Falcon Cycloidal

And you almost certainly will, with care. We are working on similar concepts (3D printed cycloid gearbox) to lift a robot.

This is a dangerous rule of thumb. In real life application, test for harmonics / cyclic loading before assuming no backdriving, or even trusting your suppliers datasheet that it doesn’t backdrive.

I can’t share details, but think a mess in the millions of dollars.

Fortunately FRC robots are considerably cheaper than that. And as we don’t usually test for efficiency to begin with on any gearbox, it’s a rule of thumb that lets you know how poorly your gearbox is functioning without actually measuring the power in and out of your gearbox.

EDIT: sudden shock loads and nonlinear friction dynamics will change the equation from a simple 50/50 over/under test, but this is true for any gearbox - not just cycloids. We’ve all seen situations where a stuck mechanism works after giving it a couple firm whacks, and gearboxes can be no different.


I bought one of those 10 or 15 years ago thinking I would use it for an FRC robot. I never did. It’s still sitting here. It was an interesting thing to study, but it wasn’t light, and the reduction it provided was never in quite the right range for my application. Now, if we had something purpose-designed for easier FRC interfacing, I would be interested again.

Didn’t we open that up one year, and discover that it was just a chunky planetary?

1 Like

If you mean this, it really does have a cycloidal reducer inside. See the picture linked in my earlier post.

This is correct. This older Harbor Freight unit is truly a cycloidal. The thing we bought more recently @CarlosGJ was just a chunky planetary.

Perhaps you have seen Jame Bruton’s posts on 3d printed cycloidial drives, but if not - check it out … 3D Printed Cycloidal Drive V2 - Much Better! - YouTube

This isn’t actually an update of any real value, just figured out how to get high resolution screenshots out of Inventor so I made this. Figured I’d share

edit: and a version where you can actually see the offset


Just about finished printing a version 3 of a 12 to1 and 11 to 1. Using a mini-cim for testing. Very encouraged that a mostly 3d printed cycloidal gear box can be designed to take FRC abuse. So far the failures have been at the eccentric shaft and rotor bearings. V3 has more robust bearings and a beefier eccentric shaft.

1 Like

We just assembled a cycloid this evening. Haven’t really put it to the test yet and needs some more work but initial impressions are promising.


Okay, this is looking great. Please share more!

I’ll take more pics and video as we do more testing. Have this pic and a few short videos for now.


Hijacking this thread back from Tim’s awesome real-life cycloidal (which I would love to see a thread on with results once you get them!), I have an actual update from the land of Myth, Make-believe, and Computer Aided Design.

Presenting V2: The One Where I Solved A Bunch Of Problems By Just Splitting Things Into Multiple Parts™

It’s a bit longer than V1 at 16mm longer than a stock Falcon, but all of the parts are now manufacturable with just one setup. Inventor also says it’s slightly heavier, but 1.5 lbs vs 1.56 lbs :man_shrugging:

Unlike V1, this gearbox could be sealed (assuming sealed bearings are used) to be greased without damaging the motor.

Additionally, I removed the 6700 bearing supporting the “Falcon shaft” in favor of a 6704 bearing in a more typical place for the shaft support bearing. This does mean that the cam shaft is cantilevered, but the cant distance pretty small compared to the distance between bearings (6704 that I put in and whatever bearing supports the back end of the Falcon) and this means there are only 2 types of bearings.

Here’s a picture for a quick peek…


and a link to the cad for those who want to dig in a bit more.


Made a variant of a_cool_username’s cycloid. Started with his first CAD, modified it so that it would be mounted to an unmodified Falcon. Adds a plate on the Falcon, screw into the holes on the 2" diameter, and gives new threaded holes at a 2.25" diameter. Then used a Falcon spline to Hex adater with the outer half turn down to fit in the 6700 bearing. The Camshaft was modified to have a .5" hex bore. Video is linked here:


Very cool. What’s the vibration like and more importantly on that, is there an rpm you found that resonates?


Quite smooth, you can see that it is not attached to anything and it is not dancing across the table. There was no detectable resonance, also it is easily back-drivable
The real test will be when we put it under load.


What load tests do you have in mind?

I expect the Falcon will not break a sweat getting to the stress limit of onyx.


Do you have a CAD model you can share of this version you implemented?

Link to Onshape file:

and to Richard: The intent is to use as a climber winch, I suspect it won’t hold up, in which case we will make an aluminum version or at least for the parts that don’t hold up. A fun summer project nonetheless.