well, I HAVE read much literature on bearing design and engineering and disagree with you re: side vs. radial forces on a ball bearing - if ball bearing were designed to have the same side forces as the "normal" forces, then there would be no thrust bearings, conical bearings, etc.
Think about it - in simple terms, instead of having the force riding along a large portion of the ball (as it is "normally") you're taking the force and having it ride along the edge of the race as it contacts the ball - so all the force is riding on two very narrow channels.
On the other hand, a conical bearing spreads that force out by moving the races' force structure. A thrust bearing goes even farther in that direction.
Here are a couple of illustrations where you can see how the design of the conical (tapered) bearing has been designed specifically FOR a thrust force, as has a thrust bearing.
Note that ball bearings we use are rated ONLY for radial forces (forces that are perpendicular to the shaft) and NOT for thrust forces (forces that are parallel to the shaft)
Now, that being said, I've tried conical, thrust, and "normal" bearings and the conical just doesn't work. You can't apply force to it and get the same friction response that you get with a normal bearing. I've also tried thrust bearings and, when they work, they work amazingly well - but that was ONLY on 1/12. I could NEVER get the danged thing to work on 1/10th - could not get a consistent force set on it and it would not behave consistently.
So my only conclusion was that our abuse of a normal bearing is the only way to do what we're doing with them.