SPC - Methods to collect data

[luckyboyvn85]

My company has NC with the following content: "the sample size is 4 pcs/ group, but current the PIC (inspector) take 4 pcs from a shot (including 4 cavities) instead 4 pcs per each cavity from 4 consecutive shots"
But SPC Manual (page 162) Method 1: Subgrouping by column (Cycle). Can help me know this NC is correct? Thank you very much.

 

[Miner]

I am not certain what you mean by NC. Please define.

This is a concept called Rational Subgrouping. You want to include items in the subroup that are similar, but exclude items that are dissimimilar. In my experience, items from consecutive shots of the same cavity are more similar than items from multiple cavities in the same shot. That is: cavity-to cavity variation is usually greater than shot-to-shot variation. In fact, the different cavities often constitute different process streams. Mixing those different streams can increase the within subgroup variation. There are several ways that you can approach this. One would be to assume that the cavities are different and select your subgroup from consecutive shots from the same cavity. The second would be to perform a hypothesis test (e.g., ANOM, 1-way ANOVA) between cavities. If the cavities are not statistically different from each other, you may then safely select your subroup samples from those cavities within the same shot.

 

[luckyboyvn85]

Thanks for your answer
I think that if Appendix A has instructions, I can use this sampling method (method 1) and I'm not sure about that. Thank you very much, Miner. information about ANOM, 1-way ANOVA.

 

[Golfman25]

Who issued the NC? This is minutia into your SPC system. Whether you do shot to shot (same cavity) or just 4 cavities may give you different results. Which sampling is best will probably depend on what you are trying to actually do. Good luck.

 

[luckyboyvn85]

Who issued the NC? This is minutia into your SPC system. Whether you do shot to shot (same cavity) or just 4 cavities may give you different results. Which sampling is best will probably depend on what you are trying to actually do. Good luck.

NC of 3rd party audit. Thank Goflman.

 

[Bev D]

As Miner said, the subgrouping is not dependent on what any manual says. The correct supbgrouping depends on the process and it’s non-homogeneity. This is the basis of rational subgrouping. SPC is not based on random sampling, convenience sampling or willy-nilly choices.

In order for us to know which subgrouping scheme is correct we would have to see the data and be able to properly analyze it.

Regardless of the advisability of your subgrouping scheme, Golfman is correct that this NC seems to be beyond the scope of any auditor.

 

[John Predmore]

One thought I had was whether there were Work Instructions how to correctly gather samples, and the operator(s) consistently did not follow company procedures, that might warrant a non-conformance. The basis of the NC should have been enumerated as the section number of the quality standard which was not adhered to.

 

[sahar khorasani]

My company has NC with the following content: "the sample size is 4 pcs/ group, but current the PIC (inspector) take 4 pcs from a shot (including 4 cavities) instead 4 pcs per each cavity from 4 consecutive shots"
But SPC Manual (page 162) Method 1: Subgrouping by column (Cycle). Can help me know this NC is correct? Thank you very much. View attachment 25632

Hi
Would you do me a favor and i have been your book (spc manual)?
[email protected]