Sum of the energy

neufneufneuf

I tested this device on Ansys Spaceclaim, I poste a video here:

At start, the orange disk don't turn around itself. The friction is very low. I start the red arm, the disk will increase its angular velocity around itself:




If the angular velocity of the red arm is [latex]\omega_0[/latex] then the angular velocity of the disk will be at [latex]cos(\alpha)\omega_0[/latex] with [latex]\alpha[/latex] the angle of the axis of the disk from the vertical. In the last example the angle is at 62°.

I understood why the disk turns around itself at start. The orange disk has more energy when it turns around itself, so the red arm must lost an energy.

I would like to know if the torque to the support is applied like that :



I'm not sure because I can change the thickness of the disk. And even the support gives a torque, if I explode the disk after it rotates around itself, the linear momentum is not conserved.

Second point, if you look at the video, a small black circle is fixed on the orange disk to look at the angular velocity. So, if the device is put on a table, a fixed point on the orange disk can move up. So I can attach a spring (the spring will follow the disk in the plan (x,y)) to the disk, the spring will increase its kinetics energy because its length increases. The disk rotates more around itself, the kinectic energy increases. The support don't receive any torque so it can't lost an energy:





So, the support must lost an energy, have you an idea where ?