I've had this puzzle for a couple of years but never been able to figure it out. Regulator manufacturers for typical 3-phase alternators provide the 14.4V by 'shorting' the phases. Obviously this comes with great losses but I cannot understand why this design is still in use. My electromagnetic theory is basic but I assume that somehow after a certain angular velocity back emf induces positive torque into the rotor which lessens the braking effect. I did intend to test that on a bench but I assume some of you at least have tried it or know the theory behind it. Is there a graph somewhere that will show braking torque vs RPMs? This could also apply to wind generators.
Edit: The stator is on the bike, contains several wound inductors connected into a 3-phase delta configuration. The flywheel connected to the engine/crankshaft has permanent magnets. The configurations I've seen had magnets inside the flywheel, which would then spin outside of the wound stator. Three wires from the stator go into the 'black box', another (usually) two wires come out.
The rectifier part is a standard 6-diode bridge. Regulator part is a zener-like part that triggers the shunt to all three (incoming) phases simultaneously to the ground after the bridge. Shorting is done via thyristor-like parts, so the zener triggers the gate for the SCRs. Typical generator power is about 400W@9000rpm, typical regulator/rectifier rating is about 20A. They get really hot, especially when no consumers are turned on. I suspect also consumers (headlights) also play a role in smoothing the resulting waveform.