GR&R - Little to no part to part variation in single part number

[Mikey324]

This may be a stupid question, but here goes. We conducted an R&R with poor results and ndc of 2. With a small 300 part PPAP run, we randomly sampled 10 parts. Cpk is very high, the process is very capable. I increased sample size, same result. I increased gauge resolution. Same results. The parts are very similar. Just no change. In order to do a proper study, can I use samples from existing products with similar features? For example, parts with hole sizes from from 1mm to 2.5mm, as this is the typical sizes we see in day to day operations? Then set the tolerance to represent that value, ucl 2.5, lcl 1.0?
Do R&R studies have to use only 1 production part number, or could I select 3 part numbers of the same part, just different sizes?

Thanks!

 

[Miner]

Key question that will drive my response to your question. For what purpose is the gage used? That is, will you use this gage for inspection, or for process control (i.e., SPC)?

If the gage is used only for inspection, focus on the % Tolerance metric only. Ignore ndc, % Study Variation, and % Process Variation.

Only if the gage will be used for process control do you worry about ndc, % Study Variation, and % Process Variation.

In the scenario that you are using this gage for process control, you can have a situation where the process is highly capable and the gage cannot resolve what little variation there is in the product. The correct solution here is not to artificially create variation, but to determine whether you have any reasonable options to improve the gaging. For example, if you were using calipers to measure, it would be reasonable for you to investigate micrometers. However, let's say that you were already using the state of the art measurement equipment, or that the next level of measurement equipment costs $250,000 and you are making $3 product. In that case, you do not have a reasonable alternative. At this point, you should put together a good case and have a discussion with your customer about your options.

Here is a long discussion on the topic at a sister forum.

 

[Mikey324]

You are correct in many ways. The device is used for inspection, not spc. You are also correct saying $1k device for $2 part is overkill. We can’t use a different device, a micrometer can’t check this dimension.
To clarify:
This was an audit minor NC. IATF certification.
We explained %tolerance is acceptable, as the device is for inspection.
We provided customer approval, we had the same discussion with them.
We demonstrated approval through a signed PSW.
The finding was still ndc less than 5.

 

[Bev D]

appeal the finding?

 

[Mikey324]

I thought about appeal, but see this as a good learning experience for my QC tech and sample selection.

My question is still as above. If we want to test our measurement process, can we measure parts with similar features? This allows us to make sure our gauges/operators can detect the variance??

 

[Golfman25]

Unfortunately your toast. They probably didn't accept the signed PSW as "approval" because "nobody checks those things." You're appeal will likely fail.

You're stuck with the MSA manual which I don't believe distinguishes between inspection and spc. And if it does, the general guideline language is frequently ignored. You get specific forms and numbers -- good luck hitting them. You'll need to get a separate letter from your customer documenting their approval.

As for the MSA technicalities, I leave that to miner. But it does sound like you are gage limited.

 

[Miner]

Speaking of the MSA manual (4th ed.), see page 77, Section D Analysis of the Results, Width Error, Acceptability Criteria - Width Error.

It states "The criteria as to whether a measurement system's variability is satisfactory are dependent upon the percentage of the manufacturing production process variability or the part tolerance that is consumed by measurement system variation. The final acceptance criteria for specific measurement systems depend on the measurement system's environment and purpose and should be agreed to by the customer."

See also page 79, Comments on the Application and Gage Acceptability.

It states: "When looking at GRR and measurement variation it is important to look at each application individually, to see what is required and how the measurement is going to be used." It goes on to give an example of how a household thermostat could have a GRR up to 30% while a laboratory thermostat must be held to a tighter standard based on their application.

Bottom line: the MSA manual clearly states and expects you to use judgment as to which criteria to use and what level of acceptability to apply.

 

[Mikey324]

The customer approval was a separate document. I agree fully about the PSW being overlooked. I just want to add value while working out the C/M

 

[Mikey324]

Speaking of the MSA manual (4th ed.), see page 77, Section D Analysis of the Results, Width Error, Acceptability Criteria - Width Error.

It states "The criteria as to whether a measurement system's variability is satisfactory are dependent upon the percentage of the manufacturing production process variability or the part tolerance that is consumed by measurement system variation. The final acceptance criteria for specific measurement systems depend on the measurement system's environment and purpose and should be agreed to by the customer."

See also page 79, Comments on the Application and Gage Acceptability.

It states: "When looking at GRR and measurement variation it is important to look at each application individually, to see what is required and how the measurement is going to be used." It goes on to give an example of how a household thermostat could have a GRR up to 30% while a laboratory thermostat must be held to a tighter standard based on their application.

Bottom line: the MSA manual clearly states and expects you to use judgment as to which criteria to use and what level of acceptability to apply.

I know you read the manual, if not you wouldn’t know the thermostat reference.
I think we can satisfy the manual intent, satisfy the auditor, and make sure our system works.

As far as us being gauge limited, I don’t believe we are. We don’t need extreme resolution for parts with wide tolerances. 10:1 is plenty according to the manual. We are using benchmarked gauges with our industry. Cost of the gauge has to play into the equation vs the measurements being made.

 

[Golfman25]

The customer approval was a separate document. I agree fully about the PSW being overlooked. I just want to add value while working out the C/M

If you had a documented customer approval, then how do you have a non-conformance?