pic: Small Swerve

small module to go with the coax power unit from the other day

2.5" colsons


errrr… how is this different from this? : https://www.chiefdelphi.com/media/photos/46562

Its very similar but I changed a few small details with the spacers that go in between the bent plate and I also changed the spacing on the wheels and basically put it together as a finished set. So this one is ready to be made, the other green one wasn’t finished.

The frame that this goes into is almost done as well so next week I hope to publish the finished drive system to show how the pivot mount works

Why shaft collars over a washer/bolt or clip rings?

Why a block of delrin (I assume) over a normal bearing setup?

I just put shaft collars on it mostly because I am lazy and was strapped for time at my good cad computer today, irl the shaft would be tapped and secured with a washer,

The pivot/clamping blocks are uhmw or some other semi hard plastic and are pretty easy to machine, four bolts go into the frame and two bolts clamp the blocks together at a tightness that holds but still allows rotation and are sacrificial parts that would have to be maintained or replaced as needed.

Can you make a full chassis with these models on them? I’m still failing to see how this particular swerve concept works (like with motors and such)

Its still a work in progress but the frame with motor setup and modules in it is on CD right now
This is everything minus chain and hardware.

The drive motors setup is here https://www.chiefdelphi.com/media/photos/46572?

I haven’t added the steering motors yet.

Thanks! If it’s easy enough to manufacture I may look into prototyping with one

This particular model is set up for a crabdrive, but the modules themselves are coaxial so you can do independent if you wanted, provided you supply power to each one your own way.

If you have access to a brake and a lathe you can make this fairly easily, the pivot would be turned down from a 4" rod that is available on McMaster and the pivot clamping blocks could be 3d printed or modified to be just flat rectangles if you have a manual mill.

Is this something your team is looking to make or is this just a concept your playing with? It’d interesting to see this in the real world, at least to FRC standards.

Im working on this on my own as a concept but this is probably atleast one season away from reality.

I’m hoping to make a prototype module or two in this off-season with the machines i have available to me at my work, I’m pretty confident in the pivot concept that is a major part of this design. For frc level competition I would be surprised if it did not hold up to the stresses it would see(assuming a relatively flat field).

I could make all the parts for this setup on my laser in the matter of an hour or two if I wanted, it would just require committing to buying all the cots parts and an investment of time to actually build it irl.

So, your “crab drive” concept is a bit confusing to me. What is the value of having all 4 wheels drive at the same speed from a single, 6 motor drive system? It seems that with a single drive unit powering all 4 modules, you would only be able to turn by stopping any translational motion and rotating the wheels to spin in a circle about the centroid of the wheels. The other possible motions that this drive base can do is translations in any direction but with the chassis always oriented in the same direction. Any combinations of rotation and translation require some of the wheels to be spinning at different speeds than others which a single drive system cannot do. This seems to be a very constraining design choice.

If you divided the 6 motors into 2 drive units and used each drive unit to power two of the modules, then at least you could steer like a tank drive in addition to the translations in any direction. This might give you a decent ability to mimic most of the motions of a swerve drive. But, again, this constrains your design and I am not sure why this would have an advantage over a pure swerve drive.

If you feel that 4 drive motors is not enough power if you put one per module in a pure swerve drive, you can always have two of the swerve modules have 2 drive motors each and the other two modules only have a single drive motor. This would allow you to put the power of 6 motors into the wheels without the added complication of all the chains that you are going to need with this setup. If one end of the robot is heavier than the other, I would put the 2 motor modules at that end as the more heavily loaded as the wheels would be less likely to slip at high power.

It is crab drive not a swerve, it has pros and cons compared to a fully independent swerve, for instance it could be driven effectively with minimal programming and sensors, but would require more coordination with whatever game piece manipulator you put on it.

It can turn in place by rotation one side of the modules 180 from the other and powering all the wheels in a forward or backwards direction or putting them all at 45° from each other, you just have to be concious of which side is the “front” of the modules.

And example of a good fit for this drivetrain would be a game that has a static scoring location that faces towards the drivers like in 2011 or 2016, your robot design would also need to be able to intake from both sides with something like pink style arm or an over the back elevator like 254 this year.

It could be split from a single 6 motor unit to two smaller three motor setups to get the extra degree of movement very easily. And if the game requires a lot of turning in place or scoring to be done in a different plane than the front of the alliance wall splitting the power would be the way to go.

Edit: this can also rotate by just turning one side the modules inward and powering everything forward, it would not rotate around the center of the frame and there would be some scrubbing though.

I’m not sure I agree with the comment about minimal programming. The programming required to determine what orientation to put each wheel in to be able to accomplish motion with all wheels spinning at the same speed seems to be fairly complicated.

Relative to sensors, you are still going to need steering encoders on each module so you haven’t saved anything there. Many swerve drives work quite well without drive motor encoders (by assuming the relationship between voltage and wheel power rather than controlling it with a feedback loop).

OK, I see that there are some other ways to accomplish rotation. But none of these allows you to turn while you are moving forward in a smooth arc. And, in fact, all of these require the orientation of the wheels to be changed significantly from the forward driving orientation. So you can not smoothly transition from translation to rotation. You pretty much need to stop and then rotate the wheels before you start your turning motion and then stop again and re-orient your wheels before you can translate again.

Maybe you could accomplish rotation while translating by turning the one side like you said and relying on scrubbing to slow down that side. If you turn both of them a small amount, you can get a little bit of scrubbing to do a gradual turn. Turning them more gives you a tighter turn with a bit more scrubbing. Turning them 90 degrees gives you the sharpest turn with a lot of scrubbing. But that seems like a really inefficient way to go. And the tread life of your wheels is going to suffer.

I agree that if the orientation of the robot on the field is always facing in one direction, this might be a good drivetrain. If you don’t have to turn at all but you want to be able to move diagonally or sideways while still facing the same direction, this drive will allow you to do that. But you can also accomplish these same motions with a pure swerve drive or even a crabbing tank drive if you wanted to. I’m not sure you would have been able to play the 2011 or 2016 game well without turning, but I agree that you could have played them at some level.

I think this would be a good compromise if you want to keep the programming relatively simple while retaining the ability to turn. I also think it would be mechanically easy to implement with the drive on each side feeding the two wheels on that side. I don’t see a downside this configuration relative to the single drive motor system.