HD20

<how-it-works/>

Three parts do the work. A wave generator, an ellipse riding on eleven loose 5 mm balls in a printed cage, pushes a flexible toothed ring (the flexspline) into an oval, so its teeth engage a rigid circular spline at only two points. The circular spline has two more teeth than the flexspline, so a full turn of the motor walks the output around by just two teeth. That tiny slip per revolution is the 20:1.

New to strain wave gearing? We wrote a plain-language explainer. And the tooth profile in this drive came out of our own free harmonic maker. You can use it to generate profiles for your own designs.

<specs/>

Reduction	20:1
Tooth profile	Cycloidal, module 0.8
Measured torque	2.1–2.6 N·m @ 0.7 A · 3.0–3.3 N·m @ 1.4 A (peak 3.5 N·m)
Efficiency	48–58%
Backlash	measurement in progress; we'd rather publish nothing than a guess
Motor	Any NEMA 17 with a 5 mm shaft. Tested with an OMC StepperOnline 17HE12-1204S: 42 × 42 × 30 mm, 26 N·cm, 1.2 A, 4-wire.
Lubrication	Super Lube synthetic PTFE grease
Printed parts	10 pieces across 9 unique parts, including 2 shear pins
Fasteners	8 × M3×0.5 × 6 mm button head screws per drive
Materials	PLA throughout, PETG for the flexspline
Version	v1.0 (RevA), 9 August 2026
License	CC BY-SA 4.0

<what-you-need/>

Beyond the bill of materials below, you supply:

• A NEMA 17 stepper with a 5 mm shaft. We used an OMC StepperOnline 17HE12-1204S (26 N·cm, 1.2 A) under closed-loop FOC control. An open-loop stepper on a basic driver will behave differently.
• 4 × M3 screws to bolt the motor to the interface plate. These are separate from the 8 × M3×0.5 × 6 mm screws that hold the drive itself together, which are in the bill of materials.
• A soldering iron for the heat-set inserts, and hex keys.
• PLA for everything, and PETG for the flexspline.
• Super Lube on the wave generator and the tooth mesh. It's in the bill of materials below. Don't run it dry.

<bill-of-materials/>

loading the bill of materials…

<printed-parts/>

Part	Qty	Material
Circular spline (base)	1	PLA
Circular spline (output)	1	PLA
Base preloader	1	PLA
Output preloader	1	PLA
Interface / motor mount	1	PLA
Wave generator	1	PLA
Ball cage	1	PLA
Flexspline	1	PETG
Shear pin	2	PLA

Both shear pins are required. They carry shear load across the output joint directly, so the connection doesn't have to rely on friction from the preloaded screws to resist it. The screws clamp; the pins take the sideways load.

Print settings

No custom temperatures and no special bed prep. The tested units were printed on a Bambu Lab P1S with Bambu PLA (and PETG for the flexspline), using the Arachne variable-width wall generator and one extra wall loop for strength.

Two things are not the default. The output circular spline and the interface both need supports; everything else prints unsupported. And we set seam position to random on all parts, so a single seam line doesn't stack up into a weak spot or a visible ridge on the round surfaces.

The nylon MJF option

There's also a Multi Jet Fusion nylon version of the flexspline, which we had made through PCBWay. Nylon has far better fatigue life than PETG, so it should last considerably longer. Ours cost $24.89 for five pieces, about $4.98 each, before tax, shipping and any import duty, which on a small order can easily cost more than the parts themselves.

Only the flexspline actually has to be made in nylon MJF, because that's the part fighting fatigue. The MJF cage and wave generator have their own geometry, but you can still print them yourself in PLA. So a service bureau order can be just the flexspline, and you print the other two from the MJF set at home.

<the-files/>

Everything lives in robrotics/hd17. Grab the packaged bundles from the releases page. The STEP files are large, so you probably don't want to clone the whole repo.

There are no pre-sliced files. Import the STEP into your slicer and use the print settings above — that way the geometry you print is always the current one.

<assembly/>

<known-behaviour/>

Read this before you conclude you assembled it wrong.

• Torque drops about 10% after running in. A freshly printed unit measured 2.32 N·m; the same unit after further running measured 2.08 N·m at 0.7 A. Plan around the run-in figure, not the fresh one.
• Output spline preload is your biggest tuning knob. Further tightening it took one unit from 2.37 to 2.55 N·m.
• Unit-to-unit variation is real. Two units of the same design differed by up to 18% at the same current. Printed gearboxes are not precision parts.
• Roughly half your motor torque becomes heat at 48–58% efficiency. Size the motor accordingly.
• PLA creeps under sustained load. Holding a heavy static load for hours will slowly deform the circular splines.
• The 1.4 A figures are above the test motor's rating. The 17HE12-1204S is rated 1.2 A, so our 3.0–3.3 N·m numbers were measured over-driven. Treat them as a short-burst ceiling, not a continuous rating. At 1.4 A that motor will get hot.

Hit something we haven't listed? Please open an issue and tell us your version, filament and printer.

<test-data/>

We measured this on a load cell with a 100 mm lever arm rather than guessing from motor specs: seven configurations, 1,500–3,000 samples each, with a no-gearbox baseline to compute efficiency honestly.

read the full torque report →

<make-your-own/>

The full parametric model is public on Onshape. Copy it into your own workspace and change whatever you like: a different motor face, a different output interface, a different ratio. If you build a variant we would genuinely like to see it: open an issue on the repo or tag @robrotics.

open the model in onshape →