Compact Falcon Swerve

This design basically had one goal. Fit a swerve with two falcon 500s into a 5" by 5" by 7" rectangular prism.

Having met that mark, I’m not sure if it was really worth it. But I’m still pretty happy with how the design came out in the end.

Here’s the onshape link.

As always, questions and thoughts are welcome.


That top plate is making me raise a lot of eyebrows…


How many of them do you have?


At least one, perhaps even two. I hypothesize that there may be more.

1 Like

You should rework the name, doesn’t quite have the ring of Snake Bite Swerve or JellySwerve.


Yeah, it might be underkill. It could be beefed up pretty easily, but I think with the swerve below it and some sort of belly pan, the chassis would be stiff enough.

Jeffry swerve, Duke of Snodsburry


What? No spinning render?

It looks really solid. Lots of small changes like tucking the bevel gear pinion into the azimuth plate and removing the absolute encoder gear to take out height. You really spent some time refining the design.

I see you are using a limit switch to zero the module instead of an absolute encoder. I like the way you integrated the cam lobes into the module plate. I might be a little concerned with the switch cam follower arm being subjected to a lot of abuse when the modules are spinning while driving. Have you guys uses this switch on the swerve modules you have used in competition?

1 Like


No, I don’t have any experience with a limit switch zeroed swerve. I just saw the snake/ jelly swerves, and thought “Huh, if you took out the need for an absolute encoder, I bet you could squeeze a falcon into a 7” tall package."

It would certainly be a reliability concern. however, I don’t see why a zeroing routine couldn’t be performed in the que or the pit. This way it wouldn’t take time out of auto, and if anything went wrong, the robot could warn the drivers to manually put the module at its home location.

1 Like

Yeah, with brushless motors having built in encoders, the need for an absolute controller for control is no longer necessary. You just need a way to initialize the position and you can live without an absolute controller after that if you want. So a side effect of the brushless revolution is to open up more options for that initialization. It was quite clever of you to use this option to make your module more compact.

A mechanical limit switch will certain work for initializing. I’m just worried about the duty cycle that the switch will be subjected to for the remainder of the match. I have seen a lot of bent follower arms on those limit switches.

I think we are going to see a lot more people exploring contactless solutions like the hall effect sensor or optical sensors.


It would certainly see a lot of clicks in a match, but hopefully they would all be pretty consistent and not all too violent. A limit switch failure also wouldn’t cause a loss of control unless the robot also lost power during the match. Obviously, having to replace limit switches more than a couple times in a season would quickly become inefficient though.

Something like a beam break for zeroing is a fun idea. With an enclosed module like this one, at least you wouldn’t need to worry about ambient lighting.

I believe 114 of 192 used a limit switch to zero their swerve some year. Would have to look through CD history to find it.

A couple of unrelated notes:

I was inspired by the cheap swerve thread to make a quick BOM for this design. With a lot of 3d printed parts and no absolute encoder, I figured it might come out pretty cheap.

Here’s a link: Printed Swerve 4 BOM - Google Sheets

I got a total of $192! That’s not including bolts (because I was too lazy), Falcons (for easier comparison), and whatever else I may have forgotten (hopefully not too much).

I also color coordinated the parts by manufacturing difficulty. Green is COTS, yellow is 3d printed, orange is simple machining like cut to length, and red is full custom like the CNCed plates.

@marcusbernstein I don’t know if you’re the right ping about MKCAD stuff, but it seems that the falcon model with an 8mm round output is a different length than the normal simplified model. You can see the difference in the renders at the top of this thread.


Thanks for the heads up. Not sure how we muffed that one, I apologize if this caused problems with anyone’s models or robots. I’ve updated the models (creating a new version) and corrected the dimensions based on the official Vex drawing. If there are any more discrepancies, let me know and I’ll fix em right quick.


Thanks, that was quick! It didn’t bother me, I just happened to notice since I had both motor models right next to each other.

1 Like

i really love the cam-switch add on for auto reset.
very well done!
how do you make the wheel ?

Thanks! The wheel would be 3d printed. I think the tread could use a more aggressive pattern though, if it was made with normal tpu filament.

1 Like

that what i though TPU!
but i don’t know the usual worn of tpu for that kind of work??

Have you used a 3D printed TPU tread before? How well did it hold up? Any specific TPU?

No I haven’t used it before. I did print a test tread for fun, but the main thing I learned from it is that it doesn’t get great traction on carpet.


Did you try different tread patterns then the one above?