3D printed gears: how to make them

A gear is a shape that transmits torque through contact between two tooth flanks. Printing builds that shape layer by layer, so the tooth is only as strong as the bond between layers. That single fact explains most printed gear failures. Under low load the tooth flank carries the force across solid plastic and the gear runs for thousands of cycles. Under high load the flank delaminates and the tooth shears off at the root.

Use printed gears for motion transmission, not power transmission. Timing pulleys, dial drives, camera sliders, robot joint mock-ups, and prototype assemblies that need to spin before the real gearbox arrives. Do not use them as the final drive in a machine that runs eight hours a day at temperature.

Before you model anything, answer three questions. What torque does the tooth actually see? How hot does the housing get? What happens if the gear strips mid-cycle? The answers decide the module, the material and the print orientation. Nothing else matters until those three are fixed.

Printed gears still follow involute rules. The module sets tooth size, and two gears only mesh if they share the same module and pressure angle. Stick to 20° pressure angle, the standard for almost every commercial gear. A 14.5° angle is stronger at the root but less tolerant of the backlash a printer leaves behind.

Add backlash in the model, not on the machine. Printers hold ±0.1 mm on a good day, and a printer that runs fast can be worse. Build 0.2–0.3 mm of backlash into the profile so the teeth do not bind when the pitch circles drift. A tight printed pair will run hot, squeal, and wear the flanks flat within an hour.

Keep the tooth count above 17 for a 20° pressure angle. Below that the root undercuts and the printer cannot resolve the fillet. For small pinions, either raise the module or accept a lower ratio. Do not shrink the module to hit a target center distance, because you will lose the tooth shape that makes the mesh work.

• 1
  ModuleMatch both gears. Module 1–4 prints cleanly on a 0.4 mm nozzle.
• 2
  Backlash0.2–0.3 mm in the model, then check with a feeler gauge.
• 3
  Face widthKeep it between 6× and 12× the module. Wider adds little strength.
• 4
  Bore and hubBore at least 2 mm larger than the shaft for a printed fit, or ream after.

Layer adhesion is the weak axis. A tooth printed flat on the bed has its layers running across the load path, so the flank peels. Stand the gear on its face, or print it on a raft with the axis vertical, and the layers run parallel to the tooth. That change alone can double the torque a printed gear carries before failure.

Material choice follows the same logic. PLA is stiff and cheap but creeps under sustained load and softens near 60 °C. PETG is tougher and handles moisture better, though it is gummy and strings. Nylon with carbon fiber holds the best combination of stiffness and toughness for a working gear, and POM prints well when you want low friction and good wear.

For gears that rub constantly, add a lubricant or pair dissimilar plastics. A nylon gear against a POM gear wears far better than nylon against nylon. It is the same trick used on steel gears with different surface hardness.

If the part will see real torque, abrasive grit, or more than about 80 °C, a printed gear is the wrong answer. Move to machined POM, nylon, or aluminum instead. We machine gears from 6061, 7075, 304 stainless and POM daily, and the same tooth profile that fails in plastic will last indefinitely in aluminum with a proper hub.

Slow the outer wall. Tooth flanks print best at 30–40 mm/s because fast extrusion leaves the corner rounded and the flank undersized. Most slicers default to 60 mm/s or more, which is fine for a bracket and wrong for a gear.

Use a 0.4 mm nozzle or smaller, and set the extrusion width to about 0.4 mm. A wider line rounds the tooth tip and changes the effective pressure angle. Four perimeters on the tooth is a reasonable starting point. More walls help, but the flank geometry matters more than the wall count.

Print a test pair before the real part. A short run of two gears, a few minutes of hand turning, and a check on the center distance will tell you more than any simulation. If the pair binds, open the backlash by 0.05 mm and print again. If it rocks, close it.

Check the first article against the CAD. Measure the over-pin dimension or the span across a few teeth, not just the outside diameter. Outside diameter tells you the tip, and the tip is the least important surface on a gear.

• 1
  OrientationPrint on the gear face so layers run along the tooth.
• 2
  Wall countFour perimeters minimum on the tooth region.
• 3
  Infill60–100% for loaded gears; gyroid or cubic both work.
• 4
  Post-processReam the bore and deburr the tooth edges by hand.