Variable study
• INRODUCTION
• METHODS
• PROCEDURE TO BE FOLLOWED
• WORKED OUT EXAMPLE
• DISCUSSIONS
3rd Edition - 2002
Variable study
Define the objective of the study
Define the components of the study
Make the Measurements
Calculate the Gage Repeatability and Reproducibility
Take actions to improve the measurement process
3rd Edition - 2002
Variable study
Define the Objective of the Study
Make the Purpose of the study very clear.
The purpose is to evaluate the total measurement
system.
3rd Edition - 2002
Variable study
Define all the components of the study
The various components of the study are
People (Appraiser)
Methods (Test Method)
Environment (Test Condition)
Material (Test Specimen)
Gage (Equipment)
3rd Edition - 2002
Variable study
People (Appraiser)
1. The No. of appraisers to be decided (2 or 3)
2. Select people who routinely use the gage.
3rd Edition - 2002
Variable study
METHODS
Procedure to be followed is to be clearly defined.
Define how measurements should be made ?
How will the study be conducted ?
When it will be conducted ?
How will the data be recorded ?
Who will record the data ?
3rd Edition - 2002
Variable study
ENVIRONMENT (TEST CONIDITON )
Use the same environmental conditions in which the
gage is used. (Eg.. If gage is used in lab, do gage
repeatability & reproducibility study in lab only)
3rd Edition - 2002
Variable study
MATERIALS
• What will be used for the study ?
Min. 5 parts and this should be from actual
production parts. Typically studies are taken for
ten parts.
• Identify the parts.
Make sure the numbers are readable.
• Mark the parts exactly where you want to measure.
(so that part variation can be avoided)
3rd Edition - 2002
Variable study
MATERIALS
The sample must be selected from the process and should
represent the entire sample production operating range.
3rd Edition - 2002
Variable study
GAGE
• Ensure that the gage is calibrated
• Gage that is selected should have least count, that is one
tenth of tolerance or process variation.
3rd Edition - 2002
Variable study
MAKE THE MEASUREMENTS
• The Measurements should be made in random order,
to ensure that any drift or changes that occur will
spread randomly through out the study.
• Randomly select a part and measure it.
• Ask one appraiser to check all the parts and record.
• The appraisers should be unaware of which numbered
part is being checked in order to avoid any knowledge
bias.
3rd Edition - 2002
Variable study
• In reading the equipment, measurement values should
be recorded to the practical limit of the instrument
discrimination.
• Once all the parts have been measured, mix them up
and measure them again.
• Each appraiser should use the same procedure -
including all the steps they normally use to obtain
readings. (The diff. Methods used by diff. appraiser
will affect the gage R & R).
• Repeat the readings based on the number of trials
specified for the study.
3rd Edition - 2002
Variable study
• The No. of appraisers, No. of parts and No. of repeat
(Trials) readings should be determined in advance.
• Critical dimensions require more parts and / or trials.
• Bulky or heavy parts may dictate fewer samples and
more trials.
• Have each appraiser take all measurements before
moving on the next trial or appraiser.
3rd Edition - 2002
Variable study
REPEATABILITY
Repeatability is the variation in measurements obtained
with one measurement instrument when used several
times by one appraiser while measuring the identical
characteristic on the same part.
Inherent variation or Capability of the equipment
Referred to as within appraiser variability.
3rd Edition - 2002
Variable study
REPEATABILITY
Repeatability is within - system variation when the
conditions of measurements are fixed and defined - Fixed
part, instrument, standard, method, operator, environment.
Is the Common Cause (random error) variation from the
successive trials under defined conditions of
measurements.
Commonly Referred to as Equipment Variation.
3rd Edition - 2002
Variable study
REPRODUCIBILITY
Reproducibility is defined as the variation in the average
of the measurements made by different appraisers using
the same measuring instrument when measuring the
identical characteristics on the same part.
Commonly referred to as “between appraisers”
variability.
3rd Edition - 2002
Variable study
Gage R & R or GRR
Gage R & R is an estimate of combined variation of
repeatability and reproducibility.
GRR is the variance equal to the sum of within-
system and between-system variance.
2GRR = 2repeatability + 2reproducibility
3rd Edition - 2002
Variable study
ANALYSIS OF THE RESULTS
The results are to be evaluated to determine if the
measurement device is acceptable for its intended
application.
A measurement system is to be stable before any
additional analysis is valid.
3rd Edition - 2002
Variable study
ACCEPTABILITY CRITERIA
WIDTH ERROR
Criteria as to whether a measurement systems
variability is satisfactory or not is dependent upon the
percentage of the manufacturing production process
variability or part tolerance that is consumed by
measurement system variation.
3rd Edition - 2002
Variable study
ACCEPTABILITY CRITERIA
Under 10 % error - Generally considered to be acceptable
for measurement system.
10 - 30 % error - May be acceptable based upon
importance of application, cost of
measurement device, cost of repair, etc.
Over 30 % - Considered to be not acceptable - Every
effort should be made to improve the
measurement system.
The number of distinct categories (ndc) the process can be
divided into by the measurement system ought to be greater
than or equal to 5.
3rd Edition - 2002
Variable study
The final acceptance of a measurement system should
not come down to a single set of indices. The long term
performance of the measurement system should also be
reviewed using graphical analyses.
3rd Edition - 2002
Variable study
CONTROL
Can be used for control only if :
The process variation is small when compared
to the specifications
The main source of variation causes a mean
1 Data Category
shift.
ANALYSIS
This is Unacceptable for estimating process
parameters and indices.
Only indicates whether the process is
producing conforming or non conforming parts.
Fig: Impact of Number of Distinct Categories of the process distribution on
Control and analysis Activities.
3rd Edition - 2002
Variable study
CONTROL
Can be used with semi variable
control techniques based on the
process distribution.
ANALYSIS
Generally unacceptable for
estimating process parameters
and indices since it only
provides coarse estimates.
Fig: Impact of Number of Distinct Categories of the process distribution on
Control and analysis Activities.
3rd Edition - 2002
Variable study
CONTROL
Can be used with variable control
charts.
5 Or more Data Categories ANALYSIS
Recommended.
Fig: Impact of Number of Distinct Categories of the process distribution on
Control and analysis Activities.
3rd Edition - 2002
Variable study
The variable gage study can be performed using a number
of different methods.
RANGE METHOD
AVERAGE AND RANGE METHOD
ANOVA METHOD.
3rd Edition - 2002
Variable study
RANGE METHOD
• Quick approximation of measurement variability.
• Provides only overall picture of measurement system.
• Does not decompose the variability into repeatability and
reproducibility.
• Used as a quick check to verify that the GRR has not
changed.
This approach has the potential to detect an unacceptable
measurement system 80 % of the time with a sample size
of 5 and 90 % of the time with a sample size of 10.
3rd Edition - 2002
Variable study
RANGE METHOD - Methodology
Select two appraiser and five parts.
Each appraiser measure each part once.
Range for each part is to be calculated. (It is the
absolute difference between the measurement obtained by
appraiser 1 and measurement obtained by appraiser 2).
Average range R is to be calculated.
Total measurement variability is to be found by
multiplying the average range by 1 / d2*
d2* - Found in table for m ( No of appraiser = 2) and No. of
parts (g )
3rd Edition - 2002
Variable study
RANGE METHOD - Methodology
GRR = R / d2*
GRR
% GRR = 100 X
Process Standard deviation
3rd Edition - 2002
Variable study
AVERAGE AND RANGE METHOD
This method will provide an estimate of both repeatability
and reproducibility for a measurement system.
It will not tell their interaction.
3rd Edition - 2002
Variable study
AVERAGE AND RANGE METHOD – Methodology.
1. Obtain a sample of n > 5 parts that represent the
actual or expected range of process variation.
2. Select the appraisers and number the part so that the
numbers are not visible to the appraisers.
3. Ensure that the gage is calibrated.
4. Let appraiser 1, measure all the parts in random manner
and enter the readings of this first trial of appraiser 1
in the corresponding row of the table.
3rd Edition - 2002
Variable study
AVERAGE AND RANGE METHOD – Methodology.
5. Repeat the same for appraiser 2 and 3 and record the
reading under trial 1 in the corresponding row of the
table.
6. Repeat the steps 4 and 5 for different order of
measurement and record the readings obtained from
trial 2 and 3 for appraiser 1, 2 and 3 in the corresponding
column.
7. Ensure that readings of the part are recorded under the
corresponding part which is numbered.
3rd Edition - 2002
Variable study
AVERAGE AND RANGE METHOD – Methodology.
8. When all the parts are not available, then each appraiser
can check the same part and relevant reading to be recorded
in the corresponding column. E.g Let appraiser 1 measure the
first part and record the reading under row and appraiser 2
measure the same part and record the reading under
corresponding row etc.
9. If appraisers are of different shifts, then appraiser 1 can
measure all the parts and reading recorded under trial 1, and
then appraiser1 check all the parts in different order and the
readings to be recorded under trial 2 and in different order
for trial 3. Do the same with appraiser 2 and 3.
3rd Edition - 2002
Variable study
Appraiser PART
AVERAGE
/ Trial 1 2 3 4 5 6 7 8 9 10
A 1
2
3
Average avg Xa =
Range avg Ra =
B 1
2
3
Average avg Xb =
Range avg Rb =
C 1
2
3
Average avg Xc =
Range avg Rc =
Part
Xp =
Average
(Xp) Rp =
Average of Range => ]
= [ avgR a +avg R b +avg R c / [Link] A ppraiser R =
Difference of Average = [ Max. X - Min. X ] => XDIFF =
[ R X D4 ]
Range - Upper Control =Limit => UCLR =
[ R X D3 ]
Range - Lower Control =Limit => LCLR =
3rd Edition - 2002
Variable study
1. GAUGE REPEATABILITY AND REPRODUCIBILITY DATA SHEET
Appraiser PART
AVERAGE
/ Trial 1 2 3 4 5 6 7 8 9 10
1A 1
2 2
3 3
4 Average avg X a =
5 Range avg Ra =
6B 1
7 2
8 3
9 Average Xb =
10 Range Rb =
11 C 1
12 2
13 3
14 Average Xc =
15 Range Rc =
Part
16 Xp =
Average
(Xp) Rp =
17 Average of Range = [ Ra + Rb + Rc / [Link] Appraiser ] => R =
18 Difference of Average = [ Max. X - Min. X ] => XDIFF =
19 Range - Upper Control = [ R X D4 ]
Limit => UCLR =
[ R X D3 ]
20 Range - Lower Control =Limit => LCLR =
3rd Edition - 2002
Variable study
NUMERICAL CALCULATIONS
1. Subtract the smallest reading from the largest reading in
rows 1, 2 and 3; enter the result in row 5. Do the same for
rows 6,7 and 8; and enter the results in rows 10 and 15
respectively.
2. Entries in rows 5, 10 and 15 are ranges and therefore
always positive values.
3. Total row 5 and divide the total by the number of parts
sampled to obtain the average range for the first
appraisers trials Ra. Do the same for rows 10 and 15 to
obtain Rb and Rc.
3rd Edition - 2002
Variable study
NUMERICAL CALCULATIONS
4. Transfer the averages of rows 5, 10 and 15 ( Ra, Rb, Rc) to
row 17. Add them together and divide by the number of
appraisers and enter results R ( average of all ranges).
5. Enter R (average value) in rows 19 and 20 which is
obtained from row 17 and multiply by D4 to get the
lower and upper control limits. D4 = 3.27 for 2 trials
and D4 = 2.58 for 3 trials.
6. The value of upper control limit ( UCL R) of the
individual ranges is entered in row 19 and the value of
(LCLR) is entered in row 20.
3rd Edition - 2002
Variable study
NUMERICAL CALCULATIONS
7. Repeat any readings that produced a range greater than
the calculated ( UCL R) using the same appraiser and
part as originally used or discard those values and re
average and re-compute R and the limiting value ( UCL R)
based on the revised sample size. Correct the special
cause that produced the out of control condition. If the
data were plotted and analyzed using a control chart, this
condition would have corrected and would not occur
here.
8. Sum the rows 1, 2, 3, 6, 7, 8, 11, 12, 13. Divide the sum in
each row by the no. of parts sampled and enter the values
under average column in each row.
3rd Edition - 2002
Variable study
NUMERICAL CALCULATIONS
9. Add the averages in rows 1, 2 and 3 and divide the total
by the number of trials and enter the value in row 4 in the
Xa block. Repeat this for rows 6, 7 and 8; and 11, 12, 13, and
enter the results in the blocks for Xb and Xc in rows 9
and 14 respectively.
10. Enter the max and min averages of rows 4, 9 and 14 in the
appropriate space in row 18 and determine the differences.
Enter the differences in the space labeled X diff in row 18
11. Sum the measurements for each trial for each part and
divide the total by the number of measurements (number
of trial times the number of appraisers) and enter the
results in row 16 (Indicated as part averages)
3rd Edition - 2002
Variable study
NUMERICAL CALCULATIONS
12. Subtract the smallest part average from the largest
part average and enter the result in the space labeled
Rp in row 16. Rp is the range of part averages.
13. Transfer the calculated values of R, X diff and Rp to the
blanks provided on the report side of the from.
14. Perform the calculations under the column entitled
“Measurement Unit analysis” on the left side of the form.
15. Perform the calculations on the right side of the form
entitled % Total variation.
16. Check the results to make sure no errors have been
made.
3rd Edition - 2002
Variable study
NUMERICAL CALCULATIONS
Number of distinct categories ( ndc)
= 1.41 PV
GRR
PV - PART VARIATION
GRR - GAGE REPEATABILITY AND REPRODUCIBITLITY
This is the number of non – overlapping 97 %
confidence interval that will span the expected
process variation.
ndc ought to be greater than or equal to 5.
3rd Edition - 2002
Variable study
2. GAUGE REPEATABILITY AND REPRODUCIBILITY REPORT
(Based on Total Variation)
Measurement Unit Analysis % Total Variation (TV)
Repeatability = Equipment Variation (EV) % Repeatability
EV = R X K1 % EV = 100 [ EV / TV ]
Trails K1
= 2 0.8862
= %
3 0.5908
Reproducibility = Appraiser Variation (AV) % Reproducibility
AV = 2 2 % AV = 100 [ AV / TV ]
[(XDIFF X K 2) -(EV /nr)]
= = = %
n = Number of Parts Appraisers 2 3
r = Number of Trials K2 0.7071 0.5231
Repeatability & Reproducibility (R & R) % Repeatability & Reproducibility
R&R = 2 2 %R&R = 100 [ R&R / TV ]
(EV + AV )
= = %
Part Variation (PV) Parts K3 % Part Variation
2 0.7071
PV = Rp X K3 3 0.5231
% PV = 100 [ PV / TV ]
4 0.4467
= 5 0.4030
= %
6 0.3742
Total Variation (TV) 7 0.3534 Number of distinct Data categories (d)
8 0.3375
TV = GRR2 + PV 2 9 0.3249
d = 1.41 [ PV / R&R ]
10 0.3146
= Data Categories
3rd Edition - 2002
Variable study
CORRECTIVE ACTION TO BE TAKEN
HIGH REPEATABILITY E.g:
• Improve equipment maintenance.
• Use a more rigid gage or gage fixture.
• Use a higher discrimination gage.
• Reduce the with in part variation.
• Improve the lighting.
3rd Edition - 2002
Variable study
HIGH REPRODUCIBILITY E.g
• Review the data and its calculations.
• Identify the methods used by different appraisers.
• Retrain the appraisers.
• Change the gage.
• Build a fixture for measuring the part.
3rd Edition - 2002