Gravity seems to yield unequal results.

[isodecryptor]

if you would like the best results for a simple setup, make the outside weights equal to the middle weight, and expect no more than a theta of 45 degrees in respect to the height.

[isodecryptor]

I have some strange results in my simulation for physion. I used the same weights on middle and outsides, a rolling circle and two springs that are holding the on the outside to the ground. I am not sure how realistic this software is but it is moving perpetually. I dont think it is trickery, or at least I hope not!, otherwise, it is very interesting and I am going to post it now. NO motors are in this system and all settings were on default. I re ran the simulation several times and the results came back to this perpetual movement. At least it is very interesting and worth taking a look. I will upload the file so you can have a look. I really hope the springs have some realistic calculations in there.

https://www.mediafire.com/?dk97mbnc9n3bdv5

[isodecryptor]

I have changed the spring constants k to many different values and it still remains in perpetual motion. What could this software not be calculating. There is the spring constants, elasticity, and friction.

[isodecryptor]

Here it is with the higher spring constant of 50. It works with the other but very slowly. https://www.mediafire.com/?dk97mbnc9n3bdv5

[isodecryptor]

Here is another one using just springs instead of outside weights! I really think that they need to update their software or I have an idea why it is doing this. I really think their is another force that is assisting this setup. The cscx has strange asymptotic behavior and springs force change with it. .  https://www.mediafire.com/?dk97mbnc9n3bdv5