PPAP Rejected - Ambiguous Specification

[WKHANNA]

I just had a customer reject a PPAP because, Quote:

“There are TWO Special Characteristic listed on the drawing, dimension .250 + -.002 and note (Hatched area to be free of burrs). Only the Dimension .250 has a control plan and CPK requirements. Please provided controls plan and CPK for the note (Hatched area to be free of burrs).”

Just how does one calculate a Cpk for a Go/NoGo condition that can only display two conditions?

BTW, the ‘Hatched Area’ is on the same .250 +/- .002 diameter that is already called out as ‘Special Characteristic!
Frankly, this condition is addressed by our Standard Shop Practices manual and any part we make would be rejected if it displayed a burr on critical diameter with a tolerance of +/-.002.

OK, I'm all done venting now.

[Jim Wynne]

The diameter and freedom from burrs are separate, albeit possibly related, conditions. If the freedom-from-burrs thing is identified as a special characteristic, it should have been addressed in your PFMEA and control plan. As for capability data, the time to discuss that sort of thing with a customer is in the beginning. The PPAP manual has default requirements for x-bar/r charts and capability analysis (Cpk/Ppk), but also says that those things might not be appropriate for all processes/characteristics. In those cases you should look to the customer for guidance in what's expected.

[WKHANNA]

In Reply to Parent Post by Jim Wynne

The diameter and freedom from burrs are separate, albeit possibly related, conditions. If the freedom-from-burrs thing is identified as a special characteristic, it should have been addressed in your PFMEA and control plan. As for capability data, the time to discuss that sort of thing with a customer is in the beginning. The PPAP manual has default requirements for x-bar/r charts and capability analysis (Cpk/Ppk), but also says that those things might not be appropriate for all processes/characteristics. In those cases you should look to the customer for guidance in what's expected.

Yes, Jim. Of course you are right, I should have at the very least made note of the characteristic on the PPAP.

Sometimes though, I have to wounder how some of these designers come up with choices of special characteristics. I used to work for this customer and designed some of the very parts my current employer manufactures for them.

Thanks again, Jim.
I'll straighten up and fly right now.

[bobdoering]

In Reply to Parent Post by Jim Wynne

The PPAP manual has default requirements for x-bar/r charts and capability analysis (Cpk/Ppk), but also says that those things might not be appropriate for all processes/characteristics. In those cases you should look to the customer for guidance in what's expected.

<knee slap> As if they know the appropriate statistical treatment for your process. Most have the same training as AIAG - which means they won't know either. Anywhere the AIAG SPC book says "ask your customer" it is their weasel words for they don't know. In fact, even AIAG analyzed their "golden data set" incorrectly, so who are they to say?

[WKHANNA]

In Reply to Parent Post by bobdoering

<knee slap> As if they know the appropriate statistical treatment for your process.

Ahhh, sweet redemption. That's just what I needed!

[bobdoering]

In Reply to Parent Post by WKHANNA

Ahhh, sweet redemption. That's just what I needed!

Sure, they rubber stamp that 1.33 garbage on everything because they are normalcentrics. Burrs are an attribute - either they exist or they do not. Measuring a length or girth of a burr is irrelevant. So, being attribute, you could use an attribute control chart BUT to pass, it must show zero. Not very exciting. It would be nice to show zero from the process, but, chances are you will need to check them after deburring to see if you overall process is effective.

On top of that, if you have burrs, I might assume you are machining. If so, and you are controlling it correctly, you should have a non-normal distribution and X-bar R would actually be the worst possible charting methodology. Think your customer knows that? Only a rare few do. Again, Cpk is not applicable. They strike out 2 for 2.

[Jim Wynne]

In Reply to Parent Post by WKHANNA

Yes, Jim. Of course you are right, I should have at the very least made note of the characteristic on the PPAP.

Sometimes though, I have to wounder how some of these designers come up with choices of special characteristics. I used to work for this customer and designed some of the very parts my current employer manufactures for them.

Thanks again, Jim.
I'll straighten up and fly right now.

The only thing you did wrong was assume that you were dealing with normal people and not automotive people. It's at times like these that I always think of the famous line from Animal House. For our younger or more sensitive viewers, there is strong language involved.

[WKHANNA]

In Reply to Parent Post by bobdoering

I might assume you are machining. If so, and you are controlling it correctly, you should have a non-normal distribution

Yes you are correct, we are typically CNC machining steels such as 4140, 12L14 & 86L20 to a total tolerance of +/- .002, and often the tolerance is to apply after heat treating & trivalent plating which includes a hydrogen relief bake afterwards. Piece of cake! Even when they have class 2 threads, too.

Oh, and they require 1.66 for any new part submissions.

They have no idea the care we take in tooling choice, set-up & process control. It’s not uncommon for material lot variation to be the largest varialble.