
The hardest number in a prototype order is not the price. It is the quantity. Order too few parts and the test proves nothing — one part that fits tells you nothing about the next one. Order too many and the prototype run becomes a small production run, paid for at prototype prices, with most of the parts never teaching you anything. A prototype quantity is a judgement about what the test needs to prove, and the buyers who get it right decide the quantity from the test plan, not from a round number. Five parts, ten parts, fifty parts — each is right for a different test, and the RFQ should say which test the parts are for.
What a prototype run is meant to prove
A prototype proves different things at different stages, and the quantity follows the stage. An early prototype proves the concept: does the part fit, does it assemble, does the mechanism move? A later prototype proves the process: can the part be made repeatedly, does it pass inspection, does it survive the function it was designed for? A prototype that is only checked for fit needs fewer parts than a prototype that must also survive a life test, because a life test consumes parts. Before ordering, write down what the prototype must prove, in order: fit first, then function, then durability, then repeatability. A part that is ordered to prove fit and then used for a durability test will fail the test for the wrong reason — not because the design is wrong, but because the prototype was never meant to be tested that way. The test plan decides the quantity; the quantity does not decide the test plan.
Fit testing: how many parts are enough
Fit is the first thing a prototype must prove, and it is the cheapest thing to test. One part proves that the design can be made. Two or three parts prove that the fit is not a one-off — that the second part fits as well as the first. But fit testing has a trap: if the mating parts are also prototypes, a good fit between two prototype parts may not mean a good fit between a prototype and a production part. Test the prototype against the real mating parts wherever possible — the actual housing, the actual shaft, the actual assembly — rather than against another prototype. If the mating part is also new, say so in the RFQ, and consider ordering the mating parts together so the fit is tested as a pair. A small number of parts, tested against the real assembly, teaches more than a large number tested against each other.

Function testing: parts that work, not parts that fit
A part can fit and still not work. Function testing asks whether the part does its job: does it carry the load, seal the joint, hold the thread, turn freely under the conditions it will meet in service? Function testing needs more parts than fit testing, because function testing consumes parts — some will be tested to failure, some will be assembled and disassembled repeatedly, and some will be kept as references. Decide in advance how many parts each function test needs, and how many are spares. A test that needs three parts and has only one will stop halfway through, and the prototype run will have to be repeated. Write the test list down: which tests, how many parts each, and what counts as a pass. If a test has no pass criterion written down before it starts, it will be judged by feeling afterwards, and feeling is not a test result.
Repeatability: the test that needs the most parts
Repeatability is the most expensive thing a prototype can prove, and it is the one most often skipped. A single part that works proves the design can be made once. It does not prove the design can be made ten times, or a hundred times, the same way. Repeatability testing needs enough parts to show variation: parts made at the start, the middle, and the end of a run, measured the same way, compared against each other. If the prototype run is very small, variation cannot be seen at all — every part looks the same because there are too few of them to differ. This does not mean every prototype must be large. It means the buyer should know what the prototype run cannot prove, and should not mistake a small successful run for evidence of repeatability. State in the RFQ whether repeatability is part of this prototype’s job, or whether it belongs to a later run. Honesty about what is not being tested is as valuable as the testing itself.
How prototype feedback should be recorded
Prototype parts are only useful if what they teach is written down. Feedback that lives in one person’s memory will not survive the gap between the prototype and the production order. Record four things for every prototype: what was tested, what the result was, what changed as a consequence, and what still needs to be tested. Photograph the parts, including the failures — a photograph of a part that did not fit is more useful than a paragraph describing it. Keep one prototype part as a reference, marked with the revision it represents, so that the production parts can be compared against it. If a dimension changed after the prototype, record the old value and the new value and the reason; a change without a reason will be questioned later, usually by someone who was not in the room. The record is the bridge between the prototype and production, and a prototype without a record is a test that has to be repeated.

From prototype to production: what changes, what stays
A prototype is not a production part made in smaller numbers. The process may be different, the material may be a substitute, the inspection may be lighter, and the price will certainly be different. Before ordering production, ask what changed between the prototype and the production plan, and whether each change was tested. A material change, a process change, or a supplier change after the prototype is a new variable, and it should be named as one. The most dangerous sentence in this transition is “it is the same part, just more of them” — it is rarely true, and when it is not true, the production parts will behave differently from the prototypes. A short transition check — what changed, what was re-tested, what is still unproven — is the cheapest insurance a buyer can buy before a production order.
Ordering workflow, from test plan to production decision
A workable prototype workflow starts with the test plan, written before any parts are ordered: what must be proved, in what order, and what counts as a pass for each test. Next, the quantity is set from the test plan — enough parts for fit, function, and spares, with repeatability either included or explicitly deferred. Then the RFQ is sent with the drawing, the test plan, and the quantity, and the quote is requested — Precimora promises a quote in 24–48h when the RFQ contains the information an engineer needs. When the parts arrive, they are tested in the order the plan sets, and the results are recorded with photographs and a kept reference part. Finally, the production decision is made from the record, not from memory: what passed, what changed, and what is still unproven. A prototype order that follows those steps costs less than one that does not, because it is ordered once, tested once, and learned from once.
Common pitfalls and how to avoid them
The most common pitfall is ordering a round number — ten, fifty, a hundred — without a test plan, so the quantity proves nothing in particular. The second is testing a prototype against another prototype and calling it a fit test. The third is using a fit prototype for a durability test and blaming the design when it fails. The fourth is skipping repeatability testing and then being surprised when production parts vary. The fifth is feedback that is never written down, so the production order repeats the prototype’s mistakes. The sixth is changing material or process between prototype and production without naming the change, so the production parts behave differently for no apparent reason. The seventh is keeping no reference part, so there is nothing to compare the production parts against. Each of these is avoidable, and each is cheaper to avoid at the test-plan stage than to fix after production.
Buyer checklist
- The test plan is written before ordering: what must be proved, in what order, and what counts as a pass.
- The quantity is set from the test plan, not from a round number.
- Fit is tested against the real mating parts, not only against another prototype.
- Function tests have enough parts, including spares and parts that will be tested to failure.
- Repeatability is either tested with enough parts to show variation, or explicitly deferred to a later run.
- Feedback is recorded: what was tested, the result, what changed, and what still needs testing, with photographs of failures.
- One prototype part is kept as a marked reference for the revision it represents.
- Before production, what changed since the prototype is listed, and each change is named as a tested or untested variable.