A2.2 — Prototyping techniques
Key concepts
A prototype is a tangible version of an idea — something you can hold, test, show to users, and modify. Prototyping is how designers learn whether an idea actually works, before investing in production. Every prototype answers a specific question: does this fit the hand? is this strong enough? do users understand this interaction?
Low-fidelity vs high-fidelity
Low-fidelity prototypes are rough, quick, and cheap — paper sketches, cardboard mock-ups, foam models, masking tape. They test big questions: is the form right? does the scale work? is the basic interaction sensible? Low-fidelity prototyping belongs early in the process, when ideas are still flexible and cheap to change.
High-fidelity prototypes are close to the final product — 3D-printed parts in correct materials, working electronic circuits, functional mechanisms. They test detailed questions: does the latch feel secure? is the grip comfortable over 20 minutes? does the battery last long enough? High-fidelity prototyping belongs later, when the design direction is settled.
Bad design practice is to jump straight to high-fidelity. A glossy prototype feels like progress, but if the underlying form is wrong, all that effort is wasted. Good designers iterate low-fidelity first, many times, until they are confident in the direction.
Rapid prototyping technologies
FDM (Fused Deposition Modelling) — a heated nozzle extrudes plastic filament layer by layer. Cheap, accessible, good for functional parts. Visible layer lines, limited surface finish, limited to thermoplastics.
SLA / resin printing — a laser cures liquid resin layer by layer. Very fine detail, smooth surfaces, excellent for display models. Brittle parts, needs post-processing (washing, curing), resin is messy and toxic.
SLS (Selective Laser Sintering) — a laser fuses powder. No support material needed, strong parts, complex geometries. Expensive, industrial scale.
Laser cutting — a focused laser cuts or engraves flat sheet material. Extremely fast, precise, repeatable. Limited to 2D profiles (though layered assemblies are possible).
CNC milling — a computer-controlled cutter removes material from a block. Produces parts in real engineering materials (metal, wood, rigid plastic). Slow setup, wastes material, limited to shapes a cutter can reach.
Hand-built physical models — foam, cardboard, clay, MDF, 3D-printed composites. Still essential — they are cheap, fast, and force the designer to make form decisions physically.
Iteration
A single prototype is an experiment with a sample size of one. Iterative prototyping — building, testing, refining, building again — is how designers convert assumptions into knowledge.
Each iteration should:
- State the question being tested
- Build the simplest prototype that can answer it
- Test with real users or against real conditions
- Document what was learned
- Modify the design and build again
Three iterations is usually the minimum for a credible IA. More is better. Each iteration narrows the gap between assumption and reality.
Testing prototypes
Prototypes reveal nothing until they are tested. Tests should be:
- Specific — targeting one question (grip comfort, not "is it good")
- Measurable — produce data you can compare (minutes until fatigue, task completion rate, error count)
- User-involved — watch real users, not the designer trying their own prototype
- Documented — photos, videos, quotes, numerical results
Case studies
Dyson's first vacuum prototype — James Dyson famously built 5,127 prototypes over 15 years developing his cyclonic vacuum. Most were small variations testing one parameter. This extreme iteration allowed gradual refinement impossible to achieve in a single design.
IDEO rapid prototyping of the Palm V — designers sculpted aluminium blocks at different sizes to find the ideal pocket form. Users carried blocks for days, reporting how they felt. No electronics, no screen — just form. This low-fidelity approach defined the product before any engineering began.
SpaceX Starship — visible iterative prototyping at industrial scale. Each prototype tests specific failure modes (tank pressure, engine relight, heat-shield performance). Accepted failures produce more knowledge per dollar than a single attempt at perfection would.
Cardboard aided design — many furniture and product designers still build 1:1 cardboard mock-ups before any CAD work. A chair you can actually sit on teaches more than a perfect render ever will.
Glossary
- Prototype — a tangible, testable representation of a design idea.
- Low-fidelity prototype — a rough, quick, inexpensive prototype that tests broad questions early in the process.
- High-fidelity prototype — a refined prototype close to the final product that tests detailed performance.
- FDM — Fused Deposition Modelling; 3D printing by extruded plastic filament.
- SLA — Stereolithography; 3D printing by laser-cured liquid resin.
- CNC — Computer Numerical Control; automated removal of material by a programmed cutter.
- Iteration — a single cycle of build, test, learn, refine.
- Iterative prototyping — the practice of refining a design through repeated cycles.
- Rapid prototyping — any fast, computer-controlled fabrication method (3D printing, laser cutting, CNC).
Check your understanding
1. When is low-fidelity prototyping more appropriate than high-fidelity?
Early in the design process, when the direction is still being decided. Low-fidelity prototypes are fast and cheap, encouraging experimentation. They let you test big questions (form, scale, interaction) without the sunk cost that makes designers reluctant to change high-fidelity prototypes. High-fidelity comes later, when the direction is stable and detailed questions need answering.
2. Describe three rapid prototyping technologies, giving one advantage and one limitation of each.
FDM: advantage — inexpensive, widely accessible, produces functional thermoplastic parts. Limitation — visible layer lines and limited surface finish. SLA: advantage — very high detail and smooth surfaces, excellent for display models. Limitation — parts are brittle and need post-processing with toxic resin. Laser cutting: advantage — extremely fast and precise on flat sheet material. Limitation — essentially 2D, limited to material compatible with the laser.
3. Why is iterative prototyping more effective than building a single prototype?
A single prototype tests one set of assumptions, which may be wrong. Iteration exposes each assumption to reality, one at a time. User feedback at each stage reveals problems the designer cannot anticipate. Without iteration, the final product depends on guesswork; with iteration, each decision is evidence-based. Dyson's 5,127 prototypes are an extreme example of this principle in practice.
4. A student builds one 3D-printed prototype at the end of their IA to "show the final design." Critique this approach against the course's expectations.
This is a model of the final design, not an iterative prototype. It tests nothing, demonstrates no learning, and produces no evidence for Criterion C (Developing Ideas) or the iteration expected in the design process. A credible IA shows multiple prototypes at increasing fidelity, each answering a specific question, with documented testing and refinement. A single end-of-project model fails to demonstrate design thinking.
Teacher's notes — additional examples and activities
Fidelity increases over time
A useful mental model: plot fidelity on the Y-axis and time on the X-axis. Early iterations sit at low fidelity (sketches, cardboard). Later iterations climb toward high fidelity (CAD, 3D prints, functional prototypes). Iteration is not only "more prototypes" — it is prototypes of increasing resolution as the design direction stabilises.
SOAR framework for ideation drawings
A memorable acronym for presenting design ideas:
- Sketch
- Outline
- Annotate
- Render
Use SOAR to ensure every IA sketch communicates clearly. A sketch without outline is unclear; without annotation the reasoning is hidden; without render the form is flat.
Sketching vs drawing
- Sketching = low-fidelity, fast, exploratory. Used early.
- Drawing = higher fidelity, more precise, communicative. Used later.
Students often confuse the two, producing over-finished "sketches" that cannot be iterated. Encourage genuinely rough exploration early.
CAD vs CAM
- CAD (Computer Aided Design) — designing digitally. Examples: Fusion 360, SolidWorks, Tinkercad.
- CAM (Computer Aided Manufacturing) — converting digital models into machine instructions (G-code, DXF, STL).
A student designs in CAD → exports to CAM → machine produces the part.
Surface vs solid modelling
- Surface modelling — photorealistic appearance, no internal data. Good for visualisation.
- Solid modelling — complete geometric data including internal volumes. Required for manufacture, FEA, and simulations.
Students should know when each is appropriate.
Virtual prototyping with FEA
Finite Element Analysis (FEA) predicts how objects react to physical forces without needing to build and destroy physical prototypes. Benefits: - Lower cost (no material, no tool wear) - Faster iteration (change geometry, re-run in minutes) - Reveals stress concentrations that may not be obvious
Physical testing remains mandatory for regulatory compliance and safety-critical products.
Emerging prototyping technologies
- Haptic technology — applies forces to the user giving tactile feedback. Allows designers to test ergonomic interactions in VR before building physical prototypes.
- Digital humans — virtual figures with motion capture and biomechanics simulation, used to test ergonomic assumptions.
Classroom tasks
Isometric drawing progression: Start with a single cube on grid paper, progress to multiple cubes, then to a cantilever chair drawn freehand with protractor. Add rendering to convey material and depth.
Iterative CAD exercise: Model a simple product in low-fidelity CAD. 3D print or laser cut. Evaluate with a target user. Re-design, re-print. Document each iteration.