Rapid Prototyping

Rapid prototyping makes physical parts quickly to test form, fit, or function before committing to production tooling. The common options are 3D printing (FDM, SLA, SLS or MJF) for speed and low setup, and CNC machining for production-equivalent materials and tolerances. The right choice depends on the question the prototype is meant to answer, not on a default process.

Matching the prototype to the question

A prototype exists to answer a question, and naming that question first is the cheapest decision in the whole process.

Geometry questions

If the question is about geometry, does the part fit in the assembly, does the handle feel right, does the form read as intended, then an additive print answers it at low cost, because the geometry is the point and the material does not need to match production.

Function questions

If the question is about function, will the bracket hold the load, will the gear run true, will the housing survive the heat, then the prototype needs the production material and often the production tolerance, which points to CNC machining or metal additive in the real alloy.

Variation questions

If the question is about variation, will the parts assemble consistently across a run, then a small batch at the production process answers it, because a single prototype cannot show variation. Matching the method to the question avoids both over-spending (a CNC part where a print would do) and under-testing (a print where the load matters).

Process options

The prototyping processes split by what they are good at.

Additive processes

FDM prints common thermoplastics (PLA, ABS, PETG) at low cost and with no setup, which suits concept models and fit checks; its parts are anisotropic, weaker across the layer lines, and show visible layer marks. SLA prints photopolymer resin at fine detail and a smooth surface, which suits appearance models and fit checks where finish matters; its parts are more brittle than the thermoplastics. SLS and MJF print nylon powder into functional parts that need no support removal, which suits moving or loaded prototypes that a filament print could not carry.

Subtractive processes

CNC machining cuts the real production alloy or plastic to tight tolerance and a real surface finish, which suits functional prototypes that must behave like the production part. Metal additive builds complex metal geometry that machining cannot reach, for prototypes where internal channels or lattice weight savings matter. Each process trades cost, speed, material realism, and tolerance, and the choice follows which of those the question needs.

Materials for prototypes

The material decision follows the same logic as the process decision. For a geometry or appearance prototype, the material can be whatever the printer runs (PLA, resin), because the part only needs to show the form.

Production-equivalent materials

For a functional prototype, the material should be the production alloy or its closest practical stand-in: aluminum 6061 for housings and brackets, 304 or 316 stainless for corrosion-critical parts, Ti-6Al-4V where strength and weight both matter, and the production engineering plastic where the part is plastic. A functional prototype in the right material gives a realistic test of strength, stiffness, and heat behavior, which a stand-in material cannot.

Stand-in alloys

Where the production material is hard to machine or expensive, a close alloy (6061 standing in for a harder grade) can answer many questions at lower cost, with the final material confirmed on the last prototype before production.

Cost factors

Prototype cost is dominated by setup, not by per-part cycle time, because a prototype run is short.

Setup dominance

That makes the setup-free processes, additive printing and laser cutting, the low-cost choice for one or two parts, since they avoid the programming and fixturing that CNC pays for. CNC machining earns its higher prototype cost when the part must be in the real material at the real tolerance, because the alternative of testing on a stand-in material would mislead the function test.

Quantity and the cost lever

Quantity matters too: a second or third prototype adds little to the additive cost (mostly machine time) but adds full cycle time to the CNC cost, so the per-prototype economics diverge as the count grows. The cost lever is to use the cheapest process that still answers the question, and to reserve the expensive process for the prototype that actually needs it.

From prototype to production

A prototype is not a draft of the production part; it is a different object answering a different question, and the move to production is a fresh design pass. A part printed or machined as a prototype often carries geometry that suits the prototype process (no draft, thin walls machined from billet, supports on a print) but not the production process (which may need draft for molding, uniform walls for casting, or stamping-friendly bends). The production design adds those features, simplifies the tooling, and often changes the material form (from billet to casting, from bar to stamping), and it has to be qualified in its own right, not assumed to match the prototype. Planning that pass early, with the expected volume and production process in mind, is what keeps a program from carrying prototype-cost geometry into a production run, where it would cost far more per part than a production-designed version.

Prototyping across the design stages

A development program moves through several stages, and the prototyping approach shifts with each.

Concept stage

At the concept stage, the goal is to explore geometry and communicate intent, and cheap, fast additive prints (FDM, SLA) serve it, often with several iterations in a week.

Fit stage

At the fit stage, the goal is to confirm the part assembles with its neighbors, and a more accurate print or a first CNC part in a stand-in material answers it, with the mating parts printed or machined together to check the fit.

Functional and validation stages

At the functional stage, the goal is to prove the part does its job, and the prototype moves to the production material, often CNC machined, to test strength, stiffness, heat, or wear realistically. At the validation stage, the goal is to confirm the design at the production process, and a small batch run at the intended process answers questions of variation and assembly before the full launch. Each stage uses a different process and a different material, and a program that recognizes the stages spends its prototype budget on the right question at the right time, rather than over-building early or under-testing late.

Choosing the prototype quantity

How many prototypes to make is itself a design question, and the answer follows the question the prototypes must answer. A single concept print is often enough to prove a geometry or communicate a form. A functional build may need two or three, to test the part in its assembly and to keep a spare if one fails a test. A fit check across a mating set may need one of each part. A variation study needs a small batch, because variation is a property of a run, not of a single part, and only a batch can show it. The temptation to make a round number, ten or a hundred, is worth resisting until the design is stable, because prototypes made before the design settles are often scrapped when the design changes. The discipline is to make the number that answers the question, and to make the next number when the next question is clear.

Working with prototype suppliers

Working with a prototype supplier is about clarity on the inputs and the question. A useful request states the question the prototype must answer (form, fit, function), the process and material preferred, the quantity, the tolerance the part needs to carry, and the deadline. A useful response confirms what the supplier will make, in what material and process, to what tolerance, and by when, and flags anything in the request that is ambiguous or that will change the outcome. The gap between a clear request and a vague one is the gap between a prototype that answers the question and one that does not, because a supplier builds to the inputs they receive. For a functional prototype especially, confirming the material and the tolerance up front avoids the common failure of a part that looks right but tests wrong because it was machined from a stand-in material or at a looser tolerance than the production part. This page does not assess any supplier; it describes the inputs and questions that make a prototype build productive.

Prototype tolerance versus production tolerance

A subtlety worth tracking is that a prototype tolerance often differs from the production tolerance, and treating them as the same leads to surprises. An additive prototype runs looser than a machined production part, so a fit that assembles on a print may bind on the production part, or a clearance that works on a print may close up in production. A CNC prototype in the production material can match the production tolerance, which is why it is the choice when the function depends on the fit. The practical step is to confirm the critical fits at the production process before launch, not only on the prototype, because the prototype answers the geometry and material questions and the production process answers the tolerance question. Reading the prototype with its tolerance in mind, and not assuming it predicts the production tolerance, is what keeps a fit surprise from appearing at launch.

Checklist

• The question the prototype must answer is named (form, fit, function, or variation).
• The process is matched to that question (additive for geometry, CNC for function).
• The material is matched to the question (stand-in for appearance, production alloy for function).
• The quantity is matched to the question (one for concept, a few for fit, a batch for variation).
• The cost is read against setup dominance (additive for low-count, CNC where the material warrants it).
• The production process is planned, so the prototype design is not carried into production unchanged.

Common mistakes

• Defaulting to one process for every prototype, which either over-spends (CNC for a concept model) or under-tests (a print for a load-bearing part).
• Testing function on a stand-in material and assuming the result carries to the production alloy, when stiffness, strength, or heat behavior differs.
• Reading prototype tolerances as production tolerances, when additive parts run looser than machined parts and a fit that works on a print may fail in production.
• Carrying prototype geometry into production without a redesign for the production process, which raises per-part cost across the run.
• Making too few prototypes to catch variation, or too many before the design is stable, both of which waste the prototyping budget.
• Forgetting that the cheapest prototype is the one that answers the question, not the one that costs the least overall, since a cheap part that answers nothing is wasted spend.
• Skipping the validation stage and launching on a functional prototype alone, so variation and assembly issues appear only at full production, when they are most expensive and slowest to fix.