From: Stan Hilliard
Date: 12 Oct 2000
Time: 01:27:02
Greetings Dave,
You said: "1.) Is our assumption regarding non-destructive testing correct?"
You can determine the answer to this question experimentally. Select a pair of sample specimens that were made at the same time under the same conditions. Test one non-destructively. Then test both specimens destructively. If the lower bond strength of the two specimens is from the one that was tested non-destructively, then it was degraded.
In order to be confident in your conclusion, repeat this process on 20 or so pairs and perform a paired t-test on the results. Then you will know for sure whether or not the non-destructive test degrades the bond strength.
You Said: "2.) How do we determine an adequate sample size for whichever test method is appropriate."
I suggest that you develop either a fixed-n or a TSS sequential sampling plan for the mean of bond strength with known (historical, within lot) standard deviation. The oc curve will allow you to make a trade-off between sample size and the size of the difference in bond strength that you want to be able detect. The method is the same for destructive and non-destructive tests. Our program TP414 will perform the calculations. You can find a description of the program, the method, and an example starting on the following web page:
www.samplingplans.com/programtp414.htm
Sincerely, Stan Hilliard