From: Stan Hilliard
Date: 01 Apr 2000
Time: 23:01:41
Greetings koklee,
Many people mistakenly think that the zero acceptance criterion offers the best protection against passing off-grade lots.
I used the attribute sampling program TP105 to show that your plan (C=0, n=5) has a 10 percent chance of passing a lot that contains 37 percent defectives. This is not a very good plan, considering that a 37% lot contains 0.37*30,000=11,100 defective pieces.
I would design a plan that offers more protection to both the consumer and producer by using the operating characteristic curve. See:
www.samplingplans.com/usingoccurves.htm
You can replace the sampling plan for dimensions with another (better) attributes plan or with a variables plan (better yet).
See the tutorial at: www.samplingplans.com/modern3.htm
Software to develop attribute (TP105) and variables (TP414) sampling plans is described here: (I recommend looking at both in order to compare alternatives.)
www.samplingplans.com/software_oc.htm
Sincerely, Stan Hilliard