- there is no bearing in the design from ricoh, a bearing could introduce a wobble so is not desired

  - the torque ripple depends on the mass and speed you r spinning at.  angular kinetic energy increases squarely. 4 times less energy is only twice less speed. Friction increases with speed if the mass is unbalanced (which is the case here)..

- the prism cannot move in the z direction if gravity is stronger, if it moves in the x,y direction this does not matter (see my images in my blog on optical weak weakness) as long as the rotational speed is constant. It should stabilize due to the gyroscopic effect

 I aim to improve the setup and measure the torque ripple and check for instance if it reduces at higher speeds or by increasing the weight of the rotor.  rotationally it has to be very stable, so it is likely I have to go for cnc parts in the end..

If I'm understanding the design then the metal central axis isn't fixed to anything on the board (e.g. a bearing).  This means the magnetic fields generated by the coils are going to push the prism in a 3-dimentional vector since the bearing would normally absorb the undesired directional force.

You could address this issue computationally but it's seems like compensating for torque ripple with extra dimensions.  The alternative is to keep it on a magnetic track.  I don't know if either of these options are better than a bearing but they are far more complicated.

There is also the fact that this is supposed to function when positioned vertically, which seems like it would either ruin or complicate the computational method.