Omni-Wheel Balancing Robot?

Last time I built an omni-directional wheel which had powered wheels around its circumference which drive in a perpendicular axis to the main drive axis.
I looked at two existing designs for this, the first was OmBURo which uses helical gears to convey motion to the smaller wheels from a motor aligned with the main axis of the wheel. This allows differential drive between two motors, one driving the main axis, and one driving the smaller wheels. This works pretty well and allows the complete machine to balance on one wheel.
I also looked at the Honda U3-X, and the Honda patent for their active omni wheel. This is a very interesting design, using a set of rollers mounted at 45 degrees either side of the small wheels, which cause the small wheels to rotate as the two side mounted hubs are rotated against each other.
I made a couple of changes to this design, mainly adding a central hub to hold all of the small wheels around the circumference in place. The original patent showed them all mounted on a flexible shaft, but this is quite hard to reproduce with no stretch and keeping high torsion resistance.
I used thrust bearings, and ball bearings to keep everything aligned and to keep tension on the central hub.
Even though all my rollers and wheels have straight edges, so they don’t follow a perfectly round contour, I found that the natural flexibility in my wheels helped me out because as force is applied to the top of the wheel, it squashes the wheels against the side rollers at the bottom, and this seems to give us quite consistent traction.
I wanted to improve on this slightly though, and also reduce friction a little, so this time I’ve made new wheels with a rigid hub. These snap into place due to a feature in the middle, and this allows the wheel’s rigid hub to run on the axle rather than the TPU, but still gives us some friction to grip the driving surface. I’ve mounted each axle in a slot shaped feature so it can still move and slide to comply with the rollers on each side as force is applied to the wheel.
That seems to run ok, so the plan is to attempt to make a one-wheel balancing robot with it, just like OmBURo and the Honda U3-X.

Links:

You can support me on Patreon or buy my Merchandise:
***************************
***************************

Affiliate links - I will get some money of you use them to sign up or buy something:
***************************
***************************

Other socials:
***************************
***************************

Huge thanks to my Patrons, without whom my standard of living would drastically decline. Like, inside out-Farm Foods bag decline. Plus a very special shoutout to Lulzbot, Inc who keep me in LulzBot 3D printers and support me via Patreon.

HARDWARE/SOFTWARE

Below you can also find a lot of the typical tools, equipment and supplies used in my projects:

XROBOTS

Former toy designer, current YouTube maker and general robotics, electrical and mechanical engineer, I’m a fan of doing it yourself and innovation by trial and error. My channel is where I share some of my useful and not-so-useful inventions, designs and maker advice. Iron Man is my go-to cosplay, and 3D printing can solve most issues - broken bolts, missing parts, world hunger, you name it.

XRobots is the community around my content where you can get in touch, share tips and advice, and more build FAQs, schematics and designs are also available.

[preddy]

Pretty impressive how quickly you build complex electro-mechanical prototypes. As for the wheel my feeling is there's too much play in many places for the control system to reach stability. Those big motors are are causing large inertial changes.

[FoxSock]

you are the one person i actually get recommended videos for and im glad, your stuff is always fun and gives me something to watch that is actually informative while still entertaining

[LucGendrot]

13:05 the Pareidolia is strong on this one. He looks so tormented.

[thequickestfox]

James. The ridges you have added to the wheels have hurt you here. The force is being transferred by friction between the two sets of wheels, not by any tooth engagement. So by having ridges you are decreasing the surface area of the contact patch between the two surfaces. Having a lower durometer material for each so the contact patch expands and removing the ridges so you have more of the area entrained in the contact region will result in a higher maximum torque between the inner wheels and outer wheels.

[MarinusMakesStuff]

This is one of my favorite projects you did so far. I really think the flexible shaft is mandatory. I have done some play with a similar wheel back in 2015 when I was graduating and we had to connect up the small wheels on the main hub to get a decent result. This is why I already commented this on the first video. The flexible shaft will help a lot, probably also maintaining creep a bit better. Good luck, I can't wait to see #3!

[ddrjoe]

A 3rd Encoder on the thread which is sticking out of the sides could actively cancel out the creep, as the controller will be given the exact position of the middle wheel hub.

[H...]

You could say that belt tensioning method involved... putting a spanner in the works.
I'll see myself out

[danielrisueno]

James: "Next video, we'll break apart a kids toy"
Me, a 30 years old: I'm SO into it

[setSCEtoAUX]

I really love the bass line in the music that accompanies the printing footage around 2:00. The tune in general is great too.

[andyking]

I like the wobbly face the robot looks like at the end, wicked videos mate

[lpbaybee]

Honestly, I think I enjoy the videos where you have difficulties getting something to work the most! Seeing you work through a problem and discuss the problem you encounter is fascinating!

[Roboticlay]

This project is turning out to be quite difficult. Keep it up! I can't wait to see how the toy accomplished it, it's got to be the flexible drive shaft..

Thank you all of James's sponsors! You're the best!

[SullenSecret]

12:39 The robot looks like a person who is uncertain about his existence.

[UB]

12:39 The robot's face seems to sum up this experience pretty well.

[daves]

I like the way you are totally happy to show your try, fail, try, fail, work a bit, work a bit better and finally work really well design process!

[TheBeefiestable]

I love the heartful analysis of a difficult problem, thanks a bunch

[BrenoLuna]

Do you reckon having diamond-shaped corrugations on the main wheels would allow better gripping with the smaller ones on either side?

[zippythinginvention]

Why not simply angle the tread on all the little side wheels to mesh with the wheels that interact with the ground?

[GunGryphon]

Could you add some sort of encoder to the center wheel so that your software can accommodate for the drift by actively tuning the motor speed? A combination of position data for the outer and inner wheels can act like a traction control system.

[WurdBendur]

Would it be possible to fit a pair of bevel gears between each pair of rollers so they all turn together? Or u-joints if there's not enough room. It's a lot of parts, though.