We have a test method to evaluate if a produced batch gives defects after application or not.
Product is applied on 30 substrates and in total 3 surface defects can be accepted to approve the batch.
The method requires a lot of product and takes quite some time and consumes a lot of substrates. Only one team of the same two persons does the test and evaluation, so no comparison between operators or team are needed. We are confident that the defects are always detected, so that's not a question.
We are not confident that only the product itself is the cause of the defects. It might also be application conditions, debris on the substrate or something else.
We plan to take out samples of one batch we have concidered as good where we haven't seen defects, and one bad.
The question is how do we set up a test and evaluate the test method in a case like this? How many test do we need to run?