Fuel pump motor driver fed by current source?

[treez]

Hello,
Is it the done thing to supply an 750W fuel pump motor with a 25 Amp current source switch mode power supply?
That is, the fuel pump motor is a Brushless DC motor which is driven by a H bridge, which in turn is supplied by a Current source SMPS...(the output of the SMPS is 25A maximum at 32v).

They say that the current source is partly adjusted by the tacho feedback from the motor....via a Frequency to voltage converter which goes into an error amplifier feeding the SMPS.

What I am wondering is, do the engineers know what they are doing?....I mean, surely all one needs to drive the fuel pump motor is the H bridge driver?...why are they wanting to use the current source SMPS upstream of the H bridge?

Is this a common technique?

 

[mtwieg]

I don't know about fuel pumps in particular, but there are some applications where controlling motor torque (current) is more useful than controlling speed (voltage). So sometimes a current regulated supply is appropriate, though there should also be a voltage limit in place to keep the motor safe under light load conditions.

 

[kam1787]

detect fault conditions
increase efficiency - ie. drain on [battery] power supply
alter/control performance/behaviour of the motor

 

[treez]

thanks, though surely one simply does the current limitation in the bridge motor driver?....I mean...the current source smps would have an output capacitor bank, so even if the average current output is regulated, sharp , high spikes of current could still be drawn from the output capacitor bank.?

 

[mtwieg]

thanks, though surely one simply does the current limitation in the bridge motor driver?....I mean...the current source smps would have an output capacitor bank, so even if the average current output is regulated, sharp , high spikes of current could still be drawn from the output capacitor bank.?

An H bridge is only really necessary when you want bidirectional control of the motor, otherwise you can get away with a DC chopper circuit which effectively is buck converter (in which the motor inductance is your choke). Filtering requirements are usually dictated by the motor, not the driver. At high voltages it's often standard practice to put a LC filter between an H bridge output and a motor.

 

[treez]

thanks, i see what you mean, this place is not using the motor as the buck inductor though, they are just regulating the current flowing from the output of the smps, into the motor...I appreciate this can be done with a dc motor, but surely not with a BLDC motor, as they rely on the coils being alternately switched by the h bridge?...or were you not referring to a BLDC motor?

Also, the engineer told me that the current would be varied to match the speed of the motor, but you are saying that its voltage variation that does this. I believe you are correct, but im wondering what the guy I spoke with was on about?

 

[mtwieg]

Oh, I completely missed that it was a BLDC motor. For that a current regulated loop is common, especially when you need to run at very variable speeds. Think about what happens when you want to advance a single step. If you just drive the coils with a fixed voltage, then once the rotor position settles there will be no back emf and your current will just rise indefinitely until the next step (at which point that current may have to ramp all the way down to zero, wasting that energy). Regulating current ensures that you always have a good amount of torque, but not excessive if rotor commutation happens to not be smooth.

 

[treez]

thanks, is that a "current source inverter"?....I read a few research articles on them and it seemed to be like that.

 

[mtwieg]

Not sure if that's a standard term, but it sounds reasonable. Most often a single shunt resistor at the common terminal of all three legs of the inverter is used, rather than sensing current of each individual leg.

Here's a decent looking app note explaining current regulation for BLDC:
https://www.st.com/st-web-ui/static...ical/document/application_note/CD00075973.pdf

 

[treez]

Thanks, the block diagram on page 10 of your link shows what the engineer kind of described to me...but in that link, i don't see a current regulated SMPS upstream of the BLDC driver. Do you know if this practice of having a current regulated SMPS upstream of the BLDC driver is actually done anywhere.

The link that you kindly linked, seems to suggest that the DC link current can be inferred by sensing motor coil windings, and thus an external current regulated smps is not required?

 

[mtwieg]

Using a current regulated supply to bias a SMPS is a reasonable strategy. In the end, the result is the same, the difference is whether current regulation is handled digitally by the MCU via PWM of the inverter, or by a separate PWM control IC while the inverter itself is driven without PWM. So it's basically a tradeoff between hardware and software cost/performance.

 

[treez]

by a separate PWM control IC while the inverter itself is driven without PWM

...do you mean that the inverter is just driven by a constant duty cycle oscillator?

Using a current regulated supply to bias a SMPS is a reasonable strategy

...thanks, but I think you meant "Motor driver" not "smps"?

 

[mtwieg]

...do you mean that the inverter is just driven by a constant duty cycle oscillator?

Well you would probably drive each leg with a SPWM, since you want sinusoidal output current. But the point is that amplitude of the SPWM (in terms of duty cycle range) does not need to be controlled with a software feedback loop. The amplitude of the inverter current is instead controlled by varying the bias current from the SMPS. So it's basically of like a high level modulation strategy (ie the Kahn technique). The frequency of the SPWM will still have to be controlled, however.

...thanks, but I think you meant "Motor driver" not "smps"?

Correct, my mistake.