Here is 5205’s off season CAD release for our Traction Swerve Design. This is our 4th iteration. We have shown that with the right wheels (70A, 4") we can generate about 160lbs of traction. Current gear ratio is about 6.7:1 which is the highest that we can fit with NEO’s. With Falcon’s and the smaller pinion gear, we believe we can get it to a 10:1, maybe a little higher.
The swerve design wasn’t designed to be very offensive. It is incapable of swiveling 360 degrees. We choose to limit cost and complexity and not include a slipring in the design. It is also not a coaxial design.
We believe this design can be as maneuverable as a standard swerve design in defense and also have as much traction as tank steer bots. It is a hybrid drivetrain that may be ideal for defensive heavy games.
Pros:
Steering gear ratio can be somewhat controlled independent of driving gear ratio by moving wheels further apart. This has been a fundamental road block to diff swerve designs.
Low cost with no machined components. All components are designed to be laser cut and formed.
Cons:
Probably a little heavy. We built ours our of steel so we can’t tell for sure that the unit will weigh out of aluminum. Out of steel, the Pod is about 15lbs each. Out of aluminum we are estimating about 10lbs per pod.
70A wheels create tons of traction but probably need to be changed out often. Probably one set for quals and one set for eliminations.
I’ve been waiting to see this since the first upload you did of this with the testing rig in Sept. 2018. This was one of the first posts I saw on CD (way back at the old site when I was still kind of a rookie), and man is this a cool swerve implementation.
Are you actually planning on using this on the playing field in 2020? Also, will you be willing to share the CAD of that testing rig?
This design is intriguing. I’m glad to see that you are still developing this design.
So, I assume you are using a 6 pole slip ring with 2 poles for pos and neg for one motor controller, 2 poles for pos and neg for the other motor controller and 2 CAN wires? Does that limit the amount of current available per motor? Or did you find a slip ring that could carry higher currents?
Your model shows a pretty heavy duty module, supported on a spring suspension which seems like overkill for FRC. If you used this for an FRC application would you remove the spring suspension, or was that added for a purpose?
Is the small black cylindrical object shown in the renderings on the top of the module the absolute encoder or is that part of the slip ring?
You mentioned in your CONS that you felt the wheel tread wear would be high. How easy is it to change wheels? It is not clear whether the bolts holding the dead axle can be removed without removing the large first reduction gear or not. You clearly added a hole in the outer plates to access those bolts, but it looks like the bolt head is going to hit the gear teeth.
Keep us posted on your progress. This design offers an interesting direction in swerve and I would like to see where it goes.
A slipring was part of our design in the first iteration. We found that it was more trouble than it was worth. It was a high cost item and its reliability was questionable. So for all subsequent design iterations we removed it. It is a design compromise but in driving the inability to swivel all the way around doesnt seem to limit functionality that badly. It will only be proven in competition.
The suspension was to ensure that all wheels stay on the ground. But the linear guides jam and I think they are more trouble than they are worth as well. We will probably remove then in the next iteration.
The black cylinder is the absolute mag encoder. We started using one from P3 that we thought had a bit more robust housing and cable. It only provides a pwm signal and the only way to get it to work with the rio is via a CTRE canifier. I was told that a newer version of the RIO image might support pwm inputs via the digital in channel. But current versions do not.
Wheel axle drops pretty easily. We should be able to change all wheels in 30 min. Haven’t timed it yet and again competition is the true test.
Here is our code for our Differential Traction Swerve Design.
Features a custom PID algorithm, field centric driving, and also because we have 2 wheels per pod, each able to rotate a different speed and each not centered on the pod, we created an algorithm to calculate each individual wheel speed.
