Variable Diameter Wheels

[Dave Scheck]

Originally Posted by KenWittlief

as you enter a train station the tracks would be wider, the train would slow down and have more torque (stopping and starting power) (no need to tip the wheels)

then out in the great far-betweens the tracks would be placed closer together, and the train would take off like a bat out of you-know-where

If I'm not mistaken, wouldn't the track need to be changed in two axes? It needs to move laterally to be positioned under a different radius, and it also needs to be raised/lowered to make smooth contact.

This reminds me a little of a roller coaster called 'X' at Magic Mountain. The cars move along a 2-railed track like a normal coaster, but there is a third rail between them that changes height. The wheel that rides on this track is attached to a rack and pinion system that causes the car to rotate forward and backward. It's a pretty wicked ride...

Anyways, I would imagine that the same could be done in the case of your wheels. You could have a third track that rotates the wheel assembly. I don't think this would be the greatest of ideas, but it would be an option nonetheless.

[Madison]

Originally Posted by Dave Scheck

If I'm not mistaken, wouldn't the track need to be changed in two axes? It needs to move laterally to be positioned under a different radius, and it also needs to be raised/lowered to make smooth contact.

This reminds me a little of a roller coaster called 'X' at Magic Mountain. The cars move along a 2-railed track like a normal coaster, but there is a third rail between them that changes height. The wheel that rides on this track is attached to a rack and pinion system that causes the car to rotate forward and backward. It's a pretty wicked ride...

Anyways, I would imagine that the same could be done in the case of your wheels. You could have a third track that rotates the wheel assembly. I don't think this would be the greatest of ideas, but it would be an option nonetheless.

X has two cam rails rather than one, but essentially, you've got it right. It also has been a technical nightmare to operate and maintain.

I don't believe that the rail height would need to change as Ken's described things. Instead, the height of the trainset above the ground would vary with the track guage.

Railroad trucks have flanges on the inside of each rail that keep the train going in the correct direction around curves and through turnouts, however. Without a flange immediately abutting the rail, it seems that the trainset Ken's describing might tend to be thrown from a curve by its centripetal force. I suppose you could always super-elevate the turns so there is no lateral force at all, but that may be a bit extreme for riders.

[alphastryk]

Our team (1002) has actually been looking into the hemispherical wheels thing since we figured it out over the summer, but we haven't figured out how to get the wheels to have enough friction to move the robot, but little enough that the ratio can change.

To make sure your ground clearance stays constant, you pivot the "wheels" around a diameter.

I think we decided on some sort of rubber-coated aluminum spheres.

Don't know if this helps anyone with figuring this out, but it will be a really neat innovation for whoever can get it working.

[Madison]

Originally Posted by alphastryk

Our team (1002) has actually been looking into the hemispherical wheels thing since we figured it out over the summer, but we haven't figured out how to get the wheels to have enough friction to move the robot, but little enough that the ratio can change.

Given that the ratio change -- or pivot -- of the hemisphere is going to happen under power, isn't that motion effectively the same as rotating another wheel? If so, I don't see that generating sufficient friction is any more a concern here than it is to any conventional drivetrain.

[sanddrag]

For a rigid wheel (non deformable tread surface) a wheel contacing the ground plane will give you a line whereas a hemisphere contacing a plane will give you a point. This may be cause of concern in the traction department.

[Madison]

For as long as we're running on carpet, nevermind what tread or material the wheel is made of, there will be significant deformation at the interface between the hemisphere and ground.

[ChrisH]

Originally Posted by sanddrag

For a rigid wheel (non deformable tread surface) a wheel contacing the ground plane will give you a line whereas a hemisphere contacing a plane will give you a point. This may be cause of concern in the traction department.

Or it might make things better. It depends on the surfaces involved. If the playing field has a clean metal surface, metal wheels might actually be a good bet. Metal on metal actually results in a very high coefficient of friction and a small contact area enhances this. That was a key principle in the CVT that 190 did a few years back. We also used in on our slide drive in 2004.

However, if we are back to the standard carpet then the line contact would probably be better. Because of the carpet, not the wheel.

[greencactus3]

Originally Posted by ChrisH

Or it might make things better. It depends on the surfaces involved. If the playing field has a clean metal surface, metal wheels might actually be a good bet. Metal on metal actually results in a very high coefficient of friction and a small contact area enhances this. That was a key principle in the CVT that 190 did a few years back. We also used in on our slide drive in 2004.

However, if we are back to the standard carpet then the line contact would probably be better. Because of the carpet, not the wheel.

within a clean gearbox, metal/metal friction may be high but on a field. its hard to keep it clean. THAT clean. and even a small scratch may affect it.well for the highest gear its easy to have a "Conventional flat faced tire.... its just the low gear... which i guess is also possible but it can also generate a torque not just longitudally but laterally? or... well. a sideways torque... which may actually work too. kinda like toe-in (on the rear axle of a car would be a good example). aggressive tread would work for that. basically clawing into the carpet. kind of.

[Madison]

Ignore for a moment the horrid material attached to this rendering as Solidworks does a poor job with plastic.



Green represents the robot chassis and how each wheel assembly may be mounted.

Red represents the variable pivot mechanism. This doesn't presume to explain how the pivot happens -- but one could use sprocket and chain, gears, rack and pinion, or a cam to achieve the required roll.

Blue represents the drivetrain. The small blue shaft is the input to the bevel gear driving the semi-spherical "wheel".

The wheel, as it were, is not a full hemisphere. The pivot point passes through the center of the sphere and thus, a full half sphere would have interfered. It represents at its largest a wheel of 8" diameter and, at its smallest, a point that has no diameter.

[mechanicalbrain]

I believe team 116 did something similar. Im not talhing drive train wise but in the mounting of the wheels and the point contact. Maybe someone might mention how good the traction was? Actually wouldn't you get better ratios with concave disks rather then hemispheres? less rotation would be needed to make a large gear shift.

[greencactus3]

Originally Posted by mechanicalbrain

I believe team 116 did something similar. Im not talhing drive train wise but in the mounting of the wheels and the point contact. Maybe someone might mention how good the traction was? Actually wouldn't you get better ratios with concave disks rather then hemispheres? less rotation would be needed to make a large gear shift.

do you mean convex? because if its concave then the low or high will be the only points that can touch the ground with nothing inbetween.

[mechanicalbrain]

Yeah after I posted that I was kicking myself because I knew that I had gotten them mixed up. Another nice side effect which Epsilon Delta capitalized on is that this drive set up allows omni directional movement. Anyways, yeah I believe a convex dish would give better ratios with a minimum amount of movement.

[Madison]

Could you better explain what would be different about a "dish" as compared to a hemisphere?

I think you're suggesting that the wheel be made with an elliptical cross-section rather than spherical, but am not certain. An elliptical cross-section would still require a 90* rotation to move through its maximum ratio, but would reduce the movement needed to pass through the most practical ratios. It may be harder to manufacture, however.

[Tristan Lall]

Originally Posted by M. Krass

Could you better explain what would be different about a "dish" as compared to a hemisphere?

I think you're suggesting that the wheel be made with an elliptical cross-section rather than spherical, but am not certain. An elliptical cross-section would still require a 90* rotation to move through its maximum ratio, but would reduce the movement needed to pass through the most practical ratios. It may be harder to manufacture, however.

Wouldn't it also suffer from a variable ellipsoidal radius, meaning that the whole trunnion assembly would have to pivot to counteract the vertical motion of the centre shaft? It could be done with some sort of bar linkage, I think, but keeping the spherical radius constant is far easier.

[Adam Y.]

Originally Posted by KenWittlief

another application for these cone shaped wheels would be railroads and subways.

as you enter a train station the tracks would be wider, the train would slow down and have more torque (stopping and starting power) (no need to tip the wheels)

then out in the great far-betweens the tracks would be placed closer together, and the train would take off like a bat out of you-know-where

Believe it or not the wheels on a train are cone shaped even though you can't tell by staring at them. Don't ask me how I know this because it's from watching too much television. I think they are shaped like that because otherwise the train would derail easily.