As everyone else I am having hard time to implement most of the requirements of the MSA, especifically on Gage R&R for attributes using the analytical Method, since my company is in the automotive division, this is the one recommended to be used by the MSA.

The problem I have is that I need to know the criteria to accept this kind of studies, in a variable Gage R&R we have that the %GRR should be less then 10% and if it's less than 30% might require improvement, and will be rejected if is above 30%.

But for the analytical method on page 136 it gives you only criteria, that the "t" value should be less than 2.093 to know if the Bias is significantly different from zero, but at the end does not say what to do with the repeatability, for this you have the formula with a value, how to judge when the repeatability is small or what criteria we should use to considered it high or low or when to rejected the study?

Thanks in advance for your help?

Regards....

PS Very nice page I found it by Luck and I think it will help me to improve my MSA and I will be able to help others.

I too have tried to search a lot for information but it is not really well documented anywhere. Unfortunately, I have not been able to find any definite rule for acceptance criteria. For attribute measurement systems, repeatability can only be compared with tolerance to find a %. IMHO, the interpretation of repeatability for attribute measurements ahould be the same as that for variable. The error would be critical when the parts are near the tolerance limits.

See Chapter V, Section C of the MSA manual, that explains the Gage Performance Curve. Given the bias and GRR, you can calculate the probability of accepting a part with a certain value.

Also see: MSA_J.PDF (p.97-100) (http://elsmar.com/pdf_files/MSA_J.pdf)

Thanks for the reply, really appreciated.

I continued to find specific description of the criteria for the repeatability value, even with advisor that are providing training at a training Site in DEtroit and they don't even know it.

I am using Minitab to calculate repeatability and the values that brings out are 0.0027 and most of the tests I ran are around this value.

Do you know if this value calculated by Minitab needs to be converted to percentage?

I checked Minitab Help, and it only says the this values should be low.

Thanks again!