Hexagonal Chassis + 6 CIM swerve drive

[TheHolyHades1]

Hello.

As you probably know, swerve drive is cool. I've been fascinated with it ever since I discovered other teams using it (namely 1640 at MAR), and took it upon myself to try to create some CAD files for our team. We've recently been discussing using 6 CIMs on the drivetrain; as an added challenge, I wanted to see how I could use all 6 in a swerve drive. My answer was to use 2 cims on each of 3 swerve modules, instead of using 1 cim on each of 4 and having 2 left over. However, rectangles don't sit/balance very well on triangles; thus, I decided that a hexagonal chassis would be better as well. Here's my current (basic) progress. The gear ratios aren't correct, but I've placed gears everywhere necessary and the appropriate sizes can be determined and bought.

It is a coaxial design; the sprocket at the top of the drive shaft above the wheel is attached to the drive shaft, while the sprocket below that is bolted directly to the pivot case and is attached to the pivot module. Between this lower gear and the plate which has the toughbox and P60 gearboxes attached (the lower of the 2 large horizontal plates) resides a thrust bearing which is inserted into the sprocket attached to the top of the pivot module.

Please let me know if you have any questions/comments/concerns about the design, attached concepts, or anything else. I'd like to know what I should improve on this design, and (perhaps more importantly) if it will work. I do know that I need to add some bolts between the top and the bottom horizontal plates, add some more support horizontally between the two vertical supports for the wheels, make an actual mounting system (which will probably just be bolts through the bellypan of the chassis in this sort of setup), and then I need to swiss cheese everything. I also need to put in the shafts of proper lengths and add keys/retaining rings where necessary. However, I'm sure there are things that I'm forgetting; if you find anything, please let me know.

Also, I would actually appreciate some assistance figuring out the gear ratio needed from the RS550 motor we will be using; I have no idea how to figure out the torque required to turn the wheel under load. I also need to correctly gear the cim motors, but that shouldn't be too much of an issue.

Thanks,
Hades

Module:



HexBot:

[heuristics]

This is a really neat idea. Did you consider having 6 different swerve modules? I suppose the fact that non-CIM motors are limited to 4 would pose some additional difficulty.

[jgerstein]

This looks quite interesting, and I like it.

My only concern with the 3 swerve drives on a hexagonal chassis is that this might be a bit unstable and prone to tipping under heavy defense or with a tall robot design. The design looks like it already has plenty of power to drive the robot, so maybe an unpowered omniwheel in each of the corners without a wheel could help with stability without impacting maneuverability.

[TheHolyHades1]

Originally Posted by jgerstein

This looks quite interesting, and I like it.

My only concern with the 3 swerve drives on a hexagonal chassis is that this might be a bit unstable and prone to tipping under heavy defense or with a tall robot design. The design looks like it already has plenty of power to drive the robot, so maybe an unpowered omniwheel in each of the corners without a wheel could help with stability without impacting maneuverability.

That makes sense, thanks. I was thinking either casters or omni wheels; omni wheels would probably be easier and more effective.

[cbale2000]

What if you mounted the motors where the wheels are now, and had a set of wheels on either side of motors? It would give you stability by having a wheel in each of the 6 corners, but the simplicity of only needing 3 modules.

[TheHolyHades1]

Are you saying to put 2 wheels on each module?

Edit: I did it the way I have it so that when the bot drives forward, there will be a flat side instead of a corner.

[asid61]

Your design looks pretty good. However, as others have said, you might face tipping problems even with what looks like a pretty sweet setup.
You can still use 4 modules on a hexagonal chassis I think.

A note about the swerve modules you have setup: I highly recommend using the swerve design from team 221 if you go coaxial. Basically, they sell a really nice looking coaxial that is already optimized for weight. Download the CAD and use those ideas in making your own coaxial swerve.
Mainly I suggest this because it's already super light and they have very nice looking plates that you can emulate. Do you have access to a CNC?

[asid61]

Originally Posted by cbale2000

What if you mounted the motors where the wheels are now, and had a set of wheels on either side of motors? It would give you stability by having a wheel in each of the 6 corners, but the simplicity of only needing 3 modules.

This seems pretty good.
Basically cbale says to have one gearbox for each two independent coaxial modules. That way you only need 3 gearboxes.

[TheHolyHades1]

Originally Posted by asid61

Your design looks pretty good. However, as others have said, you might face tipping problems even with what looks like a pretty sweet setup.
You can still use 4 modules on a hexagonal chassis I think.

A note about the swerve modules you have setup: I highly recommend using the swerve design from team 221 if you go coaxial. Basically, they sell a really nice looking coaxial that is already optimized for weight. Download the CAD and use those ideas in making your own coaxial swerve.
Mainly I suggest this because it's already super light and they have very nice looking plates that you can emulate. Do you have access to a CNC?

Thanks. I'll add the omni wheels shortly and post a revised drawing.

As for 4 modules, the reason I'm not using 4 is so that I can have 2 cims powering each wheel setup.

[TheHolyHades1]

Originally Posted by asid61

This seems pretty good.
Basically cbale says to have one gearbox for each two independent coaxial modules. That way you only need 3 gearboxes.

If you've got 6 independent drive modules, you'd need to have 6 steering motors; for consistency (and performance) sake I want to use the same motors on each setup. You can't use 6 of anything except CIMs, which means you'd have to change motors for the last 2 steering motors, or do a 3/3 setup; I fear this may cause more problems than it solves, and that the issue will be much more easily remedied by adding non-driven wheels.

[Ether]

Originally Posted by cbale2000

What if you mounted the motors where the wheels are now, and had a set of wheels on either side of motors? It would give you stability by having a wheel in each of the 6 corners, but the simplicity of only needing 3 modules.

If you want a 3DoF swerve, you have to drive and steer each wheel independently.

[Tyler2517]

One more thing to think about is the pure amount of power a swerve drive uses in the first place. We have had a lot of problems with that alone. Adding in more cims may make that worse. I would honestly focus more on how to make it more efficient.

If you are running a 3 wheeled 6 cim triangles are in my opinion a great drive base. With a corner you could penetrate a defense pretty fast. 1425 had a great kiwi drive, i can not imagine them if they would have had a swerve drive. plus 3 corners makes the rolling maneuver swerves are famous for easier.

For the gearing for the steering. we have found that a 132/1 for final reduction is pretty good. We messed around with faster/slower and found that any faster we might have problems with turning the wheels in a pushing match where we receive a lot of thrust.

On a side note make sure the module is mounted well if it becomes cantilever it makes it almost impossible to turn when the robot is not moving.

For weight saving you can look in to what we did where we put the steering motor in the belt from the drive belt. We managed to make the module super strong and insanely light doing this.

[cbale2000]

Originally Posted by Ether

If you want a 3DoF swerve, you have to drive and steer each wheel independently.

You're correct, I had to think about it a while to realize that issue. You could potentially do it with half of the wheels as omnis (alternating every other), though that's not the most practical.

I'm still of the opinion though that using omnis on a swerve drive defeats the purpose of a swerve drive (Why not just build a kiwi drive if you're going to use all omnis?).

[TheHolyHades1]

Originally Posted by cbale2000

You're correct, I had to think about it a while to realize that issue. You could potentially do it with half of the wheels as omnis (alternating every other), though that's not the most practical.

I'm still of the opinion though that using omnis on a swerve drive defeats the purpose of a swerve drive (Why not just build a kiwi drive if you're going to use all omnis?).

Well, for one swerve will have a lot more traction. Divorcing orientation of the chassis from the wheelbase is also a feature not found in omnis. In this case, the omnis will only be used to support the robot weight to prevent it from tipping over, instead of as steering. Many teams use this approach in tank drive to allow for easier turning (though they tend to power the omnis ...)

[Ether]

Originally Posted by TheHolyHades1

In this case, the omnis will only be used to support the robot weight...

To the extent that the unpowered omnis are supporting weight, they are robbing traction from the driven wheels.