I have problem that how to perform a Gage R&R-test. We have almost ten production lines and in every line there is a same kind of automated measurement system. Measurements is saved automatically to test files. So in this case we don´t have an appraiser variation, right? I think that I can´t take another production lines to appraiser two and three, because gages can´t be exactly identical.
The question is how do I perform Gage R&R-test "right" in this kind of situation?
It sounds like your variation could be in the test software. If you don't have an actual human variation you can do other types of MSA. You could do a stability test to satisfy the requirement for Measurement System Anylisis in QS9000. You might be able to do some type of variation study base on your test data.