CNC Prototype to Production: Where Machining Fits on the Process Ladder

You have a working prototype, a purchase order for 800 units, and a supplier asking whether you want to tool for die casting. Tooling would cut the unit price by more than half. It would also take eight weeks, lock the design, and cost more than the entire first order if the market moves sideways in month four.

The decision is not "CNC or tooling." It is a sequence: which process proves what, at which volume each one stops being economical, and what has to be true about your design before you commit to a tool that cannot be changed cheaply. CNC sits in the middle of that sequence, and for a surprising number of products it is not a bridge at all.

The ladder, and what each rung proves

Each process answers a different question, and skipping a rung means paying to answer it later with harder tooling.

The ratios are relative to the functional prototype and move with part size, alloy and complexity. What matters is the shape of the curve: it falls fast to pre-production, slowly through bridge volumes, then drops sharply once tooling is amortized.

What CNC proves that printing cannot

A printed prototype validates shape. It does not validate material behaviour, surface finish or the tolerance stack in an assembly. An SLS nylon bracket and a machined 6061 bracket of the same geometry are different products.

Machining the functional prototype in the production alloy is the cheapest way to find out three things: whether the tolerances are achievable in that material, whether the assembly actually fits, and whether the finish is acceptable.

Printed parts typically hold ±0.2 mm at best on small features, which is enough to hide a stack-up error that appears immediately at ±0.02 mm. Anodize colour, bead blast texture and machined surface marks are visible decisions, and they are cheaper to change on five parts than five hundred.

The crossover math

The decision between CNC and tooling is arithmetic, and it is worth doing explicitly.

Break-even quantity = tooling cost ÷ (CNC unit price − tooled unit price)

Example: a housing machined at $46, die cast and finish-machined at $18, with a die costing $38,000 including first article and a trim fixture. Break-even = 38,000 ÷ 28 = 1,357 units.

Below about 1,350 units, CNC is cheaper in total spend. Above it, the tool wins on unit cost but not necessarily on total cost, because the calculation usually leaves out four things.

Inventory risk. A tool commits you to a minimum run, often 2,000 to 5,000 pieces. If demand is 900, you are carrying 1,100 units or paying for a mold that sits idle.

Change cost. A design change after tooling means a mold modification, typically $2,000 to $8,000 and two to four weeks, versus a program change on a CNC part that costs an afternoon.

Quality ramp. A new die needs several hundred shots before dimensions stabilize. The first production lot is often the worst lot, and it arrives after you have committed.

Secondary machining. Cast parts usually still need machined bores, faces and threads, and that recurring cost stays in the unit price. Include it and the crossover often moves from 1,300 units to 2,500 or more, which is exactly the range most hardware products live in.

Design freeze discipline

The largest cost saver in this sequence is a real design freeze before tooling. Not a meeting where everyone agrees the design is probably fine, but a documented revision that is locked, with a change process behind it.

The rule of thumb is that the cost of a change multiplies by roughly ten at each stage: cheap on screen, moderate in a prototype, expensive in a fixture, very expensive in a hard tool. That is a rule of thumb rather than a measured law, but the order of magnitude holds.

A proper freeze includes a released drawing and model at a named revision with a change log; all interfaces frozen, including fastener callouts, connector positions and cable routing; finish and colour decided rather than deferred, because a colour change after tooling can mean reworking a whole lot; regulatory requirements identified, since they constrain material and marking after the fact; and a defined change path for anything discovered later.

Freezing late does not make the design better; it moves the discovery into the most expensive phase. If the design is genuinely uncertain, stay on CNC longer.

When CNC is the end process

Some products never cross over, and forcing them is a mistake. CNC stays as production when volume is genuinely low or uncertain; when the part is large and simple, so tooling plus minimum order outweighs the saving; when tight tolerances across many features mean the casting needs full machining anyway; when the revision cycle is faster than the tooling cycle; or when the material does not cast or mold well.

In those cases the right optimization is not "find a cheaper process" but "make the CNC part cheaper": fewer setups, standardized tooling, lights-out pallet running, and consolidating multi-part assemblies into single machined components.

When you do move, the handoff determines how the ramp goes. Send the released drawing, the STEP model at the same revision, the quantity ladder you actually expect including the pessimistic case, any inspection or documentation requirement, and the known problem areas from prototype testing.

At JSHS, a move from prototype to a 500-piece run usually keeps the same fixture concept and adds a second pallet, so the first production parts match the prototypes that were tested. Send the drawing, model and volumes and we will quote each stage separately.

Frequently asked questions

At what volume does CNC stop being economical?

For parts in the 50 to 200 g range, the crossover against die casting typically falls between 1,500 and 3,000 units per year once secondary machining is included. Large parts, tight tolerances and uncertain demand push that number higher.

Should I prototype in the production material?

Yes, wherever the budget allows. Printing in a surrogate hides tolerance, finish and stiffness problems that appear immediately in aluminum or stainless. Machining costs more per part but removes a whole class of late surprises.

Can CNC parts be made without any tooling cost?

For most geometries, yes: standard vises and soft jaws cover the first batch. A complex part may need a dedicated fixture, typically a few hundred to a couple of thousand dollars, and it is usually reusable across orders.

How long does a die casting tool take?

Six to ten weeks for an aluminum die, including design, machining, trial shots and correction. Plan the transition so CNC bridge stock covers demand until the tool is proven, not just delivered.