Minitab Gage R Studies

[ramore1]

Hi everyone,
I am new to this forum, however have been reading tons of great stuff here. I have a question I have not been able to find the answers to in regards to Minitab. Is the software capable of calculating a gage r study with the following inputs:

Parts: 1
Operators: 2
Trials: 8

This is the study that we use extensively for sheetmetal to determine if the gage is repeatable. The only way I have been able to make this work thus far is to lie to Minitab about the parts. In other words:

Parts: 4
Operators: 2
Trials: 8

Operator Part Trial Data
Operator A 1 1 -719.972
Operator B 1 1 -719.977
Operator A 2 2 -719.980
Operator B 2 2 -719.988
Operator A 3 3 -719.973
Operator B 3 3 -719.970
Operator A 4 4 -719.967
Operator B 4 4 -719.959
Operator A 1 5 -719.969
Operator B 1 5 -719.972
Operator A 2 6 -719.971
Operator B 2 6 -719.967
Operator A 3 7 -719.967
Operator B 3 7 -719.973
Operator A 4 8 -719.980
Operator B 4 8 -719.971


The results for "% of Tolerance" are similar to how we calculate this in Excel, however what I am trying to accomplish in this instance is to evaluate the measuring device, and operators. I have 3 different measuring devices, and have performed the 1x2x8 study on all. I used the same operators, same fixture, and same part in all studies, the only difference between studies is the measurement device. Any help would be appreciated.

Thank you,
Rich

[Miner]

The short answer is No.

Minitab needs multiple parts to calculate the part variation required to calculate the %GRR. Your post seems to indicate that you only require the P/T Ratio. Minitab does not have an option for the P/T Ratio only that would allow using a single part. If you are limited to a single part, your best bet is to create your own spreadsheet, which you seem to have done.

However, what is your rationale for using a single part? If multiple parts are available, it would be better to use multiple parts. This allows you to see if there may be an operator x part interaction as well as verifying that the gage works well on more than one part.

[ramore1]

Thank you for the quick response...

I am not sure of the rational, as it involves that our people FAR above me.. However this is my guess as to why, all we are after in our case is to prove out the locating strategy of the gage, one would "assume" that even a bad part will "repeat" if the locator strategy is solid. Of course, this depends on just how "bad" the part is because the part also has physical properties that have to be contended with. We run in excess of 500 Gage R's a year in my facility and have always found that the better the part quality, the better the results, and have proven this several times. I often refer to a part that I was once asked to Gage R for a firewall blanket for the Cadillac CTS, it was literally a "blanket", the gage was a 6mm feeler with a handfull of nets. The expectation was that it would always "fall" into exactly the same place. We do our studies far ahead of production, and normally even aquiring one part is troublesome.

Again, Thank you for the quick response, and the information.

Rich