We did that with the setter. One blank is longer overall and repeats consistently. Another has a damaged end and does not repeat. It has been excluded and retained for inspection, not hidden in the good sample set.

Two different problems. One offset would have been an enthusiastic guess

Who will check that end condition when new blanks arrive, Owen? The setter found it before automatic loading this time, but that needs a place in the incoming check rather than becoming their private trick.