If the coils are separate it doesn't matter in the slightest which way they are wound, you have 2 windings each supplying a load, if you reverse one coil you will reverse the phase of the external emf, but each load will still contribute in opposing the rotor mmf.
Think of a transformer with 2 secondaries, if you connect them opposing you get nothing out. If you connect them in parallel and in phase and ran 2 10W bulbs you would have exactly the same thing as each secondary running a 10 watt bulb. If you wound 0ne secondary the other way the thing would still see the same load with separate windings.
Flux[ Parent ]
If you are rectifying then the rectifier presents lagging pf. You may see marginal improvement by adding capacitors across the lines to bring pf leading. Unfortunately again saturation will prevent you seeing much improvement and the higher leading current will increase the resistive loss.
Now as you are using high voltage and high frequency I am going to offer you one ray of hope, if you use SERIES capacitors then if you can get the right value you should form a series resonant circuit with the leakage inductance and it should cancel leaving only R. I can't promise that this will work but it does in the case of a phase converter and you can raise the third leg voltage quite a bit. Normally the capacitor values are beyond reality but in your case it may not be so bad.
How much of the inductive voltage limiting has to do with efficiency is unclear to me.
Certainly for constant power output the effect is not a lot, but if you want to extract the maximum power out at the 50% efficiency point, it will certainly drop the output power away from the KVA = ~rpm^2 curve.[ Parent ]
