As long as it's not the first thing you do, there is nothing wrong with using the epa to manage the total power of the system. With brushless motors the fets already have to switch on and off several times every rotation, so they are never at peak efficiency like a brushed esc would be when holding full throttle.

The kv constant of a motor is a physical parameter that is established by the flux linkage between the rotor and stator, and the number of turns per phase. This is not something that can be altered by the ESC.

The ESC delivers a voltage waveform to the motor winding and controls the phasing ("timing") relative to the rotor position. Those two parameters are effectively the only thing that the ESC can adjust to affect the performance of a permanent-magnet brushless motor. And since there is one "best" timing for a given speed and load, this really leaves the voltage as the only adjustment. For any delivered winding voltage that is less than the battery voltage, the ESC is running at a duty cycle (PWM rate) less than 100%.

Running at a reduced ESC duty cycle isn't the worst thing in the world, but it does cause more heat to be generated in the ESC, motor, and pack. Because of this, it is best to select a motor kv, battery voltage, and pinion size that maximizes the amount of time spent at full output.

The switching behavior at 100% output is determined by the rotational frequency of the motor. For example, with a 4-pole motor at 30,000 RPM, the rotational frequency is 1 kHz. Compare this to the typical PWM carrier frequency of 8 kHz that the FETs are switching at during anything less than full throttle.

I have no idea how Castle implement their "Torque Limit" feature mate

I cant imagine it is the same as simply winding back the EPA

Agree that gearing and motor selection is the best option

Cheers

8 khz only at half throttle. Its not a step function by any means and since the switching frequency it never is close to 0 the resulting difference in impedance from half throttle to full throttle doesn't waver much. Current plays more in creating heat at that point.

The carrier (switching) frequency is maintained up until the point where the "modulation index" (another term for the commanded duty cycle) goes to 100%. This happens regardless of the motor rotational frequency. Put an oscilloscope on one and take a look sometime...

I^2R losses (caused by current flowing across the effective drain-source resistance) are indeed the main source of heat at high motor loads. But if you're just putzing around at part throttle, the switching losses (caused by charging and discharging the gate capacitance) can be quite considerable.

The command duty cycle is maintained, but that doesn't mean the fets are cycling at that rate.

The switching losses are only considerable relative to the current losses. They are still very small.

??? Explain this statement further, please. As written, it does not make sense.

Have you ever calculated them for a BLDC controller? They are not trivial.

Lets assume a 10khz duty cycle and linear on/off cycle to throttle % command.

At 99.99% throttle every cycle except 1 would be "on"
At 99% throttle every 100th cycle would be off.
etc.
at 50% throttle every other cycle would be off.

The change from on to off is the main loss, so the fewer times the fets actually go from on to off the lower the losses are regardless of the duty cycle.

Yes, I've calculated and measured all sorts of losses. I purposefully italicized "relatively" because while the switching losses are a good chunk, they are not the root cause of heat in an esc by any means.

My head is spinning trying to understand electric terms lol.

In a MOSFET, there is the same amount of power required to turn from "off" to "on" as there is the other way around, since you're either charging or discharging the gate capacitance. IGBTs are indeed asymmetrical, but not used in hobby controllers.

Regardless, your statement simply confirms the point of this thread - run the controller as close as possible to 100% in order to maximize controller efficiency.

If you have indeed done the calcs, then you should also realize that there are significant losses in the FETs during the "off" portion of their switching cycle, as the intrinsic diode forward-conducts as it provides a path for the winding current to "freewheel". Yet another reason to minimize running the ESC below 100%!

Don't mind the technical talk - all you need to know is that you should properly select the motor kv, pack voltage, and gearing in order to achieve the desired performance. Don't just simply turn down the travel volume, especially if the ESC is marginal for the application.

Its not just about the power used to turn the gates on and off. I'm talking about the impedance from stopping and starting the current going to the motor. Also, I never said you shouldn't avoid running your system as efficiently as possible. My very fist post explicitly said "as long as it's not the first thing you do" when referencing cutting back on the epa. Of course you should pick the most power you need for your track rather than limit an overly powerful motor. My point has been that the losses of backing off the epa is nominal in regard to the overall efficiency of a system that typically only sees full throttle for a few seconds a lap.

What are you talking about the winding current "freewheeling"?

Yep, and my apologies go out to you (and any others reading the thread) for beating this issue into the ground. I'm going to try to walk away after this post - we'll see how successful I am

The motor winding has inductance. The current in an inductor cannot be changed instantaneously, and as such it will attempt to find another path (if it can't find another path, the voltage will rise in the circuit until a path is created; this is the "flyback" principle that is used in things like automotive ignition systems).

So, the FETs in the ESC turn on and current starts to flow through the winding - that part is pretty intuitive. Then the ESC commands the FETs (at least the low-side ones) to the off state in order to limit the current in the winding to some desired level. At this point, the motor winding current can no longer follow the same path (which is the whole point of turning off the FETs), but it cannot instantly stop flowing. It ends up flowing through the intrinsic diode that is a by-product of the FET design (look at the FET symbol in a datasheet or schematic and you'll see a diode connected across the drain and source; this is the one I'm talking about). Since it takes several tenths of a voltage to get this diode to conduct, the VI losses can be quite significant.

This thread CA NOT DIE! You guys are hilarious! ....eherrmmm...but the conversation seems far too simplistic and linear...pray tell...how did you factor frequency variance into your calculations...?
JK pleeeaaaaase don't answer that!