BLDC do not use 3phase switching, they only switch 2 of the 3 windings at any one time, commuted via hall or equivalent sensors, they represent a DC motor turned inside out, hence BLDC.
3 phase P.M. synchronous motors are constructed identically but the 3 windings are fed with true 3 phase.
Max.

I didn't want to hijack the other thread. So I opened this new one because I find this fascinating.

Max, how exactly do they switch the coils of a BCD? You said that only two of them are switched on at one time. Are they being PWM'd. And while two of them are being switched, suddenly the third one is switched while another one is turned off? Or is there some sort of transition in which the first coil is gradually "phased out" while the third one is "phased in"?

 

There is quite a few links out there that show the timing sequence, the pairs are switched in sequence.
http://www.quantumdev.com/timing-a-bldc-motor-to-an-optical-encoder/
https://www.digikey.ca/en/articles/...-most-effective-way-to-commutate-a-bldc-motor

I have the Picdem MC LV Development board, they show the schematic and code on the Pic site.
If fitting a hall device or hall encoder you have to know the pole count, this can be done by shorting the leads and turn the shaft one rev and count the 'bumps'
This way you determine the number of electrical revolutions per mechanical.
To time or align them, I use the backfeed method by turning the motor with another at around 100rpm.
And obtain the scope reading as per the PDF. which shows the one Electrical revolution and the commutation point of any two windings.
I use a 10k resistor on each winding with all the ends shorted and this point is my D.B. 'scope GND or zero ref point.
As you can see from the 360 deg electrical rev. timing, switching occurs every 60deg.
Max.

 

There is quite a few links out there that show the timing sequence, the pairs are switched in sequence.
http://www.quantumdev.com/timing-a-bldc-motor-to-an-optical-encoder/
https://www.digikey.ca/en/articles/...-most-effective-way-to-commutate-a-bldc-motor

I have the Picdem MC LV Development board, they show the schematic and code on the Pic site.
If fitting a hall device or hall encoder you have to know the pole count, this can be done by shorting the leads and turn the shaft one rev and count the 'bumps'
This way you determine the number of electrical revolutions per mechanical.
To time or align them, I use the backfeed method by turning the motor with another at around 100rpm.
And obtain the scope reading as per the PDF. which shows the one Electrical revolution and the commutation point of any two windings.
I use a 10k resistor on each winding with all the ends shorted and this point is my D.B. 'scope GND or zero ref point.
As you can see from the 360 deg electrical rev. timing, switching occurs every 60deg.
Max.

Very interesting... every 120° only a single phase is active.... I'm interpreting the up and down slopes on the lower part of the graphic as current being raised or drained due to the inductive nature of the load, and not to actual controlled slopes... am I right?

 

Correct, the actual switching is via a square wave. controlled via the PWM module with Picmicro.
Max.

 

Correct, the actual switching is via a square wave. controlled via the PWM module with Picmicro.
Max.

And that's where I get a little lost... I'm assuming the square waves are pwm'd to control the amount of current reaching them. And hence, the power and torque. But the frequency of the square waves determines the rotational speed. Correct?

 

The way I have always understood it, is that at the point the different stator coils are switched via the sensor on for that particular coil, it is PWM'd for the duration of the commutation time dictated by the hall sensor.
The resulting wave form is seen as trapezoidal due to the BEMF which is sinusoidal in nature, if an increase in RPM is required then the pulse width of the PWM is increased, and like a DC motor the higher energy results in increase in rpm.
The commutation is every 60° for a duration of 120° in every electrical cycle.
Max.

 

The way I have always understood it, is that at the point the different stator coils are switched via the sensor on for that particular coil, it is PWM'd for the duration of the commutation time dictated by the hall sensor.
The resulting wave form is seen as trapezoidal due to the BEMF which is sinusoidal in nature, if an increase in RPM is required then the pulse width of the PWM is increased, and like a DC motor the higher energy results in increase in rpm.
The commutation is every 60° for a duration of 120° in every electrical cycle.
Max.

Right ... the pulses have to be delivered in sync with the motor's rotation ... if not, heating occurs because the out of phase fields cannot convert the electrical energy being input into mechanical motion. Similar (in an opposite way) to slipping when a 3ph motor is overloaded ... this is very enlightening, thank you very much...

 

Also the same way a DC comutated motor is always commutated at the same point every time with relation to the field, and the voltage is then varied to change the RPM.
Max.