SOP For Measurement System Analysis

1) Scope
Scope of MSA of measuring equipment is from furnace to customer dispatch.

2) Purpose
1.To assure manufacturer, suppliers and customers that selected measurement system is capable of delivering reliable results with repeatability and reproducibility.
2.Purpose of the Measurement system analysis is to reduce the manufacturing process variation by reducing measurement variation and improve the collected data
quality.
3.MSA help us to verify the quality of collected data which will use in quality studies with ultimate aim to improve manufacturing process and reduce waste.
4.To identify measurement system errors.

3) Flow chart for Measurement System Analysis

Step 3
Step 1 Step 2
Evaluate stability of
Prepare for MSA Prepare a plan for MSA STUDY
measurement system

Step 6 Step 5 Step 4

Evaluate Accuracy of measurement Evaluate Resolution of Evaluate Resolution of
system measurement system measurement system

Step 7 Step 8 Step 9

Evaluate Linearity of measurement Evaluate Linearity of Determine repeatability and
system measurement system reproducibility

PREPARED BY REVIEWED BY APPROVED BY

 4)Detail flowchart Selection of Measurement System Analysis

Determine the
objective of
MSA study

Evaluate Stability Evaluate Resolution

Correct the YES Is Gauge Get the new

NO Is Gauge
Stability resolution gauge or
Stable?
for the gauge adequate? improve
resolution

Find out the accuracy

of gauge

YES

Evaluate Linearity

If data is attribute If data is continuous

Prepare attribute Prepare for R and R

Get the new gauge or prove resolution
agreement study study

NO Measurement
IS R&R system needs
okay? improvement

PREPARED BY REVIEWED BY APPROVED BY

 Yes

Write Gauge R&R

report

IF the Results unacceptable then improve, else prepare the report

PREPARED BY REVIEWED BY APPROVED BY

 5) MSA Process
Step 1: Preparation of MSA study
1) Select the measurement system for which we need to perform MSA.
2) Verify the type data obtained from measurement system to select the MSA method.
3) Determine the number of operator, number of sample parts, number of repeated readings before starting the study
5) Now the green zone covers the 50% tolerance and 2 yellow zone covers 25% tolerance each.

Step 2: Evaluate the stability

1) Measure the sample and collect data 3 to 5 time within prescribe time frame. Data collection time frequency should be as large as
possible but greater than 24 hours required.
2) Plot the data on control chart for X bar and R in there respective charts.
3) Examine the chart and find out uncontrolled condition.
Control Charts Analysis:
I. You had to examine the control chart and verify that the measurement system is out of control. For the instance if the control chart is not in
control then you had to examine the control chart and find out the root cause, solve the issue and take the readings again.
a. If x bar in unstable/ out of control means there is shift reading. this maint happen due to environment, gauge change, fixture issue.
b. If R chart I unstable/ out of control means variability is present within subgroup.
ii. Before starting any new measurement, follow this procedure. If there is any out of control scenario then correct it before start actual measurements.

Step 3: Evaluate Resolution

1) Measure the sample and collect data 3 to 5 time within prescribe time frame. Data collection time frequency should be as large as
2) Plot the data on control chart for R in range chart.
3) Examine the chart and find out uncontrolled condition.
Control Charts Analysis for resolution:
a. If there are only one, two, or three possible values for the range then give instrument resolution is inadequate.
b. There are only four possible values for the range when the subgroup is greater than or equal to 3.

Step 4: Determine the Accuracy

1) Measure the sample 15-25 within short period of time while taking reading
2) Compute the average of the readings, calculate the bias as per formulae: Average - Reference value
3) Calculate the upper and lower confidence limit:
UCL = X Bar + t*s/(Square root of n)
LCL = X Bar - t*s/(Square root of n)

Take the value of t from the given chart.

Control Charts Analysis for Accuracy:

a. If the reference value is in between lower and upper 95% confidence limits, then the bias is not statistically significant and we can ignore this bias.
a. If the reference value is lower than 95% confidence limits and higher than upper 95% confidence limits, then the bias is statistically significant and the measurement
system must be recalibrated.

Step 5: Calibrate the gauge

PREPARED BY REVIEWED BY APPROVED BY

 1) Calibrate the measurement gauge as per prescribe method of calibration. Calibration should be done from independent lab.

Step 6: Evaluate Linearity

1) Take the production part, take 15 readings and average it to obtain reference values.
2) Measure the sample standard 15 to 25 times in as short a period of time as possible using the same operator, same equipment, and same set-up to obtain the data.

3) Compute the average of the readings. Compute the bias for each of the samples.
4) Prepare an X, Y graph for reference value vs bias. On that graph use best fit regression. Compute the slope of best fit line
5) Compute linearity and linearity % :
Linearity = |Slope|* Process variation
Linearity % = ( |Slope|/ Process variation ) * 100
6) Compute the correlation coefficient
as per formulae

PREPARED BY REVIEWED BY APPROVED BY

 Analysis for Linearity
a. The slope of the best fit linear regression line gives an indication of how much the measurement instrument is inaccurate across the entire process window for the
characteristic at hand. The closer to zero the slope, the “better” the instrument.

a. If the reference value is lower than 95% confidence limits and higher than upper 95% confidence limits, then the bias is statistically significant and the measurement

Step 7: Determine the repeatability and reproducibility

1) Completely randomize the order of sample and operators sequence to collect measurement data. Repeat this process at the beginning of each run to avoid any
operator bias
2) Ask the operator to measure given reading as per randomize order and ask engineer to record all readings. While taking the readings, please make sure operators do
not know each others reading to avoid knowledge bias.
3) Then put the data in MSA master format excel or Minitab to evaluate gauge R&R consisting all formulas. Refer format no. F:CQ:96.

6) Rule to follow while performing MSA:

1) The data should be collected randomly. Purpose behind this is that drift or any changes occurs due to random factors will be spread throughout the population.

2) The operator should be unaware about the number of part he is checking to avoid any measurement bias. However the person who conducting the MSA study should
be aware of which no. of part is getting check so he can collect the data accurately.
3) Reading should be estimated to nearest number possible.
4) Person conducting MSA study should be train enough to he understand the importance of taking the care require while conducting MSA. Training should be given to
operators as well.

7) Criteria for Gauge R&R value:

1)If the % Gage R&R is under 10%, the measurement system is generally considered to be an adequate measurement system.
2) If the % Gage R&R is between 10 % to 30%, the measurement system may be acceptable for some applications.
3) If the % Gage R&R is over 30%, the measurement system is considered to be unacceptable.

8) Responsibility
Respective process quality engineer are fully responsible for effective working of the system.

7) Safety
Safety Shoes, Apron and Hand gloves are compulsory while doing process monitoring with pre control chart

PREPARED BY REVIEWED BY APPROVED BY