Weird Hall-effect position sensor readings

[Fallingwater]

I bought an e-scooter second hand. It has a sensored motor, but the previous owner modified it with a 25A sensorless e-bike controller. Said controller is huge and overpowered, so I want to replace it with a more reasonable 18A scooter controller I have kicking around. Unfortunately this other controller can't run sensorless, and when I connect the sensors it gives me an E-01 motor error.

I went hunting for problems and figured out someone'd been inside the motor and screwed up the soldering. My guess is the previous owner already had issues with the hall sensors, so he opened the motor but didn't know how to solder and messed them up further, then gave up and used a sensorless controller instead.

I resoldered them properly and stuck that sensor that had come out of its notch back in it, but kept having E-01 errors - though this time more intermittently.

On probing the sensors, all three do change state as I spin the wheel manually, so none of the sensors are stuck or shorted, but - and here is the weird part - sometimes all three read triggered, and sometimes all three read not triggered. It is my understanding that this isn't normal sensor readings; I've watched several videos of people using testers, and it seems at least one sensor is always triggered, or not triggered. This makes sense, as otherwise the controller can't know where the rotor is.

Indeed if I power up the controller in one of these all-on or all-off states, I get E-01 immediately. If I spin the wheel so they're in a state where at least one is on, then I don't get E-01 immediately - but I do get it as soon as I spin the wheel manually, or try to apply power.

My question is: is this a sane sensor reading that just works differently than usual and my controller is just not programmed for it, or is something wrong with the sensors - and if something is, what could it be, considering all three sensors do seem to work?

 

[TommyCat]

Hi Fallingwater,

From your description you are describing the differences between a 60-degree and 120-degree hall sensor application.
As most controllers are 120 degree you will probably have to make a change.
Luckily, from what I understand. This may be as simple as reversing the direction that the middle hall's sensor face is pointing. (I.E. turn the sensor 180 degrees in the slot) This reverses its output and hopefully changes your 60-degree output to 120.

Here is a thread with a little description on it.

GENERAL - HALL SENSORS: 60 VS 120DEGREES



Regards,
T.C.

 

[Fallingwater]

You know what it was?

That sensor I'd found outside of its notch and stuck back in had come out again.

Yeah, that was that. I took the motor out again intending to test your proposed change and found it happily floating in the breeze. Stuck it back in with some temporary glue and the display wasn't giving E-01 errors anymore, but the motor was still twitching and not spinning. I tested the sensors and found they weren't going all-off or all-on anymore, but they were "illuminating" in a weird order. So I moved the wires around testing different combinations until at last, after days of cursing at this stupid thing, I hit the right combination and it spun right up.

I've reopened the motor, removed the temporary glue and epoxied the sensors in place with two-part resin. I'm waiting for it to dry, then I'm finally going to reassemble the whole thing and hopefully by tonight I can test it on the road.

 

[TommyCat]

Hey that's great news. Always nice to have the answer stick out for you. Job well done on the fixes and configuration!
for a good test result.