Engineering Answers / Prototyping & Validation
The wrong prototype route does not just cost time. It produces evidence you cannot use — a part that passes a fit check and tells you nothing about whether the production process will hold tolerance.
Match the route to the question you are answering. A resin print answers form and fit only. Machined silicone answers material feel for one to five parts. A 3D printed mold gives 10–100 parts in real compound within days. An aluminium bridge tool covers 500–5,000 parts and is the first route that can prove a production cycle. Steel is for volume and for any validation that has to stand up to an auditor.
Prototyping goes wrong in a specific way: the team picks a route by habit, gets parts, and then discovers the parts cannot answer the question they were built for.
A resin print is the fastest way to hold a shape, and it will tell you nothing about how silicone feels in the hand. A machined silicone block gives you the real material and tells you nothing about flash. A printed mold gives you real compound and tells you nothing about cycle time or tool wear.
So the first decision is not "which route" but "which question".
| Stage | Route | Quantity | Lead time | Proves | Does not prove |
|---|---|---|---|---|---|
| Concept, form and fit | Resin 3D print | 1–10 | 1–3 days | Shape, assembly, ergonomics | Material, softness, sealing |
| Feel and durometer | Machined silicone | 1–5 | 3–5 days | Real compound, real hardness | Flash, parting line, thin walls |
| Design iteration | 3D printed mold, cast | 10–100 | 5–10 days | Real compound in a real cavity | Tool wear, cycle time, production tolerance |
| Pre-production | Aluminium bridge tool | 500–5,000 | 3–5 weeks | The production cycle, at lower tool life | Long-run tool wear, steel-tool dimensions |
| Production and validation | Steel tool | 5,000+ | 5–8 weeks | Everything, including audit evidence | — |
Resin print. Buy it for ergonomics and assembly. A wearer trial of a mask frame, a check that a bezel fits a housing, a photo for an investor. Do not use it to judge softness: printed resin at 80 Shore A is not silicone at 80 Shore A, and the difference is obvious in the hand within seconds.
Machined silicone. Cut from a block or a molded sheet, then milled. This is the only no-tool route that gives you the actual compound at the actual hardness. It is excellent for a sealing test on a flat gasket, a grip test, or a colour sample. It is poor for anything with a thin wall — the material deflects away from the cutter — and it cannot produce an undercut or a sharp internal feature.
3D printed mold. Printed in resin, filled by casting or by low-pressure injection. This is the workhorse of silicone development: ten to a hundred parts, real compound, a real parting line you can inspect. What it cannot do is run at production temperature and pressure, so the flash you see is printed-mold flash, and the cycle time is irrelevant to the real tool.
Aluminium bridge tool. The first route that behaves like production. Aluminium cuts faster and costs less than hardened steel, and it will run 500–5,000 shots before the shut-off faces go. Use it when the design is stable but the volume is not yet contracted, or when a launch date cannot wait for steel.
Steel production tool. Where the tolerance, the tool life and the audit evidence all become real. Anything that has to satisfy a notified body — process validation, capability studies, biocompatibility on production parts — belongs here, on the compound and the process that will actually ship.
Printed mold to steel. A part dimension that holds on a printed mold does not automatically hold on steel. Printed cavities are less rigid and the clamp behaves differently, so the same nominal geometry produces a slightly different part. Treat the printed-mold dimensions as indicative and re-baseline the drawing at first article on the real tool.
Bridge tool to steel. This one is usually smoother, but only if the bridge tool was built to the same cavity count and gate layout. A single-cavity bridge tool feeding a four-cavity steel tool changes the fill balance and the pressure at each cavity, and the dimensions move.
Running a wear test or a biocompatibility test on parts from a route that will not be used in production. The data looks like progress and it is not transferable. If the result has to survive a customer's quality review or an auditor, it has to come off the production process — and that is worth knowing before the test is booked, not after.
Most projects that go wrong skip step two. They jump from a resin print to a tool, discover the part is too hard or too soft, and pay for a tool change to fix a decision that a five-part machining run would have settled in a week.
You can run the test, but the result will not transfer. A printed mold is usually paired with a cast or RTV compound cured at room temperature, while production may be LSR injected at 170 °C with a different cure system and a post-cure. Extractables depend on the compound and the cure, not on the shape. Biocompatibility and food-contact testing belong on parts from the process and the compound that will actually be used.
Around 100. Below that, printing the mold and casting is faster and cheaper than cutting aluminium. Above it, the printed mold starts to wear — the cavity surface degrades and the parting line stops closing cleanly, so flash appears and the parts stop being comparable. If the number is 200 and the deadline is tight, an aluminium bridge tool is usually the better call.
Almost nothing about the production process. Machining a silicone block gives you the right material and the right hardness, which is genuinely useful for a feel test or a fit check. It cannot reproduce a parting line, flash, gate location, cycle time, or the dimensional behaviour of a molded part. It also cannot produce undercuts or a thin wall reliably, because the material deflects under the cutter.