Alternatives to Permanent Magnet Motors in EV traction applications

By Jeffrey Jenkins

In my very rst article for Charged over 10 years ago, I opined that the Switched Reluctance Motor, or SRM, would eventually come to dominate the EV traction market, if for no other reason than the fact that it is almost as cheap as dirt to manufacture. I won’t sprain my shoulder patting myself on the back for a prediction that took 10 years to (partially) come true, but it does seem that the days of the Permanent Magnet Synchronous Motor, or PMSM, and the AC Induction Motor, or ACIM, dominating EV traction applications are coming to a close, with both Toyota and even stalwart ACIM fan Tesla introducing reluctance motors into the mix (albeit not ones that purely employ magnetic reluctance, but more on that below).

Despite the fact that the PMSM and ACIM types dominate the EV traction market, neither technology is particularly well-suited to the job, mainly because of their limited overload and starting torque capabilities. Torque is a function of the magnetic eld strength of the motor’s eld structure, and that is obviously xed in a PMSM, because its eld is being supplied by literal magnets, but it is also limited in the ACIM, because its eld is induced into its rotor via transformer action, and like any electromagnetic structure employing a high-permeability core material, it is subject to saturation, or the point at which a further increase in current fails to increase the magnetic eld strength thus developed. Since traction applications require maximum torque at 0 rpm (to overcome both the inertia and static friction, or stiction, of the load), having a hard and fast torque limit is de nitely a downside. Furthermore, the PMs invariably employed in the PMSM are rare earth types—which, as the name implies, tend to be expensive—and most avors of these magnets—particularly the neodymium type—have a very low upper temperature limit before they start to lose their ability to resist demagnetization (in some cases as low as 80° C), an especially unfortunate weakness, because the only way to extend the constant power speed range of a PMSM is with eld-weakening, in which an opposing (i.e. demagnetizing) eld is applied to the eld magnets to reduce the back EMF, or the voltage produced by the rotation of the rotor, to below that of the battery voltage. at brings up another potential downside to the PMSM: since BEMF is proportional to rpm, if the rotor is spinning fast enough to require eld-weakening and the inverter shuts down (or misjudges the timing of the stator currents) for any reason, then it (and possibly the motor) will almost certainly be destroyed from the huge jump in BEMF causing an equally huge jump in current from the now rapidly-braking motor (a situation that is somewhat euphemistically termed uncontrolled generation). is failure mode can’t occur with the ACIM, of course, because the rotor’s eld can only exist through the action of the stator currents (and a di erence in the relative rpm of both), but the ACIM is much less power-dense than the PMSM (especially the interior PM version), so achieving a given power output requires a much larger motor, and that’s a signi cant drawbacks itself in an EV traction application. With all these downsides of both technologies, it’s fair to wonder how the PMSM and ACIM ever came to be used in EVs in the rst place—much less how they became the most popular choices—and a good chunk of the reason is that they both use the same inverter hardware that has been used to control the speed and torque of industrial motors for decades, with only minor di erences in the so ware and rotor position feedback to better adapt them to traction service. e least-radical alternative motor type—and one which has already been used by BMW, in fact—is the wound-rotor synchronous motor, or WRSM. ere is some irony in this, both because supplying current to the rotor requires brushes and a slip ring, which is almost anathema to those EV purists who shun the king of all traction motors, the series eld DC type, because of its brushes and commutator (to be fair, the latter wears out far more quickly), but also because it is downright ancient technology, predating even the induction type. Basically, the PMs of the PMSM are replaced by electromagnet coils in the rotor, which means the strength of the eld can be directly controlled, something that isn’t possible with xed-strength permanent magnets, of course. is makes eld-weakening to achieve high rpms a non-event, rather than something akin to juggling chainsaws, and makes minimizing the amount of reactive current that would otherwise slosh back and forth between the inverter and motor possible (this current does no useful work, but does use up some of the RMS current rating of the inverter’s semiconductor switches and cause heating from I2R loss). e downsides to the WRSM are a somewhat lower power density and a higher assembly cost for the rotor (a lower material cost, however, since copper is likely to remain much less expensive than any of the rare earths used in high-strength magnets). Also, of course, the eld needs a separate power stage to supply it, and the controller so ware is a bit more complex, but this is well-established technology, and the WRSM is eminently suitable for 4-quadrant traction applications (i.e. motoring and braking in forward and reverse).

Being able to use the rich ecosystem of existing electronic components, so ware and knowledge behind every industrial variable frequency drive, or VFD, is a huge advantage that cannot be overstated, and this is no doubt one of the major reasons that the axial- ux variation of the PMSM has received any attention, because it is de nitely a contender for the “most expensive way to construct a motor” award. As the name implies, the axial- ux PMSM orients its eld magnets parallel to the output sha on a disc that is sandwiched between the stator phase windings (and the “back iron” to complete the magnetic path, if necessary). is naturally results in a motor with a relatively large diameter-to-length ratio, and since torque is proportional to the square of rotor diameter, but only directly proportional to rotor length, a wide-in-diameter but short-in-length motor—a disc shape—is optimal for high-torque applications, such as EV traction. Securing the eld magnets in a disc- shaped rotor requires some form of banding on the circumference of the disc, so it is arguably about as prone to disintegration at high rpm as the surface PM construction, but it is undeniably less e ective than burying the magnets, as in the IPM construction. Furthermore, the push and pull of the magnetic elds from the stator poles exert torsional, or wave-like twisting, forces on the axial- ux rotor, and resisting these forces is not the strong suit of a disc shape. Consequently, the parts of the rotor that hold the eld magnets in place must have a very high strength-toweight ratio, which basically means using exotic (read: expensive!) materials like woven carbon ber composite. e combination of exotic construction materials and rare earth magnets in the rotor is the reason that the axial- ux con guration is one of the most expensive types being considered for EV use, and there doesn’t seem to be much that can be done to lower its cost in the future. For a more in-depth explanation of the axial- ux PMSM construction, see my article in the May/June 2020 issue of Charged. For those wondering if there is an axial- ux analog to the ACIM; well, it’s theoretically possible, but doesn’t really o er any compelling advantages over its conventional radial- ux counterpart.

Another alternative motor construction that leverages the existing VFD ecosystem is the Interior Permanent Magnet-Synchronous Reluctance Motor