I found two photographs attached to old angle-bracket inspection sheets and the bench supports aren't in the same places. We're reviewing the fixture for FR5 welding, but the records blame springback without saying how the released part was supported. I'd like to check that measurement difference first.
Compare the support arrangement with the drawing datums first, because changing where a part rests can change the reading without anything new happening in the welding fixture.
One arrangement supports the base near the holes, the other near the ends. Same inspector initials. I'm asking quality whether those were intended for different part variants.
That is plausible, and I'd keep the photographs tied to their exact sheets rather than assuming the initials mean the measurement method was identical.
Do the sheets identify the thickness as well as the bracket name? We have similar-looking parts that use different inspection supports, though I don't know if that's relevant here.
Quality found a revision note introducing a different inspection setup after the base feature changed. Both sheet groups had been exported under the same short description.
Then split those groups before drawing a process trend, because comparing them directly would blend a design and measurement change into your supposed welding behaviour.
Preserve the full drawing revision and inspection-method reference in the new data structure; a short description is a poor join key when several different physical parts share it.
And did any current-revision records include readings both before and after physical correction? Those would help distinguish the original released shape from the accepted result.
Only a small current-revision set remains, and two entries were taken after straightening. We're not using them to estimate untouched released dimensions. More measurement work is needed.
I updated the review pack with the two inspection methods and their drawing revisions. The fixture decision is on hold while quality defines a consistent current-part trial.
Please bring back the trial outcome, because it will be interesting to see how much of the old scatter came from mixed records and how much remains in the actual parts.