These guys seem to have a complex solution in search of a problem. Single-speed EV transmissions are elegantly simple and reliable; I see no evidence of a high-speed issue.

Anyone who's ever seen a tens-of-kW alternator the size of a coffee cup knows the potential of up-speeding electric motors.

Using a multi-speed transmission allows the peak torque of a motor to be slashed and also run it at higher RPMs at low vehicle speeds, making it both lighter and more efficient.  Running a motor at high torque requires high current and high resistive losses, while low speed means low back EMF and low real power.  This adds up to reduced efficiency.  A multi-speed transmission relaxes the constraints.

Go back to the 20 kW coffee-cup alternator.  That is what I saw attached to a gas-turbine sustainer engine in a vehicle lab once.  Operating at turbine speeds, even the tiny torque of a turbine carries a lot of power and a tiny high-frequency alternator is all that's needed to turn it to electricity.  The power of an electrical machine is directly proportional to frequency.

The upshot is that you cut the size and torque of the motor by a factor of 3, and use an additional 3:1 reduction for low-speed operation.  When the motor gets near its redline speed, you upshift.  The smaller motor is more efficient, lighter and costs less.  The greater efficiency gets you more range out of the same battery, to list just one knock-on effect.

induction and cvt

New extra hard coatings being developed could probably solve the extra wear from very high speed e-motors and transmission gears?

It is something to be developed.

Wonder why TESLA is not interested.

Michelin tried a similar approach will smaller in-wheel e-motors but didn't market it.

Single-speed drives require the motors to have a high speed range in order to allow low-speed maneuvering as well as high-speed highway cruising.
This requires the motors to run at high speeds. In order to make a motor run at high speed you need to overcome the resistance of the magnetic field by weakening the field. This field weakening requires additional power and larger inefficient power electronics, the company says.

Or so the company says, eh. No one putting their name to this statement Huh ?

Well the first two sentences are true but the weakening of the field bit is a no,no. Clearly the author is referring to a brushless AC servomotor where you are fighting not only the unrelenting back emf of a powerful permanent magnet system but also the machine's stator/rotor reflected inductance whose impedance increases with frequency.

You can of course avoid the problem entirely by gearing the system to make base speed for the motor coincide with the maximum vehicle speed. No field weakening required there.The drawback to this is that then maximum power can only be reached at top speed. But supposing the marketing department specifies top speed to be 120mph and at the same time infers that the 0 to 60mph benchmark is also going to be an important sales feature ? Then clearly even if acceleration requires maximum current to be drawn, the motor itself will only be able to reach half its rated voltage at 60mph and therefore will be developing just half of its rated power.

Just one thing, the frame size of the motor, which governs how much torque it can produce, will need to be sized so that it can produce the required power at just half its maximum rpm. However in this article what the mechanical engineers want to do is to take a motor, with a smaller frame size, and let it run to max revs at 60mph and then change the gear ratio by half in order to reach maximum revs again but this time it will occur at the vehicle's top speed.

This is not so easy to do at high power and presents a severe challenge as Tesla engineers discovered in 2008, where even a split second counts.

So continuing on with a larger frame size motor, achieving the required power at half maximum speed obviously requires twice the torque and this torque has to come from doubling the stator current which means choosing upgraded transistors with twice the original current ratings. Fortunately the power drawn from the battery will be exactly the same in both cases during the 0 to 60mph ramp even though the larger motor will be carrying twice the stator current of the smaller. What happens subsequently to achieving the 60mph ramp up is a different matter entirely as it will depend on the current limiting applied to the larger motor, since this motor will begin to draw more battery current as its stator voltage continues to rise even though it accelerates at constant stator current. Stator current limiting will probably be applied since this motor could pull up to twice its 60mph rating if acceleration is allowed to continue. Most battery inverters will have two types of current limit BTW. One circuit protects transistors from overcurrent while another limits the battery draw.

Well, it would be nice if it were the case that the scenario thus described with the larger motor was actually used. Sadly, the reality is that cars which start off with an acceleration feeling like they were 10 sec cars (1/4mile) don't quite deliver on the promise. What's with that ? First I should say that performance vehicles aren't necessarily meant to be track cars. Nor should they be made to tow a caravan on an airstrip at 125mph until the engine blows as Top Gear has done.

So if they are not delivering on the promise then compromises are being made but where ? Not with the motor where it is straightforward enough to arrange for all motor dimensions to be increased by 26% in order to double the torque output. However, regarding semiconductor upgrades, reasonable economics would suggest limiting transistor devices to 1000 amps. To accomodate such a shortfall in available stator current will require compensation elsewhere. It will be found in changing the stator winding such that the Volts/Hz coefficient of the motor will be increased. Consequently the full stator voltage will arrive earlier, somewhat below top speed, with a commensurate reduction required in the stator current excitation, although still able to deliver the original maximum torque.

The question now becomes: how large should the Volts/Hz be allowed to be ? To answer that the limits must be explored. If the V/Hz is made sufficiently high enough the motor could develop maximum power at 60 mph with the use of relatively low current semiconductors. OTOH selecting a similar frame size motor possessing a low V/Hz that causes topping out at 120mph will require the semiconductor ratings to be doubled. So it appears, at first sight, that increasing the motor V/Hz, usually by changing the stator wiring paths, is a good thing as it allows the use of cheaper semiconductors without compromising performance to 60mph. It should be mentioned that going beyond 60 mph at max power will cause torque to decline with increasing speed. Eventually motor impedance, as speed progresses beyond 90mph, will have a greater effect and only half max power will be available at 120mph.

You might compare that performance to the motor with the low V/Hz which likely will be developing twice maximum power at 120mph.

In either case, providing aerodynamic losses do not exceed the motor power, acceleration will continue until the 120mph speed regulation kicks in.

Here's where it gets risky with this V/Hz business. The question now becomes : how far below 60mph can maximum power be reached were we to make the V/Hz even larger. The expected result would a power peak with a flattish top as 60mph is approached.

First up, let me say that I don't favor power peaks that occur in the 0-60 pass band. A power peak on the way to 60mph infers that for the next period a constant power path is experienced before entering a state of steady decline which we could call negative alpha territory. It should be needless to say that having power diminish during acceleration is something to be avoided but it is done.

This is the point at which I reflect on design philosophy. System designers are often inclined to get too absorbed in that world of amps and volts associated with alternative stator wiring schemes along with their interactions with a specific topology of electronic devices and maybe lose sight. I find it more useful to focus on the electric motor, centering on the fact that there is a peak magnitude in the magnetic flux at optimum slip (or rotor skew for a BLDC) that can be preserved even as the motor continues to spin up thereby producing a steady maximum torque. There is no power peak from the motor's point of view. The faster it spins the more power it develops. It's simply rpm multiplied by torque - the way horsepower has always been. But if on independent testing a powertrain appears to have passed through a power peak, then clearly something has intervened and compromised that ideal.

Let's imagine a motor that achieves rated power at 6600rpm but we would prefer that specific power to be attained at 4400rpm instead. We therefore require maximum voltage to be applied at the lower speed which means raising the V/Hz by 1.5 or 50%. The downside is that the new maximum current has to be 2/3 of the original to avoid saturating the stator iron since there will be 50% more turns around the poles. Since power = motor volts x motor amps then clipping the motor current to 2/3 also clips the available power by a similar amount. This is in line with my earlier statement -The faster it spins the more power it develops. So in this case maximum power can be reached at 4400rpm, but it won't compare with the maximum power it could have reached had the same stator been wired for the lower V/Hz and run to 6600rpm. Just to be quite clear on that.

Gasoline engines do experience a power peak OTOH. The limitation on them is usually to maintain piston velocity somewhere approaching 20m/s (4032ft/min) therefore it is necessary to hold them within that boundary by substituting gear ratios that allow the engine to remain on its power peak until it experiences part load at which point an even lower ratio gear (overdrive) is selected to reduce needless mechanical sliding friction. An electric motor needs no such considerations. That is not to say that it couldn't benefit from a higher ratio when accelerating at lower speeds. It's just that when you factor in the mass and cost of the multi-ratio-with-clutch style gearbox against using a larger frame size motor then the larger frame size appears to be the more elegant and seamless solution.

But so what if an overly high Volt/Hz characteristic is apt to promote negative alpha on the upper reaches of the power curve - who cares ? After all, if at any point the motor is developing more power than actual aerodynamic losses being experienced then acceleration of the vehicle is going to prevail anyway. Turns out it does matter when an EV is competing in the quarter mile alongside vehicles with turbo-charged I.C. engines.

Leaving quarter mile performance aside for the moment, this is all well and good, but Toyolla2, you may say, despite all this a uniform acceleration ramp is not particularly efficacious since on an 8 second 60 mph ramp the vehicle reaching 30mph has had invested in it only one quarter of the kinetic energy needed to get to 60mph but has taken up exactly half the time. In other words during the next half of the ramp the vehicle powertrain will need to work three times as hard. Actually I am cognizant that a good launch is paramount to achieving a favorable elapsed time. But it is what it is. Only two solutions present themselves when torque is to be king. One way is to install a larger frame size motor. The other is to employ a larger gear ratio. The latter is preferable since it comes without the penalty of having to install a heavier motor. In either case the volt/Hz of the motor should allow for the 60mph ramp maximum power to be reached close to that speed but if the more stringent quarter mile is targeted then the maximum power point should extend to at least 90 mph.

During the acceleration ramp towards maximum power the motor voltage would be seen to be rising proportionally with speed with the motor running in constant current which means the transistors are conducting with their maximum rated current. The only two variables are speed and voltage. So we can say that all other parameters are of minor importance providing the voltage/mph gradient is preserved. It turns out that this statement is profound in defining EV performance. For example, sometime in 2014 a Tesla Model S P85 was modified by a certain Mr Saleen by altering the gear ratio to improve torque at the expense of top speed. The intent was to tweak quarter mile performance. The results were less than spectacular since this changed the volts/mph ratio which in turn lowered the mph at which the powertrain traversed the point of maximum power, and it was low enough already I might add.

Since raising the system voltage is not a low cost option, the motor should have been rewound for a lower volt/Hz rating to retain the original volts/mph ratio. That option will also require compensatory increase in transistor current to maintain the original motor torque. Even if that was possible there still remains the problem of the rewinding itself. Theoretical work I have done myself tells me that that is easier said than done since you are already dealing with one conductor per stator slot in order that a relatively low voltage supply will be able to force currents into the motor at frequencies of 480Hz or higher.