EVLearner wrote: I am a bit puzzled why we need to have a higher voltage for AC motors.
It has to do with the fact that industrial motors can generally stand around 415v ac (or a little more). You can get them at nominally lower voltages. For example, ABB motors have can be ordered with an "S" in position 13 of their code number instead of "D", for 220-240v delta / 380-415 star.
Edit: I meant to suggest that the "D" stands for delta 400v operation, and the "S" for star operation at around 400v.
(There are other code letters, but they are all higher voltages than 400v delta, except for "X" which is custom, and may be much more expensive). It turns out that because a motor wound for 240v delta can take 415v, then a nominally 15kW 50Hz motor is really also a 26kW 87Hz motor, and it will take that 26kW all day every day, just as it will take 15kW at 50Hz. The reason is that the current is essentially the same, and the heating is almost only proportional to the current. So even if you have a low voltage motor, assuming you have the inverter drive current, you still want a high voltage.
You could of course just use a 26kW 50Hz motor and a lower voltage, higher current inverter, but then the motor weighs a lot more, and industrial VFDs are mostly expecting 380-440vac input. To get the extra current, you will need a much bigger VFD, and acmotor's is already almost the width of his truck!
So for industrial drives, you get best value out of a high voltage pack. The AC24LS series, which expect around 320-350v packs, are actually low voltage (about half) of what conversions using industrial drives will use (600-720 vdc).
Also, the industrial controllers often have a quite high bus voltage at which their braking resistor transistor attempts to keep the bus voltage at; it's around 750-780 vDC. You want that to be around the charging voltage for your cells, say 3.6-4v for Lithium, or 14.4-15v for lead acid. That will dictate a high voltage pack. You could ignore the braking resistor option on the VFD (dangerous), or provide your own.
From my side I have been seriously considering 4 banks of 48 V = 200 V nominally, and that is about 70 V rms neutral to phase - or about 120 V rms between phases.
I've been waiting for an excuse to show off my neutral wobble image:
This is supposed to show two phases of the VFD output (red and green), and the difference between these is the blue sine wave. Note how the blue wave has a higher amplitude (587v peak, 415v RMS) than the flattened sinewave outputs from the VFD (293v peak, sqrt(2) less than 415v).
Edit: was "Note how the blue wave has a higher amplitude (576v peak, 415v RMS) than the flattened sinewave outputs from the VFD (340v peak, same as 240v RMS sine wave)."
All the drive manufacturers do this; otherwise, you can't get a 415v sine wave from a 415v rectified input! So in fact, for 200v DC input, you can (neglecting losses and real-world voltage drops across IGBTs etc), you can get 200 / sqrt(2) ~= 141v AC RMS output (phase to phase, as 3 phase is usually specified). If you don't do the neutral wobble thing, you'd only get 200v / 2 / sqrt(2) * sqrt(3) ~= 122v RMS phase to phase. (Edit: was 115v via spurious math). For true disbelievers, I can provide the C program that generated the data for this graph.
[Using lower voltage motors] Who knows, this could be a major development in AC motor manufacturing in the very near future (a few months), if and when electric cars get really popular.
Yes, indeed, I expect that using a 240v delta motor could well become standard for conversions. But I don't expect lower pack voltages.
Sure, when it comes to lower voltages it also means higher currents, but again I don't see that as a problem as the currents (I believe) are to a large degree circulatory
Circulatory? You mean reactive current? Generally, the reactive current is less than the real current, by a factor of at least 2:1, often 3:1 (please correct me if I'm wrong).
and even then we are only talking about the same amount of current as in a DC motor (a few hundred amperes being high frequency switched)
Yes, but the DC inverter has just 1 switch, and a big diode. The AC controller has 6 switches. That's one reason that the AC controllers like to keep the current down, and the voltage up: less devices paralleled. I read today that a common Curtis DC controller has 35 MOSFETS paralleled.
- but the braking effect can be used to recharge the batteries, and that is a real definite plus compared to DC motors which really struggle to be efficient series wound generators!
Yes, a series wound DC motor doesn't like to regenerate, while an AC induction motor does it naturally. Separately excited DC motors can regenerate (2 switches are then needed, so the controller is more complex). AC inverters almost can't help providing regeneration as a feature; the bridges are inherently bidirectional.
However, I would not say that regeneration by itself is enough to justify AC on its own. Depending on the application, you might get 15% extra range from regen (acmotor may dispute this figure), but the built for EV controllers and motors are perhaps twice the price of the DC controllers and motors. The main advantage of AC (in my opinion) is the torque at high speed, sometimes obviating the need for the gearbox and clutch. Also, when you have a low voltage motor, there is a cheap sort of gearbox available in the form of the star/delta switch. It's like a fixed 1.73:1 (sqrt(3):1) gearbox, available by using a big contactor and some smarts with the inverter (you have to tell it that the 50Hz motor has gone away, and now there is this 87Hz motor with different characteristics, and would you please do a flying start in a fraction of a second, thank you).
You may have figured out that I'm in the AC camp