I hope everyone has been having a great start to the school year! This summer, Team 9312 has been cooking up a new swerve learning minibot that we hope to share with the FRC community:
TurdSwerve is a miniature, fully 3D-printed swerve chassis designed to be the perfect off-season bot for teams hoping to transition into a swerve drive next year or teams who can use it to train rookies and develop code. The pods can be used with Neos, Falcons, or Krakens (CIM support coming soon), and thrifty encoders, CANCoders, MAG encoders, or no absolute encoders. It uses the electronics base as a full swerve drive does but replaces the expensive COTS pods with 3D-printed PLA ones for a total cost of only ~$40 USD if you use your spare electronics and motors. All of the 3D printed gearing is replaceable in under 2 minutes without changing the motors, allowing programmers to minimize the number of screws that they are screwing… sigh what we do for programmers’ convenience…
The robot is compact, with configurations as small as 15" wide. It is designed for convenience and ease-of-use, so that a single programmer can use it as a complete all-in-one learning tool. Teams can also import their existing swerve codebase or TunerX generated code and simply switch their robot specs to TurdSwerve specs if they want to use the robot to develop competition code.
NERDSwerve is the new and improved version of TurdSwerve that features the same chassis design but with a more robust pod design that features a super cheap swerve bearing that we found while still costing under $100 per robot.
The older TurdSwerve design is still available if you want to save a few dollars.
We have released a few tutorials that cover the Onshape model’s configurations, robot assembly, and (coming soon!) a full step-by-step programming guide using NERDSwerve. The tutorials are listed on Team 9312’s YouTube and other resources can be found on the brand new NERDSwerve website!
Check it out if it seems like it could be useful to your team!
It seems like you’re driving it pretty hard in the demo video, how long does it last? As in- is it something that you can substantially drive? I’m thinking about something like an alpha bot, or off season drive practice
That is actually only around 50% speed. It is using a full FRC battery and the Krakens are really overpowered for the 35 lb robot…
Driving around at that speed usually lasts for 30-40 minutes before we start to brown out. A 2-pod robot would likely last for longer.
If you’re planning on using it for drive practice, you will have to keep in mind that it is so small and much easier to drive than a full robot. You could mount it to a bigger main plate if you want to make it full size. Bumpers might be necessary too
As for longevity, we have been driving that particular robot for around 20 hours and taken it to a few outreach events, and we haven’t had any issues. The pod gearing is easily replaceable though and can be fixed very easily if there are any problems.
We have driven it for around 20 hours and haven’t had any issues. We haven’t put it through anything as destructive as drive practice can be though. I can feel some slop building up in the drive gears though. I expect we will have to replace them in another 5-10 hours or so of driving.
We have been using MDF as the baseplate material so I wouldn’t pile more than 30 more pounds on top of the current 35 lb robot. If you use Aluminum or mount it onto Aluminum tubing, it should be able to handle the weight of a full 125 lb robot. Also keep in mind that if you make it wider, the baseplate will need to be stronger too.
Looks like the NERDSwerve BOM has lock nuts, but the CAD have some areas where a lock nut is too tall and all of the nuts are jam nuts. What’s the move here?
Hi, is the CIM option still in the plans? This could take the expense to almost zero for our team, which is rather enticing.
Either way, thanks for sharing all this!
I believe the jam nuts are only necessary for the extra electronics deck supports 4-pod & horizontal battery configuration. They are redundant though, so I forgot to put them into the BOM. You can just skip the corner stilts if you don’t have jam nuts available.
There should not be any other places where a jam nut is necessary. The CAD nuts look like jam nuts because it was easiest to use the Onshape nuts and they only had jam nuts.
Yes, there are still CIM and MiniCIM options in the CAD. The pod design is identical to the Neo design. However, the CIM is slightly wider than the Neo, so there seems to be about 1/16" of interference between the CIMs when they are mounted. It should probably be fine… hopefully?
We have never tested the design with CIMs so we don’t have any example code. It should be pretty easy to reconfigure our Neo example code though.
While using CIMs is technically possible, I would definitely recommend using a brushless motor since CIMs’ lack of motor encoders makes odometry impossible (which hinders autonomous programming stuff). If you have CIMs lying around though, NERDSwerve could still be useful for vision and teleop programming development and I would definitely go for it.
How long does it take to print all swerve modules? What would need to be done to make this just as competitive as metal swerves? This is really cool to see, could help many new teams!
On our Bambulab X1C, it takes about 48 hours to print 4 pods out of PLA.
The swerve pods have identical functionality as metal swerves but are likely to have durability issues when used on a 120-lb robot. The forks held up pretty well under our driving and collision tests but the gears seemed to start to wear after 20 hours of high-load driving. If you print the gears out of PA-CF or another stronger material, you might be able to play around with these drive pods on a full competition robot.
