Destructive Attribute (Crossed) Gage R&R - Need Help!

Hi forum users,

I'm trying to design a Gage R&R Study for Attribute (Pass/Fail) data, where the test is a destructive (HiPot Voltage test).

The tester is an automated tester, so there is no Operator component.

My questions are:

1) If the "Fail" rate is low (~2-4%), what should the sample size be?
2) Is it possible to do a Attribute Gage Study with destructive testing, and how would one do this? Is there such a thing as a Nested Attribute Gage study?

Thanks a bunch,

Andrew

AndrewRSundaresan said:

Hi forum users,

I'm trying to design a Gage R&R Study for Attribute (Pass/Fail) data, where the test is a destructive (HiPot Voltage test).

The tester is an automated tester, so there is no Operator component.

My questions are:

1) If the "Fail" rate is low (~2-4%), what should the sample size be?
2) Is it possible to do a Attribute Gage Study with destructive testing, and how would one do this? Is there such a thing as a Nested Attribute Gage study?

Thanks a bunch,

Andrew

Click to expand...

I remember that there are some guidelines in MSA Blue book fro destructive test.
Are these equipment included as measurement system in your control plan?

1) Attribute GR&R studies require more data. Sample size should be at least 20 parts assessed at least 3 times by each of your appraisers.

2) Where in certain circumstances measurement cannot be repeated (destructive testing) you can use the GR&R (nested) function in Minitab. This is an advanced technique which relies on the careful selection of parts, so that small batches of parts that are taken close together can be assumed to have identical properties. The batches are then selected at different times in order that the longer term process variation will be present between them. You will need experienced support from a Minitab guru to use this technique.
Hope this helps
Jim G

Hi Jim,

Thanks for your reply.

Right now, the HiPot Voltage test is not part of our control plan. We are not bound by AIAG or Customer Quality requirements, so we can define GR&R for ourselves.

My case is actually ATTRIBUTE + DESTRUCTIVE, and I have yet to see an example of this anywhere.

How would one approach a Nested GR&R study in Minitab with Attributes? I thought it was only possible to do Nested GR&R for Continuous Data.

Thanks,

Andrew