Unbalanced lever: does it rotate, and how?

[Tuber]

Here's a puzzle that I recently saw, that sparked quite a discussion:
http://www.interfacts.org/unbal.jpg

A lever can pivot in the vertical plane about a fixed position. The lever carries two equal weights, one at each side of the pivoting point. There are 2 slots in the lever, that allow the weights to slide along it.

An asymmetric guiding rail limits the movements of the weights. The left part of the rail is in the form of half a circle, the right part is half an ellips. The idea is to have a lever with its right part longer than its left part, so as to create a torque in positions 1, 2, and 3; there is no torque when the beam is in the vertical position (4).

If the beam is released in position 1, what will happen? Interestingly, people gave 3 different answers: the beam will not move, the beam will rotate anticlockwise, the beam will rotate clockwise.

What's your opinion? Better yet would be if someone could model this thing in software

[broli]

This has been modeled to death already. And can be explained by simple force analysis. If you pretend the oval shaped rail wasn't there then surely one can apply logic and use the lever theory? Well that's WRONG logic. You cannot ignore the force acting on the rail. Because if you decompose the gravity force in a force that is normal to the rail and normal to the lever (this is the force that causes the torque). You will see that this torque causing force is suddenly bigger than if there was no rail.

Edit: I made a video of a simulation that shows this. The left weight can move up and down and can also move along the lever back and forth. If I asked you, purely using lever logic, what way the lever would fall. What would you say? Run the video and you will see the solution.

http://ziosproject.com/NJ/exvid65.avi

But stopping people in their tracks is not my thing. I always want to know whether the new thing we learn can be used to our advantage.

[Tuber]

This has been modeled to death already. And can be explained by simple force analysis.

Sorry, but you didn't answer the question 
Will it rotate by itself, when released at position 1? And if yes, in which direction: CW or CCW?

[broli]

The direct answer is no. But that will teach you nothing now will it. Check the video I added to my previous post.

[AB Hammer]

Sorry, but you didn't answer the question 
Will it rotate by itself, when released at position 1? And if yes, in which direction: CW or CCW?

Greetings Tuber

It may go counter clockwise for just a little bit. The design you have posted is a good reference to placement effects on a rail in a wheel. No matter what you do the contact with the rail will complicate things a bit due to how the weight disperses. In armour there is a thing called laminated plate which mean that if you hit one it will disperse the impact through several pates making the impact of less effect. The same with any contact to a non moving device in a wheel. So your design will not work, it has been tried by hundreds of people with the same outcome. Lost time. Sorry on this one.