MCSDK Motor Profiler always fails with speed feedback error

[Quentin Ch]

Hello,

I'm using the Motor Profiler tools v5.Y.3 to try to profile a hooverboard BLDC motor (15 poles pairs, 24V, low speed typically 300rpm), with a Nucleo F401RE + IHM08M1 setup.

After having found a configuration for the hardware that made the motor spin:

• Capacitors C3, C5, C7 removed on IHM08M1, for FOC mode according to UM1996
• Solder bridges J5 and J6 on 3sh side according to UM1996
• Resistor R181 removed for Motor Profiler according to Motor Profiler's hint 'Remember to configure correctly the board..."
• Jumpers J9 open, JP1 and JP2 closed for FOC according to the hint

When I start the profiling, the motor starts to spin few seconds, then stop, then restart, etc... until it systematically triggers a "Speed feedback fault" when the speed reaches about 1/3 of max speed, and restarts from scratch the electrical profiling (and fails again, and so on). So the profiler never completes the electrical identification.

At higher speeds and torques setting, the motors sometimes starts to oscillate (vibrating without turning) but the speed read is like 200 rpm...

Looks like there is really a speed estimation problem.

This happens with and without Hall encoders and jumper J3 connected.

This happens for all the ranges of max speed and current I tested (3A to 5A, 100 to 1500rpm)

The motor is on a test bench, power by 24V batteries, with only a wheel as load.

Why is the speed feedback fault always happening ?

Could it be because the motor is not loaded, as mentioned slide 24 of this doc ?

Is the speed too low for the profiler ? (as asked here)

I attached below a screenshot of the configuration I used.

Thanks for your help,

 

[MK.1]

Motor Profiler works without Hall sensors. Try disabling hall sensors.

[Quentin Ch]

Hello,

I got the same issue even if the Hall sensors are not connected to the board...

[Quentin Ch]

Hum, I look at the Hall signals (not connected to the board, only powered) while running Motor Profiler, this is what I got:

1. The Hall signal does not look like squares at all.
2. In a previous setup I drove the motor with a cheap 6-step controller, and I saw nice squares...
3. ... but I did not paid attention at the time to the fact that I need to power and drive the motor to get signal from the Hall sensors
4. The Hall signals may look like some modulated FOC command ?

So I'm starting to wonder if my motor is actually not a Permanent Magnet Synchronous Motor at all...

Being quite new to the BLDC field and my motor being undocumented, I may have missed that.

So, is this motor a PMSM motor ?

If not, could it be the reason of failure of the Motor Profiler ?

> Hall outputs when driven by Motor Profiler (FOC-driven ?):

> Zoom on the signal:

 

[MK.1]

Hello. You are measuring the signal from the sensors incorrectly. The outputs from the sensors must be pulled to the power supply through pull-up resistors or take measurements from sensors that are connected to the board.

[Quentin Ch]

Ok, you were righ about the Hall, I got some misunderstanding about their output type. Once connected with pullups (the ones of the IHM08M1 through JP3) the signal are ok.

[Quentin Ch]

As said, the Motor Profiler does not use the Hall sensors.

I used another BLDC motors with a much higher nominal speed (3000 rpm vs ~200rpm), and got not issue.

I suspect the back-emf to be too low (because of the low speed) and noisy for the Motor Profiler and the observer to estimate proper parameters and speed.

[Prajwal]

Hello Quentin Ch,
Could you please explain why R181 should be removed?
We are also getting the same issue and the motor's nominal rpm is 1200. Is it too low for the profiler?

Thanks