Bosch performance line adding halls

My friend and I have been working on modifying the bosch gen 2 motors to accept a VESC as the controller to be able to run third party batteries and remove the speed limit.
For their size, the bosch motors are pretty stout and will run all day @30 phase amps, 20 batt amps and 36v, making them a nice, reliable alternative to the tsdz2.

For better startup, he tried to add hall sensors in the motor. We used the ss413A sensor and mounted them like so.



But the signal on the hall sensor looks weird to me and the motor runs pretty badly from it

https://youtu.be/I7BdWSo1QeA

It is not a problem with the VESC because it works fine for startup and will just switch over to sensorless @ low ERPM, but we considered switching to the KT controllers with open source firmware for more ebike features such as speed limit and easier management of assist levels.

Does anyone have any idea why we got those results?

- Victor

Up

And new (maybe more interesting) title.

The waveform from the hall sensor with the motor spinning should be 50% duty cycle as the rotor magnets toggle it on and off.
Is the vesc able to detect the hall table?

Yes it Will detect IT and run the motor sensored, but it runs pretty crappy. The transition to sensorless is noticeable and then it has a Lot more power.

But Does anyone know why the duty cycle is so far from 50% ?

I'd say there's a time delay as your not sampling the phases next to eachother, any halls I see are never across a motor like that.

If you are referring to the timing between the halls, I think we are talking about two different things here.

The waveform is only from one hall sensor.

Vbruun said:

If you are referring to the timing between the halls, I think we are talking about two different things here.

The waveform is only from one hall sensor.

Click to expand...

The halls are spread out over the motor im no expert but all the motors I open have them next to eachother and a spare set opposite if I mix them the motor runs like turd that's the issue you seem to have.

The spreading out of the halls should make for the same timing (120 degrees electric) as is common in most other motors.

What I am asking about is the duty cycle of 1 hall sensor. When we measure it, it is not 50%, which is to be expected.

The other two and the timing between them is not the topic of this discussion yet

That’s odd.

The 120 degree placement is ok, it’s not the issue.
I think problem could be using the wrong sensors and/or location so there’s a large hysteresis.

Ss411 honeywell have worked well for many here including me. 413 is less sensitive, higher hysteresis.

Try the location also, it seems too far out, ideally it’s placed in the slots, at airgap height. flux amount needs to be correct at the location you use. You can try manually rotating the rotor to see switching times are roughly correct before actually driving motor.

I guess the conclusion is that there is no simple answer as to why this happens.

There is room for a rotary encoder inside the casing, so if I someday do want to run OSEC in stead of a VESC (primarily to gain speed limiting from a mid-drive), I would have to either measure up the Rotary encoder that came from the factory or buy one that puts out hall sensor signals.

The motor does actually run very well sensorless with the VESC, though.

larsb said:

That’s odd.

The 120 degree placement is ok, it’s not the issue.
I think you are using the wrong sensors and/or location so there’s a large hysteresis.

Ss411 honeywell have worked well for many here including me. 413 is less sensitive, higher hysteresis.

Try the location also, flux amount needs to be correct at the location you use. You can try manually rotating the rotor to see switching times are roughly correct before actually driving motor.

Click to expand...

I think you are onto some of the right things here. Probably the flux is wrong ejere the are placed.
The motor is however extremely difficult to disassemble so I don't expect that er Will do more rnd in a while

larsb said:

That’s odd.

The 120 degree placement is ok, it’s not the issue.
I think you are using the wrong sensors and/or location so there’s a large hysteresis.

Ss411 honeywell have worked well for many here including me. 413 is less sensitive, higher hysteresis.

Try the location also, flux amount needs to be correct at the location you use. You can try manually rotating the rotor to see switching times are roughly correct before actually driving motor.

Click to expand...

I think you are onto some of the right things here. Probably the flux is wrong ejere the are placed.
The motor is however extremely difficult to reassemble so I don't expect that we Will do more rnd inside the motor at the moment.

Funny enough I was gonna say there's no tooth cut in the lams they look floating just catching a weak field from its position but like I said I'm no expert but I'm damn good emulator so I would have followed common trend.

Doesn´t HFI work on this motor? In this case you wouldn´t need Hall sensors. Alternative would also be waiting for A.S.S feature which will be released this or early next year from Vedder. source: https://forum.esk8.news/t/vesc-hd-a-credit-card-sized-twin-motor-controller/43178