From prototype to mass production: what to freeze before tooling
- Prototype and production are different processes with different economics. Getting from one to the other is a transfer, not a scale-up.
- Moving to tooling before the design stops changing converts every later change from a drawing revision into a tool modification.
- The usual mistake is the opposite of the expected one: not tooling too early, but leaving the part on the machining route long after the volume justified a tool.
- Prototype and production routes can run in parallel. Low-volume variants, spares and revised versions stay on machining after the main volume has moved to tooling.
The cost curve, and where the decision sits
Unit cost falls as volume rises, but not smoothly. It falls steeply while fixed setup costs are amortised, then flattens, then drops again when a tool replaces machining. The decision to tool is a bet that the volume will reach the point where the tool has paid for itself before the design changes.
| Annual volume | Typical route | Tooling | Unit cost index | What changes at this step |
|---|---|---|---|---|
| 1–50 | CNC machining | None | 100 | Nothing; the part is made to the drawing and can be revised at will |
| 50–500 | CNC machining, optimised | None | 85 | Setup amortised, fixturing and programme improved, batch sizes raised |
| 500–2,000 | Investment casting plus finish machining | Medium | 55 | Tooling committed; changes now cost a new pattern |
| 2,000–10,000 | Die casting plus finish machining | High | 30 | Tooling committed and amortised over a larger base; changes cost a tool modification |
| 10,000+ | Die casting, process optimised | High | 22 | Cycle time, wall thickness and gating optimised for the specific part |
The break-even calculation behind the third and fourth rows is set out on the process comparison page.
What to freeze before cutting a tool
A tool commits geometry. Everything that is fixed at the same time as the tool cannot be changed cheaply again, so these items should be settled first.
| Item | Why it must be frozen | Consequence of changing it later |
|---|---|---|
| Nominal geometry and all interfacing features | The tool is cut to these dimensions | Tool modification or a new tool |
| Datums and the tolerance scheme | Machining fixtures and the inspection plan are built on them | New fixturing and a re-issued inspection plan |
| Material and finish specification | Shrinkage, gating, draft and finishing all depend on them | Sampling round repeated in a different material |
| Draft angles and wall thickness | Set at the tool design stage and hard to add later | Tool rework; sometimes a new tool |
| Part marking and labelling | Markings are cut into the tool | Engraving operation added, or tool rework |
| Assembly and interchangeability requirements | Determines where tolerances must be tight | Parts that do not fit the assembly they were designed for |
| Expected volume over the tool's life | Determines whether the tool pays back at all | A tool that never recovers its cost |
Why the prototype process should not become the production process by default
The route that made the prototype was chosen for speed and flexibility, not for unit cost. Carrying it into production is a common and quiet source of overspend, because nothing fails — the parts are just made the expensive way for years.
| Dimension | Prototype stage | Production stage |
|---|---|---|
| Primary goal | Learn whether the design works | Repeat the known-good design at the lowest sustainable cost |
| Process choice | No tooling, maximum flexibility | Lowest cost per part at the expected volume |
| Change cost | A drawing revision | A tool modification |
| Tolerance strategy | Prove the tight features are achievable | Hold them; loosen everything else |
| Inspection | Full first article, close measurement | 100% before shipment, with the critical dimensions reported |
| Cost driver | Engineering time and speed | Cycle time, material yield and tool life |
Run both routes in parallel
The transition does not have to be a switch. The usual arrangement is:
- Prototype on the machining route. Test the design in the real material. Iterations cost a revision.
- Quote both routes together at the volume you actually expect, including the finish machining that tooled parts will still need.
- Tool when the volume justifies it, not when the design is merely good enough. The trigger is the break-even quantity, not a milestone in a project plan.
- Keep machining for the tail. Low-volume variants, service spares, custom options and any revision still in flux continue on the machining route. A tooled part with a revised variant usually costs more to support than a machined one.
- Finish-machine the critical features on the tooled part. Casting or moulding brings the part to near-net shape; machining brings the fit to tolerance.
Ten-point transfer checklist
| # | Item | Done when |
|---|---|---|
| 1 | Volume forecast agreed over the tool's life | A number exists, with an honest confidence range |
| 2 | Both routes quoted at that volume | Machining and tooling prices are on the same page, including finish machining |
| 3 | Break-even quantity calculated | Tooling cost divided by saving per part, with the hidden costs listed |
| 4 | Design frozen against the Table 2 list | Every item signed off by the design owner, in writing |
| 5 | Draft, wall thickness and radii reviewed for the new process | Geometry suits casting or moulding, not just machining |
| 6 | Critical features identified for finish machining | Each one has a named process and an inspection method |
| 7 | Tolerance stack reviewed across the new process | The assembly requirement is met by the tooled route, not assumed |
| 8 | First article inspection plan agreed | Which dimensions are reported, and to which revision |
| 9 | Change control process in place | Revision numbering, written confirmation and disposition rules agreed |
| 10 | Route for the tail defined | It is decided how spares, variants and revisions will be supplied |
Where transfers fail
| Failure | Root cause | Prevention |
|---|---|---|
| Fit problems at first production batch | Tolerance stack never rechecked for the new process | Re-run the stack against the tooled route, not against the machined prototype |
| Two or three sampling rounds | Geometry not reviewed for draft, shrinkage and wall thickness | DFM review before the tool is cut, not after the first samples |
| Tool never pays back | Volume forecast optimistic; finish machining left out of the calculation | Break-even calculated with the hidden costs included |
| Revision after tooling | Design not frozen against an explicit list | Written sign-off on the freeze list before tooling is released |
| Spares become expensive | Tail volume left on a tooled route with high minimum quantities | Keep the machining route live for low-volume demand |
| Splitting the route across suppliers | Prototype shop, caster and finishing shop are separate companies | Keep machining, casting and finishing in one place so the tolerance chain and schedule have one owner |
Questions about transferring to production
When should a part move from machining to a tooled process?
When the expected volume exceeds the break-even quantity — tooling cost divided by the saving per part — and the design has stopped changing. Typically that is 500 to 2,000 parts for investment casting, above 2,000 for die casting and above 5,000 for injection moulding, but it is specific to the part and should be calculated rather than assumed.
Can I keep making the prototype version after tooling?
Yes, and it is usually the right answer for low-volume demand. Machining has no tooling and no minimum quantity, so variants, spares and revised versions stay economic on the machining route while the main volume runs on the tool.
What causes most first-batch fit problems?
A tolerance stack that was validated on the prototype process and never rechecked for the production process. Machining and casting have different tolerance behaviour and different set-up structures, so a stack that worked for the machined part is not automatically valid for the cast one. See tolerance stack-up.
Should the prototype and production supplier be the same company?
There is a strong argument for it. When they are different, the prototype supplier's process knowledge does not transfer, the tolerance chain is split across parties, and nobody owns the fit at the handover. One supplier for machining, casting and finishing keeps the route, the stack and the responsibility in one place.
Cite this page
Boyang Hardware. "From Prototype to Mass Production: What to Freeze Before Tooling." www.bycncmachining.com, published 17 September 2026, updated 17 September 2026.
Dated and versioned so it can be quoted with a reliable source date.
Related pages
Get both routes priced together
Send the drawing and the expected annual volume. Machining and tooling options come back on the same page so the transfer decision is made on numbers.