2SOLID AI — Product Design
Mechanical & 3D Prototyping
Mechanical CAD, 3D prototyping, and design validation — parts that assemble, enclosures that ship.
Scope a prototype§01 — Instrument
ω₂ = −(N₁/N₂)·ω₁
the meshing relation the hero enforces tooth-for-tooth — external gears counter-rotate, and the minus sign is the physics
gear.js · nine gears, four chains, one drive angle
Grab any gear and the whole train follows your hand — every rotation derived from the drive through the exact meshing relation, phase-locked so the teeth genuinely interlock.
§02 — The gear train
Datasheet — kinematics, not illustration.
The mechanism above obeys the same definitions a gear designer writes on a drawing. Each row is computed in the scene, not drawn by hand.
| 01Module | m = d / Nmm | pitch diameter per tooth — the size unit of gearing |
|---|---|---|
| 02Pitch radius | r = m·N / 2 | every gear sized from tooth count and module |
| 03Pitch circles | tangent | centre distance = m·(N₁ + N₂) / 2 |
| 04Mesh relation | rot_b = −(N_a/N_b)·rot_a + C | phase term C keeps the teeth interlocked |
| 05Tooth thickness | ≈ 0.5 × pitch | half the circular pitch — seats into the mating gap |
| 06Tooth height | ∝ m | uniform teeth — big gears don't grow long teeth |
| 07Train in scene | 24→13→17 · 24→34→12 · 16→11 · 30→14 | tooth counts, four chains off two roots |
§03 — The argument
Mechanism design is arithmetic you can hold.
Module, tooth count, pressure angle, backlash — a gear pair either meshes or it doesn't, and the drawing tells you before the part exists. We run the same discipline on every mechanical assembly: the CAD is the claim, the prototype is the test, the validation record is the verdict.
§04 — Scope of the practice
From mechanism to validated part.
Model
Parametric mechanical CAD
Solid models driven by parameters — module, wall, clearance — so a design change propagates correctly instead of being re-drawn.
Stack-up
Tolerance & fit analysis
Datums first, then fits — worst-case and statistical stack-ups, ISO 286 fit classes, GD&T to ASME Y14.5.
Prototype
Additive & machined prototypes
FFF, SLA, SLS, or CNC — the process chosen to answer the open question, with the accuracy that question requires.
Verify
Assembly & validation
Fit checks, interference review, functional test, first-article inspection — evidence recorded against the requirement.
§05 — Datum
The numbers mechanical parts answer to.
Standard values of the discipline — cited because they are real, and because a prototype is only as honest as the figures behind it.
| 01Pressure angle — standard involute | 20° | full-depth tooth form, ISO 53 / AGMA datum |
|---|---|---|
| 02Addendum / dedendum | 1.0 m / 1.25 m | standard full-depth teeth — clearance 0.25 m |
| 03Fit classes | ISO 286 | H7/g6 clearance · H7/k6 transition · H7/p6 press |
| 04GD&T | ASME Y14.5 | position, runout, profile — datums before tolerances |
| 05FFF layer height | 50–300µm | fused filament — the working range per nozzle |
| 06SLA layer height | 25–100µm | resin process — fine features, brittle |
| 07SLS — typical material | PA12 | nylon powder — functional parts, no supports |
§06 — Deliverable classes
What a prototype iteration leaves behind.
Representative record of one iteration — the class of documentation that makes hardware iteration engineering instead of tinkering.
Prototype cycle — the stated question
representative record · not a client file
- QUESTION
Does the carrier mechanism hold alignment through the full stroke?
every iteration is built to answer one thing
- MODEL
Parametric assembly — module, clearances, and fits as driven dimensions
- PARTS
Process selected per question — FFF for form, SLA for features, machined for loads
- MEASURE
Fits checked against ISO 286 classes — critical features measured, not assumed
- VERDICT
Pass / revise / re-ask — recorded with the evidence, fed back into the model
- OUTPUT
Controlled CAD revision + test notes — the next iteration starts from record
Hardware to iterate?
Mechanical CAD, tolerance analysis, and prototypes built to answer real questions.