Finding a defective part on the bench is the best outcome available once the defect exists. The alternative is finding it at the customer, in service, or after a batch has shipped. A shop that treats a reject as a failure of the day rather than a success of the inspection will spend its time hiding findings instead of using them.
What follows a reject therefore matters more than the reject itself. The sequence of actions taken in the next hour decides whether the problem is contained or repeated, and whether the record is useful or decorative. Most of that sequence can be written down in advance, which is the point.
The first action is physical. The part is tagged, moved to a defined location and separated from conforming work. Nothing about it is adjusted, dressed or re-welded until someone has looked at it in the condition it was found. A part that gets "tidied up" before the investigation destroys the evidence that would have explained it.
Containment also covers the batch. If the same joint, fixture or procedure produced this part, the parts around it are suspect until proven otherwise. That does not mean scrapping them; it means holding them and checking the specific feature. A contained batch is a decision waiting to be made, which is a manageable state.
The written finding should say what is wrong and where, in terms someone else can verify. Position, size relative to the joint, appearance, and the inspection method that revealed it. A description that starts with the remedy has skipped the diagnosis and will usually be wrong about the cause.
There is a discipline in this that pays off later. Two findings described in the same vocabulary can be compared; two findings described in the language of their fixes cannot. When the same description appears three times in a month, the pattern is visible, and the pattern is what leads to a change in the process.
Every non-conformance ends in one of four places. Rework brings the part back to the drawing by removing the defect and redoing the work. Repair restores fitness for purpose without meeting the drawing exactly. Accept-with-deviation ships the part as it is, with agreement. Scrap ends the discussion.
The choice depends on what the part has to do. A cosmetic deviation on a heatshield is a different proposition from the same deviation on a pressure boundary. The important thing is that the decision is recorded with its reasoning, so that a later reader can see whether the judgement was sound rather than merely convenient.
Disposition authority has to sit with someone who understands both the defect and the function, and it should not sit with the person who made the part. That is not about distrust. It is about the fact that the person who made it has already made one judgement about acceptability, and a second opinion is the whole point of inspection.
Small shops often skip this because everyone is busy and the answer seems obvious. The cost appears later, when a deviation accepted informally becomes a precedent, and the next part with the same flaw is passed without anyone deciding anything at all.
The useful question is not who is at fault but what made the error possible. If a joint was welded with the wrong gap, the answer might be a fixture that allows the gap to vary, a drawing that does not specify it, or a sequence that closes the gap before the tack. Each of those has a different fix and none of them is a person.
Blame-driven investigations produce short reports and repeat findings, because everyone learns to describe events in ways that avoid attention. Cause-driven investigations take longer and change the process, which is the only thing that stops the same reject returning next month.
When a customer finds the problem first, the shop's non-conformance process is already late. What matters then is speed and honesty: what was found, what is being checked, what will be different. A supplier who describes the cause clearly and shows the containment is more credible than one who reports that everything is fine.
Our custom work is quoted with the acceptance criteria stated, so a deviation can be discussed against an agreed rule rather than argued after the fact. That single habit removes most of the disagreements that turn a small finding into a commercial problem.
The record is what makes the whole process repeatable. It should be short enough to complete at the bench and complete enough that a reader six months later can follow the reasoning without asking anyone.
The last field is the one that turns a stack of rejects into an improving process. Without it, the shop knows what went wrong and nothing about why.
Ask how many non-conformances were raised on the last similar job and what changed afterwards. A supplier who says none is either perfect or not recording. A supplier who can describe a finding on a welded titanium part, the containment and the process change is telling you the system works.
Trace that answer through the product range. The same discipline should apply to welded aluminium tanks, where a leak path is a safety matter, and to fabricated titanium components, where a contamination finding cannot be dressed out. Our additional parts pages show the breadth of that work.
A useful test is to pick one part at random and ask for its story: what it was made from, who welded the joint, what inspection it had and what was found. A shop that can answer without disappearing for an hour has a working system, and the answer is more informative than a folder of certificates. The plant where those records originate is described on our company profile.
The reject is not the enemy of a healthy shop. Concealment is. A shop that raises findings freely, records them plainly and changes something as a result is easier to buy from than one with an unblemished record and no explanation for it.
