What is your outcome?
It is hard to get a good % Gage R&R. Your samples MUST represent the entire tolerance interval. If there is no variance within the parts, your Gage R%R suffers.
If I were you, I'd get 15 samples and choose 10 of the data that represents the overall tolerance interval.
If cpk>2.5 , you may also consider using tolerance % result
Are you working on tan_delta/ loss/storage modulus? You can also obtain help from your equipment's manufacture (such as TA etc)