The difference of Back EMF waveform shape of BLDC and brushless PM generator

[vunteng]

I am a bit confuse on the back EMF waveform shape of the brushless permanent magnet machine.

As I know brushless DC permanent magnet motor has a trapezoidal shape back EMF waveform. If I use this motor as a generator, do I still get the same waveform shape or I will get a sinusoidal waveform?

 

[DragonPetter]

I am a bit confuse on the back EMF waveform shape of the brushless permanent magnet machine.

As I know brushless DC permanent magnet motor has a trapezoidal shape back EMF waveform. If I use this motor as a generator, do I still get the same waveform shape or I will get a sinusoidal waveform?

I've had this confusion too. The back EMF should be based on the winding distribution along the stator, from what I've read. See the attachment for what I'm talking about.

The reason I am confused is because I was using a BLDC motor before, and to determine the order of the hall magnets, I hand spun the rotor and looked at the back EMF on an oscilloscope, and I saw sinusoidal waveforms. However, when I drove the motor with a 6-step commutation, it was a trapezoidal waveform. This was a delta configuration though, so I might have missed something.

 

[Yuri B.]

It is a mystery to me that it is possible to see the "back EMF" by someone.

 

[jim hardy]

However, when I drove the motor with a 6-step commutation, it was a trapezoidal waveform.

when it's driven won't terminal volts will be sum of counter-emf plus IR drop of windings ?

What did current wave look like?

 

[DragonPetter]

It is a mystery to me that it is possible to see the "back EMF" by someone.

when it's driven won't terminal volts will be sum of counter-emf plus IR drop of windings ?

What did current wave look like?

These are the 3 phase's terminals as I drive it with the BLDC driver I made, which was standard 6-step commutation. The spikes are my PWM voltage control. If I disconnect the 3 phases from my driver, and spin the rotor with my hand, those 3 waveforms looked sinusoidal. I didn't measure the current waveforms unfortunately.

 

[jim hardy]

looks like each phase enjoys an "off" period between high and low time.

Are your traces each phase to common?
Would that mean the phases themselves(you said they're delta) get the difference between your voltage traces?, ie ab feels top minus middle, bc feels middle minus bottom, etc ?
Can you acquisition system display that?

ESIT - NOw i don't know what's going on either, just thinking along with you,

But remember a square wave voltage applied to a pure inductor results in triangle wave current
di/dt = constant
maybe that has something to do with it.

 

[DragonPetter]

looks like each phase enjoys an "off" period between high and low time.

Are your traces each phase to common?
Would that mean the phases themselves(you said they're delta) get the difference between your voltage traces?, ie ab feels top minus middle, bc feels middle minus bottom, etc ?
Can you acquisition system display that?

ESIT - NOw i don't know what's going on either, just thinking along with you,

But remember a square wave voltage applied to a pure inductor results in triangle wave current
di/dt = constant
maybe that has something to do with it.

This was last summer, so I don't remember all the details, but I am pretty sure that all traces are with respect to common ground because the oscilloscope I used internally connects all of the probes references together.

The last part you said makes a lot of sense, and I remember thinking that the back EMF must be dependent on the signal the motor is driven with, but in the literature I was reading, they said that the trapezoidal back EMF is determined by the winding distribution. If that is the case, then that must be an independent factor that determines the EMF, even though when I measured it like I described, I saw sinusoidal phases. That contradiction is what confuses me and probably the OP too.

 

[Yuri B.]

I am a bit confuse on the back EMF waveform shape of the brushless permanent magnet machine.

As I know brushless DC permanent magnet motor has a trapezoidal shape back EMF waveform. If I use this motor as a generator, do I still get the same waveform shape or I will get a sinusoidal waveform?

I think so. Sinusoidal form of the voltage induced in the windings of the most power generators is the most natural form to occur there. Voltage induced, when one rotate permanent magnet motor, has the same form because, in principle, this motor is built like a power generator.

DC permanent magnet motors do not "have trapezoidal shape EMF", just contrary - they are powered by such signals. To obtain (so to say "to shape") such a signal, one needs LRC (and of course semi-conductors ) elements in a not simple circuit. The special drives, which have been mentioned here, do produce such form voltages - and it is not possible to have one by simply rotating the rotor of a BLDC motor.

 

[vunteng]

I think so. Sinusoidal form of the voltage induced in the windings of the most power generators is the most natural form to occur there. Voltage induced, when one rotate permanent magnet motor, has the same form because, in principle, this motor is built like a power generator.

DC permanent magnet motors do not "have trapezoidal shape EMF", just contrary - they are powered by such signals. To obtain (so to say "to shape") such a signal, one needs LRC (and of course semi-conductors ) elements in a not simple circuit. The special drives, which have been mentioned here, do produce such form voltages - and it is not possible to have one by simply rotating the rotor of a BLDC motor.

So what if I run a back to back testing? Which the BLDC motor is act as a turbine to rotate the BLDC generator's shaft? Whether I will get the trapezoidal back EMF waveform?

 

[Yuri B.]

Being turned by an external force the BLDC will produce a sinusoidal wave form emf that will not be very suitable to power another BLDC. BLDC's get powered by simple alteration of the stator poles being something like positive-negative-neutral-etc. However (a viable) drive to produce that power is not of course very simple.

 

[Mike_In_Plano]

I've been designing brushless motor drives off and on since the nineties.

The back-emf refers to the voltage waveform seen in respect to the "neutral" as the motor rotates at a constant speed. Since the motors don't typically have a neutral coming out, you can synthesize one using three resistors, one to each phase, joining together in a wye configuration.

There are two general schools of thought regarding BLDC motors. One has the magnetic flux increasing linearly for 120 degrees of rotation then settling out at max positive flux and reversing for about 60 degrees, and then linearly decreasing for another 120 degrees, settling out at max negative for 60 degrees and starting over. This gives the trapazoidal back-emf, which is consistent with your waveform.

The other school prefers to replicate the classical sine wave machine, though few of these motors actually come out ideal. They tend to either look a little more like triangles (third harmonic in phase) or less often, a little flattened.

I preffer to use trapazoidal motors myself, with constant current control, where the DSP guys like to use PWM (voltage control) waveforms with sine wave motors.

Either is good for a generator through the addition of a simple three phase rectifier (6 diodes). Personally, I think it would be a bit better with the trapazoidal motor since the output is practically free of ripple when rectified.

The only issue you're left with is what to do with a DC voltage that varies to the rate of your motor RPM. If your using an appliance motor, it's probably capable of putting out 300V at operating speed. If you exceed this, it's going to want to do more, and may destroy your circuitry.

Thus two additions may prove useful:
Using a less than ideal speed - in case of overages.
Adding a boost-type regulator circuit to provide constant output voltage for your load.

Best Wishes,

- Mike in Plano

 

[DragonPetter]

Hi Mike in Plano. When you say 2 schools of thought, do you mean that some people are winding the motors differently than others?

One thing I have remembered since I first explained my confusion is that when I measured the back EMF by rotating the rotor with my hand, I was looking at the EMF between 2 phases, and this was sinusoidal. When I measured the back EMF as in the waveform, this was done with respect to common ground.

 

[jim hardy]

it takes careful attention to make a good sine wave.
Recall your flux wave has to be cosine shaped. That won't happen by accident.
The electric company goes to a lot of trouble to make a sinewave.