Question about an interesting experiment video (two magnetic balls in tubes)

[Swamp Thing]

In this video...

https://www.physicsforums.com/media/magic-magnetic-balls-defy-physics.6951/

... a pair of magnetic balls in two vertical tubes, roll upwards when the tubes are rotated, irrespective of the direction you rotate them in.

A couple of questions...

[1] Say the tubes are mounted on nearly ideal bearings. We give them a little spin and release them, imparting enough kinetic energy to raise the balls through some height h. Clearly, the tubes will slow down as they exhaust their kinetic energy. But what will happen when height h is reached? Will they just stop turning, or will the balls start to roll downwards and force the tubes to rotate in the reverse direction?

[2] The presenter throws out a couple of hints that this phenomenon may be somehow philosophically interesting vis-a-vis time's arrow. Is this the case, or is it a trivial curiosity?

 

[A.T.]

[1] Say the tubes are mounted on nearly ideal bearings. We give them a little spin and release them, imparting enough kinetic energy to raise the balls through some height h. Clearly, the tubes will slow down as they exhaust their kinetic energy. But what will happen when height h is reached? Will they just stop turning, or will the balls start to roll downwards and force the tubes to rotate in the reverse direction?

I don't see why they would roll all the way down. Why would ideal tube bearings prevent them from finding a new equilibrium position at a new height? There might be some small oscillation around that new equilibrium tough.

[2] The presenter throws out a couple of hints that this phenomenon may be somehow philosophically interesting vis-a-vis time's arrow. Is this the case, or is it a trivial curiosity?

Around 11:06 he explains why the system has no time symmetry. I think this is similar to a one-way-valve or a ratchet, where you can also tell the time direction from kinematics.

 

[Swamp Thing]

I think this is similar to a one-way-valve or a ratchet,

I am not too sure if my understanding is correct as to why the balls' N-S axis should tilt away from the plane radial to the tube, when the tube is rotated. I am guessing that it is the total inertia of the ball, i.e. a combination of mass and moment of inertia, resisting the action of the wall's surface. Is this true?

If this is true, then there should be no upward drift in the following condition: Start with an initial condition where (a) the balls' magnetic axes are aligned within the radial plane (b) the balls have angular momenta around their respective magnetic axes such that their rolling would exactly cancel the movement of the wall's surface (c) they have no linear velocity (d) the wall is rotating at a uniform angular velocity such that it cancels out the rolling as stated in point (b).

Is the above true? If so, it also means that if we start with the tubes stationary, and apply a very, very small angular acceleration to the tube, then there will be very little deflection of the balls away from their relaxed / radial alignment, hence the vertical rolling will be quite negligible even after the tubes reach a significant angular velocity.

In other words, the vertical velocity that is ultimately achieved would be very much a function of the initial conditions and the history, with a spell of acceleration being required to really kick start the process.

Pheww..

So... any errors in the above thought process?

 

[A.T.]

I am not too sure if my understanding is correct as to why the balls' N-S axis should tilt away from the plane radial to the tube, when the tube is rotated.

I think this is because the friction acts below the effective magnetic force, so you have a torque around the radial axis, which tilts the N-S axis away from the radial plane.

I am guessing that it is the total inertia of the ball, i.e. a combination of mass and moment of inertia, resisting the action of the wall's surface. Is this true?

If this is true, then there should be no upward drift in the following condition: Start with an initial condition where (a) the balls' magnetic axes are aligned within the radial plane (b) the balls have angular momenta around their respective magnetic axes such that their rolling would exactly cancel the movement of the wall's surface ...

They would still have rolling resistance, that would shift them from their rest positions and tilt the N-S axis as described above.

 

[Baluncore]

Is there a short period, when you reverse the direction of rotation, when the balls change orientation, before they begin to climb again?

 

[Swamp Thing]

Is there a short period, when you reverse the direction of rotation, when the balls change orientation, before they begin to climb again?

I will watch at slow speed and try to see if that happens.