~Rukmi Dutta*, Lester Chong, and Fazlur M. Rahman~
Rotor and stator losses can be divided into hysteresis losses and eddy current losses. The eddy current loss is proportional to the frequency square, so the slot harmonic losses dominate compared to the fundamental component. The other side of loss, hysteresis, depends on the material loss coefficients. You need to know the max flux density, hysteresis and eddy current loss constants of the PM, rotor and stator material, which I definitely don't do for the Scorpion HK I-IV. You also need to know the Steinmetz loss constants. Once you have answer for all of these things, you need to examine the loss at each harmonic order's frequency in and then add it all up for total loses. But without all that Stefan, intuitive vision should relay
none is simply
better. As you can see knowing this number accurately depends a lot on knowledge of the core and PM materials, but if I just throw out a number out there on a good core material. Let's say a fundamental frequency of 300Hz. You will likely lose several watts each to the stator and rotor, but even more at higher fundamental frequencies and the higher orders. This was just the base fundamental frequency and has nothing to do with the additional eddy losses associated with the inverter PWM carrier.
Maybe you know someone that will share a FEM model with you of the HK series you have with the traditional DL wrap. You could possibly make some rotor and stator loss determinations easily from that. I suppose you could isolate these figures simply by calculating your winding,frictional,and windage losses. Less that would be your iron, stator, and pm losses. Another PMO may be just examining the temperature of the two after prolonged use. If you could watch a temperature rise over time u may be able to determine the power needed for that to occur in a particular material.
Without the new wind the only things you could do to reduce the losses is rotor and stator re-design with new material selection or pm segmentation or skewing. Have you ever seen the loss calculation differences at your loading points between I-IV?
Generally speaking at 500hz the loss of a material like m19 might be as much as 12% of the total output power and this would result in as much as a 10% difference in efficiency.