Designing 3D-Printed Gears and Mechanisms: Module, Pressure Angle, Backlash, and Print Orientation

3D printing makes it trivially easy to generate a gear shape that looks right — most CAD packages and several free web generators will spit out a gear profile in seconds. Getting a gear that actually meshes smoothly, transmits load without stripping teeth, and doesn't bind is a different problem, and it's one that trips up a lot of otherwise-solid FDM designs. This guide covers the parameters that actually matter — module, pressure angle, backlash — and the printing-specific choices that determine whether a printed gear survives real use or strips its teeth on the first load cycle.

The Parameters That Define a Gear

ParameterWhat It MeansWhy It Matters Module (metric) or Diametral Pitch (imperial)Tooth size, expressed as pitch diameter ÷ tooth count (module) or the inverse ratio (DP)Two gears must share the same module/DP to mesh at all — this is the single most common cause of "printed gears that just don't fit together" Pressure angleThe angle of the line of force between meshing teeth, almost always 20° in modern designs (older 14.5° designs still exist in some legacy hardware)Mismatched pressure angle between two gears causes binding or excessive backlash even if module matches Number of teethCombined with module, determines the gear's pitch diameter and, with a mating gear, the reduction ratioVery low tooth counts (under ~17 for a 20° pressure angle) get undercut at the base of the tooth, weakening it — a well-known limit in gear design, not a printing-specific issue Backlash allowanceThe intentional small gap between mating teeth so they don't bindFDM parts need more backlash than machined metal gears because of dimensional variance between prints — designing for zero backlash on a printed gear reliably produces gears that don't turn Face widthHow wide/tall the gear is along its axisDirectly trades off against print layer strength — a wider face spreads load across more layers but adds print time and material

Backlash: Design More Than You Think You Need

This is the single biggest adjustment moving from "a gear that would work if machined" to "a gear that works printed." FDM parts vary from nominal dimensions more than machined parts — typically a few hundredths of a millimeter from over-extrusion, elephant's foot at the base, or general dimensional drift — and that variance eats into whatever backlash you designed in. A metal gear pair might run comfortably with 0.05-0.1mm of backlash; a printed pair usually needs 0.2-0.3mm or more to account for real-world FDM tolerance stack-up. Most gear generator tools (including the OpenSCAD-based ones below) expose a backlash parameter directly — don't leave it at a machinist-appropriate default.

Generating the Tooth Profile

Freehand-drawing an involute gear tooth profile in a general CAD package is possible but tedious and error-prone — the curve is mathematically specific, not a simple arc. Three practical routes:

• OpenSCAD's gears library (several well-established community libraries, searchable as "OpenSCAD involute gear") generates parametric involute gears from module, tooth count, pressure angle, and backlash inputs, and is the most flexible option since it's plain code you can parameterize and version-control alongside the rest of a mechanism design. This site's OpenSCAD coverage has the basics of the parametric workflow if you're new to it.
• Fusion 360's spur gear add-in generates a proper involute profile as a native Fusion feature, which is convenient if the rest of your assembly already lives in Fusion.
• Web-based gear generators (searchable as "involute spur gear generator") export an SVG or DXF profile you can import and extrude in any CAD package — the fastest path for a one-off gear, though less convenient for iterating on a full gear train.

Whichever tool you use, always generate both mating gears from the same module/pressure angle/backlash inputs in the same session rather than eyeballing a "close enough" second gear separately — small profile mismatches between two gears that were supposed to match are a common source of binding that's hard to diagnose after the fact.

Print Orientation and Layer Lines

Gear teeth loaded in normal operation experience bending force at the tooth root — exactly the failure mode FDM layer adhesion is weakest against. Print gears flat, teeth pointing up out of the bed (the standard orientation for any gear disc), so that layer lines run parallel to the gear's face and perpendicular to the direction of tooth-root bending stress. Printing a gear standing on its edge (teeth pointing sideways) puts layer lines running straight across the tooth root in the worst possible direction for bending load, and is a reliable way to strip teeth under load that a flat-printed identical gear would have survived.

Material and Settings for Load-Bearing Gears

SettingRecommendationWhy Wall count / perimeters4+ for small gears, more for larger onesGear teeth are often thin enough that "infill" barely exists inside them — perimeters do most of the structural work Infill40-60%+ for load-bearing gears; low infill is fine for display/prototype gearsUnder load, tooth-root stress propagates into the gear body, not just the tooth itself MaterialPETG or nylon for real mechanical load; PLA is fine for light-duty or prototype fit-checksPLA's brittleness and lower heat tolerance make it a poor choice for anything transmitting meaningful torque or running near a heat source (a geared extruder drive, for example) Layer heightMatch to your general dimensional-accuracy layer height, not a speed-optimized coarse settingTooth profile fidelity degrades visibly at coarse layer heights on small-module gears

Fit-Testing Before Committing to a Full Gear Train

Before printing a full multi-gear assembly, print just one mating pair on a test shaft spacing and check three things: they mesh without binding when turned by hand, they don't have so much backlash that there's audible/visible slop under light finger pressure, and the shaft-to-shaft center distance in your design actually matches the printed pitch diameters — a center distance that's even 0.5mm off from what the pitch diameters call for is enough to cause either binding (too tight) or tooth skipping under load (too loose). Iterating on this one-pair test print is far cheaper than discovering a mismatch after printing an entire gearbox housing built around a fixed center distance.

Printed gears are genuinely capable of real mechanical work — geared extruder drives, print-in-place mechanisms, small robotics gearboxes, clock mechanisms — but only when the module, pressure angle, and backlash are treated as real engineering parameters rather than whatever a generator tool defaults to. Get those three right and orient the print correctly, and an FDM gear will outlast most people's expectations for a "just plastic" part.