From: Stan Hilliard
Date: 25 Apr 2000
Time: 00:44:08
Greetings Mark,
> Another note is that the current monthly > quality is for Supplier A=95%, > Supplier B= 99.5%, and Supplier C= 99.2%
The 95% confidence limits calculated by TP105 show supplier A to be approximately 92% to 97%. I would carry out the sampling plan separately for each supplier, but combining the part numbers.
Even though you don't plan to form lots for rejection purposes, you could use days to form lots for the purpose of detection and reporting. You could start at 100% inspection during a trial period, using sequential sampling plans from TP105 to portray the results and use them to justify switching to sampling. This will show the point where a decision would have been made if sampling. See:
www.samplingplans/outputbinomial.htm
An Excel spreadsheet could display the data and chart. After some experience and demonstrating success, it could be given to the OEMs to execute. Then the OEMs could supply the sample test results with the repaired drives.
I have been assuming that you have attribute data. If variables data is involved, you could use TP414. see:
samplingplans/outputvariablesisl.htm
Sincerely, Stan Hilliard