WELCOME

TO
MSA
TRAINING PROGRAM
OBJECTIVE
To provide
 Basic concept of measurement system
 Detail application methodology of all
statistical studies for variable and
attribute data
 Methods to meet QS9000 & ISO/TS-
16949 requirements
Introduction
Uses of measurement data are
 Decision to adjust a manufacturing process
 Determine the significant relationship
between two or more variables using
statistical studies
 The result of analytical studies is depends
upon the Quality of data produced by
measurement system
Why MSA required?

Defect prevention
Prediction SPC
Statistics
Data
Inspection

MSA is mandatory prior to SPC

PURPOSE OF MSA
 Quantify Variation present in the
Measurement system
 Compare the Consistency of
inspectors
 Provide methods for Validation of
Measurement System
 Measurement Systems
Process to assign a number or decision to
a characteristic
 Process
 Appraiser
 Checking Method
 Instrument
 Environment
 E.g, OD checking, Concentricity checking
Terminologies
 Standard
 Effective resolution
 Location variation
 Width variation
 Uncertainty
Statistical properties of MS
Ideal Measurement system is the one
 Which produces always only the
correct measurements
 Always agree with the standard
 Statistical data produced over time
Statistical properties of MS
Properties of “GOOD” Measurement Systems :
 Adequate discrimination & Sensitivity
 System is always under statistical control
 Variability of the measurement system is
small compared to Manufacturing process
variation and Tolerance
Criteria for MS design
selection
 Team members
 Purpose of use
 Frequency of use
 Part specification
 Sensitivity required
 Contact? Non contact/
 Auto? Manual?………………
Statistical properties of MS
Classification Measurement system errors are:
 Bias
 Linearity
 Stability
 Repeatability
 Reproducibility
Statistical properties of MS
Determination of these errors will help us to:
 Criteria to accept new MS
 Compare one against another
 To evaluate a suspect gage
 To evaluate the effectiveness of repair
 QS9000 / ISO-TS16949
Requirements
 Define a process for implementation of MS
 Identify all types of measurement systems
 Cover all systems covered in Control Plan
 Perform Appropriate statistical study
 Use MSA ( or Customer approved) Manual
 Obtain Customer approval for use of
methods other than MSA manual
 QS9000 / ISO-TS16949
Requirements
Types of MS is based on
 L.C
 Range
 Construction
 Parameter
 Manufacturer
 QS9000 / ISO-TS16949
Requirements
Vernier Vernier
L.C – 0.02 mm L.C – 0.02 mm
Range – 0 – 200mm Range – 0 – 200mm
Construction- Construction-
Ordinary Ordinary
Parameter- O.D Parameter- Width

Is it Same or different type of MS ?

 QS9000 / ISO-TS16949
Requirements
Bore Dial Gage Air gage
L.C – 0.001 mm L.C – 0.001 mm
Range – 18-35 mm Range – ---
Construction- Construction- Dial
Ordinary Parameter- Bore
Parameter- Bore
Discrimination
 Ability to detect the small changes in
process
 Traditional rule
 1/10th of Tolerance
 1/3rd Considering the cost/ criticality
 MSA/ SPC requirement
 1/10th of Process variation
 Resolution 1/10th of Process variation
 ACCURACY & PRECISION
ACCURACY PRECISION
 Closeness to reference  Ability of MS to repeat
or master value the same reading
 Required where two or  Required where MS is
more MS measuring a repeatedly used to
same characteristic assess and adjust the
 Same parameters are process
checked at Suppliers  In process inspection as
end or at Customer end per control plan
 ACCURACY & PRECISION
ACCURACY PRECISION
Captured by Captured by

 BIAS  REPEATABILITY
 LINEARITY  REPRODUCIBILITY
 STABILITY
BIAS – WHAT IT IS ?
Ref. Value
BIAS= X bar - Ref.Value

X bar
BIAS - STEPS
1. Capture reference value
• Use masters calibrated traceable to
standards
• Measure by better instrument than the
instrument being measured
• By Layout inspection

2. Measure the part/ master by Min. 10 times

 BIAS - STEPS
3. Calculate BIAS for individual reading & draw
histogram
4. Calculate ,  bias, t limits

APPENDIX C t distribution table

 ACCEPTANCE CRITERIA
Zero must lie between t Limits
of Bias distribution or
Bias should be within
acceptable gage error as
defined in Calibration Proc.
BIAS – VARIOUS CASES
Ref. Value
Ref. Value

X bar
BIAS – VARIOUS CASES
Ref. Value
Ref. Value

X bar
 CAUSE OF LARGE GAGE
BIAS
• Error in Master
• Instrument
- Worn
- Made to wrong dimension
- Measuring wrong characteristic
- Improper use by operator
- Improper calibration
APPLICABILITY OF GAGE BIAS
• Not a Problem like R&R
• Required where Accuracy to be captured
• Being Carried out as part of Calibration
• Applicable for Gages designed by the
Organization
 LINEARITY
Difference in Bias over the Operating range
of the Gage
+ve Bias
- ve Bias

X bar X bar
 Linearity – Steps
1. Collect 5 parts / masters over the
operating range of the gage through process
variation
2. Capture reference value for the 5 parts
3. Measure at least 12 times in a random
manner to cover all 5 parts

4. Measure at least 12 times in a random

manner to cover all 5 parts
 Linearity – Steps
5. Calculate Average, BIAS and enter the data
in a Spread sheet

6.
Draw Best fit line for the average, Lower limit and Hi
gher limit of Bias
 Linearity – Acceptance criteria
ZERO LINE MUST BE CONTAINED WITHIN THE
TWO LIMIT LINES OR WITHIN ACC.CRITERIA OF
CAL ERROR
+ VE Upper t limits

0
Lower t limits
- VE

Is this MS acceptable ?
 LINEARITY - CASES

0 0
 Causes for Gage Linearity
•Error in Master
•Instrument
-Worn
-Made to wrong dimension
-Measuring wrong characteristic
-Improper use by operator
-Improper calibration
 Applicability of Gage Linearity
• To Capture Accuracy over the operating
range to assure its suitability of gage for the
entire process variation

• Required for Special gages designed by the

Organization

• Can be performed as part of Calibration

 Stability study – What?
Variation of Bias w.r.t time when measured over a
Same Master

T1

T2

T3
 Stability – How?
No.of Samples/ 1 or 3 at Minimum, Mid and Higher
Masters Specification limits

Subgroup Size 1 or 3 times over a period in the

Morning, Noon and Evening
Frequency Daily,Weekly, Monthly based
on the stability assessment of
instrument
Charts One chart for each sample, X-
MR , or X bar& R or X bar &
S
 Stability – Acceptance criteria
• No out of Control conditions in Control charts
with careful interpretation
• No drift
• No trend
• Zero must lie within two limits of bias variation
 Stability – Applicability
• To compare two or more measurement systems
measuring a same characteristic
e.g Bore dial gage & Air gage

• To establish or evaluate the Calibration

Frequency
Constitution of total variation
 Total variation
 Manufacturing process variation
 Part to part variation
 Within part variation

 Measurement system variation

 Appraiser variation
 Equipment variation

 Appraiser part interaction

Methods to capture variation
Elements of Range X bar & R ANOVA
Variation Method Method Method
TV
PV
WIV
AV
EV
App & part
interaction
Repeatability & Reproducibility
Repeatability Reproducibility
• One Appraiser • Same equipment
• One Equipment •Same Parts
• Same part •Several trials
• Several trials •Different Appraiser

This Variation is This Variation is

represented by represented by
Equipment Appraiser
 Range method
No.of Samples 5 nos. gives 80 % chance of
detecting a wrong MS.
10 nos gives 90% chance
No.of Appraisers Min. 2

Formula Sigma = R bar / d2*

Range method - Exercise
Appraiser Appraiser Range
A B
0.02
1. 15.03 15.05
2. 15.04 15.04 0.00
3. 15.02 15.03 0.01
4. 15.05 15.05 0.00
5. 15.02 15.02 0.00
Tol = 0.05, TV(SPC)= 0.120
Range method - Exercise
Average R = 0.006

Std.deviation = 0.006 / 1.128 = 0.0053

where n=2 d2 is1.128, n=3 d2 is 1.693
R&R ( 1 sigma) = 0.0053
% R&RTV = (0.0053)/ (0.120/6) * 100 = 26.5
% R&RTOL = (0.0053)/ ( 0.050/6) * 100=
66.5
Interpretation = MS IS NOT ACCEPTABLE
Range method - Exercise
COMPARE AGAINST TOLERANCE WHEN
• TV IS MORE THAN TOLERANCE
• WHERE SPC IS NOT IMPLEMENTED

COMPARE AGAINST TOTAL VARIATION

WHEN
• TV IS LESS THAN TOLERANCE
• WHERE SPC IS IMPLEMENTED
 Range method – Acceptance
criteria

% R&R < 10 % OF TOL OR TV

% R&R 10 – 30 % OF TOL OR TV ACCEPTABLE

SUBJECT TO ANALYSIS AND JUSTIFICATION
W.R.T COST OF REPAIR AND CRITICALITY
 Range method – Applicability
Being the short study it is suitable for studying MS
where it is time consuming to measure one
measurement e.g Concentricity, R/O

Being the Short study this can be used for

Periodic verification of R&R to detect any changes
rather than average range method
 Average & Range Method
No.of Samples 10 nos. Randomly selected
representing actual process
variation
No.of Appraisers 3 From personnel actually
performing on a day to day
operation
No.of trials Min 2 . Preferably 3
 Average & Range Method-
steps for conducting study
1. Identify the parts from 1 – 10
2. Mark the Place where to check to eliminate
Within part variation
3. Communicate the purpose of the study to all
appraisers
4. Conduct the study in a Random manner
covering all parts for required no.of trials
through all appraisers
5. Monitor all appraisers are following the same
method
 Average & Range Method-
steps for conducting study
6. Record the observations in a manner which is
not seen by the appraisers – FORMAT-
7. If any abnormal readings are observed ask
the appraiser to repeat the reading
8. Maintain the samples till the completion of
analysis
9. Draw R chart and X bar chart and interpret

10. Calculate EV,AV,PV & TV and it Percentage

against TOL or TV and ndc
 X bar & R chart Method –
Acceptance criteria

% R&R < 10 % OF TOL OR TV

% R&R 10 – 30 % OF TOL OR TV ACCEPTABLE

SUBJECT TO ANALYSIS AND JUSTIFICATION
W.R.T COST OF REPAIR AND CRITICALITY
No. of distinct data categories
 ndc > 5 where SPC is used
ndc >2 where no SPC is used
R chart interpretation
CONDITION INTERPRETATION
 One or more than one  His method is
point of one different from others
appraiser out of UCL
R
 One or more than one  Measurement System
point of all appraisers is sensitive to
out of UCL R appraisers skill
 In one part all  Part is deformed or
appraisers points are Damaged
out of UCL R
X chart interpretation
CONDITION INTERPRETATION
More than 50 % of Measurement
readings are out system is adequate
of control limits enough to capture
process variation
Causes for R&R
When repeatability is large compared to
reproducibility
 Instrument needs maintenance
 Redesign gage for more rigidity
 Improve clamping or location of gauging
 Excessive within – part variation

( Conduct R & R with within part variation )

Causes for R&R
When reproducibility is large compared to
repeatability
 Appraisers need better gage use training
 Need better operational definition
 Incremental divisions on instrument are not
readable
 Need fixture to provide consistency in gage use.
 X bar & R method -
Applicability
• To capture Precision, Mandatory before SPC
study
• Measurement Systems which are used
repeatedly on line to control the process based
on the measurement data
• Where individual readings are affected by
the precision of the instrument

• Applicable for all MS used in In process

where the MS can be repeated
 ATTRIBUTE MEASUREMENT
• COMPARES EACH PART TO A SPECIFIC SET OF LIMITS
AND ACCEPTS THE PART IF THE LIMITS ARE SATISFIED
• IS DESIGNED TO ACCEPT/REJECT A SET OF MASTER
PARTS
• CANNOT INDICATE HOW GOOD OR HOW BAD A PART
IS,ONLY WHETHER THE PART IS ACCEPTED OR
REJECTED (PASS/FAIL)
• VISUAL STANDARDS MAY RESULT IN 5 TO 7 DISTINCT
DATA CATEGORIES. MSA MANUAL DOES NOT
PRESCRIBE ANY METHOD OF EVALUATION OF THE
SAME
• GPC METHOD IS MOST APPROPRIATE ANALYSIS
• RISK ANALYSIS METHOD IS CUMBERSOME, BUT IT IS
PREFERRED BY SOME CUSTOMERS (Eg, FORD)
Gage performance curve
method
LSL USL HIGH
RISK
AREA

I II III II I
GPC KEY INSTRUCTIONS
 Select the gage to be studied
 Calibrate the gage to know the actual size of the
gage
 Understand the distribution and location of the
process
 Decide the on which side of specifications
samples to be taken and the measurement method
 Select samples (8) and repeat minimum 20 times
on a random manner and count no.of accepts ( a )
GPC KEY INSTRUCTIONS
 Select Minimum 8 samples out of that
 One sample on lower than specification, no.of
accepts a = 0
 One sample on higher specification, no of accepts
a = 20
 Six parts on the gray area where, a is more than
or equal to 1 and less than or equal to 19
 Select sample repeat the study till we meet the
above criteria
GPC KEY INSTRUCTIONS
 Calculate Probabilities of acceptance using
formulae
 Plot Probabilities of acceptance as % in
normal probability paper
 Calculate Bias & Repeatability
 Conclude the decision on the acceptability
of gage
RISK ASSESSMENT – Key
steps
 Take 50 parts representing entire process
variation having bad, marginally bad, good
& marginally good parts
 Mark as 1 to 50
 Select 3 appraisers
 Conduct study in a random manner and
record the decisions, 1 as OK and 0 as Not
ok
RISK ASSESSMENT – Key
steps
 Get the reference value for all the 50 parts
using layout inspection/ variable
instrument and decide its status as 0 or 1
 Compare each trial of each inspector with
the another inspector for their decision
 Complete the cross tabulation table
 Calculate Kappa for
 A Vs B, A vs. C , B Vs C
 A vs. Ref, B Vs Ref., C vs. Ref.
Risk analysis – Acceptance
criteria
Decision Effectiveness
Acceptable for the More than 90 %
appraiser
Marginally acceptable More than 80%
for the appraiser
Unacceptable for the Less than 80 %
appraiser – Need
improvement
MSA TRAINING

ANY QUESTIONS ?
Thank You