Hello,
I'm new to the ALLSTAT mailing list and I have a question regarding sample
size for Cpk studies. I'm curious to find out what's the best acceptance
criteria for small samples when performing Cpk studies. Our current
acceptance criteria
is all process capability studies are based on a 30 piece sample. In the
event that time and/or costs do not permit a 30 piece sample, the following
table which is equivalent to 1.33 Cpk, apply:
Max. centered range as percent of the total tolerance:
Sample Size Bilateral Tolerance Unilateral Tolerance
5 31% 36%
10 40% 47%
15 44% 52%
Are there better ways of doing this? Is this statistically sound? If yes,
are the numbers correct? If not, please comment. Appreciate all comments.
Jack-Chinh Huynh, P.Eng.
Quality Engineer
John Deere Welland Works
Tel: 905-734-4501 ext. 3301
Fax: 905-734-2300
E-Mail: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|