A couple of thoughts:
- You seem to have tried both solutions at once. Either try the rigid link with a Cartesian constraint; or a cylindrical theta displacement applied to the outside edge of the end surface (I think the units of theta are mm). I'm not sure how a point will respond to a theta displacement, given that it has zero radius.
- I've had problems with points on the model geometry (they seem to become nodes of solid elements) - try moving the point out into free space.
