t_partlow said:
We presently use an electronic differential level for calibrating granite surface plates. My original thought was to "calibrate" a surface plate "X" number of times with "Y" operators and base the calculations on the results. However, this will not cover the entire measurement range of the levels. The other way I came up with is to use a sine plate and gage blocks to generate angles covering the entire measurement range of the levels. I'm not really sure which would be better. Any insight would be a great help.
Actually, if you are using the levels that I am thinking of, the gages themselves lend very little to the variations in the process of calibrating granite surface plates. The process of moving the levels to each point winds up being 80% or more of the variation. Surface inperfections of granite, variations in placement, dust on the plate, temperature fluctuations, etc. will be you major contributors. When you measure a plate, check your closure errors. Those will, by and large, be very close to being your process variation.
Granted, I've never done a GR&R, as I've always had to perform full uncertainty budgets, which takes quite a bit more into account. But as far as variation goes, your tool won't be the cause.
Ryan