GR&R - Gage R&R for 1 test Stand, 10 locations, 10 parts

[ann5750]

In reading the MSA book, the ANOVA method as a general rule of thumb identifies up to 3 appraisers, 10 parts & up to 3 trials.

I've looked for information relating to GR&R of Test Stands and was unable to find any useful tips. We have a single Test stand designed to locate and automatically test 10 parts at the same time. The test stand is operated by one person who loads the parts into each of the test locations and presses a button to start the test sequence. The test stand measures output voltage.

For each of the "Studies" 3 trials are intended. My assumptions would be:-

# Trials = 3, therefore K1=3.05
# Appraisers = 1, AV=0?


Study 1 -#Test stand locations=1, #Parts=10, K3 = 1.62

Study 2 -#Test stand locations=10, #Parts=1, K3 =?

Study 3 - #Test stand locations=10, #Parts=10, K3 = 1.62

Is Study 1 & 2 sufficient and what would the testing of only 1 part in all 10 locations tell me? and what value would I assign for K3?

Should I modify Study 2 to include 3 Test locations for 10 parts and use Appraisers = 3 (assume each of the 3 selected Test Stand locations is an "Appraiser"?

Or should I just go ahead and complete study 3 first of all. Completing Study 3 would take care of completing Study 1 & 2 - the data could be just extracted as needed but it would also generate a lot of data that could not easily be entered into the standard ANOVA format.

I'd like to understand why I may be making this more complicated than is probably necessary.

 

[ann5750]

Revise selection of Parts & Appraisers

I have re-considered how I should approach this GR&R study and would appreciate feedback. Initially, I thought that I could use fewer parts to assess the Test Stand, which has 10 independent test locations.

To overcome the difficulty of assigning K values for a single part or using AV=0 for a single appraiser (the person loading the components into the Test Stand), I will be using 2 people to conduct 3 trials each, using all 10 parts.

Test Sequence:-
10 parts will be selected and placed into the Test Stand in positions 1 through 10. After Test 1, each of the 10 parts will be placed into the adjacent location and the Test repeated until 10 test cycles have been completed. This ensures that all 10 parts have been tested in all 10 positions.

This will result in basically 10 GR&R reports, 1 for each of the test locations.

It will be interesting to see if any one of the test locations performs better than the others.

I had considered using 1 of the 10 parts in a specific test location throughout eg Use Part #1 in Test Loaction #1 for all 3 Trials by both Appraisers and randomize the other 9 parts with the remaining test locations. I can see some logic to doing this but haven't fully figured out what benefits this may yield.