I have hand-loaded brass sleeve blanks with different stick-out readings in our FR5 planning sheet, but no matching overall lengths or repeated readings on the same blank; what should quality and the setter compare before anyone calls the spread a robot-position problem?
Have them agree the measuring reference, identify each blank and measure its overall length and locating-end condition, then compare repeated seating of those same blanks rather than treating every reading as a different part.
Setter and quality have agreed one reference and added blank identity, length and end condition to the sheet; repeated loadings keep separate entries under the same blank, and automatic loading remains outside this comparison.
Keep the ordinary range of accepted blanks in it too, so the first nicely repeatable example doesn't become the allowance for every sleeve you actually issue.
Has the comparison produced any repeated readings yet, Robin? I'd like to hear whether the same blank varies or whether most of the spread follows the different lengths. Could be both, of course.