Most prototype material decisions are made by habit: aluminium because the last bracket was aluminium, ABS because the production part will be ABS. Both habits can be right, but neither is a reason. The question is what the prototype has to prove before the next design freeze.
Start with the test, not the material
Write down the one thing that would make the prototype useless if it failed. The answer usually falls into one of three buckets, and each points somewhere different.
- Fit and ergonomics — the part has to assemble and feel right. Almost any material works; speed and surface finish matter more than strength.
- Function under load — hinges, clips, threads, pressure. The material modulus and fatigue behaviour are the test, so substitute materials invalidate the result.
- Manufacturing intent — the part must behave like the moulded or cast production version, including wall thickness, draft and sink marks.
When plastic is the right call
Plastic wins when the design is still moving. Machined ABS or POM is fast, forgiving and cheap to iterate, and a printed nylon part can carry a snap fit that would be impossible to cut. For cosmetic shells, a painted SLA part gets close to a moulded surface without a tool. The limitation is stiffness: a 2 mm plastic wall will flex where a 2 mm aluminium wall would not, so a plastic prototype can hide a stiffness problem until the production part arrives.
When metal earns its cost
Metal is the honest choice when the part carries load, holds a thread that will be used repeatedly, or has to survive heat. Al 6061 is the default for machined housings and brackets: stable, weldable, anodises predictably. Al 7075 buys strength at the cost of corrosion resistance. Stainless is for medical, marine and food-contact work, and titanium only when mass is the constraint that cannot be met another way. Metal also gives you a true datum surface, which matters if the prototype will be used to check a fixture.
The crossover is quantity and load
At one to twenty pieces, machining plastic is usually cheaper than machining metal of the same geometry, and printing is cheaper again. Above roughly fifty pieces, the calculus changes: a machined aluminium part can approach the cost of a printed one because print time scales with volume while machining time scales with setups. The crossover is not a fixed number — it depends on the envelope and the number of faces — but it is worth asking for both prices on the same drawing before committing.
A decision order that holds up
- If the part will be moulded, prototype in the production plastic where possible — or accept that the test is about shape only.
- If the part carries a thread, a seal or a preload, use the production metal.
- If the part only has to look right, print it, paint it, and spend the money on the next iteration.
- If you are unsure, ask for one of each. Two parts on one drawing is a cheap way to end the debate.
The expensive mistake is not choosing plastic or metal. It is choosing a substitute material, passing the test, and discovering at tooling that the production material behaves differently. State the substitution on the drawing, and treat any result from it as provisional.