600W BLDC speed controlled by current output regulated buck-bost converter.

Is this a feasible way to control a BLDC pump?…, ie by having “Buck-boost SMPS” -> “3 PHASE INVERTER” -> “BLDC”
(the buck-boost SMPS is output current regulated, and the current is regulated to that level which makes the BLDC speed go to the desired speed).

So, we wish to supply a 600W BLDC with a buck-boost converter which has output current regulation. The Buck-boost’s output current will be regulated to that level which allows the BLDC to run at whichever the desired speed is.
We thus have no current sensing of the coil current of the BLDC, because the current is being sensed (and regulated), at the output of the buck-boost converter.
There is a three phase inverter between the buck-boost converter and the BLDC. This obviously commutates the current from coil to coil of the BLDC so that the BLDC can go round.

Maximum current in the BLDC is 20 Amps. At maximum current in the BLDC, the voltage of the DC bus, which supplies the three phase inverter, tends to be around 30V.
The applicatin is a fuel pump.


Here is a (now closed) thread of the same system, but there was no conclusion as to whether or not it was possible…
https://www.edaboard.com/threads/311275/

..is the above BLDC control method nonsense?
(there is no coil current sensing in the BLDC motor)

The current from the regulator is obviously split between the three phases of the BLDC motor. The current basically determines the torque or output power. The three phase converter determines the overall speed. These three quantities are coupled: current, output power and the speed- in a way that is best seen graphically. If the frequency is constant, the speed can only be regulated by the current (or voltage, if you like it) at the cost of the torque. As the motor loses sync with the frequency, the torque suffers badly.

IMHO, sensing the coil current does not help much.

IMHO, sensing the coil current does not help much.

Click to expand...

Thanks, -But surely sensing of the motor coil current is essential? ..unless one is doing a current source inverter and the current is being fed to the BLDC and inverter by an upstream Buck converter which operates with pulse to pulse peak current regulation. (ie such that the buck’s output current is very tightly controlled unlike it would be with a buck-boost converter)

If the frequency is constant, the speed can only be regulated by the current (or voltage, if you like it) at the cost of the torque. As the motor loses sync with the frequency, the torque suffers badly.

Click to expand...

Thanks, but as you know, it is the frequency that determines the speed, with a BLDC, we cannot let the rotation loose sync with the frequency, BLDC’s just don’t work like that.

The question I ask in post#2 is a simple one, but the lack of any answer to it despite huge internet searches simply shows just what an area of secrecy that the world of electric motor drives really is. Just imagine if I had asked if an flyback SMPS could be done without sensing the primary current….there would be hundreds of answers saying that primary current must be sensed….

…but I ask an analogous question about a BLDC, that is, should the coil current of a BLDC be sensed, and there will be no answers, and nowhere on the internet is it explained. A simple question that nobody can answer, anywhere on earth.

I fully agree with the sentiments expressed in #4. It is important to know the current in the coils- at least, if for nothing else, to prevent it from hurting itself- and do something about it when it is out of limits. In the same way, it is very important to control to speed by changing the frequency and it is essential not to lose sync.

My suggestions to your very relevant point:

1. Produce full characteristics about current vs torque and identify the linear region (for several frequencies)
2. Produce the frequency vs current graph at different torque and identify the linear region
3. Produce the torque vs frequency at constant current for different currents

The resonance (mechanical vibrations) must be critically damped with different tools.

These graphs must be included in the three phase controller so that when the frequency is reduced the current is kept constant and the voltage is reduced and the load is sensed by the current.

In the simple way, the three phase driver must implement these features but at present these drivers are really primitive (in my opinion) but there is little interest within the industry because the features are rather dependent on the mechanical design of the motor. The only way to do this is by using look up tables...

A brushless dc motor works exactly the same as a dc commutator motor, but without the mechanical brushes or the mechanical commutator.

We don't need to know the drive frequency, because the switching points for the motor windings are determined by the physical position of Hall sensors fitted inside the motor.

Now suppose we have a rather special application of a positive displacement (fuel) pump required to create a fairly constant output pressure, independent of flow.
A motor that delivers constant torque over a very wide speed range would produce a fairly constant output pressure from a positive displacement pump.

If a dc motor is current fed, that is the type of characteristic you will get.

We could make a really nice current source from quite a few different switching power supply topologies, especially if run in current control mode.

Hook the whole thing up together, and it should produce a pretty constant output pressure without much need for overall output pressure feedback, although that would be the real icing on the cake.

Thanks, so basically, are you saying that in the same way that we can supply a string of LEDs with a constant current buck-boost SMPS, we could do exactly the same thing with a BLDC (& of course its inverter) replacing the LEDs?

That is the general idea.
Constant current motor drive, which in this case delivers a constant torque right down to totally blocked flow, where the motor only creeps around slowly.

Warpspeed said:

Constant current motor drive, which in this case delivers a constant torque right down to totally blocked flow, where the motor only creeps around slowly.

Click to expand...

This is not good for health; yes, you must have current control for the output torque and also must have frequency control for the speed. If the motor is slowing down because of load, it is overloaded and slipping (unsync with the freq)- very bad of the motor. If the motor stalls, anything may happen (it may even die). You need to ask for the torque needed (and decide on the current being fed) and also the speed required (so that you can decide on the frequency to be used) and also ensure that the total power handling capacity of the motor is not exceeded.

Thanks, so basically, are you saying that in the same way that we can supply a string of LEDs with a constant current buck-boost SMPS, we could do exactly the same thing with a BLDC (& of course its inverter) replacing the LEDs?

Click to expand...

That is the general idea.

Click to expand...

…Thanks, but I don’t believe that a Brushless DC motor and its inverter can be driven from the output of a output_current_regulated buck-boost converter (there being no actual current sensing at the BLDC, but rather the current just being sensed at the buck-boost output)..
I think this is backed up by the fact that if you search the whole internet and bookshops, you will not find a single case involving an output_current_regulated buck-boost converter supplying an inverter/BLDC.

…you can however, find examples of peak current regulated “buck” converters supplying inverter/BLDC’s, and this as you know, is called a “Current Source Inverter”. However, this always involves a Buck converter because in a buck converter, the dynamics are such that the output current can be easier controlled by having a comparator “Looking” at the buck inductors peak (and/or trough) current. A “Current source inverter” cannot be done with a buck-boost converter, and never is.

The reason that an output_current_regulated buck-boost converter cannot be used to drive an inverter/BLDC is because all you would end up doing is creating an LC oscillator, with the BLDC motor coil inductance resonating with the DC link capacitor. The reason for this is that for dynamic reasons, the BLDC coils are low inductance, and since there’s never room for an enormous DC link capacitor, the LC resonant frequency discussed ends up being higher than the BLDC/inverter commutation frequency…..that in turn means that you get several resonant LC periods in a single BLDC coil commutation interval….and thus you end up with overly high resonating current. This is brought about by the sudden switching in of a BLDC coil across the DC link capacitance….you can see this pulsing effect in the BLDC’s DC link current waveform below in the linked BLDC document.

This resonant effect is not so bad in a Brushed DC motor, and this is because brushed DC motors tend to have much higher coil inductances than brushed DC motors, and also, the commutating from coil to coil in a brushed DC motor is not so “pulsey” in a brushed dc motor (as you can see from the brushed dc motor current linked below). Therefore, you just don’t get this resonance happening within the commutation periods of a brushed DC motor. The LC resonant period of a brushed motor coil and the DC link capacitor tends to be always much greater than any coil’s inter-commutation interval, so you just don’t get the resonance referred to here. Also, as discussed, the pulsing from coil to coil in a brushed DC motor just isn’t as unsmooth as it is in a BLDC…in a BLDC there is literally “dead time” as you commutate from one coil to another, and in this dead time, there is actually no current moving from the DC link capacitor to the Inverter/BLDC. –whereas in a brushed DC motor there isn’t a particular “dead time” of the current moving from the DC link capacitor to the brushed DC motor..the current is more or less continuous..(this is seen in the wavefrom below). Thus the brushed DC motor doesn’t provide such violent switching of current in the LC circuit, and so the resonance problem is less likely to be excited in the brushed DC motor.

As you know, sensing of current in a BLDC’s coils is problematic, since the sensing can be afflicted by noise. If life were so easy that one could just avoid BLDC coil current sensing altogether, and have the BLDC coil current regulation done by an upstream buck-boost converter, then everybody would do it like that. It is testimony that it cannot be done like that nobody is doing it like that. Indeed, if it were possible, then high frequency PWM’ing of the BLDC coils within the commutation PWMs would not even be necessary….and such an ideal world would mean less hysteresis losses in the BLDC motor iron, as well as less switching losses in the inverter’s IGBT’s…alas, an idea world is all this is…….an inverter/BLDC cannot be driven by an upstream output_current_regulated buck-boost converter.


BLDC motor current: (fig2)
https://www.nmbtc.com/fans/white-papers/dc-brushless-cooling-behavior/

Brushed DC motor current:- (fig5)
**broken link removed**

…you can however, find examples of peak current regulated “buck” converters supplying inverter/BLDC’s, and this as you know, is called a “Current Source Inverter”. However, this always involves a Buck converter

Click to expand...

Yes exactly, its the simplest and best solution.
Although any forward converter topology would also work, with the advantage of having a transformer.

because in a buck converter, the dynamics are such that the output current can be easier controlled by having a comparator “Looking” at the buck inductors peak (and/or trough) current. A “Current source inverter” cannot be done with a buck-boost converter, and never is.

Click to expand...

Have to agree with that too.

Part of the fun of designing switching power supplies is in finding unusual solutions to uncommon problems. Just because it has never before appeared in any book does not mean it cannot be done, or will not work.

There is not always the requirement to monitor the load current in a constant current supply. Sometimes you can monitor what that current is doing instead, and that may not only give much better control of the final process, but be easier.

The whole concept if an output_current_regulated buck-boost converter supplying a inverter/BLDC fascinates me a little.

This is because I once worked in a large department of engineers that was working on this very thing, despite objections and technical reports written by myself and other engineers who could see that it was a no-hoper. However, the senior-most engineer had called for it, so it was being developed. Then one day this senior-most engineer suddenly strangely left the job with no other job lined up. Then the entire department, and most of the engineers and managers that had been working in it were layed off.
The Chief Engineer of the organisation had come over from another country to see us, and I believe he could see we were doing it wrong, and he wanted to see it for himself. At a late stage meeting, the Chief engineer said nothing, but just sat at the back and observed us all…when asked for his opinion on our buck-boost solution he just walked to the front and displayed us a “current fed full bridge” solution, then took his picture down, and silently walked to the back of the room. Then he left to go back to his country, then we were closed down.

In this case the BLDC coil current is obviously not sensed in the normal way because the buck-boost is taking care of its current, and increasing the current to the inverter/BLDC as the speed increases. This is of course bogus.

As you know, the standard way to drive an inverter/BLDC is by presenting it with a voltage regulated bus, and then the inverter gives both “commutation PWM”, as well as “high frequency PWM”, to the BLDC coils…….this situation of the two different PWM’s present in a standard “voltage source inverter” is strangely not described in any BLDC literature anywhere………and it is the “high frequwncy PWM” that actually regulates the BLDC coil current, because it effectively treats the BLDC coils as the inductor of a buck converter, and PWMs the current in them so that current gets regulated to whatever the required level is.

The huge secrecy surrounding the electric drive world meant that our entire department got this wrong, and lost our jobs.

treez said:

The Chief Engineer of the organisation had come over from another country to see us, and I believe he could see we were doing it wrong, and he wanted to see it for himself. At a late stage meeting, the Chief engineer said nothing, but just sat at the back and observed us all…when asked for his opinion on our buck-boost solution he just walked to the front and displayed us a “current fed full bridge” solution, then took his picture down, and silently walked to the back of the room. Then he left to go back to his country, then we were closed down..

Click to expand...

The problem is two fold- we need speed control and for that we need frequency control. The commutators in a DC motor automatically converts to a frequency that is proportional to its speed. For the BLDC, we need to sense the speed and feed the correct frequency. As the speed changes, the voltage must be adjusted to keep the current at reasonable level so that a decent torque is produced. This needs to be done in software.

We "usually" prefer constant speed for most common motor applications. But for some things a constant motor torque is much more desirable, even a necessity.

For instance, a winch motor that maintains constant tension, or a positive displacement pump that delivers constant output pressure.

For constant torque applications driving a motor from a current source is a much better solution.
And it can be either a commutator type dc motor, or an induction motor, provided the induction motors drive frequency is derived from rotor position as it always is in a BLDC motor.

Another very common application for a current sourcing switching power supply is for many types of welding.

Yet in the several switching power supply text books I have here, the topologies all seem to be for generating closely regulated output voltages.