John in CR said:
LFP,
One thing is still bugging me in this thread, and that is the idea that I can run the same voltage with a speed wind and a slow wind motor and get the same efficiency at low speeds at the same rpm and load, so same power at the wheel.
I appreciate that bucking to the lower apparent voltage is very efficient, but it seems to me that the speed wind motor would run at a lower duty cycle, and that would cause phase currents to be higher for the speed wind motor than they would be if it ran the proportionately lower voltage required for identical performance of the low speed motor.
Let's look again at the simple case, the motor with double the Kv of the other. With the speed wind using a battery pack of half the voltage of the slow motor, and the speed wind set to double the current limit, it makes sense that they would both reach full duty at the same rpm pushing the same load. If we take the speed wind and give it the same pack voltage as the slow wind, wouldn't it be at 50% duty cycle at the same rpm and load? Doesn't that double the current compared to full duty at the lower voltage....though it would be only on for half the time? Since heat increases by the square of current, doesn't that create more heat in the windings?
I can measure battery side, but once the controller does it's thing to feed the motor, I'm unclear what exactly happens on the phase side at partial duty. I'm good at the practical side though, and when I went up in voltage on my system and before I completed my cooling mods, I definitely noted a hotter motor even riding at low speeds. After I noticed it, I even tried riding very easy to try to make sure the heat increase wasn't from using the greater performance on tap, and still got a hotter motor.
John
As the controller bucks, the current doesn't stop flowing when it's not switched on, it just flows from the coil's field collapsing back into the controller's mosfet body diode. The inductance of the motor does for current what a capacitor does for voltage, this keeps it essentially smooth and regular, even while the FET is pwm'ing at 20kHz or whatever, and the voltage waveform looks like a mess of ugly square waves (even on a sinus controller).
This makes the penalty of starting from higher voltage and bucking down come from increased time spent conducting through the MOSFET's body diode while the FET is in the 'off' state of the PWM'ing cycle.
Let's look at what sort of bucking already occurs with a controller starting from a stop. Let's say our motor has say 10mOhm phase resistance. Let's say the controller phase current is set to ~200A. This means taking off from a stop, whatever pack voltage you started from, it's going to buck it down to an average of ~2V. If you started from a 100V battery, and bucked 100V to 2V, it means the controller was running something like a ~2% dutycycle to create the average 2V voltage required to keep 200A of current in the winding while starting from a stop. If you ran a 50V battery, it would be something like ~4% duty cycle. If you range a 25V battery it would be something like 8% duty cycle.
Once you start rolling at the wheel gets BEMF, the duty cycles for current control all grow rapidly as the generated BEMF of the motor climbs with RPM.
Because the 'off' state of current flowing through the body diode has higher loss (because it's conducting through the diode which has a ~0.8-1.5Vf), it enables slightly reduced controller losses to drive from a lower pack voltage. In practice this is going to be unlikely to ever make even a <1% system efficiency difference (from slightly reduced controller losses), and no difference in the motor performance itself.
It would make no sense to run more pack voltage than you need to spin your motor through the RPM range you need if you were planning a new ground-up system build. That would just be extra battery management hassles, cost and complexity and reduced safety for no benefit and a likely unnoticeable tiny efficiency hit.
If you happen to already have a bike or vehicle battery at a higher voltage already built and BMS'd in the vehicle, I don't think it would be worth changing it just for an extremely tiny efficiency difference in the controller. If you're building a new system, then of course it would be silly to go any higher pack voltage than needed.
The reason is that the speed wind motor's torque curve is broader and flatter. That means greater and more constant acceleration. This enables you to dial back the current into the speed wind motor for a smoother launch and less heat in the first rotation or two of the wheel, and still have greater torque and acceleration (of course more heat too, since they make the same heat for the same torque) all the way up to a greater top speed. Even if you like to stay slow and don't use the extra speed, it's good to have the overhead, so you get to your cruising speed quicker. I can't imagine suffering with an ebike ridden at WOT with every puff of wind or slightest grade slowing me below my desired speed, not to mention all that extra distance needed to get to the steady state top speed. 
