Jude: Thanks for the bearing explanation, and the Texas A&M paper link.
Regarding the yaw control:
You have uncontrolled gyroscopic precession forces pretty much under control with the two pairs of two spools spinning opposite each other. Great idea. Why not use gyroscopic precession forces to your advantage?
In the image of the new design you're contemplating:
You have four spinning gyroscopes in the form of your cyclorotor spools. Look specifically at the upper left spool and the (so far) unconnected right end of its axle shaft. If you were to use a servo to raise the end of that axle slightly you would generate a counter-clockwise yawing moment due to gyroscopic precession. Lower that end and you get a clockwise yawing moment. And when the craft yaws, moments from the other precessing gyroscopes (other spools) virtually cancel out, provided you match the movement on the opposite end of the axle on the bottom right spool - and you get twice the yawing moment for a given movement of the axle end. If you only control one of a pair of spools then you will get some rolling moment produced. Further, you are actually tilting the net lift vector in the correct direction on each of the opposing spools to give additional yawing moment, though I think the gyroscopic precession moment could be significantly larger than the tilted lift depending on the angular momentum of the spools, the rate of movement, and the amount you actually tilt the spools.
You don't quite get it for free - you would need to add two servos (only one servo if you are REALLY clever with the mechanical linkages), plus linkages if you want to locate servos closer to the center of mass, maybe bellcranks to change direction, and some sort of secure pivot or hinge on the opposite end of each spool axle, but I think it would be a relatively clean method to achieve yaw rotation without hanging some other spinning part off of the vehicle. You would also introduce new forces on the blade pivot arms with this control method, but they would probably be relatively small in comparison to the centripetal force needed to keep the blades from pulling the whole thing apart as they spin.
Note that the yawing moment generated is proportional to the *rate* that you raise the axle end (plus the effect of the tilted lift vector). There will likely need to be some "wash-out" logic where, in order to stop the yawing and return the spool to "flat", the movement will likely need to be fast outward from neutral to get the yawing moment, with a slow return to neutral to be ready for your next yawing maneuver. Note this only works well when you don't need a steady state yaw rate - just quick bursts to change or stabilize heading. I guess if you raised the axle enough (trim out a yaw imbalance?) you would get a small steady-state yawing moment by the opposite tilting of the spool lift vectors.
Basically, the yawing torque generated is the cross product of the angular momentum (moment of inertia * spinning rate of the spool) and the angular rate at which you are forcing it to precess. If anyone is interested, please check me for errors, but this does track with some observations and quick measurements I have made.
And I call dibs on being made a co-author on your paper if the method works ;-)
