Hi everyone,

I am looking for help concerning gage R&R described in MSA 3rd edition.

If I read this book, we are asked to select 50 samples in the global range of measurement. This batch must include 3 different zones :
- some samples out of tolerances
- some samples inside the tolerances
- some samples near the limits

My problem is to find these bad samples! We use gauges to check products made with plastic injection machines.
I do not find samples near the limits and I am told by producer that they cannot produce such pieces in production without a incredible cost.
How must I do in this case?

Thanks for your help...and sorry for my english (hoping that it is comprehensible!) :confused:

Hi everyone,

I am looking for help concerning gage R&R described in MSA 3rd edition.

If I read this book, we are asked to select 50 samples in the global range of measurement. This batch must include 3 different zones :
- some samples out of tolerances
- some samples inside the tolerances
- some samples near the limits

My problem is to find these bad samples! We use gauges to check products made with plastic injection machines.
I do not find samples near the limits and I am told by producer that they cannot produce such pieces in production without a incredible cost.
How must I do in this case?

Thanks for your help...and sorry for my english (hoping that it is comprehensible!) :confused:
I assume you're referring to an attributes study. If you don't have nonconforming parts, you have no way of knowing for sure whether the measurement system will detect them, or whether "good" parts will be identified as nonconforming. This knowledge should be a requisite for use of an attributes gage. I say "should be" with full knowledge that it's rarely the case. You should discuss the situation with your customer, and explain what you plan to do, which should be using parts as close to the specification limits as reasonably possible.

Is it possible to create your own bad parts? Perhaps drill out a hole or file down a surface or even hit it with a hammer a few times? The trick would be to do it in a way that isn't obvious to the person doing the measurements. It would kind of defeat the purpose if the operator says "Hey, this one has the paint scratched scratched off - it must be the one that is out of spec."

Tim F

of course the dilemma posed is quite common for new molds: plastic or stamping. Teh molds typically produce parts at the target point of the specification by design. The process variables (time, pressure, temperature, equipment, etc) typically contribute fairly small amounts of variation - especially compared to what happens to the mold over time as it wears. Thus finding parts near the either tolerance limit for a new mold just doesnt' happen (unless there was fatal flaw in the design or construction of the mold; in which case you have a far bigger and more expensive problem!)

It would be extremely expensive to naturally 'produce' parts at the limits - unless you simply wait for months or more typically years for the mold to wear. And although it's tempting to think that limit samples can be 'artificially created' by shaving or compression, the results are often sub optimal and noticable.

The more effective approach is to verifythe R&R wiht similar features on different parts. Also discuss the situation with the Customer. Any supplier engineer worth their salary is well aware of the variation - or lack of it - that you describe.