A prototype is a physical model built to answer a question about a product before it goes into production, and the question decides the method. A model for judging size and grip can be printed quickly in almost any plastic. A part that has to survive a fatigue test needs to be made in the production material, or something very close to it, by a process that gives similar properties. Choosing a method before the question is clear is an easy way to spend prototype money learning the wrong thing.
Most programmes need more than one kind. Early models test form and ergonomics; the last builds before launch are made as close to production as possible, with fit and function tested in between. Each round costs more per part and answers harder questions.
What a prototype is for
- Form. Size, proportion, appearance and ergonomics. Material properties barely matter here; geometry and surface do.
- Fit. Whether parts assemble with each other and with bought-in components, cables and fixings, and whether there is room to build and service them. Dimensional accuracy matters more than strength.
- Function. Whether the mechanism works, the seal holds and the part carries its load or survives temperature, vibration and drops. Material properties and process effects matter most for this kind.
- Presentation. Models for investors, customers, trade shows and photography, finished to look like the real product. They may not work at all.
- Pre-production. Low-volume builds using production or near-production processes, made to try the assembly, supply units for testing against the specification and put early units in front of customers.
3D printing processes
FDM
Fused deposition modelling extrudes a thermoplastic filament through a heated nozzle, one layer at a time. It is the cheapest way to get a shape in the hand, and engineering filaments such as ASA, nylon, polycarbonate and carbon fibre filled grades make useful jigs, fixtures and brackets. Parts are visibly layered and much weaker across the layers than along them, so thin snap features, and parts loaded across the build direction, are better made another way.
SLA
Stereolithography cures a liquid photopolymer with a laser or projected light. Surface finish and fine detail are excellent, which makes it the usual choice for appearance models and for the masters used in vacuum casting. Standard resins are brittle and some soften when warm. Engineering resins imitate ABS, polypropylene or rubber in some properties, and most degrade in sunlight over time.
SLS
Selective laser sintering fuses nylon powder, usually PA12 or PA11, with a laser. The surrounding powder supports the part, so complex shapes and even moving assemblies print without support structures. Properties are much more uniform in all directions than in FDM, and SLS parts make good functional prototypes, snap fits and clips included, as well as low-volume end-use parts. The surface is slightly grainy and can be dyed or smoothed.
Metal powder bed fusion
Laser powder bed fusion melts fine metal powder layer by layer, in alloys such as AlSi10Mg aluminium, 316L stainless steel, Ti-6Al-4V titanium and several nickel alloys. It can make metal parts that machining cannot, such as internal channels and lattices, and after heat treatment dense parts reach static strengths comparable with cast or wrought material, depending on the alloy. Parts need supports, stress relief and removal from the build plate, and the faces that locate or seal usually need machining. As-built surfaces are rough, which lowers fatigue strength unless they are machined or polished.
CNC machining
A prototype machined from billet is made of the real material, with real strength, stiffness and thermal behaviour, and tolerances as good as production machining. For parts that will be machined in production, it is the production process. For parts that will later be moulded or cast, it gives the right material made by a different process: a machined nylon part has no weld lines or moulded-in stress, and may outperform the moulded part it stands in for. Tool access also limits machining, so features designed for a mould may need simplifying. On the fishing reel programme, the detailed models were prepared ready for CNC machining of the prototypes.
Vacuum casting
Vacuum casting makes copies of a master pattern, usually an SLA print, by pouring polyurethane resin into a silicone mould under vacuum. Resins are available that imitate ABS, polypropylene, polycarbonate, clear parts and rubbers across a range of hardness, and they can be pigmented to a colour. Each silicone mould yields a limited number of parts, commonly quoted at around 20, so the method suits batches of tens for user trials and pre-launch samples. The cast parts match the imitated material in some respects and fall short in others, heat resistance and long-term ageing especially.
Soft and rapid tooling
Where parts need to be moulded in the real polymer, a simplified mould in aluminium or soft steel produces true injection mouldings in the production material. For very small runs, a printed insert in a standard mould frame can do the same. These parts show the weld lines, sink, warpage and shrinkage the production part will have, which no other method reproduces. Tool life runs from hundreds to many thousands of shots depending on the tool material and the polymer, so rapid tooling can also bridge into early production. The trade-off is the tool itself: changes to the part mean changes to the tool.
Fabrication
For frames, enclosures, chassis and structures that will be fabricated in production, a prototype made from laser-cut and folded sheet, plate and standard sections is usually the most representative option. It uses the production materials and joining methods, shows distortion and fit-up problems early, and can be modified with a grinder and a welding set. The hydrogen production prototype is one example, a fabrication-ready prototype system of frames, process equipment and pipework.
Comparing the methods
| Method | Typical uses | Materials | Strengths | Limits |
|---|---|---|---|---|
| FDM | Form models, early fit checks, jigs and fixtures | ABS, ASA, PLA, nylon, polycarbonate, filled grades | Low cost per part, quick to change, large parts possible | Visible layers, weak across layers, modest accuracy |
| SLA | Appearance models, fine detail, casting masters | Photopolymer resins, some imitating ABS, PP or rubber | Smooth surface, fine detail, good accuracy | Brittle or heat-sensitive resins, degrade in sunlight |
| SLS | Functional prototypes, snap fits, ducting, low-volume parts | PA12, PA11, glass-filled nylon | No supports, complex shapes, fairly uniform properties | Grainy surface, narrow material range |
| Metal powder bed fusion | Complex metal parts, internal channels, lightweight brackets | Aluminium, stainless steel, titanium and nickel alloys | Shapes machining cannot make, dense metal parts | Supports, post-machining, rough surfaces, build size |
| CNC machining | Functional and fit prototypes, parts machined in production | Most metals and engineering plastics | Real material, tight tolerances, good finish | Tool access, cost rises with complexity, no moulding effects |
| Vacuum casting | Batches of tens for trials and samples | PU resins imitating ABS, PP, PC, clear parts and rubber | Production-like look, colours, several copies from one master | Short mould life, properties only approximate |
| Soft or rapid tooling | Moulded parts in the production polymer, bridge to production | Most thermoplastics | True moulded parts, shows moulding defects early | Tool investment, part changes need tool rework |
| Fabrication | Frames, enclosures, structures, test rigs | Steel, stainless steel and aluminium sheet, plate and sections | Production materials and joints, easy to modify | Distortion, skilled fit-up, less suited to small detailed parts |
Many builds combine methods. A functional prototype of a hand-held instrument might pair a machined aluminium chassis, which carries the loads and the heat, with SLA covers painted to the final colour and an SLS clip that has to flex. Each part is made the way that answers its own question.
When analysis replaces a prototype
Some of the questions a prototype is built to answer can be settled by simulation on the 3D model, sooner and for several design variants at once. Whether a bracket stays inside a deflection limit, how hot a component runs in a sealed enclosure, what a drop does to a housing, or how air moves through a duct can all be addressed by stress analysis, drop test simulation, thermal or CFD work. Where the analysis gives a clear answer with margin, a physical prototype for that question is sometimes not needed. Prototypes remain the better tool for perception and feel, assembly by real hands, sealing at real tolerances, and anything where the inputs to an analysis are themselves uncertain.
Testing against the specification
A prototype earns its cost when it is tested against something. Before building, list which requirements in the specification it is meant to show are met, how each one will be measured and what result counts as a pass. A functional prototype that works on the bench has told you little unless the loads and cycle counts were the ones the specification sets. Our guide to writing a product specification covers making requirements measurable in the first place.
How SixGate™ can help
At the Prototype & Test stage, SixGate™ provides a prototype or pilot build and tests it against your specification. Prototypes can be additive, machined or a mix of the two, in polymers, composites, aluminium or stainless steel, as functional prototypes, presentation models or low-volume pre-production builds, with finishing where it is needed. Where analysis answers the question, a physical prototype is sometimes not needed, and the parts that are built go to the questions only a physical part can settle.