SCMH 3.11.4 MSA Case Studies Rev New Dated 21MAR2018

SCMH Section 3.11.
4
Revision Letter: New
Revision Date: 21MAR2018
www.iaqg.org/scmh Section 3.11
Measurement System Analysis Case Studies
Introduction
This document is a follow up document to the Measurement System Analysis Process Flow
document. This document contains a group of case studies depicting MSA examples as well as
analysis and conclusions that were made to reduce measurement system variation for improving
product quality. These case studies follow the process flow diagram below that is from the
Measurement System Analysis Process Flow document.
From this document, you will be able to obtain a basic understanding of conducting an MSA on a
measurement system.
Multiple situations may exist where an MSA needs to be conducted and it is the company’s
responsibility to work with their customer’s and determine an acceptable method to use. Multiple
documents and systems are available that will assist in this process.
Following are the five case studies:
• Mechanical (Micrometer)
• Attribute MSA (Visual Inspection)
• Attribute MSA (Go/No-Go Gage)
• Electrical device (Multi-meter)
• Software controlled device (Coordinate Measuring Machine)
For additional case studies see SAE AS13003 Measurement Systems Analysis Requirements
for the Aero Engine Supply Chain
1
SCMH Section 3.11.4

Process Flow Diagram (Appropriate gage selection)
1
Input
- Part feature &Tolerance
- Method/systems
-Calibrated gages
- Trained operators
- Suitable environment
-
2
Determine that the
gage(s) has the
appropriate
resolution for the
feature and
tolerance
3
Perform measument
uncertanty of the
gage(s)
Yes
4
Does the gage meet
the criteria for A
Yes
- Accuracy
-Linearity
-Stability
No
4A
Is there another
gage available that
meets the criteria
No
4B
Take action as
appropriate
Return to Start
2
SCMH Section 3.11.4

Process Flow Diagram (Variation analysis)
5
Is there a same or 6
EOP
similar measurement Yes Is the measurement Yes
(End of process)
system? system stable?
No
7
Determine the
sources of variation
No
that may affect the
measurement
system
7A
8
Is your measurement
Yes Perform an analysis
system dependent on
of your software
software?
No
7B 10
Is the measurement Develop and
equipment the primary No perform a gage R&R
source of variation in the to evaluate sources
system? of variation
Yes
9
Perform
measurement
B
uncertanty analysis
on the system
(Made up of
multiple gages)
3
SCMH Section 3.11.4

Process Flow Diagram (Analyze results and take action)
11
Verify and
document that the
elements of the
MSA have been
evaluated
11A
Has critera for
Yes End of process
elements been
met?
No
12
Take action as
appropriate
Return to start
4
SCMH Section 3.11.4

Process Flow Diagram
5
SCMH Section 3.11.4
Introduction to First Case Study, Mechanical MSA

This case study will be looking at a mechanical measurement.
This case study will use a micrometer to measure the diameter of a cylinder.
From this case study, you will be able to obtain a basic understanding of conducting an MSA on a
measurement system where a mechanical instrument is used to take a measurement.
6
SCMH Section 3.11.4
First Case Study

This case study will be looking at a simple cylinder with two basic requirements. The part being
produced is a cylinder that is 1 inch tall with a 0.5 inch diameter.
Gather inputs:
Cylinder height is 1 inch tall, with a specification of ±0.1 inch.
Cylinder diameter is 0.5 inches, with a specification of ±0.005 inches.
For this case study, we will be specifically looking at the cylinder diameter.
Gage determination:
There are four measurement instruments available to choose from in the current gage inventory.
(CMM, Caliper, Micrometer, steel ruler)
When selecting the appropriate resolution of a gage it is recommended that the resolution of the gage
be 10 times more resolution than the tolerance. i.e. If the range of the tolerance is 0.01 inch; the
resolution of the gage needs to be at least 0.001 inch.
7
SCMH Section 3.11.4
4
Does the gage meet
the criteria for
- Accuracy
-Linearity
-Stability
Gage criteria:
The gage criteria includes accuracy, linearity, and stability.
• The tolerance of the part diameter is ±0.005 inches so the accuracy of the gage should be 4
times better with a tolerance of ±0.00125 inches.
• For linearity, it is recommended to select a gage where the measurement taken is in the
mid-range of the gage. In this case, we need an instrument with a range of 0-1 inches to
measure the 0.5-inch diameter.
• It is recommended that the stability be at least 10 times better than the accuracy. The stability
of the measurement instrument can be evaluated by looking at historical calibration records.
With a new instrument, the manufacturer’s calibration documentation should describe the
stability of the instrument.
For this case study the 0-1 inch micrometer was selected.
7
Determine the
that may affect the
measurement
system
Sources of variation:
Determine what the possible sources of variation are with this measurement system. A fish bone
chart is a good tool to list the possible sources of variation. With this case study, it has been
determined that the operator and the environment are the primary sources of variation. With this
8
SCMH Section 3.11.4

information a gage R&R will be conducted.
10
Develop and
perform a gage R&R
to evaluate sources
of variation
Develop and perform a gage R&R:
A gage R&R process is started by obtaining parts, using a trained operator, a gage to take the
measurement, and having a template to record data:
• Parts – Recommend 10 parts with a minimum of 5. Parts should span the normal process
variation. Identify and label each part for recording on the template
• Trained operator – All operators that are part of the measurement process with a minimum of
2 operators. The operators should have basic knowledge and skill in measurement
techniques, and in use of the selected gage
o The operator should follow the normal measurement process
• Gage to take the measurement
o Verify calibration and proper operation
• Template to record data – record each measurement of each part by each operator
o Recommend 2-10-2 measurement. Which is 2 operators, 10 samples, 2 recordings for
each part
o Record any observations during measurement
Conduct the gage R&R and take data.
9
SCMH Section 3.11.4
Once the data has been obtained it needs to be analyzed. In this case study a statistical software
package was used to analyze the data. Following is a chart showing the results of the analysis:
From this analysis, it can be observed that there is a difference between the operators’
measurements.
An investigation into the variation between the operators needs to be performed:
• Determine which operator was more consistent (Observe techniques,; use a master part with
known values, etc.)
10
SCMH Section 3.11.4
12
Take action as
appropriate
Take Action:
After the analysis and investigation, an action plan can be made to address variation between the
operators. Some examples of an action plan may include the following:
• Training
• Better documented method
• Improve skills
• Improve ergonomics
After the implementation of the action plan re-evaluate the element(s) where a change was
introduced. In this case study, the gage R&R would need to be conducted again.
11
SCMH Section 3.11.4
Introduction to Second Case Study, Attribute MSA

This case study will be looking at the surface of a cylinder for surface blemishes.
This case study will use visual inspection to evaluate the cylinder for blemishes.
measurement system where attribute data is used to make a decision.
12
SCMH Section 3.11.4
Second Case Study

This case study will evaluate the surface of a cylinder for blemishes such as nicks, scratches, and
dents. We will use the same cylinder from the first case study.
Gather inputs:
Maximum Scratch Length 2mm

Defect depth less than 2 microns
Maximum - 5 defects total per part
A Pass/fail condition is determined through a visual inspection compared to a standard.
7
Determine the
that may affect the
measurement
system
Sources of Variation:
diagram is a tool that can be used to list the possible sources of variation. With this case study, it has
been determined that the operator and the environment are the primary sources of variation. The
measurement method was reviewed for controls to minimize these sources of variation.
13
SCMH Section 3.11.4

10
Develop and
perform a gage R&R
to evaluate sources
of variation
Develop and perform an attribute agreement analysis:
(With an attribute study, an attribute agreement analysis is typically referred to as an Attribute gage R&R)
A gage R&R process is developed by obtaining parts, using trained operators, and having a template
to record data:
• Parts – Obtain 30 samples with a mix of known 50% pass, 50% fail. For each group of pass/fail
at least 25% of the group should be borderline. The borderline samples will help determine the
true effectiveness of the measurement system.
• Recommend 30 parts with a minimum of 25.
2 operators. The operators must be trained to the visual standard.
o The operator should follow the normal process.
• Calibration/ known population.
o As attribute studies typically don’t have a calibrated gage (per NIST standards) a
“known” result must be determined for each sample by a recognized expert or by team
consensus.
• Template to record data – record each measurement of each part by each operator.
each part.
o Record any observations during the measurement.
o Once these have been determined and taken into consideration an attribute agreement
analysis is conducted.
14
SCMH Section 3.11.4
Attribute Gage R & R Effectiveness
DATE: 10/17/2017
Attribute Legend5 (used in computations) NAME: IAQG
1 Pass PRODUCT: Cylinder All operators
2 Fail GAGE: Visual Inspection agree within and All Operators
BUSINESS: ACME Cylinder between each agree with

Other standard
Known Population Operator #1 Operator #2 Operator #3 Y/N Y/N
Sample # Attribute Try #1 Try #2 Try #1 Try #2 Try #1 Try #2 Agree Agree
1 Pass Pass Pass Pass Pass Pass Pass Y Y
4 Fail Fail Fail Fail Fail Fail Fail Y Y
7 Fail Pass Pass Pass Pass Pass Pass Y N
25 Pass Fail Fail Pass Pass Pass Pass N N
Typical Attribute Agreement Study Spreadsheet
As seen in the chart above, as indicated by the red highlight, the operators consistently deviated to
the “known” attribute. Upon reviewing sample numbers 7, 13, and 23 it was determined that the visual
standard was insufficient for the operators to distinguish pass from fail.
In conclusion: The visual standard was updated to include examples that better represent the three
samples that caused inconsistencies vs the “known” attribute.
15
SCMH Section 3.11.4
Statistical Report - Attribute Gage R&R Study

DATE: 10/17/2017
NAME: IAQG
PRODUCT: Cylinder
GAGE: Visual Inspection
BUSINESS: ACME Cylinder
% Appraiser1 %Score vs Attribute 2

Source Operator #1 Operator #2 Operator #3 Operator #1 Operator #2 Operator #3
Total Inspected 30 30 30 30 30 30
# Matched 30 30 30 26 27 27
False Negative (operator biased tow ard rejection) 1 0 0
False Positive (operator biased tow ard acceptance) 3 3 3
Mixed 0 0 0
95% UCL 100.0% 100.0% 100.0% 96.2% 97.9% 97.9%
Calculated Score 100.0% 100.0% 100.0% 86.7% 90.0% 90.0%
95% LCL 88.4% 88.4% 88.4% 69.3% 73.5% 73.5%
Screen % Effective Score 3 Screen % Effective Score vs Attribute 4

Total Inspected 30 30
# in Agreement 29 26
95% UCL 99.9% 96.2%
Calculated Score 96.7% 86.7%
95% LCL 82.8% 69.3%
% Appraiser 95% UCL % Score vs Appraiser 95% UCL

Calculated Score Calculated Score
95% LCL 95% LCL
110.0% 110.0%
100.0% 100.0%
90.0% 90.0%
80.0% 80.0%
70.0%
% Efficiency
70.0%
% Efficiency
60.0% 60.0%
50.0% 50.0%
40.0% 40.0%
30.0% 30.0%
20.0% 20.0%
10.0% 10.0%
0.0% 0.0%
Operator #1 Operator #2 Operator #3 Operator #1 Operator #2 Operator #3
Notes
(1) Operator agrees with him/herself on both trials
(2) Operator agrees on both trials with the known standard
(3) All operators agreed within and between themselves
(4) All operators agreed within & between themselves AND agreed with the known standard
By looking at the statistical report, the false negative and false positive data can be observed. In this
case, you can see that the operators were biased toward accepting non-conforming parts, which
could pose a significant risk to the customer. Typical reporting for the attribute analysis would be the
16
SCMH Section 3.11.4

Screen % Effective Score vs Attribute of 86.7%, as seen highlighted in red above. Screen % Effective
Scores thresholds may range from 90-100% and be considered acceptable. The actual acceptable
level is based on risk, which may be determined with the customer.
The Kappa statistic is a more robust measure than the agreement analysis used above, since Kappa
takes into account the possibility of the agreement occurring by chance. Statistical software
programs can be used to calculate Kappa statistics. Here is an example using our data above.
Within Appraisers Kappa Statistics

Appraiser Response Kappa SE Kappa Z P(vs > 0)
Michael Fail 1.00 0.182574 5.47723 0.0000
Pass 1.00 0.182574 5.47723 0.0000
Paul Fail 1.00 0.182574 5.47723 0.0000
Pass 1.00 0.182574 5.47723 0.0000
Jay Fail 1.00 0.182574 5.47723 0.0000
Pass 1.00 0.182574 5.47723 0.0000
In conclusion, all appraisers agreed with themselves every time achieving a perfect Kappa score of
1.0.
Each Appraiser vs Standard Kappa Statistics

Appraiser Response Kappa SE Kappa Z P(vs > 0)
Fail 0.732143 0.129099 5.67115 0.0000
Michael
Pass 0.732143 0.129099 5.67115 0.0000
Paul Fail 0.797980 0.129099 6.18112 0.0000
Pass 0.797980 0.129099 6.18112 0.0000
Jay Fail 0.797980 0.129099 6.18112 0.0000
Pass 0.797980 0.129099 6.18112 0.0000
In conclusion, Michael’s slightly lower Kappa score of 0.73 is attributed to his assessment
disagreeing with the “known” attribute one additional time compared with the other assessors.
17
SCMH Section 3.11.4
Between Appraisers Kappa Statistics

Response Kappa SE Kappa Z P(vs > 0)
Fail 0.963284 0.0471405 20.4343 0.0000
Pass 0.963284 0.0471405 20.4343 0.0000
In conclusion, the Kappa statistic shows that there is strong likelihood of agreement between all
assessors (without consideration for the “known” attribute).
All Appraisers vs Standard Kappa Statistics

Response Kappa SE Kappa Z P(vs > 0)
Fail 0.776034 0.0745356 10.4116 0.0000
Pass 0.776034 0.0745356 10.4116 0.0000
In conclusion, the overall Kappa statistics of 0.776 falls short of the typical minimum requirement for
Kappa of 0.8.
Either agreement analysis or Kappa statistics draw the same conclusion that the measurement
system has issues with discriminating Pass/Fail parts versus a “known” attribute. Once the visual
standard was updated, the complete MSA was re-run and the Screen % Effective Score was at 100%
and the Kappa statistic was 1.0.
The above case study illustrates the most common use of Attribute Agreement analysis using binary
(Pass/Fail) data. The same approach applies for Nominal (distinct categories, i.e. colors) and Ordinal
(ordered scales, i.e. poor, fair, good, and excellent) measures. Statistical software programs are
typically used to set up and evaluate the results which will include an agreement analysis percentage
as well as Kappa (or similar) statistic.
18
SCMH Section 3.11.4
Documentation:
This process is documented by the data taken from the attribute agreement analysis.
12
Take action as
appropriate
Take Action:
From the above data it has been determined that the operators are consistent,; however the
understanding of the standard is not clear. The standard will be clarified as a long term solution to
ensure adherence to the standard. In order to clarify the standard, further operator training will be
provided. Part magnification and better lighting conditions should be considered to reduce variation.
The attribute agreement analysis should be run again to verify the actions taken were successful.
19
SCMH Section 3.11.4
Introduction to Third Case Study, Attribute MSA

This case study will be looking at a Go/No-Go measurement.
This case study will use ring gages to evaluate the diameter of a cylinder.
measurement system where attribute data is used to make a decision.
Multiple situations may exist where an MSA needs to be conducted and it the company’s
20
SCMH Section 3.11.4
Third Case Study

This case study will use ring gages to evaluate the diameter of the cylinder used in the first case
study. In this case study the manufacturer wants to use an in-process check of the cylinder thickness
using simple ring gaging.
Gather inputs:
7
Determine the
that may affect the
measurement
system
been determined that the operator, the environment, and technique are the primary sources of
21
SCMH Section 3.11.4

variation. The measurement method was reviewed for controls to minimize these sources of variation.
10
Develop and
perform a gage R&R
to evaluate sources
of variation
Develop and perform an attribute agreement analysis:
(With an attribute study, an attribute agreement analysis is typically referred to an Attribute gage R&R)
A gage R&R process is developed by obtaining parts, using trained operators, and having a template
to record data:
• Parts – Obtain 30 samples with a mix of known 50% pass, 50% fail. For each group of pass/fail
at least 25% of the group should be borderline. The borderline samples will help determine the
true effectiveness of the measurement system. Given that the tolerance is 0.495 – 0.505,
borderline samples should be close to the upper and lower specification limits.
Borderline samples
• Recommend 30 parts with a minimum of 25.
2 operators. The operators must be trained to use the gages.
o The operator should follow the normal process.
• Calibration/known population.
o When using an attribute gage on a variable tolerance, there first must be a verified
measurement system. For this attribute agreement study, we will use the results from
the first case study (0-1 inch micrometer) and assume changes were made to reduce
the measurement system variation to an acceptable level. Using the 0-1 inch
22
SCMH Section 3.11.4

micrometer all samples were measured and translated into “known” Pass/Fail results
for the sample population.
• Template to record data – record each measurement of each part by each operator.
each part.
o Record any observations during measurement.
o Once these have been determined and taken into consideration an attribute agreement
analysis is conducted.
Conduct the study and take data.
23
SCMH Section 3.11.4
Attribute Gage R & R Effectiveness
DATE: 10/17/2017
Attribute Legend5 (used in computations) NAME: IAQG
1 Pass PRODUCT: Cylinder All operators
2 Fail GAGE: Ring Gages agree within and All Operators
BUSINESS: ACME Cylinder between each agree with

Other standard
Known Population Operator #1 Operator #2 Operator #3 Y/N Y/N
Sample # Diameter Attribute Trial 1 Trial 2 Tial 1 Trial 2 Trial 1 Trial 2 Agree Agree
1 0.5072 Fail Fail Fail Fail Fail Fail Fail Y Y
4 0.5043 Pass Pass Pass Pass Pass Pass Pass Y Y
8 0.4951 Pass Pass Pass Fail Fail Fail Fail N N
13 0.4953 Pass Pass Pass Pass Pass Fail Fail N N
Typical Attribute Agreement Study Spreadsheet
As seen in the chart above, as indicated by the red highlight, some operators consistently deviated to
the known standard. Upon reviewing sample numbers 8, 19, and 29 it was determined that the
borderline samples were difficult to distinguish. The operator technique caused the ring to bind up
and create a false negative (rejecting good parts).
24
SCMH Section 3.11.4
Statistical Report - Attribute Gage R&R Study

DATE: 10/17/2017
NAME: IAQG
PRODUCT: Cylinder
GAGE: Ring Gages
BUSINESS: ACME Cylinder
% Appraiser1 %Score vs Attribute 2

Source Operator #1 Operator #2 Operator #3 Operator #1 Operator #2 Operator #3
Total Inspected 30 30 30 30 30 30
# Matched 30 30 30 30 27 24
False Negative (operator biased tow ard rejection) 0 3 6
False Positive (operator biased tow ard acceptance) 0 0 0
Mixed 0 0 0
95% UCL 100.0% 100.0% 100.0% 100.0% 97.9% 92.3%
Calculated Score 100.0% 100.0% 100.0% 100.0% 90.0% 80.0%
95% LCL 88.4% 88.4% 88.4% 88.4% 73.5% 61.4%
Screen % Effective Score 3 Screen % Effective Score vs Attribute 4

Total Inspected 30 30
# in Agreement 24 24
95% UCL 92.3% 92.3%
Calculated Score 80.0% 80.0%
95% LCL 61.4% 61.4%
% Appraiser 95% UCL % Score vs Appraiser 95% UCL

Calculated Score Calculated Score
95% LCL 95% LCL
110.0% 110.0%
100.0% 100.0%
90.0% 90.0%
80.0% 80.0%
70.0%
% Efficiency
70.0%
% Efficiency
60.0% 60.0%
50.0% 50.0%
40.0% 40.0%
30.0% 30.0%
20.0% 20.0%
10.0% 10.0%
0.0% 0.0%
Operator #1 Operator #2 Operator #3 Operator #1 Operator #2 Operator #3
Notes
(1) Operator agrees with him/herself on both trials
(2) Operator agrees on both trials with the known standard
(3) All operators agreed within and between themselves
(4) All operators agreed within & between themselves AND agreed with the known standard
By looking at the statistical report, the false negative and false positive data can be observed. In this
case, you can see that the operators were biased toward rejecting conforming parts. Typical reporting
for the attribute analysis would be the Screen % Effective Score vs Attribute of 80%, as seen
highlighted in red above. Further analysis with a statistical software package yielded a Kappa score of
0.79 which is less than the minimum 0.80 needed to pass.
25
SCMH Section 3.11.4
Documentation
This process is documented by the data taken from the attribute agreement analysis.
12
Take action as
appropriate
Take Action:
From the above data it has been determined that the operator method caused the rejection of
conforming parts. The method that the first operator was using was documented as standard work for
the measurement process and implemented across all operators in the process.
26
SCMH Section 3.11.4
Introduction to Fourth Case Study, Electrical MSA

This case study will be looking at an electrical measurement.
This case study will take a DC Current measurement at 5 Amps.
measurement system where an electrical measurement is made with multiple instruments.
Multiple situations may exist where an MSA needs to be conducted and it the company’s
27
SCMH Section 3.11.4
Fourth Case Study

This case study will be looking at a DC Current measurement at 5 Amps.
Gather inputs:
The part feature being measured is 5 Amps DC with a specification of ±20 milliamps.
There are three common methods to measure DC Current:
1.) Measure directly with a multi-meter
2.) Measure voltage drop across a bar shunt with a multi-meter
3.) Measure voltage drop across a thermally stabilized shunt with a multi-meter
Gage determination:
Without performing an uncertainty analysis, it is unknown at this point if the resolution and tolerance
is sufficient for the measurement. A multi-meter will be used and a decision will be made at the end if
the multi-meter was sufficient for the measurement.
28
SCMH Section 3.11.4
Measurement Uncertainty Analysis:
4
Does the gage meet
the criteria for
- Accuracy
-Linearity
-Stability
Gage criteria:
With an uncertainty of ±0.0059 Amps dc, this results in a Test Uncertainty Ratio of approximately
3.3:1. With a risk requirement of 2% Probability of False Accept, this Test Uncertainty Ratio does not
meet the criteria.
29
SCMH Section 3.11.4
Choosing an alternate gage:
An alternate method will now be chosen to make the measurement. Measure voltage drop across a
thermally stabilized shunt with a multi-meter in the 1 V dc range.
Measurement Uncertainty Analysis:
30
SCMH Section 3.11.4
4
Does the gage meet
the criteria for
- Accuracy
-Linearity
-Stability
Gage criteria:
With an uncertainty of ±0.00071 Amps dc, this results in a Test Uncertainty Ratio of approximately
28:1. With a risk requirement of 2% Probability of False Accept, this Test Uncertainty Ratio is
sufficient to meet the criteria.
7
Determine the
that may affect the
measurement
system
In the uncertainty analysis above, all sources of variation, such as multi-meter, shunt, human factors
and environment, were taken into account.
Documentation:
This process was documented by the data taken from the uncertainty analysis.
31
SCMH Section 3.11.4
Introduction to Fifth Case Study, Software MSA

This case study will be looking at a software controlled device (Coordinate Measuring Machine)
With this case study the concentricity of the internal diameter to the external diameter will be checked
using a CMM.
From this case study, you will be able to obtain a basic understanding of conducting an MSA on
measurement system that uses a software controlled device such as a CMM.
responsibility to work with their customer’s customer and determine an acceptable method to use.
Multiple documents and systems are available that will assist in this process.
32
SCMH Section 3.11.4
Fifth Case Study
25 mm
50 mm
Gather inputs:
Concentricity of the 25mm
hole to the centerline with a
tolerance of 0.013mm
This case study will be looking at a ring and checking the concentricity of the internal diameter to the
external diameter.
A 50 mm diameter ring with a 25 mm hole.
Measure the concentricity of the hole with a tolerance of 0.013 mm.
Because of the type of measurement and the tolerance, a CMM has been selected to take the
measurement. A software program has already been developed to take this measurement.
7
Determine the
that may affect the
measurement
system
been determined that the part cleanliness, temperature, loading/securing technique, and CMM probe
type and calibration are the primary sources of variation. The measurement method was reviewed for
controls to minimize these sources of variation.
8
Perform an analysis
of your software
33
SCMH Section 3.11.4

Software analysis:
In this scenario a correlation study is conducted by using two different CMM’s and taking the
concentricity measurement of the part. The data between the two measurements is subtracted and
the difference should not be more than 10% of the part feature tolerance. The correlation study
showed a 96% correlation (i.e. 4% of the part tolerance) between the CMM’s.
7B
Is the measurement
equipment the primary
source of variation in the
system?
Primary Source of Variation:
At the start of this case study the primary source of variation was not known. In order to determine
the primary source of a variation gage studies were conducted.
A type one gage study is a best practice for determining gage error, but is not a replacement for a
variable gage R&R (also called a type two gage study). However, the type two gage study does not
separately calculate the specific gage error (i.e. the gage error will be combined with the operator
repeatability error)
A type one gage study can be conducted as the first step to determine gage repeatability error without
introducing other sources of variation. This is accomplished by taking 30 consecutive measurements
of the same part minimizing the sources of variation. If there is a known dimension of the part you can
also calculate bias.
A variable gage R&R (also called a type two gage study) can be started by obtaining parts, using
trained operators, a gage to take the measurement, and having a template to record data:
• Parts – Recommend 10 parts with a minimum of 5. Parts should span the normal process
variation. Identify and label each part for recording on the template
• Trained operators – All operators that are part of the measurement process with a minimum of
2 operators. The operators should have basic knowledge and skill in measurement
techniques, and in use of the selected gage (CMM)
o The operator should follow the normal measurement process including loading and
34
SCMH Section 3.11.4

unloading the part between each measurement
• Gage (CMM) to take the measurement
o Verify calibration and proper operation
• Template to record data – record each measurement of each part by each operator
each part
o Record any observations during measurement
Conduct the gage R&R and take data.
35
SCMH Section 3.11.4
Type 1 Gage Study for Concentricity

Reported by: Jay M. Vospodopoulos
Gage name: CMM #1 Tolerance: 0.01 3
Date of study: 1 0/1 7/201 7 Misc:
Run Chart of Concentricity

Ref + 0.10 × Tol
0.0054
0.0048
Concentricity
0.0042 Ref
0.0036
0.0030 Ref - 0.10 × Tol
1 4 7 10 13 16 19 22 25 28
Observation
Basic Statistics Bias Capability

Reference 0.0042 Bias 0.000034 Cg 2.52
Mean 0.004234 T 1 .0978021 Cgk 2.45
StDev 0.0001 71 9 PValue 0.281
6 × StDev (SV) 0.001 031 3 (Test Bias = 0)
%Var(Repeatability) 7.93%
Tolerance (Tol) 0.01 3
%Var(Repeatability and Bias) 8.1 5%
In conclusion, the percent variation coming from the gage repeatability (type one gage study) was
determined to be 7.93% and no statistically significant bias present. Based on this low variation from
the gauge, a complete variable gage R&R was conducted.
In conclusion, the total gage R&R percent tolerance of 13.92% showed an acceptable performing
gage.
36
SCMH Section 3.11.4
Further investigation of the gage R&R ANOVA report also did not indicate any areas of concern.
Documentation:
This process is documented by the data taken from the Gage R&R and correlation study.
12
Take action as
appropriate
Take Action:
From the above data it has been determined that no significant variation exists and no action needs to
be taken.
37
SCMH Section 3.11.4
References
SCMH Section 3.11.3 Measurement Systems Analysis Process Flow
Automotive Industry Action Group (2010) Measurement System Analysis, 4th ed., Detroit, Mi.
SAE International, AS13003, 2015-02
Wheeler, D.J and Lyday, R.W., Evaluating the Measurement Process, SPC Press, Inc., Knoxville, TN,
2006
38

SCMH 3.11.4 MSA Case Studies Rev New Dated 21MAR2018

Uploaded by

Copyright:

Available Formats

SCMH 3.11.4 MSA Case Studies Rev New Dated 21MAR2018

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SCMH 3.11.4 MSA Case Studies Rev New Dated 21MAR2018

Uploaded by

Copyright:

Available Formats

SCMH Section 3.11.

Measurement System Analysis Case Studies

Following are the five case studies:

Measurement System Analysis Case Studies

Measurement System Analysis Case Studies

Measurement System Analysis Case Studies

Measurement System Analysis Case Studies

Measurement System Analysis Case Studies

Introduction to First Case Study, Mechanical MSA

Measurement System Analysis Case Studies

First Case Study

produced is a cylinder that is 1 inch tall with a 0.5 inch diameter.

Cylinder height is 1 inch tall, with a specification of ±0.1 inch.

Cylinder diameter is 0.5 inches, with a specification of ±0.005 inches.

(CMM, Caliper, Micrometer, steel ruler)

resolution of the gage needs to be at least 0.001 inch.

Measurement System Analysis Case Studies

The gage criteria includes accuracy, linearity, and stability.

times better with a tolerance of ±0.00125 inches.

measure the 0.5-inch diameter.

of the measurement instrument can be evaluated by looking at historical calibration records.

stability of the instrument.

Measurement System Analysis Case Studies

Develop and perform a gage R&R:

measurement, and having a template to record data:

techniques, and in use of the selected gage

o The operator should follow the normal measurement process

• Gage to take the measurement

o Verify calibration and proper operation

o Recommend 2-10-2 measurement. Which is 2 operators, 10 samples, 2 recordings for

o Record any observations during measurement

Conduct the gage R&R and take data.

Measurement System Analysis Case Studies

An investigation into the variation between the operators needs to be performed:

known values, etc.)

Measurement System Analysis Case Studies

operators. Some examples of an action plan may include the following:

• Better documented method

Measurement System Analysis Case Studies

Introduction to Second Case Study, Attribute MSA

Measurement System Analysis Case Studies

Second Case Study

Maximum Scratch Length 2mm

A Pass/fail condition is determined through a visual inspection compared to a standard.

Measurement System Analysis Case Studies

Develop and perform an attribute agreement analysis:

true effectiveness of the measurement system.

• Recommend 30 parts with a minimum of 25.

2 operators. The operators must be trained to the visual standard.

o The operator should follow the normal process.

• Calibration/ known population.

o Recommend 2-30-2 measurement. Which is 2 operators, 30 samples, 2 recordings for

o Record any observations during the measurement.

Measurement System Analysis Case Studies

Attribute Gage R & R Effectiveness

2 Fail GAGE: Visual Inspection agree within and All Operators

BUSINESS: ACME Cylinder between each agree with

Typical Attribute Agreement Study Spreadsheet

samples that caused inconsistencies vs the “known” attribute.