Hadn't read that thread but gotta admit the points Jim's raised in the OP pretty much mirror my own more recent conclusions...
The trick is gonna be doing it without draining the momentum of the source frame.
If everyone currently thinking about gravity wheels instead focused on how to apply vertical rotation to leverage the same net effects as an N3 break, we'd be rolling in no time..
It's not a gravity wheel we're after, it's an effective N3 break.
There's something that an N3 break does, that vertical rotation can also do.
If you can find it, your input energy will remain constant, relative to the rest frame of the wheel's inertia, and so rising linearly with velocity, while KE from the static frame squares with velocity.
This is not merely a "theory" - it is unimpeachable logic, the incontrovertible implications of the known facts. It is all that is left, the only clear energy asymmetry that fits with and resolves all of the other clues.
Eliminating the possibility of gravitational asymmetries is not easy. Our first priority is to question everything, especially first principals. While it's easy enough to digest the obvious objections, there's always a "what if?" that places a prospective system beyond the bounds of a given classification - the potential range of counter-balancing schemes limited only by imagination. It takes time and effort to draw a line under it, but it can be done. I have, so has Jim and doubtless others - my whole point here is that it's the inevitable, inexorable conclusion of a full consideration. An objective axiom, and a waypoint that everyone will eventually reach.
The only other field left then is mass, inertia, momentum. and KE. And here, straight away we see there IS an implicit energy gradient - every mass in the universe has the net KE of the entire physical universe, relative to the sum of all possible reference frames. And this applies equally to rotating frames. A wheel rotating at any given speed is also rotating at every other speed possible, in both directions - all directions, actually - and all at the same time, without any conflict. In fact, if red shift horizon is taken into account, then every mass has infinite KE relative to those beyond it. But all of this is completely trivial and irrelevent of course, unless we can tap a little of it.
More specifically, we
don't wanna tap it. That would be a rubbish energy gradient. We already have flywheels. What we want is to
make a source frame that can be sustained indefinitely - either freely replenishing its tapped momentum (which just seems to shift the goalposts), or else finding a way to beat N3 and so extract excess work from the frame without slowing it down.
This is our endgame, as it was Besslers... so how to proceed? We know his masses moved in and out radially, so what changes with radius? CF / CP, obviously (as Jim points out), but such translations at constant velocity are just as energy-symmetrical as GPE loops. So that seems futile.
But something else changes, too, with radius - displacement; A pair of masses at the perimeter of the wheel are travelling a greater distance per cycle than when they're closer to the axis. Work performed under the same force over the same angular increment has a larger displacement the greater the radius at which it is performed, for a given RPM. So perhaps there's another angle of attack for modulating the d of an F*d integral..
As for actually tackling N3 head-on - or even turning it against itself - i'm sure we've barely scratched the surface here. If we'd expended half as much effort on N3 as we have on gravity... but i've said it already. Scissorjacks seem to provide one approach, at least momentarily, but the problem remains of how to tap the gain without equalising the FoR difference. Logically, we should rather be anticipating the exact opposite dynamic - that as net velocity grows the FoR divergence will widen, and so thus the asymmetry, our PE gradient.
So while i've no dog in whatever races might be running here, Jim's right - it's "motion from motion". The more, the better. Ultimately though, i think that a more succinct framing of the issue should center on the N3 barrier that stands in its way.. "motion from controlled counter-motion" isn't quite so catchy, but this is the task ahead.