The basic function work quite nice, for a given PWM value, the rotation speed of spinning disk convergates to an according value. However the transient response is quite long (about 6 minutes).

The damping factor in cause of friction seems for a large range constant, only for slow velocities it seems higher.

With a few more experiments and analytical consideration it should be possible to create a (open-loop or closed loop) control circuit which takes a specific speed as set point.