Work

The escape

Receiving inspection on vendor-machined hardware, a part family I had inspected thousands of times. Standard incoming inspection to the print.

One machined depth read a few thousandths off target. Close enough to pass a glance, outside tolerance on the gage. [CONFIRM: did I re-measure and check a second piece before holding the lot, or hold on the first reading?] I held the lot and wrote it up the way I write everything: the characteristic and its print reference, the requirement, the gage and its calibration status, the readings to the print's decimals, where on the part, how many pieces and which lot, where I put them, and what I recommended.

The investigation traced the condition back years. More than 13,000 suspect pieces, all of which had cleared prior inspection. Nobody else had caught it.

The supplier corrective action and the escape documentation were both built on that first write-up. Months later I read through the whole package and the description of the nonconformance was still my wording, unchanged, because nobody had needed to come back and ask what I meant. RTX recognized it with a SPOT award, which is an internal safety and quality recognition that doesn't get handed out much.

The award was nice. The thing I'd point at is that one write-up held as the record for a multi-year escape without a single clarifying conversation.

Stopping work on interpretation

After a unit move, my area picked up bearing ducts, a part family we had seen once before. The work instructions for it were a mess. Steps that didn't match the print, references that pointed nowhere.

I halted inspection and went to my supervisor and the QA tech. Both tried to work out what engineering meant and offered a reading I could inspect to. I wasn't willing to accept the part on an interpretation. If the instruction is wrong, the acceptance is wrong, and the record says I accepted it.

Engineering updated the work instructions. Then I inspected to them and the part was accepted.

Unwilling to accept on interpretation is different from refusing to budge. I gave them exactly what I needed to proceed, and I proceeded the moment I had it.

A record error, and what I did about it

Cross-checking barcodes against a lot in the MES, I marked the wrong serial as held. The nonconforming part was physically segregated the entire time and never shipped. But the record was wrong, so the serial that should have been on hold showed as bought off, and a good serial showed as held.

I caught it the following week. I went straight to my supervisor and QA, we corrected the records, and the bought-off serial was pulled back. No further fallout. It was month-end and the pressure was real, and none of that is the reason. I didn't check the screen against the tag before I committed the record.

The part being right in your hand does not mean the record is right, and the record is what the next person works from.