Getting started on prototyping comes down to three moves. Pick the one question your first prototype has to answer, build the cheapest thing that can answer it, and write down what you saw before the next build. The Stanford Biodesign guide to prototyping names one beginner error above the rest: building the whole concept end to end instead of building up from single parts. The sections below take the three moves in order, with the materials and tests that fit each.
Getting Started on Prototyping: What Question Comes First?
Write the question as a sentence before you open any software or order any material. Does this form feel right in the hand, does the mechanism work in principle, or does the part fit the space it is meant for? Each one calls for a different build, and a prototype that tries to answer all three at once answers none of them cleanly.
The Stanford guide advises building just enough to settle one or two narrow questions per round, and it sorts prototypes into four types by effort, starting with looks-like and running through feels-like and works-like to is-like. A looks-like model can be good enough to put in front of users for feedback on sizing and design. A works-like model needs just enough function to answer a key question, and most barely work.
Take a bracket that has to clear an exhaust hanger: the question is clearance, so a printed shape in the right place answers it and surface finish does not matter. Take a handheld device: the question is hand feel, so a block of foam carved to size teaches more than a working circuit board in the wrong case. Take a sculpture sized for a lobby: the question is scale at full size, so a full-size foam or cardboard mockup standing in the space answers it. We would rather throw away a model that answered its question than keep polishing one that did not.
Write down the result you expect before you build. If the prototype could only confirm that expectation and never contradict it, the question was too soft to be worth a build.
Which Materials and Methods Fit a Concept Model?
Foam, cardboard, clay, and 3D-printed parts all work for a concept model. Printing earns its place because it needs no mold, special tooling, or fixtures, so setup cost is minimal. Choose the process by the detail you need: SLA (stereolithography) resolves embossed detail as small as about 0.1 mm, where FDM (fused deposition modeling) needs about 0.6 mm wide and 2 mm high, and both vary by printer.
Dimensions come off a printer with their own error. Protolabs lists ±0.5% with a lower limit of ±0.5 mm for desktop FDM prototypes, so a concept model can show whether a part fits a space but not whether it fits to a tenth of a millimeter.
Printed PLA (polylactic acid) softens around 55 to 60°C, so do not judge heat behavior from a PLA model. We would resist printing in an expensive material at this stage, because a model that costs little is a model you will actually throw away. If the part has to fit something that already exists, start from a scan instead of a tape measure, as covered in 3D scanning for prototyping.
When Does a Prototype Need Real Material?
Once the concept holds, the next build tests function, and material starts to matter. Per Formlabs, the bond between layers in an FDM part is often 30 to 50 percent weaker than the strength within a layer, so the strength you measure on a printed part depends on which way it was loaded. Pick a material that approximates production instead: an engineering resin, a machined plastic, or cast urethane.
We would print the concept model first and commit to a functional build only after the shape stops changing, since changes get more expensive the later they land. The prototyping process post breaks down what each build in the sequence should produce, so you can tell when one stage is finished and the next should start.
How Do You Judge a Prototype?
Define pass and fail criteria before you build, not after you have seen the result. Stanford’s guide recommends a written test plan that ties every test to the specification it checks, which saves iteration cycles. A workable plan covers four things:
- Fit within tolerance.
- Strength under the loads it will see in use.
- Feel: whether the controls and surfaces work for the person holding it.
- Buildability, meaning whether it can be made and assembled as designed, and at what cost.
Make each criterion something two people would score the same way. A lid that opens with one hand can be checked, while a lid that feels easy to use can only be argued about.
Keep a record of every round, because Stanford notes that without one it is hard to tell which changes helped and which hurt. For a mechanical build, annotated photos, a short caption, a parts list, and the CAD (computer-aided design) file are enough. There is no fixed number of rounds, and we would count one as worthwhile only if its changes came from a test result. For the point where another round stops paying off, read iterative design and prototyping.
What Does Production-Ready Look Like?
A prototype is ready to hand to production when the questions it still carries are ones only production can answer. The checks, and what to do about a part that still has flaws, are in when a prototype is ready for production.
Before you get there, check your plan against the common rapid prototyping mistakes, from process choice to post-processing. Prototyping is also one stage of a longer sequence from brief to manufacturing handoff, laid out in product design basics.
Tell us what your first prototype has to answer and what you already have, such as a sketch, a CAD file, or a part to match. Use the contact page, or phone 718-557-9578 to talk it through. We work from printed concept models through CAD, casting, and inspection, so a first build does not have to change hands for the next one.
