Let assume I have an inline measurement device for which I would like to determine repeatability. I am not concerned about reproducibility. Can I just take a n-number of measurements and calculate repeatability as follows:
r = +/- 1.96 sqrt(STDev)
And what should be my sample size n?
Create a crossed gage R&R if you use Minitab and set it up for just one operator and you will get the repeatability. Get 10 distinct times in your process when it is supposedly stable and get 3 measurements from each time. The 10 distinct times are your 10 parts. It’s best to still follow the 80_% of your typical process range for finding the 10 distinct times to measure your process in line.