I have done some testing of similar ideas on my bike wheel set up, with little success, but time spent was enjoyable just the same.
I could never avoid the sticky spot
at the point of change from attraction to repulsion, although my wheel was rotating in a horizontal plane at the time and not taking advantage of any gravitational assist. In looking at your drawings two things come to mind.
1- you have a large imbalance, which I'm sure gravity would like, but bearings maybe not so much. Perhaps not as much imbalance is needed and might be accomplished by adding a second spoke opposing the first. The two sliding P-magnets could be connected by a rod and would be working together.
2- Now this one may be way out there, for I have very little electrical knowledge, and would very much complicate the build. But perhaps someone here could set me straight on this, for there are many bright thinkers here.
I see in a two spoke set up, sliding magnets moving four times in each revolution. Now here is the big IF, if these magnets were sliding inside a coil might a electrical pulse be produced, stored in a capacitor, and dumped to a solenoid to some how assist in getting past the sticky spot?
Best to you, Gene
(Thinker + Tinker + F.E.)