pic: PDP-Be-Gone Swerve



“Your engineers were so preoccupied with whether they could, they didn’t stop to think if they should.”

This crazy idea was eating at me, so I decided to sit down and crank out a CAD. It’s rough, it needs some optimization and mounting features, but it wasn’t ever intended to be a serious thing. I figured someone might find it entertaining.

It would, however, theoretically be capable of 26ft/s and have 18lb/ft of torque available at each axle assuming no losses.

Really interesting design you got there; looks like it was inspired off the (not actually) “differential swerve” design. I love the lower module design with the tabs and 4-piece layout that lets you use thinner plates.
Does this shift at all or is it just going to be 26fps flat? I think you could fit a shifter in the first reduction if you wanted to to make your acceleration a little better.
What made you go live axle over dead axle for the wheels? How would live axle even work? How does this mount to the chassis?

Going off of the differential swerve inspiration, are the wheels independently controlled? Because I’ll need a picture of the internals to understand what’s going on in the center.

How would increasing the amount of motors reduce the amount connected to the PDP? I’m a build lead and not electrical, so I’d like an explanation here. Besides, wouldn’t 16 775pros suck more juice than 4 CIMs and 4 Versa’d BAGs (or alternatively, 6 CIMs and 4 Versa’d BAGs running on WCD shifters)?

The two plates that mount the 775pros seem to be the seating method.

That said, this looks like a really heavy design, even considering the subtraction of CIMs. Other than that, it looks purty good. If it truly is a new take on the differential swerve, I’d love to see the internals, because this could be a new innovative step in FRC drivetrains.

I would be very wary of using a Lazy Susan in this way, they are perfectly capable of handling the “thrust” load but are terrible at handling side/torque loads.

If your robot is pushed sideways or the module rotates while going fast the entire module becomes a lever( with a lot of force amplification) trying to crack your bearing open( and crack it it will).

Quite a few teams have had problems with these bearings cracking open during heavy impact when used on turrets. If the bearing cracks open you will also probably end up bending the 775pro shafts.

tl;dr Rotating sushi is not the same as rotating a swerve.

As someone really trying to get a good grasp on designing turrets and swerve, I had these concerns regarding the Susan bearing being too prone to damage. That said, are there alternatives? Or are there more robust, industrial variants that would suit this purpose?

What you’re looking for is called a thrust bearing. It’s designed to take up thrust loads without breaking. Basically a lazy-susan made for industry. I’m pretty sure you can also just use a block of HDPE or other sufficiently slippery plastic, but I haven’t confirmed that (I have used it instead of radial bearings for a WCD and it worked fine).

Hey! Crazy swerve is my job :stuck_out_tongue:

There isn’t necessarily a problem with using a Lazy Susan as long as it is used correctly, that means using conventional radial bearings/ bushings to take the non-thrust forces.

There are some really nice slewing rings available from IGUS, which can take both thrust and torque/side-loading but they are generally expensive, heavy and bulky. Normal ways to do swerve include thrust bearing supported by radial bearings/bushing, custom thrust bearing using ball bearings( basically milled slots in the plates) as well as using a large diameter deep-groove bearings( 1323’s swerve from this season is a good example of this).

Otherwise, reducing the height of the module will reduce the loading on whichever method you select. A shorter module is good in other ways such as taking up less space, weight etc.

If you are doing a low down turret you can get away with a Lazy Susan, we used one last year on our turreted shooter didn’t have any problems in regards to the ring( 12" FIRST Choice). But if your turret is tall you may get problems, a good solution to this is a custom ring using deep-groove bearings which are capable of handling side-loading and thrust.

To address a few questions:

This is indeed not a differential swerve. Three motors provide power to drive the wheel, the fourth drives the rotation of the module. The lower shaft is a live shaft A timing pulley is situated in between two 4" performance wheels to accomplish a final reduction.

To address the shifting question: This drive would be traction limited if geared for 56 ft/s assuming no losses. Geared for 26 ft/s it would have plenty of torque and thereby acceleration.

The name of the photo was a joke referring to the fact that you don’t have to worry about PDP slots, because they’re all occupied by the drive assuming a 4 module configuration.

I have no intention of actually releasing this thing into the wild, as I designed it independently for fun on my own time, which is why it doesn’t have accommodations for mounting to a chassis, though that could be accomplished pretty simply. The feedback is useful though, as I can work that information into mentoring the parts of our team doing design work.

I realize one thing for certain, dynamic loading situations the drive is exposed to were severely underestimated and the bearing interface needs a re-work and the gearing to power rotation is probably inadequate.

Thanks for that input Pilleya, now that you mention it I recall seeing the top half of a robot fall off in 06’ at competition because of a lazy susan being used in a very tall, high mass turret for a shooter.

One thing I’d like to understand better, for potential future use is what the loading on a module looks like? I suppose the torque load on the bearing could be evaluated considering the distance from the floor to the bearing interface and the traction limit of the robot? That number is fairly straightforward, but turning that into a specification for that bearing configuration isn’t as clear. Maybe one half of the axial load capacity of the bearing? I’ll jump on Igus and check some parts out in the next few days. Big fan of off the shelf solutions for this kind of stuff.

I’ll probably swap the rotation over to a versaplanetary for the sake of simplicity of a high reduction and encoder availability, but what are the cool kids using for reductions on module rotation these days? Understanding of course, its a 16:72 in the chain stage.

Thank you to both pilleya and Ari423 for the helpful advice. Duly noted!

Did you check how much current each module would be pulling? I see why you called it PDP-Be-Gone, if you tried to use one you would trip your breakers immediately

Love it!