Identifier: CX-CP-1.

14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

PURPOSE
Provide instructions and guideline to facilitate the application of statistical techniques
including Gage R&R in ChampionX Chemical Technologies business units and
manufacturing facilities.

SCOPE
This procedure applies to all ChampionX Chemical Technologies business units and
manufacturing facilities that use such statistical techniques.

RELATED DOCUMENTS
CX-CP-2.14 Analysis Plan Authorization Procedure
CX-CP-1.13 Capability Study and Statistical Process Control

ACRONYMS, ABBREVIATIONS AND DEFINITIONS

Gage R&R – Gage Repeatability and Reproducibility

TRASAR – A trade name for select treatment chemicals which use fluorescent tracer

ROLES AND RESPONSIBILITIES

The leadership team in each business unit or manufacturing facility is responsible for
identifying the needs for statistical techniques and provide such trainings to employees
when necessary.

PROCEDURE
Statistical techniques are recommended to be used at all ChampionX business units and
manufacturing locations. In CX-CP-1.13, statistical techniques include Statistical Process Control
(SPC) or Statistical Quality Control (SQC) charts, evaluation of capability data for product analysis
plans, and evaluation of process and method potentials were discussed. This procedure will focus
on the GR&R study that is required for TRASAR products to confirm the analytical capability of the
QC instrumentations.

Each manufacturing facility is required to provide Statistical training as deemed appropriate by

plant management. It is recommended that Operators and laboratory analysts receive basic
Statistical training and Professionals receive advanced Statistical training.

GAGE REPEARTABILITY AND REPRODUCIBILITY

1.GR&R can be used to identify measurement system variability due to a combination of

repeatability and reproducibility.

____________________________________________________________________________
1
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

2.GR&R study is recommended for all ChampionX analytical, measurement and charging
systems.

• Procedures for completing the long and short method for GR&R’s are detailed in Appendix
II. Alternate procedures, such as computer software systems, may be used provided they
are statistically valid.

3. GR&R study is required for the analytical, measurement and charging system used for
TRASAR products.

• Specific instructions on TRASAR analytical, measurement and charging system GR&R

studies is shown in Appendix II.

• The %GR&R specification shall be less than 10%.

APPENDIX I

GAGE REPRODUCIBILTY AND REPEATABILITY

1.0 GAGE STUDY FOR REPEATABILITY AND REPRODUCIBILITY (SHORT

METHOD)

Purpose: To identify measurement system variability due to a combination of

repeatability and reproducibility.

Limitations: Gage repeatability and reproducibility cannot be isolated. The results

reflect a combination of both types of error. This study provides a
quick and easy way to determine the acceptability of gage variation
over a short period of time.

Number of Samples Required (n): 5 - 10 **

Number of Operators: 2 **

Data Collection Procedure:

1. Randomly select 5-10 samples from a given process. (In the following
example, 5 samples are used.)

____________________________________________________________________________
2
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

2. Number each sample.

3. Select an instrument for study. Check the calibration if appropriate.

4. Have each operator measure a randomly selected sample until all 5

samples are measured. (In the following example, 2 operators are used.)

5. Record the data as shown in FIGURE 1.1.

6. Re-check the calibration of the instrument if appropriate.

Data Evaluation Procedure:

1. For each sample, calculate the range among the measurements of the
operators.
NOTE: Positive (or absolute) differences are entered in the column
labeled Range.

2. Calculate the average range, which is the sum of all ranges divided by the
number of samples.
_
Average Range (R) = R/n
_
Example: R = 150/5 = 30

FIGURE 1.1

Variable Gage Study (Short Method)

Gage K F Water Date January 1, 1996

Characteristic H2O (ppm) Specs 0 - 1000

Product 9xx123 Formula 1234

Samples Operator A Operator B Range (A-B)

1 490 430 60
2 490 480 10
3 480 440 40
4 470 440 30
5 450 460 10
____________________________________________________________________________
3
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

Sum of Ranges: 150

3. Calculate Gage Repeatability and Reproducibility error (GRR) as follows:

_
GRR = 6 R/d2*, where d2* is determined from Table 1.1.

The value 6 in the above formula is the number of standard deviations that
encompass 99.73% of the data (or area under the normal curve).

EXAMPLE: d2* = 1.19 which is the intersection of the row representing 5

samples and the column representing 2 operators.
Therefore,
_
GRR = 6 R/d2* = 6(30)/1.19 = 151.3.

GRR = 151.3 represents the combined gage error.

4. To convert GRR to a percent of specification range, divide by the
specification range and multiply by 100%.
GRR x 100%
GRR as a % specification range = -----------------
(USL - LSL)

151.3 x 100%
EXAMPLE: GRR as a % specification range = ------------------ - 15.13%
1000
_
Note: For a unilateral (one-sided) specification, 2 |x - SL| can be used for
the specification range, where SL stands for upper or lower
specification limit.

Conclusion: Based on its GRR as % of specification range, this

measuring system is concluded to be acceptable based on
criteria presented in Section 3.0.

** Recommended number of samples for different number of operators is given

below:

Number of Operators 2 3 4 5
Recommended Number
of Samples 5-10 4-7 3-5 3-5
____________________________________________________________________________
4
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

TABLE 1.1

d2 Values for Converting the Average Range into a Standard Deviation

Number Number of Operators

Of
Samples 2 3 4 5
1 1.41 1.91 2.24 2.48
2 1.28 1.81 2.14 2.40
3 1.23 1.77 2.12 2.38
4 1.21 1.75 2.11 2.37
5 1.19 1.74 2.10 2.36
6 1.18 1.73 2.09 2.35
7 1.17 1.73 2.09 2.35
8 1.17 1.72 2.08 2.35
9 1.16 1.72 2.08 2.34
10 1.16 1.72 2.08 2.34

____________________________________________________________________________
5
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

2.0 GAGE STUDY FOR REPEATABILITY AND REPRODUCIBILITY (LONG

METHOD)

Purpose: To determine variations due to gage repeatability and reproducibility

(R&R) separately. Potential causes of gage error can also be identified.
This study is often performed after the Short Method has indicated that
problems in the measuring system may exist.

Number of Samples Required (n): 5-25

Number of Repeated Measures Per Sample: 2-5

Number of Operators: 2-5

Data Collection Procedure:

1. Randomly select 5-25 samples.

2. Number the samples so that the numbers are not visible to the operators.
Refer to the operators as A, B, C and so on. In the following example 10
samples and 3 operators are used.

3. Select an instrument for study. Check the calibration if appropriate.

4. Let Operator A measure 10 samples in a random order and enter the

results. See FIGURE 2.1.

5. Repeat Step 4 with each of the operators.

6. Repeat the cycle (Steps 4 and 5), with the 10 samples measured in
another random order, for the number of trials required.

7. Steps 4 through 6 may be modified for limited sample quantities or when

operators are on different shifts.

8. Re-check the calibration if appropriate.

____________________________________________________________________________
6
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

Data Evaluation Procedure:

Compare the operators’ readings for any obvious errors or trends (i.e. one
operator’s readings are consistently higher than the other operators). If obvious
errors or trends are present, investigate potential causes for these errors or trends
before continuing with this study. It may be necessary to re-measure a sample or
to completely omit a sample (or operator) from this study.

1. Calculate the range (positive difference, in case of two repeated

measures) for each set of repeated measures.
_ _ _
2. Calculate RA, RB, RC, which denotes the average ranges of operators A, B
and C respectively.
_ _ _
EXAMPLE: RA = 10/10 = 1.0 RB = 9/10 = 0.9 RC = 8/10 = 0.8
_ _ _
3. Calculate XA, XB, and XC, which denote the average of all the measures
taken by operators A, B, and C respectively.

EXAMPLE:
_ _ _
XA = 7658/20 = 382.90 XB = 7659/20 = 382.95 XC = 7638/20 = 381.90

4. Calculate the Average Range (R) of all the operators.

= _ _ _
R = (RA + RB + RC) /3
=
EXAMPLE: R = (1.0 + 0.9 + 0.8)/3 = 0.9
_
5. Calculate the range of the Averages (RX)
_ _ _
RX = Max X - Min X
_
EXAMPLE: RX = 382.95 = 381.90 = 1.05

6. Calculate the error due to repeatability (RPT)

=
sRPT = K1 x R, where K1 is obtained from TABLE 2.1; and where

A = (Number of Operators) x (Number of Samples) and

____________________________________________________________________________
7
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

B = Number of times samples were measured by each operator.

EXAMPLE: RPT = 0.885 x 0.9 = 0.7965 using A=30 and B=2.

7. Calculate the error due to Reproducibility (RPD):

_
RPD = K2 x RX, where K2 is obtained from TABLE 2.1; and where

N = Number of Operators

EXAMPLE: RPD = 0.524 x 1.05 = 0.5502 using N=3.

8. Calculate total error due to Repeatability and Reproducibility (E) is given

by

E =  2RPT + 2RPD.

EXAMPLE: E =  0.79652 + 0.55022 = 0.97

9. GRR as % specification is given by

(6E/specification range) x 100%.

EXAMPLE: GRR as % specification

= (6 x 0.97)/30) x 100%

= 19.4%

Conclusion: Based on its GRR as % of specification, this measuring system

is concluded to be acceptable based on criteria presented in
Section 3.0.
TABLE 2.1
_
Factors K1 and K2 for Converting the Average Range, R,
into a Standard Deviation X

____________________________________________________________________________
8
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

K1

A Greater
B 1 2 3 4 5 8 10 Than 10
2 0.709 0.781 0.813 0.826 0.840 0.855 0.862 0.885
3 0.524 0.552 0.565 0.571 0.575 0.581 0.581 0.592
4 0.446 0.465 0.472 0.474 0.476 0.481 0.481 0.485
5 0.403 0.417 0.420 0.422 0.424 0.426 0.427 0.429
6 0.375 0.385 0.388 0.389 0.391 0.392 0.392 0.395
7 0.353 0.361 0.364 0.365 0.366 0.368 0.368 0.370
8 0.338 0.344 0.346 0.347 0.348 0.348 0.350 0.351
9 0.325 0.331 0.332 0.333 0.334 0.334 0.336 0.337
10 0.314 0.319 0.322 0.323 0.323 0.324 0.324 0.325

K2
N 2 3 4 5 6 7 8 9 10
K2 0.709 0.524 0.446 0.403 0.375 0.353 0.338 0.325 0.314

____________________________________________________________________________
9
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

FIGURE 2.1

Gage Repeatability and Reproducibility Study

(Long Method)
Gage Viscometer Date January 1, 1990
Characteristic viscosity Spec 370 - 400
Product x1234 Formula 1234

Operator A B C
Sam 1st 2nd 1st 2nd 1st 2nd
ple Trial Trial Range Trial Trial Range Trial Trial Range
1 383 382 1 384 383 1 381 380 1
2 382 382 0 383 383 0 384 383 1
3 384 382 2 385 384 1 480 380 0
4 383 383 0 381 383 2 383 382 1
5 382 384 2 380 380 0 383 381 2
6 383 383 0 382 381 1 382 382 0
7 384 382 2 385 384 1 380 381 1
8 383 384 1 384 384 0 384 384 0
9 383 383 0 383 382 1 382 381 1
10 382 384 2 383 385 2 383 382 1
TOTAL 3829 3829 10 3830 3829 9 3822 3816 8
SUM 7658 SUM 7659 SUM 7638
AVG 382.90 1.0 AVG 382.95 0.9 AVG 381.90 0.8
= _ _ _ _
4. R = (RA + RB + RC)/3 = 0.9 7. 2RPD = (RX K2)2 = 0.3027
_ _ _ 
5. RX = Max X - Min X = 1.05 8. E =  2RPT + 2RPD = 0.97
=
6. 2RPT = (R K1)2 = 0.6344 9. GRR as % of specification range
= (6 E / Specification range) x 100%
= 19.4%

3.0 CONCLUSIONS AND INTERPRETATIONS

Results of gage studies should be carefully evaluated to determine if the
instrument is acceptable for the application under consideration. Whether or not a
measurement system is satisfactory depends largely upon the percentage of
specification range that is covered by the variation of the instrument.
____________________________________________________________________________
10
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

According to Quality Assurance for the Chemical and Process Industries, prepared
by the American Society for Quality Control, “measurement variability is often
much larger for properties important to the chemical industries (e.g.,
viscosity and purity) than it is for properties measured in the mechanical industries
(e.g., dimensions and voltages). Measurement may be responsible for 50 percent
or more of the observed variation.”

ChampionX’s criteria for acceptance of gage repeatability and reproducibility for

Product and raw Material Specifications is:

under 30% error - Acceptable.

30% to 50% error - May be acceptable.
over 50% error - Not acceptable. Identify and correct the problem.

ChampionX’s criteria for acceptance of gage repeatability and reproducibility for

Other Operational Specifications (e.g., processing parameters, meters and pumps)
is:

under 10% error - Acceptable.

10% to 30% error - May be acceptable.
over 30% error - Not acceptable. Identify and correct the problem.

NOTE: For TRASAR GR&R analytical, measurement and charging system

studies, only under 10% is acceptable.

If RPT is large in comparison to RPD, it can indicate that:

• Maintenance of the instruments may be required.

• More operator training may be needed.
• Instruments and/or methods may need to be improved or replaced.

If RPD is large in comparison to RPT, it could indicate that:

• More operator training may be needed.

• Calibrations of the instrument may need to be improved.

____________________________________________________________________________
11
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

APPENDIX II

GR&R FOR TRASAR LIQUID CHARGING SYSTEM

1. The procedure requires two drum scales and a pump similar to that used in the
actual charging of one of the TRASAR dyes. The drum scales must be able to
measure 0.5 units (lb. or kg). Ideally, the scale would be able to measure 0.1 units.

2. A drum of water is placed on the first drum scale and a series of target charges are
made to a drum on the second drum scale.

3. As in actual practice, care should be taken to assure the pump and associated
hoses are emptied into the second drum.
4. The target charge is the amount the operator is asked to charge. The recorded
charge is the figure the operator reads from the first scale. The actual charge is the
reading from the second scale.

5. The series of charges should cover the range of actual charges you would
experience in making TRASAR products at your site. For example, charges of
30-250 lbs. could be used.

6. A linear regression analysis is completed comparing the target charge to the actual
charge.

NOTE: The linear regression of target charge to actual charge will show the error in
the entire charging system (scale error, human error, and pump/hose error).

If the recorded charge is compared to actual charge, this would account for
scale error and pump/hose error. Subtracting the 2nd calculation from the 1st
would account for human error.

7. In the GR&R calculation, the sigma is obtained from the linear regression analysis.

8. The specification range used in the GR&R calculation is the absolute range of +10%
of the smallest intended charge in your actual TRASAR manufacturing. For
example, if you will produce a product containing 0.6% of a TRASAR dye with a
batch size of 20,000 lbs., the dye charge will be 120 lbs. The spec range would then
be 120 x .1 x 2 = 24.

____________________________________________________________________________
12
UNCONTROLLED COPY IF PRINTED
 Identifier: CX-CP-1.14
Document Type: Corporate Procedure
Revision: 18

Title: Gage R&R Study

9. The GR&R is calculated as follows:

% GR&R = (6 x sigma x 100)/spec range, where

sigma = the sigma from the regression analysis

spec range = +10% of the smallest expected TRASAR charge

REVISION HISTORY
Revision Date Revision Description Prepared By Approved
No By
01-15 Resides in the legacy system Rich Zwick
16 12/21/2018 New template and split CP-1.13 into two Xinjun Fan Jim Beaulieu
separate procedures
17 02/05/2020 Update company name and logo Xinjun Fan Jim Beaulieu
18 01/26/2022 Update company logo Xinjun Fan Jim Beaulieu

____________________________________________________________________________
13
UNCONTROLLED COPY IF PRINTED