I have two questions regarding the definition and computation of a
process Cpk...
Background:
My company requires suppliers to provide Cpk data with shipment of
product.
During a recent visit to a supplier, I discovered what I think is a
discrepancy in the way they are computing the Cpk data.
To maintain low inventory, the manufacturer is short-run
batches...nothing new about that.
But they compute Cpk "per lot". That is to say, today they produce a
lot of 100 pieces, pull a sample of 5 and compute a Cpk. Two months
later, they pull out their tooling, set up the machinery, and run
another lot of maybe 80 pieces. Pull a sample of 5, and compute a Cpk
based solely on that lot. Etc etc etc.
First of all, it seems to me that Cpk makes more sense for a
continuous process....and essentially represents "historical data"
over time. So I am not sure how applicable it is to this sort of shirt-
run manufacturing scenario.
But, if it *is* to be applied, then it seems to be the data should
still be historical...that is to say the mean used should be
cumulative....not just based on the present lot. Isn't that what we
mean when we talk about the capability of a "process"?
This NG is all but dead, I'm afraid. A better place for your question:
http://elsmar.com/Forums/index.php?
http://Elsmar.com/Forums/showthread.php?t=16321
http://Elsmar.com/Forums/showthread.php?t=1975
http://Elsmar.com/Forums/showthread.php?t=38229
http://Elsmar.com/wiki/index.php/Cpk
Also see Process Capability.