Team 3674 Swerve Drive CAD

This is our first time doing Swerve, but we think the CAD model turned out very well. We took heavy inspiration from team 2910 Jack-in-the-Bot. Because this is our first time doing swerve, do you guys have any tips or recommendations. We haven’t built it yet, so any advice or feedback would be useful.

The gear ratio for the steering is 21.12 : 1
The gear ratio for driving has been updated to 10 : 1

Onshape link to CAD model:

Once again thanks for any help.


Looks like a great start! Here are a few notes:

  1. The steering looks to be done through a hex shaft. This makes supporting the central drive shaft difficult, and introduces unnecessary backlash to the module steering. Assuming the top plate of the rotating portion is going to be 3d printed, then you should make it much thicker and integrate the steering gear teeth. this would allow you to just put a bearing in the center to support the drive shaft.
  2. I don’t see anything supporting the weight of the robot. I think of you pushed up on the wheel now, the whole rotating portion would slide up until the shaft collar at the top hit the NEO.
  3. Assuming the forks are also 3d printed, I would recommend you put the bearings in the wheel, and change the wheel shaft to be a standoff between the two forks. This would make the assembly considerably stronger.
  4. You will need to modify the large bevel gear and/or the wheel for them to fit together.
  5. The thin section bearing is interfering with the plate it is mated in.
  6. 14.4 to one will make this quite slow on a 4 inch wheel. I would suggest something closer to 8:1.
  7. I’m not sure why two of the frame rails stick out past the others. Is it an old habit from WCD?

I think custom swerves are a great design exercise, and its very cool to see you’re using onshape. Is 3674 going to use it for the entire robot next year?


Thanks for the advice. While modifying our design we will be sure to use your advice. In response to your questions, the entire frame was just to see how the swerve modules would look on a frame. In reality, we would most likely use a square frame.
Also, we will be using Onshape for the next year. While working on this design, we found how amazing it is to be able to work on the the assembly at the same time with other teammates.

1 Like

You have an encoder on the module, but I would recommend that you have one that’s geared 1:1 with steering. This will let you tell the rotation of the module without zeroing.

I’ll second Bryce’s advice about using dead axle for your wheel. Especially with 3D printed parts, this will increase your wheel module rigidity and strength by a lot.

Personally, I like not using shaft collars. I try to tap the ends of shafts and retain using spacers and a washer+bolt on the end. I think this is a lot cleaner and more compact.

If you want to 3D print the steering bits, printing a belt pulley is much stronger than printing gear teeth.

Out of the question, why a NEO for steering, to me it seems a little overkill? I know of plenty of teams that have use much smaller and lighter motors for the steering and have had great success with them.

As I said earlier, this is our first time doing swerve. As such, we took lots of inspiration from other teams. Most steering modules were controlled by NEOs, and we decided that NEOs would give us plenty of strength and speed. If other motors can give us the same ability, we would be open to changing.

1 Like

A smaller, faster motor with additional reduction to reach the right speed is heavier than a NEO setup with a similar output speed.


Do you think that 3d printing gearing is strong enough, especially for the gear used for turning?

We thought about doing a pulley just like 2910 did, but then we though about the issue of having a belt stretch over time and begin to slip, so we decided to go with gears. This definitely adds weight, but would keep us consistent in our matches with less need to swap out parts/modules.

The belts commonly used in FRC are made of Kevlar (the same stuff they make bulletproof vests out of). They don’t stretch over time. Gears on the other hand do wear down and increase backlash over time.


I do think it would be strong enough. I’m pretty sure 33 used printed gears to steer in 2018. But if you are worried about strength, you could use a vex gear with a bearing bore and a versa keys, and bolt it to the printed part.


That looks like a great first start!