MakerBot Print is our newest print-prepration software, which supports native CAD files and STL assemblies,
allows you to interact with all your printers via the Cloud, and many other exciting new features.

Download Now


144:1 Differential Cyclodial

by Corona688 Jan 19, 2016
Download All Files

Thing Apps Enabled

Please Login to Comment

I like this little cheap design !
I'll use that reducer to give a big torque to a little DC motor

I also have a PF35 stepper motor, im using it on a Cnc shield with a 4988 stepper driver, it seems to stay always to almost hot, even when turn down that amps to almost stalling the motor, Does yours run warm also ?? my wiring is (Black, brown one coil),(Orange, yellow 2ed coil) which is how I have them wired to the driver.

Yes, mine run hot also. Usually they're bolted to a metal case. Otherwise you might need a biscuit fan or something.

That is what I was thinking but I am making flat pack gearing system, so I can use it tucked under a 20X20 aluminum profile to make the Vat tilt on a dlp printer im building, its a tight fit to begin with, before going all local on it, ill simply try adding a aluminum heatsink to it to see if that helps. I was mainly wondering if all this heat was normal with these pancake steppers.

A tiny, uncooled heat sink in an enclosed area won't do squat.

What voltage is your motor rated for? If you've got a 5V motor, you could easily be overdriving it. A 12V motor is harder to overload.

I am almost sure it's the 12V one ( not 100% because there is no identifying label on it ) but the data sheet of yours makes me believe its identical as the dimensions are the same and its a 6 wire motor as the data sheet mentioned, I think it came out of a scanner. but then again judging by how warm it stays it could very well be a 5 V ??? I was almost thinking id I change the values of the Capacitor on the Cnc Shield that under that stepper motor if I can reduce the power and still maintain torque

Easy to tell with a multimeter, unplug the motor and measure the resistance across a winding. The 12V ones should be 90 ohms, and the 5 volt ones should be 16.

I get 78 ohms which is close enough hahah But thanks I didn't even think to check it that way! Nice Catch!

Very nice design, thank you :-)
This $2.95 stepper should also fit and help fixing the center pin: www.aliexpress.com/item/35-new-imported-stepper-motor-2-phase-4-wire-step-angle-7-5-degrees-motor/32761727778.html

The huge ratio does nicely compensate the 7.5° steps. But how fast can the output rotate ?
I am planing for a SCARA printer and am looking for a lightweight stepper. But will i be able to achieve let's say one rev per second (60 rpm) ?

What about vibrations due to the eccentric nature of the drive ? Vibrations could be canceled with two "ecc.stl" (of half height) rotating with a phase of 180°.

60*144 RPM input, or 8640RPM, is well into dremel territory -- a tin can stepper has no hope. It's rated for 1800 revolutions per minute, unloaded, absolute theoretical max, and I suspect that's a stretch.

The bigger quadcopters have multipole motors which turn that fast, though. (The smaller ones just use DC motors.) Controlling them will be a challenge, but if you're up for it, could be an interesting project.

Check out my 10:1 cycloid heli: http://www.thingiverse.com/thing:1597419 I'm not sure what RPM its input was, just "lots and lots", but it certainly made a buzzing sound from input vibration.

10:1 Cycloid Heli Gearbox

I power my bike with a multipole brushless RC helicopter motor. And they are more expensive than a Nema17.
The PM35S-048-XTC5 stepper (google for images) at 24V is shown with a max of 1000 PPS (pulses per seconds) which is only 1000 / 360 x 7.5 x 60 = 1250 rpm
For a 10:1 ratio i would use a single stag planetary gear.
Up to 30:1 i would choose a split planetary gearbox: http://www.zakgear.com/Split_Gear.html
A Nema14 can go up to 5000 pps which would yield 6250 rpm. http://www.alltronics.com/mas_assets/acrobat/14HR08.pdf
But that would not make the steppers cheaper nor a big weight reduction.

The compelling thing with your design is that it needs only 4-5 printed parts. But while the wobbling cog rolls into the base without friction, the output housing is pushed around causing friction. That is why it does not work the opposite way. And when adding little bearings that push without friction (and adding a second wobbler to cancel vibrations), the systems gets the complexity of a planetary gear :-/

So maybe a hyperbolic worm gear is the better way for me to go: http://www.thingiverse.com/thing:343922

Well.. for your outer design i am not that sure that there is frcition: https://youtu.be/hMOB9t8UE74

Hyperbolic Worm Gears

A stepper is going to have next to no power at those speeds, is the problem. That's why I suggested a multipole motor - they're actually intended to deliver force at those speeds.

Hypocycloids often don't work in reverse because the line of force is the wrong direction half the time, not because of some supposed massive inefficiency. Internally they roll, not slide. A hyperbolic worm gear on the other hand is going to have wear out the wazoo. Worms are among the least efficient gear reduction systems.

I've made reversible cycloid designs before, incidentally - they had teeth like this: https://media.licdn.com/mpr/mpr/shrinknp_800_800/p/2/005/0b4/012/2687ceb.jpg So that at any point, some part is still exerting force on the inner rotor in the correct direction.

Now i got it, this is nothing but a split/differential planetary gear with only 1 planet and no suns: https://youtu.be/C8d6fEVcT5c
And so the output only has contact to one tooth and the high resolution of a 144:1 ratio gets lost :-(

The original Cycloidal Speed Reducer http://www.thingiverse.com/thing:8348 is different as the 4 "output shafts" always have a sliding contact to the planet where half of them push the blue output forward and half prevent the output from being pushed backwards = no play : https://youtu.be/HJq8Mfd7Xv0

errata: The Nema14 with 1.8° (200 steps) has 5000 / 200 x 60 = 1500 rpm. So only reason for such a high ratio would be high torque at which the printed gearbox would break apart :-/

Cycloidal Speed Reducer in OpenSCAD

Stepper motors have near-zero torque at that speed. That's how fast they can freewheel, not how fast they can push.

and where other files?

  • The top cover with the bearing
  • Central disk (red)
    Without them the model cannot be repeated: (

The gearbox still works without the "top cover" -- that's actually my Kossack extruder. It's listed under "Remixed From". You can attach this gearbox to anything meant to bolt to a NEMA17 motor.

The red(violet) central disk is a 625 ball-bearing. You can't print that, sorry, but you can mail-order them. Just a solid disk of plastic wouldn't work.

My apologies for not answering sooner, thingiverse lost your comment somehow.

I re-checked my math and this is not 132:1, it is 144:1.

Comments deleted.