AIAG Statistical Process Control (SPC) Manual - Page 81

[Sam]

Comments Please

Please refer to the AIAG "Statistical Process Control" manual page 81, Section C - Description of Conditions and Assumptions, last bullit.

Please provide your comments/interpretations of the following statement:
"There exists a willingness to accept the computed index(or ratio) value as the "true" index (or ratio) value - i.e., to discount sampling variation's influence on the computed number (e.g.,a computed Cpk of 1.05 may be from a process whose "true" Cpk 1.40, or vice versa, due simply to sampling variation). Please see Appendix H, Referewnces 19, 20, and 21 for more on this subject."

 

[Craig H.]

Sam:

Not having a copy of the manual, this might be off base, but I'd like to take a shot.

I believe what they are refering to is the tendancy for us to see the numbers we generate as coming from a "magic truth box". I once went to a conference held by a major supplier to the auto industry (call them company A). Company A held this conference for their suppliers, and the suppliers sent mostly sales people, even thought the stated mission of the conference was supplier quality (go figure).

One of the big points company A was trying to make was that they had hired outside labs to test incoming raw material, and had found that upwards of 60% of incoming raw material was out of spec (not a typo). Yours truly raised his hand and started asking about test methods. You should have seen the blank stares. The blank stares continued when I asked about instrumentation.

My point is that we often spend too much time looking at the numbers, and accept them as gospel, when we should first be looking at just where the numbers come from, and how they were generated. Sure, you might get your shoes dirty, but it beats being made a fool of by relying on faulty data. Yeah, I have been caught myself, and it is a lesson I hope to never have to repeat.

Hope this helps.

Craig

 

[Rob Nix]

AIAG SPC Manual

The paragraphs preceding your reference (Section C) are describing conditions and assumptions. Then, when referring to capability measurements, it is ASSUMING four bulleted conditions (i.e. controlled process, normal distribution, customer specs, AND "willingness to accept...").

From this I would think they are saying that between you and your customer you are both accepting the idea that the results you get are ASSUMED to be correct, reality, or true - EVEN THOUGH you both likely agree that the actual true capability is unknown. It may be higher or lower; you are simply reaching an agreement that you both accept the results on the study.

IMHO

 

[Sam]

Craig,
I sgree. We place too much emphasis on the final number represented by the term Cpk.

Rob, I also agree. We are making an assumption based on a condition.

However, I'm having difficulty determining what the condition is.
- First, What is a "computed index (or ratio)"? Is it the same as the CR on the previous page,
- if so, can I present the CR as capability versus Cpk?

 

[Rob Nix]

Sam,

Yes, the computed "ratio" is one of the calculations on page 80.

Your customer most often dictates what equation is required to show capability, e.g. Ford uses Cp and Cpk, GM uses CR (the reciprocal of Cp). Whatever you or your customer choose is acceptable.