Variable Diameter Wheels

well offtopic but square wheels would keep the axle movement linear if the ground it were rolling upon was half circles… theres plenty of articles onthat all over the place if you’re interested… like square wheeled bicycles.

waaaaaaaay offtopic but with all the ways people have done and attempted multispeed robots, anyone try a variable wheel diameter? not in the transmission but like the wheel(or w/ tire) that touches the ground… it expands, and itll basically be like a higher gear. and also give more ground clearance… so little junk wont be in the way when its going fast but when it pushes, the center of gravity goes lower andmore torque is applied via smaller diameter wheels.
if you’re tricky, you can probably pull that off automatically with some sort of centrifugal system.

wow, I think you just blew my mind!

a sort of cone shaped wheel with a bit of a curve to it.

when one end is contacting the ground the ‘wheel’ diameter is very small, the other end large, and inbetween, infinitely variable!

all you need is pnuematics to tip the cone towards one end or the other.

Holy monkey! so if you could put them flatish, you’d be almost un-pushable! But I think i’d try to run them with a motor setup Ken. It’d be smoother, and more variable, unless you rig up a special exhaust valve.

like this!



Okay, I think I’m correct now – wouldn’t this end up being a hemisphere?

If you started with the driving shaft parallel to the ground, the wheel circumference would equal that of the hemispherical equator. As you pitch the shaft toward vertical, the effective wheel circumference approaches zero as the hemispherical pole comes into contact with the floor.

Does that make sense?

I have examined other methods of using centripetal force to engage a shifting mechanism. I would like, someday, to build such a thing to better get an appreciation for its challenges.

whoa. i was thinking a MUCH more complicated idea. but that is nice! hm… if you’rw going to have variable angles on the axles, there has to be a way to incorporate that to do a multidirectional… something different from an omniwheel or a transwheel, or a rotating thing… w/e they were called. cant rememebr it right now tho… hmmm… darnit… shoulda thought about it myself and blown everyone’s mind at the competition lol. except i didnt have the variable andgle idea… darn… woulda been cool for rookies to come out with a totally new concept like that and had it work tho. oh well’ ill think of sumthing else :rolleyes:

yup. makes sense. although i doubt anyone’ll really need it to turn 90 degrees tho… so a hemisphere will work fine but by having a “custom” curved cone shape, you can vary the surface contact patch are too. well i guess diameter and tire material play an effect into it, the skinnier it is the easier it is to turn so maybe you can have a small contact patch in “high” and a bigger contact patch in “low”. and well you can make the transition between “highest” and “lowest” at different rates too. oh. you can even use different materials so high can be slipperier than low. WOW. this idea is incredible. you can have the grippiest tires possible in the “lowest”… you dont even need to be able to turn… and have an omniwheel at the “highest” so you can manuever as fast as you want… again. WOW… the possiblities of such a system…

wanna share?

oh… sorry… this is getting WAAAAAAAY off topic… anyone want to move this into a new thread that would be more appropriate?

I’m liking the idea of hemispheres as wheels. You could actuate their tilt with pneumatic clyinders if only high and low were desired, or you could do it with a lead screw for much more variation. I’m just wondering what to make the hemispheres out of (and how) and what tread material to cover them with and how).

What’s further interesting about using hemispheres as wheels is that you can maintain constant driving power to them. You can drive all four “wheels” with a single gearbox. You do not even need to vary the power output to vary its speed.

This, to me, is far more interesting the square wheels. Who wants to file for a patent? :slight_smile:

you could also have one of these somewhere in your drivetrain, and have a variable speed transmission

instead of having the cone/hemisphere in contact with the ground, it drives an internal wheel that is part of the drivetrain.



but again, with frictiontype gearboxes, more can go wrong. i think its more worth lookinginto having the cone/himisphere as the final part rather than within somewhere.
i have a feeling something of the like has been patented tho. i have seen tilt wheel rc toys… not sure if variable speed/torque was the point of those tho…
but if not, twould be sweet.

What’s further interesting about using hemispheres as wheels is that you can maintain constant driving power to them. You can drive all four “wheels” with a single gearbox. You do not even need to vary the power output to vary its speed.

if you drive all 4 wheels with the same gearbox how can you turn? or are you saying about the angleing?
a leadscrew…hmm… or wormgear…or…

By varying the diameter of the effective wheel on one side or another, you can vary the speed at which it travels a given distance. Make the wheels on the left larger than those on the right and your robot will turn right, for example. :slight_smile:

one of the benefits of a variable transmission like this: it can be in the high rpm part of the drivetrain, right on the motor for example

where the energy is hi rpm and low torque, so you would not need as much friction between the two surfaces, and the parts could be fairly small.

This is really what a toriodal CVT is. Just with one wheel.

oh MAN! they stole our idea! darn time travellers! :^)

In 2005 we had a long talk about “Dude, think about how cool a CVT would be”, and then we had a longer talk about “Dude, think about how cool a 300 lb robot would be”

Ken, this is me >:) smiling.

Easiest way to make your hemispherical wheels tilt is to have the motor and wheel all tilt together. 10 lb tilting assembly, with the weight of the robot on it. Sorry, but I’ll pass.

But actually… it can be done. If you think about it, as the hemi-wheel tilts, it does not change the height of the robot. All you need is strong bearings on the tilting axis, and thrust bearings on the hemi-wheel. All it takes is the will to do it, and soon enough you’ll see that it isnt that hard. You know, I bet you swerve drive is harder to do.

And as for me, I would control it with pneumatics. 4 cylinders, all at the same height, digitally set. When they reach the correct height, use a double and single solenoid to lock the cylinder where it is. fun fun. I will bow to the team that pulls it off.

I can’t say i’ll bow, but I will spend every spare minute ogling their drivetrain!