pic: LEGO Coaxial Swerve Module

[philso]

This is cool and so elegantly compact. The one Steve made is cool too.

Originally Posted by bstew

The reason the 12 tooth gear is on a 3m axle is because if I extend it across between the liftarms, it rubs on the tire. The way it is now is the best solution I could come up with.

The 12-tooth gear may end up adequately constrained (for the forces involved) by the gear below it and the one above it, laying horizontally. Those 3x5 L-shaped liftarms are pretty stiff so the gears might not skip.

Did you want to use the chain or 40-tooth gears for turning the modules, for powering the drive wheels or for both? Have you considered using long axles and the knob-gears Calvin referred to? The axles aren't that great at resisting torque but then you get gear lash with a large number of gears.

For "Smart Move", one of the local whizz kids and his partner attempted to build a second robot with swerve/crab drive for the Championship in February, after the Qualfier in November. I think their school ran out of parts and they could not manage to get the software going (no mentors on his FLL team). They did end up winning Third Place in Robot Performance using a robot they started building less than 24 hours before the Championship.

[Calvin Hartley]

Originally Posted by bstew

The tires I am using must be a bit bigger than the the tires in the render because I had problems fitting an axle in. While I wish I had smaller tires, the ones I have will probably have to do. Thanks for your help anyway.

The ones in the render look like tire 89201 with (I am assuming) hub 55982. If you want to acquire some, you can find both easily on Bricklink for cheap. If you haven't used Bricklink before and have questions, feel free to ask.

I tested those parts with a one-stud gap between the wheel axle and the white axle in the image below. There's plenty of gap. Hope this helps.

[carpedav000]

Would this work with having one module in the center of the robot and a caster on every corner? That way it takes a lot less gears to drive and only one gear to turn.

[asid61]

Originally Posted by carpedav000

Would this work with having one module in the center of the robot and a caster on every corner? That way it takes a lot less gears to drive and only one gear to turn.

How would you turn?

[tindleroot]

Originally Posted by carpedav000

Would this work with having one module in the center of the robot and a caster on every corner? That way it takes a lot less gears to drive and only one gear to turn.

Having at least two modules is necessary (3 or 4 is probably best) for two reasons:

1. with only one module, swerving it would move the robot as well since there is no other anchor point for the module to rotate.

2. having multiple modules allows the robot to do tricky maneuvers - one module would allow only omnidirectional movement, and sacrifice the helpful maneuvers.

[samfruth]

I just made a swerve drive using this module and it worked great for me. The only modification I made was with the bottom axel because it kept coming lose from the L brackets. I used a gray piece that was just longer than the original light grey piece and has an end cap on it. Two modules was all I used and for motors I had three of the small motors that come with the EV3.

Thanks for sharing this!

[bstew]

Thanks for all of the replies! Thanks Nato for telling me about similar designs. They will be helpful if I run into problems.

I am progressing with the design of the robot as I await pieces to arrive. I think I will be using a set of 40 tooth gears and an NXT motor for turning the modules. I probably will use chain and the medium EV3 motor for turning the wheels. I will still have two motors left for other actions because I will control it with an EV3. After completing it, I think I will try to do some FLL missions to evaluate its viability for FLL.

[AlexanderTheOK]

Strange that theres no pictures of this outside of the render. I thought i would make it, and gear it for speed.

Heres the simplest one I could make with nxt 2.0 parts. surprisingly sturdy considering the drive motor is attached at literally at only one other point outside the drive shaft. Love the shaft catchy thingy.

[faust1706]

Originally Posted by asid61

How would you turn?

I think conservation of momentum would allow for conversion to angular momentum if going fast enough and have sudden enough changes of direction.

[AlexanderTheOK]

Hate to resurrect a thread, but a friend and I decided to have some fun with our access to some NXT 2.0 kits and made a 2 wheel swerve using this design. The stacking he came up with is the most aesthetically pleasing thing I've seen all month. Didn't have time to program it, and I doubt we could have since I believe the nxt 2.00 motors use tachometers instead of encoders.

Still fun to drive though, with one phone controlling steering and the other controlling speed it vaguely resembles swerve movement.

[bstew]

Originally Posted by AlexanderTheOK

Hate to resurrect a thread, but a friend and I decided to have some fun with our access to some NXT 2.0 kits and made a 2 wheel swerve using this design. The stacking he came up with is the most aesthetically pleasing thing I've seen all month. Didn't have time to program it, and I doubt we could have since I believe the nxt 2.00 motors use tachometers instead of encoders.

Still fun to drive though, with one phone controlling steering and the other controlling speed it vaguely resembles swerve movement.

Pretty cool stacking design. I'm glad you could make use of my module design.

My parts finally arrived from LEGO today, so I worked on it some. Currently, I am using a EV3 with a four wheel crab setup. I am working on optimizing it, but I should get a video of it sometime when I'm done. This has been a fun project so far.

[PAR_WIG1350]

Originally Posted by AlexanderTheOK

Hate to resurrect a thread, but a friend and I decided to have some fun with our access to some NXT 2.0 kits and made a 2 wheel swerve using this design. The stacking he came up with is the most aesthetically pleasing thing I've seen all month. Didn't have time to program it, and I doubt we could have since I believe the nxt 2.00 motors use tachometers instead of encoders.

Still fun to drive though, with one phone controlling steering and the other controlling speed it vaguely resembles swerve movement.

I'm pretty sure it is actually an optical encoder regardless of how it is marketed. They are relatively good at position control, not sure about speed control. I think the only thing that would get in the way would be the unpredictability of the passive casters.

[philso]

Originally Posted by AlexanderTheOK

...I believe the nxt 2.00 motors use tachometers instead of encoders...

The NXT (and EV3) motors all have optical encoders in them, not tachometers.

[AlexanderTheOK]

Originally Posted by philso

The NXT (and EV3) motors all have optical encoders in them, not tachometers.

Yup. Looked that up later. I have some pretty good pictures of it though, so not too hard to put back together when i get access to the kits again.