Gauge R & R - What is the requirement for number of operators in study?

[DJN]

We are in the process of preparing a PPAP. I have been asked to provide Gauge R & R studies for those gauges used. This I have done, but have now been told the as the study was only carried out using 2 operators, that this will not be acceptable? Apparently, this is quoted in the MSA Reference Manual? Could anyone let me know if this is true, or I I OK in using 2 operators, and whereabouts in the MSA Reference Manual that this is stated.


Many thanks

David

 

[Rob Nix]

David,

First of all, the MSA Reference Manual is just that - a reference. It is not law. Two people, Three (NOTE: the pg. 114 table specifically uses 2 or 3), or more could be used depending on the need (i.e. the likelihood that reproducibility will be an issue).

The question is: WHO told you that two is unacceptable?

If the customer, what does your customer expect? At contract review that requirement should have been clarified.

If someone else, say, an auditor, then the person should be reminded that there is no rule; it is whatever the customer states and/or what makes the most sense for the application. If the measurement result appears to be highly influenced by the user, then more appraisers (e.g. 3) should be used.

I hope this helps a little.

 

[TownDawg]

I see no one is willing to take a stab at this, but most likely it is because it is posted in the wrong forum. This is not a TS-16949 question per se'.

Interesting that the standard AIAG GR&R excel worksheet would have t-test constants (D4 = 3.27, and 2.58) for 2 operators and 3 operators.

Interesting that page 97 of the AIAG MSA reference would state that "the range method typically uses two appraisers and five parts for the study."

Interesting that page 99 of the same reference would state that "although the form was designed with a maximum of 10 parts, the total number of 'ranges' should be > 15 for a minimal level of confidence in the results."

Interesting that page 115 of the same reference would state that "K2 depends on the number of appraisers used in the gage study, and is the inverse of D2 which is obtained from Appendix C."

It sounds like you have an amateur/professional statistician on your hands. Certainly, with more samples, and more operators, and more replications, the GR&R model gets more and more accurate -- but it is statistics. It is an estimation of the truth. It is only a model.

In 20 yrs, I've never had anyone reject my PPAP over something this petty. Maybe I've been lucky?

 

[Sam]

As previously stated the Gage R&R process is set up to use two or three operators.
However, all persons responsible for accepting the part are required to participate in a Gage R&R. For example, if you have six inspectors approving the part then you would perform the R&R three times. Likewise if you have a 1st shift and 2nd shift you would do an R&R twice.
These results can also be combined to ststistically evaluate each group.

 

[DJN]

Thanks to both. We have a new Quality Engineer, specialising in the automotive industry. It his he who has said that at his last place of work, their PPAP was rejected because they did their R & R with two operators. I realise that with more operators and replications, I will get an more accurate picture. If our customer says he wants R & R with 3 operators that ain't a problem. It was just that nobody had questioned it before. Once again, many thanks