Prototypes and simulation answer different questions at different costs. How to choose between 3D printing, machined parts, soft tools, FEA and mould flow, and why the best programs use both deliberately.
A start-up once came to us after building eleven rounds of 3D-printed prototypes of a clip-on device. Each round took two weeks. Each one looked better. None of them answered the question that mattered most, whether the snap-fit arms would survive 500 cycles in the production material, because a printed nylon part behaves nothing like an injection-moulded acetal one. Two days of FEA and a single machined acetal sample answered it, and showed the arms would crack at around 80 cycles.
Prototyping and simulation are both ways of buying certainty before you commit to tooling. The trick is to buy the right certainty, at the right time, at the right price.
Start with the question, not the method
Every test should answer a specific question. The most useful ones fall into a few groups:
- Form: does it look and feel right? Does it fit the hand, the space or the mating parts?
- Fit: do the parts assemble, align and close as intended?
- Function: does it do the job, under load, temperature and repeated use?
- Manufacturability: will it mould, fill, cool and eject properly, without sink, warp or weak weld lines?
- Durability and compliance: will it survive drop tests, fatigue, chemicals and certification tests?
Different methods answer different questions well.
What each prototyping method is good for
- 3D printing (FDM, SLA, SLS, MJF): fast and cheap for form and basic fit. Poor for mechanical performance, because the materials, layer structure and residual stresses differ from moulded parts.
- CNC machining from production-grade resin: good for fit and many function tests, in the real material. No moulded-in stresses, weld lines or fibre orientation, so it can still flatter the design.
- Vacuum casting: useful for small batches of cosmetic or user-trial parts in polyurethane resins that mimic production materials.
- Soft or prototype tooling, such as aluminium or pre-hardened steel: the only way to get true moulded parts in the production material before the production tool. More expensive and slower, but it answers manufacturability and many durability questions properly.
What simulation is good for
- Finite element analysis (FEA): stresses, deflections and strains under load. Ideal for snap fits, living hinges, bosses and structural ribs, and for comparing design options quickly. Results depend on realistic material data, which for plastics should account for temperature, strain rate and, for filled materials, fibre orientation.
- Mould flow analysis: filling, packing, cooling and warpage. It predicts short shots, weld lines, air traps, sink and shrinkage before the tool is designed, and lets you test gate locations and cooling layouts virtually.
- Tolerance analysis: whether the stack-up of real manufacturing variation will still let parts assemble and function.
Simulation is fast and cheap once the model exists, and it lets you test ten options where a prototype tests one. But it only answers the question you set up, under the assumptions you gave it.
Combining them well
The most efficient development programs use both deliberately:
- Early concept: quick 3D prints for form and ergonomics; simple FEA to compare structural concepts.
- Detailed design: FEA on every critical feature; mould flow before the tool design is frozen; machined or cast samples in production-grade material for function.
- Pre-production: prototype tooling where the risk or cost of failure justifies it; then confirm simulation predictions against real parts and refine the models.
Each method checks the other. Simulation tells you where to look in a prototype; the prototype tells you whether the simulation assumptions were right.
A simple rule for choosing
Ask three questions: what decision will this test support, what is the cheapest method that answers it credibly, and what will it cost if we are wrong? A wrong answer on form costs a re-print. A wrong answer on durability can cost a recall. Spend accordingly, and build these choices into your DFM process rather than leaving them to habit.
