Hi, Annie!

It's not possible to answer your question adequately with the information you've provided.

I strongly recommend that you consult with a statistician about your specific issues, but if I'm guessing correctly, it sounds to me like you may wish to structure your test as a tolerance limit problem. Meeker, Hahn, and Escobar have an excellent treatise of the subject in their book __Statistical Intervals: A Guide for Practitioners and Researchers, 2nd Edition__, and they detail sample size calculations for distribution-free tolerance intervals.

Good luck.