I've found protective film on one of our supposed FR10 reject references. The bright streak is spectacular, but quality's note describes the bare faceplate. I want to sort this sample mix-up before anyone orders lighting to make the wrong surface easier to see.
Match each physical sample to its inspection note. Then distinguish coated or covered presentation from the surface the proposed inspection actually needs to judge. Film can produce its own marks and reflections.
Quality found the mismatch. The note belongs to another faceplate; someone reused a tray position as the sample identity. I've withdrawn both labels, not thrown the parts away.
That tray-position habit is a pain. We had a good sample move one compartment and inherit a rejection note. The parts were fine; our little map was lying.
The process stage needs settling as Leo asks. A correct judgement of the exposed metal is not automatically a usable reference for a camera looking through film. You may need different checks at the two stages, depending on what the job is meant to detect.
After removal, Leo. The filming session used whatever was in the demonstration tray. That was a terrible shortcut. Quality is rematching the parts and notes before we take replacement pictures.
We are using individually identified sample bags with the matching quality notes. The camera operator takes out one at a time. No marking on the inspected face.
Who checks the sample's condition after handling? Individual identity fixes the mix-up, but an accepted reference can acquire a scratch while being passed around.
Henry's point is easy to miss when the set is new. Let quality define when a sample needs rechecking, and give the operator somewhere to put a questionable one without relabelling it alone.
Quality has supplied that instruction. Suspect reference goes back to them, with its existing identity. I don't want the next person guessing whether a new mark was always there.
Yes, Leo. A broad reflection remains on the accepted plate, while the correctly identified reject has its actual scratch visible in the new view. We're retaining both, with the preparation and seating recorded.
The reference mix-up is closed now. Thanks Rachel for the one-part-one-identity nudge. We had good notes attached to positions instead of parts, which explains why the pictures kept winning arguments they shouldn't have won.
Keep one of the withdrawn examples with the training notes if useful, clearly outside the tuning set. It is a concrete reason for the preparation step, not just another instruction people are asked to obey.