Marble spiralling inside a cylinder

[A.T.]

The superball is changing its course by applying its own spin to the surface and meeting resistance.

The same happens with the ball rolling in a cylinder. Just continuously: infinite number of infinitesimally small course changes, with the same net effect.

You've re-befuddled the issue by introducing the spurious example of the square box.

One could use a hexagonal box, or an octagonal box. Or let the number of sides (and bounces per cycle) go to infinity, then you have rolling in a cylinder.

It will roll - but its rotation will not then transfer back into motion.

This doesn't make sense. If it has enough traction to transfer motion into rotation (start rolling), why should it not have enough traction to transfer rotation back into motion (change course)?

 

[DaveC426913]

This doesn't make sense. If it has enough traction to transfer motion into rotation (start rolling), why should it not have enough traction to transfer rotation back into motion (change course)?

Try it. Give a marble some spin when you throw it. See if it bounces off to the left like a superball.

 

[sophiecentaur]

Surely, once the ball has been in contact for a reasonable time and is rolling at the right, peripheral, speed, there is much less friction force needed to turn it/ spin it and the surface of the ball is not very relevant. What happens with a superball during an impulsive bounce is not like the original scenario at all.

 

[DaveC426913]

Surely, once the ball has been in contact for a reasonable time and is rolling at the right, peripheral, speed, there is much less friction force needed to turn it/ spin it and the surface of the ball is not very relevant. What happens with a superball during an impulsive bounce is not like the original scenario at all.

This is why I think the bouncing under the table is a red herring. It'll work with a superball.

superball: High relative mass, high friction = high transfer
marble: low relative mass, low friction = low transfer

But we're just reiterating the same argument now. No one's brought anything new.

 

[A.T.]

Try it. Give a marble some spin when you throw it. See if it bounces off to the left like a superball...

... and then try to catch it with your teeth! And if you can't then I'm right.

Seriously though, I never claimed that a marble will bounce just like a superball. And I wasn't talking about giving spin to the marble with my hand. Here is again what you said:

It will roll - but its rotation will not then transfer back into motion.

And here my question again:

If it has enough traction to transfer motion into rotation (start rolling), why should it not have enough traction to transfer rotation back into motion (change course)?

This is why I think the bouncing under the table is a red herring.

The bounce under the table is a very simple case, because the ball moves in one plane, and its spin axis doesn't change, so you have no gyroscopic effect.

The bounce around all sides of the square box already involves the gyroscopic effect, because the torques applied during the bounces are not parallel to the angular velocity.

superball: High relative mass, high friction = high transfer
marble: low relative mass, low friction = low transfer

I agree about the friction, but what do you mean by "relative mass", and how is it relevant here?

 

[A.T.]

I tried it, and as expected the ball came back consistently, after 3-5 bounces. I even hit the camera by accident. Here the video (it's a cheap camera at only 30fps so you have to watch closely):

https://www.youtube.com/watch?v=qdBL41lUzl8

This bounce version is basically a discretized version of rolling in a cylinder. Here the momentum is transferred in a few discrete steps. It might be simpler to explain/understand than the contious rolling case.

I remade that video with a better camera for slow motion, recording at 400fps, playing at 30fps.

https://www.youtube.com/watch?v=AfPhuwBItB4