Re: Gage R&R Studies - 1200 operators, 2000 measurement tapes and many characteristic
Reynald- thank you for your answer.
I have just one issue--> I understand that, that I can random choose operators and tapes but what with characteristics--> on the one station there are on average 3 characteristics and one montage station has about 9 operations ( we have 150 montage stations:mg: ) It is impossible do GRR for each characteristic..
It is some insanity
The point of the exercise is to do a reasonable
sampling to verify that the gage is a suitable gage for the job. It is difficult to use a GR&R for verifying a gage for dissimilar geometries, such as for measuring a round surface in comparison to measuring a flat surface, but for similar flat surfaces (for example), with similar tolerances one study should prove the point. Ususally, tighter tolerances are verified for GR&R, since more open tolerances of similar geometry should be OK. Also, most automotive customers like to see a verification that the gages specifically used for critical characteristics are suitable, by having a good GR&R. Sure, there are exceptions to every rule (like do what the customer wants - especially if you can not convince them otherwise - they sign the checks).
I have always held that the operators should not be chosen randomly, but should be representative of every level of operator that would be expected to use the gage - from a new shop floor operator to a QC lab technician. That way, you can tell how much variation the skill affects the results of that gage, no matter who uses the gage.
On a final note, people that require annual GR&R's really do not get it either. Gage R&R is a tool to determine if the gage is the
correct gage for the job. It has no time function. That is what calibration is for. It verifies the gage is
still good for the job.