...From what I understand, a shunt regulator shorts to ground to regulate. A series regulator only completes the circuit when voltage low/below threshold. When a series regulator opens the circuit, there is no current, but that results in massive voltage in the stator. In theory, this is [just as bad | worse | the same]?
I am thinking/hoping that the SCR circuitry might be to blame, not the fact that it's a shunt design...
Trying to put this down without writing a book is challenging. It's difficult without a chalkboard and expressive arm-waving.
Basically, just thinking out loud. A lot of things are happening here. I hope the idea gets across.
Apologies to any EE's reading this, please excuse liberties taken in this rambling.
First, an over-generalized, over-simplified review of the basics:
Semiconductor Vf (Voltage drop in the forward, conducting direction) - Silicon-based semiconductor devices (diodes, transisters, SCRs, triacs, FETs, ...etc) demonstrate a voltage drop when current flows in the forward/conducting direction, normally around 0.6 - 0.65 Vdc, sometimes as high as 1.1 Vdc for large-current rectifying devices.
This Vf is not the same as a resistor, which demonstrates a voltage drop proportional to current. As such, a semiconductor permits essentially unimpeded current flow once the Vf voltage requirement is reached.
Inductors - Current is reluctant to flow through an induction coil as it's trying to build a magnetic field. But, once the magnetic field is established, this current flow is reluctant to stop flowing, like a freight train with bad brakes on ice. Ref: Inductive kickback, Inductor freewheel effect.
Inductive heating - Occurs when the magnetic flux changes, not just in the stator core, but also in the rotor, because it's experiencing reactive magnet field changes in the nearby stator arms.
Now, what happens in a single coil of the PMA (viewing this in very slow motion):
As a magnetic pole of the rotor approaches the stator coil, the magnetic flux changes in the stator arm, and causes current to start flowing through the coil.
Also, a little inductive heating occurs, in both rotor and stator, based on the magnitude and rate of change of this magnetic flux.
When the magnetic pole of the rotor is aligned with the stator coil, the magnet flux in the stator arm is established and the current flow is established.
As that magnetic pole of the rotor departs, the magnetic flux in the stator arm remains and current continues to flow, for a short decay period.
When the opposite pole of the rotor magnet approaches the stator coil, the magnetic flux starts reversing, and current flow thru the coil slows/stops/reverses.
Also, a little inductive heating occurs, in both rotor and stator, based on the magnitude and rate of change of this magnetic flux.
And the cycle repeats.
Current delivered out of the coil to the rectifier/regulator is replaced by current sourced from the other coils of the other 2 phases, thru the common tie point of the 'wye' configuration, or simply from the other end in the 'delta' configuration.
Now, here's the theory scenario:
When the regulator shunts the output of this coil to ground (via SCR or MosFet), the current will flow while it's above the Vf of the shunt device. This current is sourced from the backside by the other 2 phase coils thru their rectifier diodes tied to ground, closing the loop, but includes their Vf. (Need a better schematic here, I hope you can visualize this).
This Hayabusa schematic uses a delta-type winding, but if you follow the current flow arrows you can see how the circuit is a closed-loop during output shunting:
If this shunted current doesn't decay rapidly enough, the residual magnetic flux in the stator arm will induce a larger-than-normal flux change when the next pole of the rotor arrives. Causing additional induction heating. And the cycle repeats.
If the shunt devices clip their grounding action soon enough, the current will be stopped before the next pole arrives, reducing the magnitude of the flux change, and the added inductive heating effect.
Factors influencing this would be:
The strength of the rotor magnets.
Inductance/reluctance of the coil/arm
The Vf of the semiconductors.
Control circuitry of the shunt devices (SCRs just conduct until a specified cutoff value)
RPMs and number of poles of the rotor (more magnets & rpm = less time to decay)).
Response rate of the shunt devices.
Plus more.
It would seem that when the engine is running at low enough rpms, that the residual flux probably decays rapidly enough to avoid this additional induction heating.
It would also seem that very efficient, long-duration-conducting shunt devices would produce this undesireable effect. Add to that very powerful rotor magnets.
A series-type regulator would clip the output current and arrest this coil freewheel effect. I can see why there's been no reports of PMA overheating with these.
Rambling mode off...