The first piece off a machining job carries more information than any drawing. It shows whether the program does what the model intended, whether the fixture holds the part where it was assumed to hold it, whether the material moves when stock is removed, and whether the tool reaches what it was supposed to reach. None of that is known for certain before the first cut, which is why the first piece is treated differently from the rest.
That is also why it usually starts as a longer piece of stock. The extra length is not waste. It is a margin for all the things that can be corrected after the first attempt, and it is cheaper by a wide margin than starting again.
The first reason is holding. A longer piece can be gripped at one end and supported at the other while the working end is cut off later. The grip is on material that will not be part of the finished component, so the jaws can mark it, crush it or clamp it hard without consequence.
The second reason is re-cutting. If the first attempt goes wrong at the working end, the part can be pushed further into the fixture and the end remade. A part cut to final length on the first operation has no second chance, and a scrap part is more expensive than a longer billet every time.
The third reason is indicating. A longer piece gives the operator room to sweep a dial along a surface and establish where the part actually sits before any material comes off. That is the step that turns a good program into a good part, and it needs real estate that a finished-length blank does not have.
The stock left on the ends is a buffer, a handle and a witness. As a buffer it absorbs the error of the first setup, which is always the least well understood. As a handle it lets the part be moved from one operation to the next without gripping a finished surface. As a witness, the material left at the end is the first thing cut off on the final operation, and the cut face shows whether the part was straight and square before it was finished.
All three of those jobs disappear once the part is cut to length. That is the moment the part becomes finished, and it is deliberately the last thing that happens rather than the first.
The first piece is measured thoroughly, not sampled. The features that carry a relationship to each other are the ones that matter: a bore to a face, a face to another face, a hole pattern to a datum. A shop that only checks sizes on the first article learns almost nothing, because the sizes were the part of the job that was already predictable.
The other thing measured is the fixture itself. Marks on the part show where it was gripped and whether the grip was where the plan said. Movement between the roughing and finishing cut shows how much stress was released. Neither of those appears on a drawing, and both change the next design.
What is not measured on the first piece is the process capability, because a sample of one cannot show a distribution. That is a limit of the method and it is worth being honest about. The first piece proves the setup, not the run.
If the first piece is good, the second is cut without unclamping, without re-indicating and without changing the program. That is the cheapest part in the whole order and usually the most accurate one, because every variable that was established for the first piece is still in place and nothing has been disturbed.
This is the reason a small batch is priced so differently from a single part. The single part carries the entire setup cost, and it receives none of the benefit of repetition. You can see the kind of short-run hardware where this arithmetic dominates on our machined and fabricated parts pages.
Many shops keep the first piece rather than shipping it. Sometimes it becomes the reference part that later parts are compared against, sometimes it goes to the customer as the approval sample, and sometimes it stays on the shelf as a record of what the setup produced on the day the job was proven.
There is a sound reason for this beyond sentiment. The first piece was made under conditions that will never repeat exactly: fresh tooling, a fresh fixture, an operator paying full attention. Later parts are made under production conditions. If the first piece is shipped and later parts differ, the difference is invisible because there is nothing left to compare against.
A customer ordering a single part is ordering a first piece and nothing else. There is no second part to amortise the setup, no run to average out the variation, and no opportunity to correct a trend because there is no trend. The price reflects that honestly, and it is not a markup or a penalty.
The practical consequence is that a one-off should be designed for it. Avoid features that need a special fixture if a standard one will do the same job. Avoid tight relationships between faces that could be machined in one setup if the geometry allowed it. And accept that a single part will sit somewhere in the middle of a tolerance band rather than at its centre, because there is no run to aim with.
The questions about first articles are really questions about how much of the process you want to see before the rest of the order runs.
You can see the sort of parts these questions apply to on our second group of small components, the brackets and adapters, our titanium fabrication components, the aluminium welding work and the titanium welded parts built beside them.
