A prototype ready for production has been measured, tested, and reviewed by the people who will make it. Its dimensions match the drawing, it is built from the material you will ship or a close stand-in, it has survived real use, and a manufacturer has told you what will be hard to produce. Miss any of them and you find out after tooling is cut, the costliest moment to learn it. Engineering changes cost more and take longer the later they land.
Four checks cover the ground, and one question handles what can safely stay open. We would run them in order, since the early ones can change the design the later ones depend on. Still deciding what your first prototype should prove? Start with the practical guide to getting started on prototyping.
Do the Measurements and Tolerances Match the Drawing?
Measure the part, because a printer or router hitting its nominal size is an assumption until a gauge says so. Check every interface, meaning any feature that mates with another part, and every dimension on the drawing against its nominal value and tolerance. A good digital caliper handles most of it, and Mitutoyo rates its Series 500 models at ±0.02 mm up to 200 mm. For complex surfaces, comparing a scan to the CAD (computer-aided design) model produces a color map of deviation across the whole part, which a caliper cannot.
Then write the tolerances down, because a drawing without them gets read the way the manufacturer prefers. Protolabs, for one, applies a published general-tolerance default to machined parts that arrive with none, ±0.3 mm on plastic dimensions between 30 and 120 mm. Tolerances also stack: five parts at ±0.5 mm each add up to ±2.5 mm in the worst case. We would tighten only the dimensions that govern fit, since tighter tolerances usually cost more to hold.
A printed prototype also carries its own process error. Formlabs specifies accuracy for its Fuse SLS (selective laser sintering) printers as ±0.5% or ±0.3 mm, whichever is larger, so a ±0.1 mm requirement is beyond what an SLS prototype can vouch for.
Is the Prototype Made of What You Will Ship?
A printed part in PLA (polylactic acid) tells you about geometry and little about how the molded or machined version will behave. PLA softens at 55 to 60°C, and plastics under steady load slowly deform, faster as the temperature rises. In an Arizona study of cars parked in the sun on days above 38°C, dashboards averaged about 69°C, so a PLA bracket that holds on the bench can sag in a car. Cosmetic parts forgive more, but anything structural needs the real material.
If production will use a different material, get at least one prototype made in it, or in a close stand-in, before you commit. Urethane casting in silicone molds is one route, since casting resins can mimic production polymers, though Formlabs notes you cannot prototype in the final injection-molded material itself.
What Has the Prototype Been Through?
Functional testing asks two things: does it work, and how long does it keep working. For some products the first pass is informal. Assemble it, use it the way a customer will, and note what breaks. Write a test plan that ties each test to the requirement it checks, which Stanford Biodesign’s prototyping guide says saves iteration cycles.
The test that counts most happens in the real assembly. For the replacement controller handles we made for Barcade’s Quick & Crash game, printed prototypes were test-fit on a deployed cabinet before any mold was cut, which confirmed the fit of the internal parts and the ergonomics. The final handles were cast in a high-impact urethane resin chosen to survive being slammed into the holster and dropped. We would not cut a mold until a prototype has done that kind of job in the real assembly.
If you plan to put a warranty behind the product, add life testing. The lifespan data should come from that, not from a hunch.
Has a Manufacturer Reviewed It for Manufacturability?
DFM (design for manufacturability) review is a manufacturer assessing your design for how efficiently it can be made while still doing its job, and suggesting changes that cut complexity without losing function. It fits a design built in CAD to the process that will actually make it.
Injection molding shows why it matters. A printed or machined prototype works fine without draft, so it can look finished while the molded version would stick in the tool, and Protolabs says 1 to 2 degrees of draft works in most situations. Uneven wall thickness is the next catch, since thick sections keep shrinking after the surface has frozen and cause warpage and sink.
Undercuts, features that stop a part from coming straight out of the mold, come third, because each one needs a slide, lifter, or similar moving tool that adds cost and lead time. Each of these is cheap to fix in CAD and expensive to fix in steel. Our design and engineering team can go through a design with you and tell you what is ready and what is not, before you commit to tooling.
Is a Prototype Ready for Production When It Still Has Flaws?
Yes, if the flaws are known, small, and ones production can settle. The same Stanford guide says to hold off on building a fully integrated, final-looking prototype until well after you think you are ready. What you need instead is a written list of what is still open and who answers each item.
Production has its own check for what remains. A first article inspection verifies a representative part from the first production run against every requirement on the drawing. The prototype does not have to prove everything, only the risky things.
Before you commit, hand the part to someone who has not seen it: an engineer, a prospective user, or your manufacturer. Watch what they do with it rather than asking what they think. Nielsen Norman Group gives the same advice for user research, because self-reported claims and guesses about future behavior are unreliable. Watching someone try to assemble the part costs very little next to finding the same problem after tooling.
Tell us what the prototype is made of, what it has been through, and which dimensions are on the drawing, using the contact page. A quick call to 718-557-9578 also works, and we will say which of these checks is still open.
