![Image](http://i.ibb.co/n8ThfDm/Alt-OB-Torque.gif)
..no prizes for guessing it'd be conservative. Turns out the work done each cycle is increasing; presumably due to the rising wheel velocity adding to that of the lifts.
The logic loop applies whatever the absolute wheel speed to the lever motors as the weights pass BDC - thus setting their speed relative to that of the wheel - whenever the following weight is in the 2nd quadrant
I've considered changing the code to control for 'acceleration' rather than 'velocity' - eg. i could set each motor to accelerate then decelerate at equal rates for each half of a lift to eliminate spikes & dissipation, but it seems unlikely to make any real difference.
Still, it's an alternative means of applying OB i suppose - whereas my usual go-to would be radial lifts with angular drops, it turns out you can equally use angular lifts only, still achieving a +/- G-time asymmetry and net momentum gain each cycle, if without yet decoupling the input FoR from the absolute frame. Surprising it never occurred to me before..
Bottom line however is that i'm looking for a momentum gain cycle that has a preferential speed, so as to be consistent with B's hypothetical slow & torquey wheel; this obviously can't do that, so another trick learned perhaps, but no cigar..