t%e p%ase cali(rators for several %ours ;$igs0 ,011 and ,01,>, it was estimated t%at p%ase variations over a time scale of a(out %alf an %our could (e appro8imated well (y linear interpolation0 T%is is t%us t%e time scale at w%ic% one needs to o(serve t%e p%ase cali(rator ; minutes>0 T%e t%ree B6A ',+ M"# cali(rators 1709-299/51830-36, and 1822-096 wit% ',+-M"# flu8 densities of *, ,9, and 1' Fy respectively, were used as p%ase cali(rators0 T%e angular separation in t%e sky (etween t%e p%ase cali(rators and t%e target field was typically 0 T%e ma8imum error in p%ase due to errors in t%e antenna co-ordinates, w%en t%e p%ases from t%e p%ase cali(rator are transferred to t%e target source was estimated to (e a few degrees ;see Dection ,0*01>0 Tests done (y p%asing t%e data from one cali(rator using periodic o(servations of anot%er cali(rator s%ow t%at t%e array is p%ased over time scales of %alf an %our using t%is procedure0 !andpass cali%ration T%e antenna (ased comple8 gains vary across t%e pass(and, primarily due to t%e antenna (ased (and s%ape and residual fi8ed delay errors0 T%ese variations in t%e comple8 gains must (e corrected (efore t%e visi(ilities from individual fre@uency c%annels are averaged0 As mentioned a(ove, t%e p%ase cali(rators were strong enoug% to provide enoug% signal to noise ratio for t%e computation of c%annel dependent antenna (ased comple8 gains0 T%e antenna (and s%ape corrections were t%erefore derived using t%e p%ase cali(rators0 An average gain was computed for eac% of t%e p%ase cali(rator scans per c%annel and t%e linearly interpolated values applied to t%e target source data to correct for t%e c%annel dependent comple8 gains0

Inversion and deconvolution of GMRT data

T%e ',!-M"# primary (eam of GMRT antennas %as a full widt% at %alf ma8imum ;$K"M> of 0 T%e central s@uare provides a ma8imum (aseline of km e@uivalent to an angular resolution of arcmin0 At t%is resolution, t%e full GMRT primary (eam can (e mapped wit%out severe degradation due to non-coplanar effects0 As a first step, wit% t%e dual purpose of gauging t%e data @uality and locating strong confusing sources, single facet images were made at a resolution of arcmin using t%e task IMAGR0 "aving identified t%e sources in t%e field of view from t%is lower resolution image, %ig%er resolution imaging was attempted0 Typically, a ma8imum (aseline of k was used corresponding to an angular resolution of arcsec0 Most of t%e fields %ad strong e8tended sources all over t%e field of view forcing t%e mapping of t%e full primary (eam0 At t%ese resolutions, t%e num(er of planes re@uired along t%e n-a8is is 90 "ence, a 'E inversion was re@uired0 T%e IMAGR task of AIPS performs a 'E inversion using t%e poly%edron imaging algorit%m0 n t%is, t%e entire field of view is divided into a two dimensional grid of facets0 A small part of t%e sky ;corresponding to t%e si#e of t%e facet> centred around eac% facet is t%en imaged (y first s%ifting t%e p%ase centre of t%e visi(ility to t%e centre of t%e facet and t%en performing t%e normal ,E inversion and C67ANing0 Dince, t%e ,E appro8imation is assumed to (e valid wit%in t%e facet, it is important to make sure t%at t%e facet is not so (ig as to re-introduce distortions at t%e edges of t%e facets0 T%e num(er of facets re@uired for 'E inversion using IMAGR and t%e appropriate RA and E7C s%ifts for t%e centre of eac% facet, were computed using t%e relatively new task in AIPS called FCSET0 7ssentially, given t%e si#e of eac% facet, t%e si#e of t%e field of view and t%e RA and E7C of t%e p%ase center, t%is task writes out a IMAGR reada(le list of field specifications ;t%e field num(er, its RA and E7C s%ifts and its si#e in num(er of pi8els along t%e RA and E7C a8is>0 Typically, t%is resulted in a grid of facets, eac% of si#e 0

After t%e inversion of eac% of t%e facets, t%e IMAGR task uses t%e usual ,E Clark C67AN ;Clark1&90> on eac% facet0 T%e class of C67AN and M7M (ased deconvolution algorit%m treats eac% pi8el in t%e image as a degree of freedom0 7ven w%en mapping in t%e Galactic plane ;or close to it>, it is clear from t%e

images t%at most of t%e pi8els do not %ave any p%ysical emission associated wit% t%em0 Reconstruction of t%e p%ysical emission in t%e field of view s%ould t%erefore (e some%ow constrained to use only t%ose pi8els w%ere t%ere is significant p%ysical emission from t%e sky0 Not doing so is e@uivalent to giving more freedom to a non-linear fitting process ;t%e deconvolution process>, t%an is Custified (y t%e data0 C67AN (ased algorit%m ;and its variants> are t%emselves unconstrained0 T%is constraint must t%erefore (e provided e8ternally (y setting (o8es around t%e dominant sources in t%e field of view at eac% cycle of C67AN0 t %as (een s%own (y 3riggs ;1&&!> t%at t%e (est results are o(tained (y putting as tig%t a (o8 as Custified (y t%e data ;essentially (y inspection>0 Euring t%e deconvolution process, t%e IMAGR task provides a facility to dynamically define (o8es for eac% field for every cycle of C67AN0 "owever, since t%e emission was usually of comple8 morp%ology making it difficult to define tig%t (o8es, simple s@uare (o8es enclosing t%e source of emission were used0 T%is was manually done for eac% facet0 T%e resulting set of facet images were put toget%er to reconstruct t%e sky using t%e task F+AT*0 T%e F+AT*ed image was t%en primary (eam corrected using t%e task P8C2R0 T%e GMRT visi(ilities correspond to t%e date-epoc% at t%e time of o(servations0 T%e final image was t%erefore rotated to t%e F,000 epoc% using t%e task REGRD0

The Images
T%is section presents t%e final images produced via t%e procedure descri(ed a(ove0 All images were corrected for primary (eam attenuation using a polynomial appro8imation of t%e GMRT primary (eam0 As mentioned earlier, t%e resolution in t%ese images c%anges from image to image due to a com(ination of declination dependent uv-coverage, c%anges in t%e num(er of availa(le antennas and t%e flagging of (ad data0 mages of fields wit% large angular si#e sources are presented at a few arcmin resolution0 "ig%er resolution images of some of t%e fields were also made w%ere necessary ;due to t%e presence of small angular si#e sources of interest in t%e field, e0g0 t%e field containing G00'0*-001>0 $ig0 )0) s%ows t%e GMRT image of t%e field containing t%e DNR G0010)-000 at t%e centre of t%e image0 Mt%er well known sources ;DNRs and " > regions in t%e Galactic Centre region are clearly visi(le in t%is image0 T%e RMD noise is relatively %ig%, possi(le due to t%e Galactic Center w%ic% lies at t%e sout%-

western edge of t%e primary (eam0 $ew of t%e GMRT antennas %ad servo related errors due to w%ic% t%ere were small oscillations in t%e antenna pointing w%ile tracking0 T%is, in t%e presence of strong sources at t%e edge of t%e fields, results in s%ort time scale differential gain variations w%ic% are not easy to correct later and also results in a %ig%er RMD noise0 $ig0 )0! s%ows t%e full primary (eam corrected images of t%e field containing t%e DNRs G00)0+-00,, G00'09T00' and t%e unclassified source G00'0*-0010 T%e low resolution image in t%e left panel was made using a single facet, w%ile t%e %ig%er resolution image was made using a grid of facets0 T%e lower surface (rig%tness DNR G00'09T00' is (etter discerned in t%e low resolution image0 T%e dominant e8tended source in $ig0 )0* is a known ?ltra Compact " region ;3ecker et al01&&)>0 A small angular si#e DNR G00)0,-000 was reported (y Gray ;1&&)a> in t%is field at t%e centre of t%is image0 "owever t%ere is no indication of t%is source at t%e level of 10 mFy/(eam in t%is image0 t is, %owever detected as a compact flat spectrum source in t%e low resolution image0 T%is sources is unlikely to (e an DNR0 $ig0 )0+ s%ows t%e GMRT ',!-M"# image of t%e s%ell type DNR G00)09T*0,0 T%e strong, marginally resolved source due west of t%is DNR is t%e well known :epler2s DNR ;$ig0 !01)>0 G00)09T*0, is again clearly detected in t%e NBDD image of t%is region ;$ig0 !0+>0 T%is DNR is also detected in t%e image made from a ',+-M"# B6A o(servation of a region close to t%is source ;$ig0 !09>0 $ig0 )09 s%ows t%e field containing t%e (arrel s%aped DNR G'!*0,-10!0 T%e 9)'-M"# image of t%is DNR (y Gray ;1&&)a> was severely affected (y artifacts due to t%e grating response of near(y sources0 T%is DNR is %owever clearly detected in t%e GMRT ',!-M"# image0 A marginally e8tended source of emission is also visi(le in t%is image in t%e nort%-eastern direction0 $ig0 )0& s%ows t%e GMRT image of t%e s%ell type DNR G'!*0,T)0!0 T%is DNR is also clearly visi(le in t%e NBDD image of t%is region ;$ig0 !010>0 T%e @uality of NBDD images close t%e Galactic plane is usually poor0 "owever, a few degrees away from t%e plane, low surface (rig%tness DNRs are often easily visi(le in NBDD fields ;3%atnagar,000P Green,001P Trus%kin1&&&>0 A careful e8amination of t%e NBDD fields, few degrees away from t%e plane is t%erefore likely to result in t%e identification of more, %it%erto unknown DNRs0 Eeep imaging of suc% o(Cects can t%en (e followed up wit% t%e GMRT/B6A0 Eetailed multi fre@uency imaging of a num(er of %ig% Galactic latitude DNRs

can (e used to possi(ly deduce t%e distri(ution of ioni#ed gas and e8amine t%e statistical significance of t%e - - relation ;Caswell O 6erc%e1&+&>0 $ig0 )010 s%ows t%e GMRT image of t%e incomplete s%ell of t%e DNR G'!900T'090 T%is is a low surface (rig%tness DNR, (ut also detected in t%e NBDD image ;$ig0 !011>0

Figure 232: $ull primary (eam image of t%e field containing t%e sources G0010)

000 at

t%e centre of t%e image0 T%e resolution in t%e image is 0 T%e RMD noise in t%e images is mFy/(eam0 T%e Galactic Centre sources ;6aRosa et al0,000>

Dgr E " region and Dgr E DNR pair, t%e DNR G0000& 001, Dgr 31 and Dgr 3, are t%e dominant sources in t%e sout%-east direction0

Figure 239: $ull primary (eam image of t%e field containing t%e sources G00'0* 001, G00'0+-00,, and G00'09T00'0 T%e resolution in t%e image in t%e left panel is t%is image w%ile t%at in t%e image in t%e rig%t panel is 0

T%e RMD noise in t%e images is mFy/(eam0 T%e e8tended emission is more clearly visi(le in t%e low resolution image, w%ic% was generated using a single facet0 T%e %ig%er resolution image was made wit% multiple facets0

Figure 23:: "ig% resolution, multi facet image of t%e field containing G00)0,

0000 T%e

resolution in t%e images is and t%e RMD noise mFy/(eam0 T%e dominant e8tended source in t%e field is an ?ltra Compact " region0 T%ere is a %int of a emission at t%e location of G00)0, 000, (ut muc% (elow t%e e8pected level, pro(a(ly

indicative of t%ermal source ;see $igure !0! for a low resolution image w%ere G00)0, 000 is detected as compact source>0

Figure 23;: $ull primary (eam corrected image of t%e field containing G00)09T*0,0 T%e weaker source, due east of t%e centre of t%e image is t%e DNR G00)09T*0,0 T%e strong source due west of t%e image centre is t%e '9 Fy :epler2s DNR0 T%e resolution in t%e image and t%e RMD noise is mFy/(eam0 Apart from ot%er

unknown sources of noise, t%e %ig%er RMD noise in t%is image is also due to t%e wo((le in t%e antenna pointing for some of t%e antennas used for t%is o(servation0

Figure 23<: Multi facet image of t%e field containing G'!*0,-10!0 T%e resolution in t%e image on t%e left is and t%e RMD noise is mFy/(eam0 Notice t%e e8tended emission nort%-east of t%e (arrel s%aped DNR in t%e centre of t%e field0 T%e panel on t%e rig%t s%ows a %ig% resolution image of t%is e8tended source0

Figure 23=: $ull primary (eam corrected image of t%e field containing G'!*0,T)0!0 T%e resolution in t%e image is and t%e RMD noise is mFy/(eam0

Figure 234>: $ull primary (eam corrected image of t%e field containing G'!900T'090 T%e resolution in t%e image is and t%e RMD noise is mFy/(eam0

hat is the d"namic range of GMRT?

Eynamic range of a telescope is its =a(ility= to detect weak sources in t%e presence of strong sources0 Commonly used definition of dynamic range is t%e ratio of t%e peak flu8 to t%e rms noise in t%e image0 T%is is true if t%e rms noise remains same at all regions in t%e image0 f due to systematics, correlated noise and cali(ration errors deconvolution is not proper and ripples appear surrounding t%e strong sources in t%e image, t%en t%is definition of dynamic range may not (e fully meaningful0 7ven in t%is kind of images, rms noise @uite far from t%e strong source will (e very small and will look =impressive=0 T%e ratio of peak flu8 to t%e noise in t%is region could (e large - (ut we will not (e a(le to detect faint sources close to t%e strong source0 T%is =source dependent noise= w%ere t%e noise is large close to t%e strong source tells us t%at all is not well0 Ke %ave attempted to understand t%e dynamic range of GMRT at 6-(and ;1,90 M"#> using t%e very strong source 'c)90 T%e source was o(served at a continuous stretc% for a(out +0! %ours at 1,90 M"# using 1* M"# (andwidt%0 Moderate flagging, (ased on single c%annel, was applied to t%e entire data0 nitially, t%e @uality of t%e image did not improve as muc% as e8pected0 6ater, all far away arm antennas are flagged ;all ARM-),!,*>0 3andpass was determined for every , min and was applied w%ile averaging a(out &0 c%annels0 T%e c%annel collapsed uv data was self cali(rated twice wit% p%ase only and once wit% amplitude and p%ase0 1lease click %ere for t%e final image0 T%e peak flu8 of 'c)9 at t%is fre@uency is 1+0,, Fy, (ut t%e grey scale of t%e image covers from 0 to 001 Fy ;linear> for (etter contrast0 T%e e8pected t%ermal noise is a(out !! microFy0 Away from 'C)9, t%e noise in my final image is 1!! microFy and t%e peaks are T*!0 and -*&0 microFy respectively0 T%is means, t%e peak to noise ratio is 0011 million0 Bery close to 'C)9, t%e noise is )00 microFy and t%e noise peaks are T,0) and -,0, milliFy respectively ;ripples>0 T%e peak to noise ratio now is a(out )',0000 K%at is t%e dynamic range of GMRT at 6-(andU is it 0011 million or a(out a t%ird of t%isU Ke are a(le to detect very faint sources far away from 'c)9, (ut we will not (e a(le to detect a source of similar strengt% ;very> close to 'c)90 K%y did t%e image @uality improve after ignoring far away antennasU 3y flagging far away antennas t%e ?B coverage (ecame less t%an )0 kilo lam(da, for larger (aselines, 'c)9 is partially resolved0 T%is may not (e t%e maCor

reason, (ecause even if 'c)9 %as structures for larger (aselines, t%e selfcali(ration will take care of it0 Ke may (e affected (y genuine errors wit% (aselines involving t%ese arm antennas0 Dince good dynamic range can not (e ac%ieved w%en t%e cali(ration is not accurate, we need to understand w%y we are not a(le to cali(rate t%e data in suc% a strong source to %ig%est accuracy0 T%ere could (e (aseline errors, (andpass insta(ility, amplitude and p%ase fluctuations at s%ort time scales ;s%orter t%an integration time>, mild R$ , etc0 could limit t%e accuracy of final cali(ration0 Ke are looking into t%e data for furt%er understanding t%ese pro(lems0 *ffect of %andpass T%e image @uality (ecomes poorer, if average (and pass was applied to t%e data, instead of estimating t%e (and pass for s%ort intervals of a few minutes0 T%is suggests t%at all is not well wit% t%e (and pass0 1lease visit Nimis%a2s %ome page for %er work on (andpass insta(ility0 0ormalised vs @nnormalised Ke %ave analysed unnormalised data to c%eck w%et%er unnormalised data improves t%e dynamic range0 T%ere was no noticea(le difference (etween t%e two0 Ke also need to c%eck w%et%er A6Cs M$$ give (etter results0

Information for -isitors a%out $omputers at GMRT General nformation

T%ere are , computers vi#0 gtac1 and gtac, placed in terminal room for astronomical data analysis specifically reserved for GTAC o(servers0 n addition, t%ere are four computers vi#0 astro1, astro ,, astro' and astro! are kept in Terminal room at GMRT central (uilding0 A computer astro) is kept in su(wing near control room0 T%ese mac%ines are similar to gtac1 and gtac, and can (e used for astronomical data analysis0 Eisk spaces of several G3 will (e routinely availa(le in /EATA/ partition of eac% of t%e a(ove mac%ine0 f t%ere is not enoug% space, please contact t%e system administrator0 1lease note t%at wit%in a day of your departure from t%e GMRT, all t%e files created (y you will (e deleted wit%out (ackup0 t is t%e responsi(ility of t%e GTAC user to take proper (ackup of t%eir data, (efore t%ey leave GMRT0 t is e8pected t%at t%e users will (ring t%eir own media for (ack-up0 "owever, in emergency, efforts will (e made to get a limited @uantity of EED-' tapes for Rs0 )!0/-0 n t%e computer room, a network printer is availa(le called 2npgcc20 $rom a 6inu8 mac%ine, command to print is. lpr -1npgcc 4files5 or 2lpr 4files50

6A1TM1 ?sers at GMRT.

T%ere are ta(les kept in t%e su(-wing near t%e control room ;passage from control room to server room>0 Aou can plugin your laptop and do t%e following settings. Name. guest0gmrt0ncra0tifr0res0in 1 Address . 1!901))0190,! Netmask . ,!!0,!!0,!!01&, Eefault GateKay . 1!901))0190,0 END Derver . 1!901))01901+

.reparing for o%servations: Dource 1arameters.

Create source list in your %ome area in similar to t%e source list in /(temp,/source/vlacal0list0 T%e spacings (etween t%e columns are important0 T%e sample command file is in

/(temp,/cmd/sample0cmd0 Mnce you are ready wit% t%e source list and command file, inform operators0 T%ey will c%eck t%ese and upload t%em to online mac%ine0

$re@uency Dettings.
A standard form availa(le in t%e control room w%ere t%e fre@uency settings can (e filled in0 7very GTAC user s%ould fill in t%is form indicating t%e o(serving fre@uency, (andwidt%, integration time and ot%er details0 Giving ver(al instructions can lead to wrong settings due to language pro(lems0 Computers astro1, astro,, astro' and astro! %as a EED' EAT tape drive wit% device /dev/nst00 Mn a EED' tape, 10 to 11 G3 of GMRT data ;GMRT raw data and/or $ TD> can (e put (y tar command or t%roug% aips0 Eefault data area w%ere t%e GMRT raw data is recorded is /rawdata for ?D3 data and /rawdata1 for 6D3 data0 n t%is area data will (e present in t%e su(directory called (y day and mont% e0g0 /rawdata/1+mayP /rawdata1/1+may0

Euring t%e M(servations00

Dome offline analysis software are availa(le to look at t%e data0 $or any pat% confusion, source /etc/(as%rc for (as% and /etc/cs%0cs%rc or /etc/cs%rc for cs%0 Dource /astro/RC for (as% s%ell or source 2/astro/RC0cs%2 for cs%/tcs% s%ell0 Run t%e software and say e8plain0 T%is will give %elp for t%e software2s usage0 T%e following programmes will (e useful for any GTAC user0 10 lta%dr . Gives info a(out t%e GMRT raw data, w%ic% is in 2lta2 format0 T%e rawdata for ?D3 comes wit% an e8tension 0lta and 6D3 wit% 0lt(0 ?sage. lta%dr -i 0 C%eck 2lta%dr --%elp2 for details0 ,0 ta8. Grap%ical interface to plot t%e data0 ?sing t%is programme, one can monitor t%e p%ase and amplitude sta(ility of t%e data w%ile t%e o(servation is in progress0 Type 2ta82 and t%en type 2e8plain2 for more details0

Converting GMRT raw data to $ TD format.

T%e GMRT data can (e converted into $ TD file using two programmes, listcan and gvfits0

10 listscan . 6ist t%e scan in t%e ltafile for converting to $ TD format0 Go to /EATA/P make sure t%ere is enoug% space0 Run listscan ;eg. listscan >0 T%is will create a file wit% 0log e8tension0 Mpen t%is file wit% your favorite editor0 T%e parameters you mig%t need to c%ange are t%e output $ TD file name, C6ADD CINMRM ;normalisation (y self>, removing unwanted antennas or scans, giving pat% to t%e $6AG file etc0 10 gvfits . Convert GMRT ;lta> data to $ TD format0 Run gvfits on t%e a(ove log file ;eg. gvfits test0log>0 T%is will take some time depending on t%e si#e of t%e input file0

Miscellaneous offline programmes.

T%e following programmes mig%t not come in use for GTAC users if t%ere are no pro(lems in t%e o(servations0 10 8tract. 78tract data from lta file in various formats0 ,0 ltacom(. com(ines lta files0 '0 ltaclean. cleans up ltafiles wit% glic%es in it0 $irst one or two records can also (e removed using t%is programme ;similar to 2V?AC:2 in A 1D>0 )0 ltasel. allows data averaging selection0 !0 relta. allows averaging in time0 *0 offstop. Dtop fringe offline (y providing suita(le ta(les antenna coordinates and fi8ed delays0 +0 tcali((p . 1rogram for (andpass cali(ration0

"ow to get into A 1D.

To use aips on t%ese mac%ines, after login do following. gtac1S newgrp aips gtac1S source /usr/aips/6MG N0CD" ;for cs%> or source /usr/aips/6MG N0D" ;for (as%> gtac1S aips 1refer to use aips id 1010, 10,0, 10'0, 10)0 and 10!0 for visitor1, visitor,, visitor', visitor) and visitor! accounts respectively0 T%ere is /EATA partition kept for aips data purpose0 Make re@uired num(er of su( directories %ere and edit 0dadevs appropriately0

!ac,up of the data:

t is t%e responsi(ility of t%e visitor to take (ackup of t%eir data0 No (ackup is taken for t%e /EATA and /analysis area and user %as to take %is own (ackup (efore %e leaves GMRT site0 All t%e files from t%e visitor account will (e deleted t%e day following t%e day visitor leaves GMRT0

GMRT>>>3A.G
Eate. 0!/,!/,00+ Biews. 1*!

GMRT>>23A.G
Eate. 0!/,!/,00+ Biews. 119

GMRT>43A.G
Eate. 0!/,)/,00+ Biews. 1!0

GMRT>443A.G
Eate. 0!/,!/,00+ Biews. 11!

GMRT>483A.G
Eate. 0!/,!/,00+ Biews. 109

GMRT>4;3A.G
Eate. 0!/,!/,00+ Biews. 100

GMRT>73A.G
Eate. 0!/,)/,00+ Biews. 11!

GMRT>743A.G
Eate. 0!/,!/,00+ Biews. 1,,

GMRT>873A.G
Eate. 0!/,!/,00+

GMRT>8<3A.G
Eate. 0!/,!/,00+ Biews. +1

GMRT>23A.G
Eate. 0!/,)/,00+ Biews. *+

GMRT>2>3A.G
Eate. 0!/,!/,00+ Biews. *)

GMRT>93A.G
Eate. 0!/,)/,00+ Biews. **

GMRT>:3A.G
Eate. 0!/,)/,00+ Biews. **

GMRT443A.G
Eate. 0!/,)/,00+ Biews. +1

GMRT4443A.G
Eate. 0!/,!/,00+ Biews. ))

GMRT483A.G
Eate. 0!/,)/,00+ Biews. +9

GMRT423A.G
Eate. 0!/,)/,00+ Biews. +1

GMRT2>3A.G
Eate. 0!/,)/,00+ Biews. +*

GMRT243A.G
Eate. 0!/,)/,00+ Biews. +)

GMRT223A.G
Eate. 0!/,)/,00+ Biews. 9)

GMRT2<3A.G
Eate. 0!/,)/,00+ Biews. +'

GMRT9>3A.G
Eate. 0!/,)/,00+ Biews. &+

GMRT923A.G
Eate. 0!/,)/,00+ Biews. 10& first previous