Engineering Answers / Tooling & Cost
The expensive mistake is not choosing the wrong tool — it is buying both, in the wrong order, and paying for the design lesson twice.
Soft tooling is machined aluminium, usually single cavity, built fast and cheap. It produces real parts in the production process — 10 to 2,000 parts in 7 to 12 days — which makes it the right choice for design validation, clinical builds and market tests. Steel is NAK80, S136 or H13 hardened to HRC 48–54 and rated 100,000 to 500,000+ cycles, built in 15 to 25 days. Steel pays back when annual volume and programme life are known; soft tooling pays back every time the design is not yet settled.
Soft tooling and steel tooling are not good and better versions of the same thing. They answer different questions.
Soft tooling answers: does this design mould, assemble, test and sell? It is machined aluminium, usually a single cavity, built quickly and cheaply, and it runs in the same process as production — real LSR injection or real compression moulding, not a 3D print standing in for one.
Steel tooling answers: can I make this part, at this tolerance, at this volume, for the next three years? It is NAK80, S136 stainless or H13 hardened to HRC 48–54, rated from 100,000 to 500,000+ cycles, with the cavity count and the shut-off quality that volume demands.
Choosing between them is a scheduling question as much as a financial one. The wrong order costs more than the wrong tool.
The soft tool figure is the one that changes programme schedules. A medical device company that needs twenty units for a bench test, or a consumer brand that needs five hundred units for a launch photo shoot and a trade show, cannot wait fifteen to twenty-five days for steel. It does not need to. Those parts can come off an aluminium tool, in the production material, with the production process, in under two weeks.
| Situation | Route | Reason |
|---|---|---|
| Design not frozen, changes expected | Soft | A design change on aluminium is cheap; on hardened steel it is a new cavity |
| Clinical or bench builds, tens to hundreds of parts | Soft | Speed matters more than cycle life |
| Market test or limited launch | Soft | Volume does not justify steel |
| Design frozen, annual volume above roughly 20,000 | Steel | Tool life and per-part labour dominate |
| Tight tolerance across a long run | Steel | Aluminium wear moves dimensions |
| Multi-cavity requirement | Steel | Cavity alignment on aluminium does not hold at count |
| Regulated programme with a frozen design | Steel | The validation batch should come off the tool you will ship from |
The last row deserves emphasis. In a regulated programme, the parts used for validation should come from the production tool, because the submission describes a process. A soft tool is a legitimate bridge, but it is not the tool you validate against if the design is already fixed.
Steel too early. The design changes after the first moulding trial — a wall is too thin, a parting line shows on a cosmetic face, a seal does not seat. The cavity has to be re-cut or scrapped. On a hardened multi-cavity tool that is the most expensive way to learn something a 7-day aluminium tool would have taught.
Steel too late. The launch date is fixed, the design is frozen, and the team spends six weeks running prototypes and re-validating because the steel tool was ordered after the marketing deadline rather than before it. The tool itself was never the problem; the sequencing was.
Both, badly. A soft tool is built for validation, then a steel tool is built from the same drawing without using the soft tool's measured shrinkage — so the steel tool repeats the same dimensional error, at ten times the cost.
The value of a soft tool run is not only the parts. It is the data:
Those five outputs go straight into the production tool's DFM. Done this way, the staged path costs one extra tooling cycle and removes most of the risk of a steel re-cut. Done badly — soft tool for parts only, steel tool from the original drawing — it costs the cycle and teaches nothing.
A tooling quote that can be compared has to state the same assumptions on both sides:
Ask for both options in one quote: the staged path and the single-step path. The comparison is usually clearer than either number alone.
"How many parts do you need before the design stops changing?" If the answer is more than zero, a soft tool is probably part of the plan. If the answer is zero — the design is frozen and validated elsewhere — steel from day one is correct and the staged path is just delay.
Soft tooling buys speed and design freedom: 10 to 2,000 real moulded parts in 7 to 12 days, at a fraction of steel cost. Steel buys cycle life and dimensional stability: 100,000 to 500,000+ cycles over 15 to 25 days of tool build. Use soft tooling while the design can still change, steel once it cannot, and carry the soft tool's measured data into the steel tool so you pay for the lesson only once.
Yes, if the parts meet the specification and the tool life covers the run. Soft tooling is used for limited launches, market tests and bridge production all the time. What it cannot do is hold a tight tolerance across a very long run, because the aluminium wears and the cavity dimensions drift. For a few thousand parts that drift is measurable but manageable; for a hundred thousand it is a recall.
The honest answer is a range, not a number: 10 to 2,000 parts depending on geometry, material and how carefully the tool is run. Sharp corners, hard grades and abrasive compounds shorten it; simple geometry and gentle ejection extend it. Any supplier who quotes a single guaranteed shot count for aluminium is guessing.
Not physically — the aluminium cavity is not a steel cavity. What can be carried over is the design data: the parting line, gate position, ejection layout and the measured shrinkage from the soft tool run. That shrinkage measurement is the most valuable output, because it replaces the 2.5–3.5% book value with the number your actual geometry produces, and the steel tool is then cut correct the first time.
Yes, when the programme is already committed. If the design is frozen, the annual volume is confirmed and the product life is two years or more, building a soft tool first adds a full tooling cycle and a validation loop for information you already have. The staged path is insurance against design change, not a universal best practice.