Ok guys. This one really baffled me. I was in a meeting with our vendor when the program manager said. We can't compute Cpk using GD&T. (I have since convinced most of them that you actually can, thanks to everyone's input.) He then said but we can compute Cp. I have read some previous posts comparing Cp with Cpk. But I can't for the life of me figure out why you can't (or at least they say you can't) compute Cpk but can compute Cp. Are there any differences that would keep you from making the math work.

Hi Tim!
GD&T is a common set of rules for creating a part drawing with form, fit, and function in mind. Every part has some limits above and below which you cannot go. These limits are called tolerances.

Cpk (and Cp) is a statistical calculation from [usually a sample of] the actual produced parts comparing that data TO the tolerance on the drawing.

In this order:
Create drawing with tolerances, etc.
Make parts to drawing
Analyze (measure) parts - compute average and sigma
Compare this calculation to the tolerance on the drawing.

Are there any differences that would keep you from making the math work?

No there are not. Both equations use mean and standard deviation, along with constants (3, 6, USL, LSL, etc.) Since you can determine mean and standard deviation, if you can compute Cp on a variable, you can compute Cpk on the same variable.

Regards,
Don

------------------
Just the ramblings of an Old Wizard Warrior.