I am soliciting advice on how to qualify a CMM inspection program. I need to establish that the program will provide repeatable results from multiple measurements of a single part. I am not talking about measuring gage blocks or masters. I certainly don't want to get into performing a full blown gage R&R study.

What repeatability is typically required? What percentage of the manufacturing tolerance can be used consumed in the inspection process?

For example: Two dowel holes with a +/-.001 center distance tolerance may measure +.00001, +.00032, & -.00021 (from nominal) when measured 3 times on the same CMM. The range of the measurements is .00053, or 26%, of the tolerance. This is a problem.

Any thoughts would be appreciated.

Kerry Murphy