Commutation speed KT controllers

[d8veh]

A friend of mine builds cargo trikes. He uses a special motor that's attached/integrated to the differential. It has an output speed of 200 rpm and 10:1 reduction ratio, so 2000 rpm motor speed.He doesn't know how many poles or magnets. He tried the 12 FET KT Sinewave controller (S12S from BMSB), but can't get it to run properly at any P1 setting. He also tried the L setting that multiplies P1 by two; however, it runs fairly well with the 12 FET Square wave version (S12P), when run sensorless, though the starting from standstill is a bit rough. I heard from Xionda that KT sine wave controllers don't work at high commutation speeds, which is why you have to buy their special controller (Lishui) with their tiny motor.

The question is whether there is a nice controller that can run with the hall sensors at a commutation speed that high. It needs to run continuously at 20 amps, so a max current of at least 30 amps. I think 15 or 18 FETs would be better because he has to use the software in the 30 amp S12P to limit the current to 20A to stop it from overheating.

 

[amberwolf]

A friend of mine builds cargo trikes. He uses a special motor that's attached/integrated to the differential. It has an output speed of 200 rpm and 10:1 reduction ratio, so 2000 rpm motor speed.He doesn't know how many poles or magnets.

That's easy to test for, by rotating the motor once and counting the pulses out of either a phase or hall signal wire. Rotate the wheel once and you get 10 motor rotations (assuming live axle), so just divide the number of pulses by ten to get the number of pole pairs.

I heard from Xionda that KT sine wave controllers don't work at high commutation speeds, which is why you have to buy their special controller (Lishui) with their tiny motor.

There may be some info on the ERPM capability of the KT controllers in the Open Source Firmware thread(s) by Casainho / Stancecoke et.al.

<snip>
He tried the 12 FET KT Sinewave controller (S12S from BMSB), but can't get it to run properly at any P1 setting. He also tried the L setting that multiplies P1 by two; however, it runs fairly well with the 12 FET Square wave version (S12P), when run sensorless, though the starting from standstill is a bit rough.
<snip>
The question is whether there is a nice controller that can run with the hall sensors at a commutation speed that high. It needs to run continuously at 20 amps, so a max current of at least 30 amps. I think 15 or 18 FETs would be better because he has to use the software in the 30 amp S12P to limit the current to 20A to stop it from overheating.

If the controller is overheating, but the motor is not, it sounds like there's a problem with the way the controller is driving the motor; or some kind of electrical feedback from the phases causing problems with FET heating. It could be the ERPM thing causing timing problems, if the motor is too high an ERPM for the controller.

If the controlelr does not get hot in sensorless mode, and it runs properly in this mode (but does get hot in sensored, and does not run properly), then it probably means a false positive on the hall/phase combo is causing both the heating and the problems running the motor. (rather than ERPM)

I'd investigate the hall/phase combo first, before the ERPM.

 

[d8veh]

Thanks for that. The overheating is because these trikes run with max power a lot of the time. The S12S is only rated at 15 amps. That's why he sets the max current to 20A.

The problem with the overheating is just down to the demand for current. The problem with the S12S is not overheating, it just doesn't run smoothly at normal running speed. I'm sure that it's a commutation problem. I've had the same problem with other high-speed motors.

I'm hoping that those guys might see this thread and comment, otherwise I'll send them a PM.