john61ct said:
Now, I thought that Hall sensors were a compromise, that the phase sensing with True FOC was done via shunts?
You are mixing up things.
The hall sensors in the motor are detecting the rotor position (but only 6 states per revolution) This has nothing to do with sinusoidal or FOC commutation.
Sinusoidal commutation means, that the phase currents are shaped sinusoidal. You can put a sinusoidal voltage on the coils and you will get a sinusoidal current. (even if the motor isn't spinning

) But you have to syncronize the voltage with the rotor position to make the motor run.
With
FOC you can control the voltage (and current as a result) in a way, that you get most torque and minimum heat from the motor. (or do field weakaning e.g....)
You need at least two phase current sensors (as the sum of the three currents is always zero, you don't need to measure the third phase current)
For sensorless FOC you need an "observer" that predicts the rotor position.
The STM8 processor in the Kunteng SVPR is too slow to do all the calculations for FOC, let alone the additional rotor position prediction.
The Kunteng SVPR uses only one phase current sensor (based on the hall effect

, not a simple shunt+opamp ) and we are doing what we called "simplyfied FOC". We syncronize the rotorposition with the zero-crossing of the one phase current signal. The rotor position is extrapolated from the motor-hall signals.
There are several open source examples how sensorless FOC can work on cheap e-bike controllers e.g. the
vesc or my
lishui project.
regards
stancecoke