Swerve Design With a Printed Worm Gear

The idea of this design was to make a swerve that would use a 775 and a USD encoder for steering so the control loop could run on the Talon SRX. I also wanted to make it fit within a 5" by 5" square in the corner of the frame, while keeping it short enough to stay below the bumpers.

Not a super serious design, so I’m not going to bother with lightening the plates.

There are several things I like about this design that might make it onto future versions, but I don’t really like the need for 4 machined plates, or the 3D printed worm reduction in general. I might make a version that uses a NEO to steer that would be considerably simpler. Questions, comments, and feedback are welcome.

One of my favorite aspects is that this has six gear ratio options from 9.9 to 19.1 FPS free speed all by swapping the first stage gears. The one in the CAD is for 15.2

Here’s the CAD: https://cad.onshape.com/documents/e3e150ee965c3b0fc24cfc62/w/50615b38721d0db1df825f8c/e/37a6fab1f252452811c2e028

For fun, here’s 4 put together into a 10" by 10" frame.


I’m now seriously tempted to 3d print the entire thing.


I was considering putting an entire FRC robot inside a 2018 Power Cube with the Microswerve drives. Then we were thinking that the robot CG would be so high because of the battery it would fall over with almost any maneuver. Regardless, someone should do this soon! Bonus points for picking up said robot with a 2018 robot and scoring it! Or having it go through a vintage exchange on its own power.

27 has a fully operational robot inside a 2018 power cube (albeit not a swerve). It was driving around sometimes during the 2019 season when we were over there hanging out with them and practicing.


Does anyone have experience with 3D printed worm gears? I would think that the much higher inefficiency of a worm gear compared to spur gears would mean you run the risk of melting the plastic. And with such a high reduction I would think the tooth load would be very high, to the point where they may shear. But those are just guesses, I’d love to hear if someone has real world experience.

FTC Team 5795 built a swerve drive bot for Velocity Vortex. We considered taking these modules and building a slightly smaller frame that would fit inside a powercube and then use the FTC control system to drive it. But, the team wanted to keep that bot assembled as it was still being used as a demo bot for the FTC club.

That bot used the FTC motors (AM Neverest) for drive and Continuous Rotation servos for steering. After initial testing they slowed the drive ratio down as it was too fast for the FTC field. But as a power cube bot, they could have used the higher drive speed. It would have been pretty fun.

Last summer I explored something similar with worms. Printed both the worm and worm gear. Biggest problem with both printed out of PETG was stiction when loaded. Would probably be oK with a brass or bronze worm. Did make a worm gear driven grabber to latch onto the lander for a FTC team but went with a spur gear design because it was more compact and mountable. I wonder how worm steering would work with the control loop.

These mirror some of my own concerns with the design. Although I’m not particularly worried about shearing teeth, because it’s not a high load application and there are four or five teeth engaged, I am concerned about building up too much heat.

I vaguely remember 5818 doing something with worm gear steering. I wonder how it turned out for them.

Team 1658 in Missouri has used worm gears salvaged from garage doors for azimuth in their swerves since at least 2017. Pretty crazy looking design, I can’t find a pic at present.

Unless I forgot, I don’t believe it was us doing a steering drive with a worm gear.

Oops, looks like I got you confused with 5817.
On the plus side, I found a picture:
Google Photos

1 Like

My favorite impractical robot idea in 2018 was a 13" cubic robot that climbs by driving into a teammate’s intake and having their elevator lift it. I called it the “Simba”. It needed to be 13" with bumpers though, which made packaging… tricky.

Edit: bonus points for the argument you’d get to have with the head ref when you drive the robot out of the exchange and put it in the vault.