Capability, FMEA and PPAP - 4.9.2 - The meaning of "high degree of capability"

Hi, I have two questions to submit:

First, I can't understand the meaning of "high degree of capability": Cpk/Ppk>=3 (see reference 4.9.2). Most of the time for us, a Ppk is higher than a Cpk because we carefully monitore all process parameters during a pre-launch. So, during full production, a Cpk has more chance to be lower than a Ppk. So, how can this equation be higher than 3 ?

Secondly, in the QS9000's third edition, are APQP, FMEA and PPAP have no significant changes. We plan doing nothing in those parts to update from second to third edition of QS9000. For me, these program elements have only been moved ?

[Don Winton]

Daniel,

"high degree of capability": Cpk/Ppk>=3

That I would like to see. Perhaps I misunderstood.

Ppk is higher than a Cpk...

This is, more often that not, a poor estimate of sigma.

So, how can this equation be higher than 3 ?

I ain’t gotta clue. As I said, a Cpk of > 3.0 would be a sight.

You may want to see http://Elsmar.com/pdf_files/CPK.pdf

Regards,
Don

[This message has been edited by Don Winton (edited 02-04-99).]

Cpk or Ppk >3 is ok. We understood Cpk divided by Ppk... It doesn't make sense. Thanks !

What about my second subject ?

[Don Winton]

I think what Don was getting at is that a process with a Cpk of > 3, in the true sense, is unheard of.

Correct. But, upon re-reading the original post, I think Daniel was looking at the RATIO of Cpk/Ppk. This is even more confusing. Some QS folks may shed light, but I am so confused.

What about my second subject ?

Sorry, Daniel, but as I have stated before, I am not QS savvy. Therefore, I will rely upon other members of the Cove to cover this aspect.

As always, if I am totally off base, someone please let me know.

Regards,
Don

[Kevin Mader]

I think what Don was getting at is that a process with a Cpk of >3, in the true sense, is unheard of. Generally speaking, a Cpk that high would indicate that the tolerance is too liberal or an error in calculation. If there is no calculation error, then the process is probably not worth monitoring (not cost effective) and you would be better off measuring something requiring your attention.

Read "Ppk/Cpk" as Ppk OR Cpk. What they are saying is that, depending upon which calculation you used to determine capability - SPC type chart or histogram, a capability index of 3 is something you should pay little attention to, assuming normal distributions and "in control" is evident.

BTW, there are circumstances where a Cpk of 3 today is a negative Cpk tomorrow. Injection molding is a good example - in fact I have a real example. You can be near the edge but have extremely tight distribution with a Cpk of 3, send the tool out for polishing and suddenly produce parts out of spec. The dimension only moved .0007. Yes, shoot for nominal is normally the rule, but if you only look at Cpk, then you could be surprised in the future...

Oops, I forgot your second question. My input is, if it ain't broke, don't fix it. If you are satisfied that you meet the intent in those sections, they do not need changed. Somewhere you are going to state that your manual meets the "Third Edition" so you may take that opportunity to review those particular sections you mentioned, but I would certainly expend more resources on the sections that did change.

[This message has been edited by Batman (edited 02-07-99).]

[Don Winton]

Good reply Batman. Thanks for the assist.

Regards,
Don

Thanks ! Your comments are very helpful.