Home / Engineering notes / Sourcing & lead time

Sourcing & lead time

CNC-Machined Prototypes as the Bridge to Injection-Molded Production

Why machine a prototype before cutting a mold, what to validate, and the data package that lets a molding partner quote a tool you can actually run.

Why prototype at all when the end goal is a mold

A molded part and a machined prototype are made by opposite cost structures. The mold is a large, fixed investment that only pays back across volume, while a machined part carries almost no setup and a higher per-unit cost. Because the mold is largely irreversible - once steel is cut, changing the part means cutting steel again - the most valuable thing you can do before committing to tooling is to prove the design in a form that is cheap to change.

That is the role of the prototype: it is the cheap, reversible step that sits between an idea and a tool. Machining a prototype does not compete with molding; it protects the molding investment by surfacing the problems that are trivial to fix in CAD and ruinously expensive to fix after the tool exists. Seen that way, the prototype is not a delay, it is the cheapest insurance on the whole project.

What a machined prototype actually proves

A machined prototype proves geometry, fit and function. You can hold it, assemble it, drop it on a bench, snap it into the housing it has to live in, and find out whether the wall you drew is stiff enough or the clip you specified actually releases. Those are the questions that decide whether a product works, and they are answerable from a handful of parts made directly from the CAD model.

It also proves the assembly, which is where most real-world failures live. A part that is perfect on its own can still wreck a product if it interferes with a neighbour, hides a screw, or blocks a cable. Machining the first few parts of a sub-assembly, rather than just the hero component, is how you learn whether the whole thing goes together - and that lesson is worth more than any single-part tolerance.

What it cannot prove, and where molded samples are still required

Machining proves the shape; it does not prove the material behaviour of a molded part. An injection-molded component is made from a melt whose properties depend on flow direction, gate position and cooling rate, so the same polymer can be stronger along the flow path and weaker across it. A machined block of that polymer has none of that anisotropy, which means a machined prototype can pass a test that a molded part would fail.

This is why a molded sample - usually from a soft tool or a bridge mold - is still the right final check before full production. The prototype earns the right to cut the real tool; the molded sample confirms the tool makes a part that behaves like the one you validated. Skipping either step, or assuming one substitutes for the other, is how a launch slips.

Plastic prototype, metal prototype, or both

The material of the prototype should match the question you are asking. If you are checking fit and assembly of a plastic housing, machine it from a rigid plastic such as ABS or acrylic so the clips and snaps behave like the real material. If you are checking a structural bracket, machine it from the production metal so stiffness and fastening are real. Picking the wrong prototype material answers the wrong question.

3D printing earns its place for complex, non-load-bearing shapes and for very fast visual checks, but for anything that has to carry load, seal, or snap, a machined prototype is the more honest test. The table below is the working comparison most teams actually use when they decide which prototype route to run.

Choosing a prototype route

Prototype routeWhat it proves wellLead timePer-unit costMaterial fidelityBest use
CNC machined (plastic)Fit, snap, clip, assemblyDaysLow at low volumeClose to production polymerFunctional housing and mechanism checks
CNC machined (metal)Stiffness, fastening, loadDaysLow at low volumeProduction metal gradesStructural and mounting parts
3D printedComplex shape, visual, ergonomicsHours to daysLow at low volumeApproximate, often anisotropicForm study, non-loaded geometry
Molded sample (soft/bridge tool)Real material, real processWeeksModerateIdentical to productionFinal validation before hard tool

The point of the table is not that one route wins. It is that each answers a different question, and the cheap mistake is to treat the first prototype as if it had answered all of them. A sensible program runs a machined prototype to fix the design, then a molded sample to confirm the material, and only then cuts the production tool.

Designing the prototype so it de-risks the tool

The prototype is most valuable when it is built to expose the things a mold cannot easily do. Undercuts, very tight tolerances, thin walls and cosmetic surfaces with a high finish requirement all make a mold more complex and more expensive, so the prototype should test whether the part actually needs them. If a feature can be relaxed without hurting the product, the prototype is where you discover that - before the tool is quoted around it.

It also pays to prototype the assembly, not just the part. Designing in a clearance, a draft angle or a locating feature that the molder will need is far cheaper to do now than to discover after the first mold trial. The prototype that is machined with production intent - real materials, real fasteners, real stack-up - is the one that produces a tool-ready design rather than a pretty mock-up.

The data package a molder needs before quoting

A molding partner cannot quote a tool from a rendering. The quote rests on a small set of inputs, and the quality of those inputs decides whether the number you get back is real or a guess with a safety margin baked in. Arriving with a complete package is what turns a quote from a negotiation into an order.

The checklist below is the minimum a serious molder expects. Partners such as MadeInDayin's ODM/OEM product development - which covers R&D design, mold manufacturing, injection molding, secondary processing and electronic assembly - will still run a manufacturability review, but the review is faster and cheaper when the basics are already on the table.

What to put in the RFQ package

InputWhy it mattersWhat to provide
3D modelThe tool is cut from it; its quality is the tool's qualitySTEP or IGES, ideally with the design intent noted
2D drawing with tolerancesDecides which dimensions are tight and which are notCritical dimensions called out, general tolerance stated
MaterialShrinkage, fill and finish all depend on itGrade and any additive (e.g. ABS, PC, PP, PA, POM, PMMA, TPE)
Volume forecastDecides cavitation and whether molding is even rightFirst order and annual volume, even if approximate
Surface and regulatory needsDrives finish, marking and compliance workFinish spec, markings, market certifications required

Two of those inputs decide everything else. The material sets shrinkage and gate strategy, and the volume forecast sets how many cavities the tool should carry. A molder who has to guess either will either quote conservatively or come back with questions that stall the project, so getting those two right is the highest-leverage preparation you can do.

Reading a molding quote once the prototype has done its job

A molding quote has two very different numbers in it: the tooling cost and the unit cost. The tooling is a sunk, upfront sum amortised across the parts; the unit cost is what you pay per piece forever after. A prototype that has already de-risked the design is what lets you trust both numbers, because the tool being quoted is a tool for a part you have already validated rather than a hope.

Watch the assumptions. A quote that assumes a tolerance or a finish you did not ask for is not a gift; it is a cost you may not need. The prototype gives you the evidence to say which tolerances matter and which do not, and that evidence is what keeps the unit cost where it should be instead of where a cautious estimator put it.

Handing off to a production partner

When the prototype and the data package are ready, the handoff is straightforward: send the model, the drawing, the material and the volume, and let the partner run a mold-flow and manufacturability review. A partner with in-house tooling and molding - such as a dedicated plastic injection molding service - can take the part from review through tooling to molded samples without the model leaving one organisation, which keeps the feedback loop short.

The advantage of a single-source ODM/OEM operation such as MadeInDayin, which runs both a China plant and a Vietnam plant and pairs molding with secondary operations like painting, silk-screen, pad printing and electronic assembly, is that the molded part does not then become a logistics problem. The prototype has done its job the moment the partner can quote a tool you trust and a unit cost you can plan around.

A practical sequence that avoids the expensive mistakes

The sequence that consistently saves money is short. Machine a functional prototype from the production material and assemble it. Fix what the assembly reveals. Send the corrected model and the data package to a molder for a mold-flow review and a quote. Run a molded sample from a bridge tool. Only when that sample passes do you cut the production tool.

Every step before the hard tool is reversible and cheap relative to the tool itself. The discipline is simply to treat the prototype as a question-asking device rather than a formality, and to keep the questions coming until the part is tool-ready. That is the entire difference between a tool that pays back and a tool that gets re-cut.

Where this leaves you

A CNC-machined prototype is not an alternative to injection molding; it is the step that makes molding safe to commit to. It answers the fit, function and assembly questions that decide whether a design is worth tooling, and it produces the evidence a molder needs to quote a tool you can actually run. Done well, it is the cheapest money you will spend on the whole program.

The mistake to avoid is treating the prototype as a box to tick. A prototype that was machined without production intent - wrong material, no assembly, no tolerances - proves almost nothing, and a tool cut from it carries every assumption the prototype failed to test. Build the prototype to be wrong about as little as possible, and the tool will be the easy part.

Frequently asked

Do I need a prototype if I am confident in the CAD model?

The CAD model proves the geometry is buildable, not that the product works. Fit, snap, assembly and material behaviour are all discovered by holding and using real parts, and those discoveries are cheap now and expensive after the tool is cut. A prototype is how you convert confidence into evidence.

Should the prototype be the same material as production?

For functional checks it should be. A plastic prototype of a housing should use a rigid plastic so clips and snaps behave; a metal bracket should use the production grade so stiffness is real. Using the wrong material answers the wrong question, which is worse than no prototype at all.

Can a 3D print replace a machined prototype?

Only for form and visual checks. For anything that carries load, seals or snaps, a machined prototype is the more honest test because it uses production-like material and processes. 3D printing is fast and cheap for complex shapes, but it does not prove structural or assembly behaviour.

What does a molder need to quote a tool?

At minimum a 3D model, a 2D drawing with the critical tolerances called out, the material grade, and a volume forecast. The material and the volume decide cavitation and shrink strategy, so getting those two right is the highest-leverage preparation before asking for a quote.

Why run a molded sample before the production tool?

A machined prototype proves the shape; a molded sample proves the material behaviour of the real process, including flow-induced anisotropy and shrinkage. Confirming the part on molded samples is the final check that prevents cutting a production tool for a part that does not behave as validated.

GE
Gopetrel Engineering

Application engineers and machinists who quote, program and inspect the parts described here. Written from production experience, not from a catalogue.

Send a part for a written quote

The price comes back with its assumptions stated, at the quantity you actually need.

Request a quote