Most of our OEM/ODM projects start exactly the same way: a prototype to prove the design, low-volume bridge runs to prove the demand, then production tooling. This page explains the four ways to get that first physical part, what each one can and cannot tell you, and how the prototype hands off into production tooling without losing what you learned.
A production mould for a silicone part is cut in hardened steel and compensated for the shrinkage of one specific compound. If a wall is too thin, an undercut is in the wrong place, or the hardness feels wrong in the hand, discovering it after the tool exists means a tool modification, a requalification, and a launch date that moves by weeks. Discovering the same problem on a prototype means a file edit and a reprint.
The economics are asymmetric by an order of magnitude, which is why we quote most new projects as a staged path: a prototype step that answers the open questions, then a production tool cut only after the design is signed off. The prototype is not a delay before the real work - it is the cheapest insurance in the whole project.
Every prototype project starts with a DFM review, not a printer. Within 24 hours of receiving your files, engineering returns a review covering mouldability, wall thickness, draft, undercut placement, gate and parting-line proposals and tolerance feasibility - so the prototype you approve is already the geometry the production tool will cut. See our DFM guide for what the review covers.
Not all prototypes are made the same way, and the right choice depends on which question you need answered - fit, feel, function, or market response.
| Route | What you get | Best for | What it cannot tell you |
|---|---|---|---|
| 3D printed replica (SLA / SLS) | The exact geometry of your CAD file, in a rigid resin or sintered polymer. | Checking form, fit against an assembly, and packaging volumes. | Nothing about silicone: no elasticity, no hardness, no tear behaviour, no feel in the hand. |
| CNC machined master | A dimensionally precise machined part, often used as the master for casting. | Tight-tolerance form checks and appearance models for photography. | Same as printing - the material is not silicone, so functional behaviour is absent. |
| Cast silicone (RTV tooling) | Real silicone cast around a master, in soft tooling, without cutting steel. | Small numbers of functional feel-and-fit parts before any tooling commitment. | Cast parts differ subtly from injection-moulded grain and cure, so dimensional and strength results do not fully carry over. |
| Soft tooling (aluminium bridge tool) | Parts injection or compression moulded in your actual production compound, hardness and colour included. | Functional testing, user validation and market-test units - the route most of our projects take. | Nothing significant - which is why it is also the most expensive prototype option. |
About to order the production tool? A soft tool first usually costs one extra cycle and removes the risk of re-cutting steel — where the two routes pay back.
The decision rule is simple. If the open question is geometry, print it. If the open question is the material - how soft, how tear-resistant, how it grips, whether it smells - the prototype must be moulded in the production compound, and a printed part will actively mislead you: a resin model of a 40A silicone case feels nothing like the real part.
The value of a prototype is lost if the production tool does not inherit what you learned. Four things carry over on our projects.
1. The approved geometry, with shrinkage compensation. The production tool is cut from the same signed-off 3D data as the soft tool, but scaled for the compound's shrinkage, cavity by cavity. First-article inspection then compares the production part back to the approved prototype before the run is released.
2. The material record. The compound, hardness, colour master and cure system validated on the prototype are locked into the production specification, so the first production batch behaves like the parts you tested - including the 200°C post-cure that removes residual volatiles.
3. The inspection criteria. Whatever you measured on the prototype - a pull-off force, a seal gap, a dimensional stack - becomes the acceptance criterion on the production first article, rather than a new argument about what "good" means.
4. The tooling plan. When a project starts in soft tooling, the quotation already states the path: prototype tool, low-volume bridge runs, and the production tool with its cavity count and steel grade matched to your volume. You are never surprised by a second tooling bill halfway into launch.
It depends on the question the prototype has to answer. If you are checking geometry and fit, a 3D printed or CNC replica is enough. If you need to feel hardness, tear strength, odour or how the part behaves in the hand, the prototype must be moulded in the production silicone compound through soft tooling. If you need market-test units in production material before committing to a multi-cavity tool, bridge tooling covers the gap.
3D printed and CNC replicas are not - they approximate the shape but not the feel, elasticity or thermal behaviour of silicone. Soft-toolled prototypes are moulded in your actual production compound, including hardness and colour, so test results carry over to volume production.
Custom prototypes are delivered within 3-7 business days from approved data, and in-stock standard samples are dispatched worldwide via DHL or FedEx within 1-2 days. Engineering DFM feedback on the design is returned within 24 hours of receiving your files.
Simple parts with no undercuts, generous tolerances and non-critical surfaces sometimes go straight to tooling after the DFM review confirms mouldability. The trade-off is risk: a geometry or material problem discovered after the production tool is cut costs a tool modification and requalification. The staged path - prototype tool first, production tool after design sign-off - is quoted as one plan so you decide with numbers rather than hope.
Yes. Soft tooling and bridge tooling exist exactly for this: low-volume production runs moulded in your actual production compound while the final multi-cavity tool is still being cut, so launch schedules are not held hostage by tooling lead time. This is how most of our OEM projects de-risk their first market entry.
Yes. We reverse-engineer physical samples, photos and hand sketches into manufacturable 3D data, then run the DFM review before anything is made. See how the process works on our develop a silicone product without CAD page.