Hello,
Its been a while since my last update, but I am pleased to say that the controller is now built and running.
With regard to performance I am still a little unsure as something does not seem quite right. I have been testing the controller by driving a 25kw 100v motor on 24 volts. I do not have a very good scope, but I am finding that the battery supply is rock steady while the motor is running and the switching of the transistors looks good when running at 100khz frequency. I am applying pwm to the low side only at the moment with the high side on.
What I am finding though is that the motor is using a lot of no load current, tripping a 10A fuse at about 65% throttle. These currents can be confirmed when clipping the battery cable. When clipping a motor phase, when the motor is turning slowly at around 1RPM my clamp meter is reading about 3A on AC current. Soon as I start to accelerate the motor this climbs quickly reaching 50amps by the time we are seeing 10A on the battery. I also noticed that the ring connects between the motor wires and the speed controller were getting hot! The controller also started to warm up quite quickly, as did the motor.
I have a current clamp plug in module for a multimeter, so I put my scope on this and I see a large spike in current each time the phase sequence changes. However these spikes are not evenly spaced. This got me thinking, what about the motor sensors. So I scoped these, and ideally a 3 channel scope would be the tool of choice, but with only a 2 channel scope I could still see quite clearly that the sensor pulses are not evenly spaced.
I then added a delay into the software between each sensor event, to delay the switching sequence, giving a period with all phases switched off. From this I found that when a phase is switched high for two changes of sensor event, the gap that appears between them which should split the switched on time in the middle, was only half the size at one side compared to the other, maybe even a 1/3rd of the size.
So it would appear to me that the motor sensors are the issue, however I would be grateful for any feedback. Unfortunately I only have this motor so I cant try a different one.
My plan is to make a flywheel for the motor for testing with slots in it for doing the feedback, but I am not yet sure how to make it for an out runner motor which seems to have 24 slots and 28 magnets after a quick count up.
Some pictures, I do have a video but can't upload it on here;
View attachment 2