My Theory! I can't find why it wouldn't work! Big suprize!

[bottorml]

There isn't really a theoretical limit to the idea just a limit for each weight and radius difference combination.

[bottorml]

Another thing to consider is the effect of having the movable weights moving perpendicular to the direction of rotation. If you get even slow rpms out of this thing it will vibrate horribly. If you want smooth motion you need to move in the plane of rotation and you need to do it as smoothly as possible. There are of course practical limits.

[nwman]

Doesn't work sorry!

[bottorml]

OK, how about this: make the spokes 2 rods that acts like guides for the sliding weight. The middle section of each weight is a roller. Put stationary ramps at the top and bottom of the wheel fitted in between the guide spokes. Make the ramp curved so that as the weight makes contact it is always rolling down the ramp at the same angle as it travels either up the guide spokes or down them. That could be a way of moving the weight back and forth on the spokes, but how do you keep them in position?

[wattsup]

@nwman - Now that I have seen your video, I have a better idea of your system. I noticed on your second try when you manually flipped several magnets to initiate the rotation, the top magnets did not flip due to the constant speed of over-travel. I think you may need to add another stationary magnet to the top and bottom, not to increase the magnetic field but to make the field longer. As the speed increases the time in the field will decrease making the flip harder to achieve.

Before I saw your video, I had an idea on such a system as shown below. Of course the distance of the A and B magnets must be ajusted to allow minimum sticking while assuring that the magnets are raised. I also added a speed adjustment lever that changes the inner and outer magnet positions to raise the rotating magnets later in the rotation.