At best, experience with real processes, more likely laziness. It allows for the fact that although a Cpk >= 1 ought to be OK, in practice your estimate of sigma may be wrong, your process may drift over the long term, and so on.
Of course it also assumes that your spec limits are meaningful. I remember that when we specified Cpk=1.67 for a new process the manufacturer immediately asked us to widen the specs. Idiots.
Cheers
Greg Locock
New here? Try reading these, they might help FAQ731-376
Thank you for the reply GregLock. I am a new Process Engineer fresh out of college in the automotive industry. I have been doing some Cp & Cpk studies recently. I do know the significance of that value in terms of "is your process capable?" or "how close to center it is?". Although, I was just curious as to why 1.33 ? How is the number 1.33 derived ?
It means use 4 standard deviations from the nearest spec limit to the process mean. As I said, somewhere between experience and laziness. If someone came back with 1.30 I wouldn't scream as I have no faith that the spec limits are that well defined typically. You might want to instigate a temporary (or permanent) SPC plan on that process tho, to increase your confidence.
Cheers
Greg Locock
New here? Try reading these, they might help FAQ731-376
