Acceptance Of Gauges > 10% < 20% Tolerance

[D Moore]

We have recently come across a number of gauges showing between 10% and 20% of tolerance.

I understand that this band of results can be accepted for use. What we would like some help with is where the line can be crossed?

We have done a lot of work on some of the gauges to get them acceptable but we are now at a point where we are thinking of accepting them at their values.

Does anyone have any guidelines on this? How can we justify to Ford that these gauges are deemed acceptable to use by ourselves.

Cheers.

[bobdoering]

What ndc values are you getting, and are any of them being used for SPC?

[D Moore]

Sorry for the delay.

Genrally the ndc values are in the teens although there are a few cases where it is below 5.

What do you mean when you say are they used for SPC? Some certainly are used for machine capability studies.

Thanks

[Miner]

There are two categories. Gages used for inspection, and gages used for any type of statistical purpose. This purpose may be SPC, capability studies, DOEs, etc. The suitability of this type of gage is dependent on the process variation. Whereas the first type is dependent on the tolerance.

Regarding your original question, the answer depends on the importance of the characteristic to me measured. A Critical To Quality (CTQ) characteristic should ideally be < 10%. A non-CTQ can be between 10% and 30%. However, this is ultimately an economic decision.

You should also read Donald Wheeler's articles on an Honest Gauge Study.

[bobdoering]

When gages are used for SPC, you need at least 10 discriminate categories between the control limits (not the PV or specifications) - that is why the question is important. Otherwise, you get "chunky" data, and makes the SPC chart data ineffective.

[Richard Pike]

Originally Posted by D Moore

We have recently come across a number of gauges showing between 10% and 20% of tolerance.

I understand that this band of results can be accepted for use. What we would like some help with is where the line can be crossed?

We have done a lot of work on some of the gauges to get them acceptable but we are now at a point where we are thinking of accepting them at their values.

Does anyone have any guidelines on this? How can we justify to Ford that these gauges are deemed acceptable to use by ourselves.

Cheers.

The 10 - 20 % numbers are thumbsuck (industry norm) numbers used in the absence of anything else.

If your MSA data is statistically "under control" then you are probably achieving the best you can without significant cash expenditure.

So "the line can be crossed" if the measuring process can be demonstrated as "under control" or "no special causes of variation present" and if any further expense cannot be justified and of courseif you have no history of problems with this or similar measurements.

Hope that helps.

[D Moore]

Thanks Guys.

Shall Take all your thoughts into consideration.

[adamsjm]

per Miner
Regarding your original question, the answer depends on the importance of the characteristic to me measured. A Critical To Quality (CTQ) characteristic should ideally be < 10%. A non-CTQ can be between 10% and 30%. However, this is ultimately an economic decision.

I agree and per an earlier posting:

Regarding your original question, the answer depends on the importance of the characteristic to me measured. A Critical To Quality (CTQ) characteristic should ideally be < 10%. A non-CTQ can be between 10% and 30%. However, this is ultimately an economic decision.A good reference source on this is RESOLVING DILEMMAS THAT OCCUR WHEN USING QUALITY MEASURES by Karl D. Majeske and Richard W. Andrews

http://deepblue.lib.umich.edu/bitstr...9.0001.001.pdf

They state, “Automobile manufacturers and suppliers that follow Automobile Industries Action Group (1995) guidelines use a cutoff of Co = 1.67, i.e., processes with Cp >1.67 are approved for production.”

With this condition a capability of Cpk > 1.67, Gage R&R most likely must be less than 30% to be confident of the process capability. With a condition of capability, Cpk > 2.00, Gage R&R most likely must be less than 10% to be confident of the capability. The confidence that the process is that capable is more complicated but these are some guidelines.

Per Majeske and Andrews, “… Therefore, the more capable a process, the more precise a gage must be to pass the correlation criteria (i.e.: Gage R&R vs. Cpk).”