The copied entry just says aligned. Original sheet search is underway. The estimate assumes a robot cycle followed by final inspection, with no separate correction operation, so I have flagged that omission now.

Keep that flag as an unknown scope item, not an automatic decision to build a straightening station. You haven't established what correction meant here or whether the new process will need it.

Original sheet found. It records tack-stage fit-up against a setup aid, not the completed weldment in its final fixture. The summary had placed it under restrained final alignment. I've corrected that stage description in our review copy.

That is worth preserving with the source reference. It removes a false before-and-after comparison, but still leaves the released failure and later correction to understand on their own terms.

Does the released failure identify the same assembly as the later accepted result?

Yes, their assembly identity agrees. Released failure gives measured dimensions; final sheet says accepted after correction. The correction work instruction is referenced but wasn't included in the pack we received.

Then ask for that instruction revision and any operation record tied to this assembly. The instruction explains intended work; it doesn't necessarily prove every step was performed on this part.

Instruction recovered. It describes an offline manual correction route with reinspection. We found a signed operation record for the assembly too, but no duration or readings immediately before and after the correction on a common setup.

That gives planning an identified operation, though not a measured time allowance or a correction amount. The new proposal should say whether it retains that route or needs a study to justify another one.

Manufacturing engineering has taken that decision into the process review. My corrected history now says tack-stage fit-up, released failure, recorded correction and final acceptance. It no longer claims measured movement from a restrained finished bracket, or a correction-free robot cycle.

Keep the old summary alongside the corrected one with the source links. That lets the next reviewer see why the interpretation changed rather than wondering which apparently conflicting document to trust.