pic: Traction Differential Swerve v2



Version 2! Embrace the strengths, disregard the weaknesses.

Onshape Link

17.26 ft/s free speed

Changes:
Absolute encoder is now on the module (theoretically allowing full-talon control)
Changed to AM Magnetic backshaft encoders because if you’re not going to build it, why not splurge?
Saved about 4.5 inches vertically

I didn’t put together a BoM or weight for this one, but it’s probably actually slightly lighter because of some material savings. About the same price.
This design only has about .25" of ground clearance, which is probably okay because I wouldn’t want to use it in a game with a non-flat floor, for reasons discussed in the previous thread.

As mentioned before, it should be possible to do 3 sensor cascade control with the Talons using auxiliary PID. I haven’t looked deeply into this, so there may be a gotcha preventing this operation.

This design probably wins for both the tightest thing I’ve ever CADded, as well as (probably) the largest bearing on a swerve (6824, 120mm id).

The AM RedLine Encoders, paired with 775Pros, seem to be capable of overflowing the velocity readings on the Talons. I’ve submitted this to CTRE and don’t know if a fix has been implemented yet.

I’d like to know if anyone has gotten this particular combo working.

Is this also true with the RS7 Encoder? That’s what I would use for a 775 in a situation like this, but I don’t know all about them or the use in the design…just a humble suggestion.

That’s what I used in the previous version… Cothron IIRC recommended these ones. The RS7 has only 12 CPR, opposed to the AM Mag Encoder which has 1024 CPR. Makes sense though. Dang it Cothron!

The RS7 should be perfectly fine though. You don’t really need the much-higher resolution when it’s behind so much reduction.

I would beef up the standoffs, but it looks good other than that!

Looks like a lot more teams have been using Onshape at least from what I could tell from posts on CD, looking forward to see how people enjoy that. Downloadless CAD is great for getting more students up and running learning CAD and design!

If that is the same slipring that’s been floating around, it has 6 30A wires. If you are trying to run CAN to those talons, you are going to be using 2 of them. What that means, is that you are left with only 30A per motor on the drive instead of all 40A just due to the limiting factor of the part. Looks cool though!

Are you referring to the spacer between the block the Talon is on and the main plates? Yeah, I’m not particularly happy with how that worked out in general, but I think it’ll be fine with a metal spacer in there and some beefy bolts.

I’m loving Onshape. There’s some missing features coming from Solidworks, and the mate system is a bit weird (someday I’ll sit down and learn it) but overall it runs really nice and having MkCad at my fingertips is super helpful.

Yep, it’s the same 180A 6-wire Missouri Wind and Solar slip ring I’ve designed with a few times before. 323 for sure used a similar slip ring on their drive this year, and with current limiting I don’t think the full 40A breaker will be necessary with 8 775s of power directly to the drive.

That’s a scary number - you’ll bulldoze lost bolts and zipties instead of driving over them…
We ran a nominal 1/2" for Steamworks, ended up with all kinds of marks on the bellypan and bolt heads.

Freaking sweet module!! :slight_smile:

Thank you!

I just checked and it ended up being 0.37" clearance to the plate. That could probably be doubled-ish with only minor changes. So not terrible. Definitely something to consider when looking at a module like this.

Any suggestions for the next version? I’m thinking individual-wheel positraction to maybe get past some of the friction issues.

Looks sharp.

What about ditching the large bearing for three or four V-grove bearings and a matching v-profile ring?

Didn’t think of that. How well would it take the thrust loads?

McMaster carr has their 1.5" OD version (here) rated to 140 lbs static for thrust loads (each).

I can’t find the load ratings for the rev robotics version (smaller and considerably cheaper) at the moment (here). Although I am confident 3-4 would do the trick w/o adverse effects. (You may be able to back up the v-groove bearing with a thrust bearing if you design it with the v-grove bearings under the bottom plate.)

Proper alignment for a system like this is key. 971 has a good visible example here.

You could also angle the bearings in such a way as to transfer some of the thrust load to axial load (at the cost of adjustability and complexity).