Engineered Software

Designing an R&R Study


Home
Reliability Problems
Reliability Software
Gage Capability Problems
Gage Capability Software
Maintenance Optimization
EVOP Software
Six Sigma
On-Line Store
Books
Contacts

Below is a step by step example of how to design a repeatability and reproducibility study.  The example utilizes the software package Measurement AssuranceClick here to download a free demo version of this software.

Click here for the theory of designing repeatability and reproducibility studies.


To design an R&R study select Procedures then Design R&R Study from the menu, and the following screen will appear.

Next specify

  • the number of parts,
  • the number of appraisers, and
  • the number of trials.

Click the Design R&R Study button, and the data entry sheet will contain the study in a completely randomized order.  This is shown below.

The study should be performed in the order specified in the spreadsheet shown in the figure above (This spreadsheet can be printed or copied into an Excel spreadsheet).

  1. Appraiser 2 should measure part 9, then
  2. Appraiser 1 should measure part 2, then
  3. Appraiser 1 should measure part 7, then
  4. Appraiser 2 should measure part 4, then
  5. Appraiser 3 should measure part 2, and so on.

Back to the Top of the Page


Theory of Designing R&R Studies

When performing a measurement system capability study, it is generally recommended to quantify as many sources of error as possible. Because of this, the default procedure is an R&R study using the analysis of variance method. Another rule of thumb is to use 10 parts, 3 appraisers and 2 trials. The number of parts, number of appraisers and number of trials to be used should be based on financial considerations and the ability of the study to be representative of the entire measurement process. If 3 shifts are used in a production facility, the measurement assurance study should include operators from all 3 shifts. If repeating a measurement is inexpensive, or if the characteristic being measured is critical, the number of trials should be increased appropriately.

All measurement assurance indices are invalid if the measurement assurance study is not conducted in random order. For example, if appraiser 1 measures part 1 twice, then appraiser 1 measures part 2 twice, then appraiser 1 measures part 3 twice, this study has not been randomized. Consider a simple study consisting of 3 parts, 2 appraisers and 2 trials. Table 1 shows this study in a systematic format. Table 2 shows this study in a randomized order.

Table 1. Non-randomized measurement assurance study.

Part

Appraiser

Trial

1

1

1

1

1

2

1

2

1

1

2

2

2

1

1

2

1

2

2

2

1

2

2

2

3

1

1

3

1

2

3

2

1

3

2

2

 

Table 4.2.
Randomized measurement assurance study.

Part

Appraiser

Trial

2

2

1

1

2

1

1

2

2

1

1

1

2

1

2

3

1

2

1

1

2

3

2

2

3

2

1

2

1

1

3

1

1

2

2

2

In some cases it may not be possible to completely randomize a study, but as much randomization as possible should be used to ensure the validity of the results.


Engineered Software, Inc.
734-502-8582
Information@EngineeredSoftware.com