pic: GBX-139, budget swerve



GBX-139 is my attempt at a cheap swerve drive while keeping the low-weight package. It is made for a 70-80lb demo chassis, although it is possible to convert it to a competition design by swapping bevel gears and motors.

Geared for 17.6fps with a 2.06" wheel, rotation/swivel speed is about 330rpm. Switching to a CIM or miniCIM can let your drive speed be far lower. The BB RS-775 24v also seems to have a free speed of only 12k rpm, so running it at 12v would be only 6k rpm, resulting in a much lower rotation speed.

This uses delrin/acetal bevel gears from SDP-SI that are 1.5 Module (~16 pitch) instead of costly steel ones. The large gears and plates are all meant to be 3d printed, however waterjetting and CNC with a 1/4" endmill both work as well.

Link to the cost sheet:


It does take the liberty of assuming you own versaplanetaries and 10:1 ratios. It also uses my AS5047P magnetic encoder to save money; the chips can be ordered as free samples from AMS or for $8 each in bulk. This results in a cost of about $90 per module, slightly more if you buy the encoder chips.
There are no drive encoders, although a 775 encoder can be added on the back if needed.

As always, CAD is available in the 2015 offseason release in the GBX-139 folder: https://workbench.grabcad.com/workbench/projects/gcBvYbFAI7TQYHiKMOq7jiSQVQxuTYPyTAXwhy5IZUFIGK#/space/gcvLTQUiZ-c9-szjJsPH3WvpxdeKOB9B1Z2b-tQWAq3gUh/link/430734
It’s well past 2015, but I still use the folder for ease of use.

I will upload a pack and go and step file later.

As a side note, this is the only swerve drive I would consider manufacturing right now. I came to the realization that most of my previous swerves would have serious issues taking side loads, or are too expensive.

Awesome design!

Do you have the drive shaft to the bevel gears going through the module turning gear?

Any evidence (either way) that the delrin/acetal bevel gears can hold up for a full event?

Also, are you sure you want a 775 as your main drive motor? Your chance of burn out is much higher and you really don’t want to take that type of risk. The way the motor is designed makes it a much smarter choice for any application where it won’t be stalled for a significant length of time. Although you probably won’t stall much in a swerve drive, you will be running close to stall conditions frequently and that will generate a large heat load.

Awesome execution either way.

“It is made for a 70-80lb demo chassis, although it is possible to convert it to a competition design by swapping bevel gears and motors.” -asid61

The OP said that this was meant for a demo bot (a light one at that). I don’t think the goal is to hold up to full on aerial assault defense.

That’s how I saw the module, correct me if I’m mistaken OP. Love the design BTW.

Yup! This is how 1323 did it last year, and is the lightest way I’ve found to make a swerve drive.

Bingo. Heck no, this won’t hold up for a competition. But if a team wants to try swerve programming, this would be my preferred module. By swapping, I meant swapping for steel gears.
A few teams did run 775pros (rip Mittens) this year for drive motors, with varying results. However, as this is made for a lighter demo bot, I’m not worried about a demo bot that won’t see defense. If I had to compete, I would convert to a CIM drive for sure.

How though? If it turns all the way around, what connects the outside teeth to the central part of the gear? Is there a slot cut into only a section of the gear?

The drive motor has a 12t pinion that is coaxial with the turning gear. When the module turns, the center-center distance between this gear and the offset one remains constant like a point going around a circle. Have you taken a look at the CAD yet? I’m not sure I answered your question right.