Hi.....there is a very real problem if you work with partial grooves.

For instance, if you had a box and the face in front of you was at an angle of 45 deg, any attempt to push the box forward would also push it sideways with equal force.

Having the balls impinging on a partial groove in the rollers is tantamount to them pushing on the roller grooves at the 45 deg angle not as a direct axial thrusting force to the frame.

The action of the screw when it's rotating forwards (whatever) is to apply force to the back side of the balls as they are deep seated in the screw thread form........ this force on the back side of the balls is applied to the back side of the roller groove, and if it is as full a form as the screw, the force will be converted to axial thrust to the frame, the roller groove acting as a fulcrum.

There will be some backwards reactionary force to the roller but with a full form thread profile not enough to consider important.

HOWEVER..........with a half form to the roller thread profile, the balls will exert the force on them at an angle to the roller of approx. 45 deg (whatever)........putting undue stress on the roller ball races.

True, the 45 deg force that acts on the screw(s) side grooves acts in both directions and the roller fulcrums the partial load in the axial direction, but at the expense of undue stress to the roller bearings.

Having a roller with a partial groove means the second roller, also with a partial groove, is acting in the opposite direction....the front side of the ball...... and so the full force from the screw, via the balls, is applied to just a partial segment of one roller profile.

I agree that it could work, but for how long, and how long would the tiny bearings last.......they are not sealed in that small size, at least the ones I have seen, so inclusions are a very real problem apart from lubrication.

My opinion is that this design is compromised due to the design calling for two grooved rollers, (to maintain ball contact with the screw)...... and as the space for two rollers against one ball limits the roller diam and subsequently the bearings too.....radial support for the rollers will also be in doubt.

BUT........the proof would have to be in the pudding......it would have to be made to this design to prove the validity of the claim.

In the very last design I proposed, the design has one set of rollers or double roller sets, as the case may be, and can be 10mm diam with a full groove profile to ball contact, and the bearings for the roller radial support are not mounted on the rollers but against them......the rollers actually float between the thrust bearings and the radial bearings........bearing sizes are therefore substantial and sealed too.
Ian.