Fuel Pump speed control
Discussion
stevieturbo said:
Whilst it really isnt something I'd be interested in myself....
If the limitations of the unit may be 20A.
I guess it wouldnt really cost much more to add a further high power driver if 2 pumps were required ?
If going to the trouble of building a controller for a high powered fuel system, there is a good chance more than one high pressure pump may be needed.
Given most cars will have an ecu....is an inertia switch needed ?
CAN....how many people are really likely to use this over conventional stuff ? Although no idea how much cost it would add.
I take it the return solenoid is actually the FPR ?
And with regards to the siwrl tank. Is there anything here to ensure the swirl tank remains full. ie, could there be a warning system, or some system in place to ensure it IS always kept full, or warning issues should the level drop for whatever reason
It would be easy (and not that expensive) to add a second 20A driver for a 2nd HP fuel pump if necessary. The idea was to use the "voltage boosted" output to enable a single pump only to be used?? (say 044 Bosch + 25% flow)If the limitations of the unit may be 20A.
I guess it wouldnt really cost much more to add a further high power driver if 2 pumps were required ?
If going to the trouble of building a controller for a high powered fuel system, there is a good chance more than one high pressure pump may be needed.
Given most cars will have an ecu....is an inertia switch needed ?
CAN....how many people are really likely to use this over conventional stuff ? Although no idea how much cost it would add.
I take it the return solenoid is actually the FPR ?
And with regards to the siwrl tank. Is there anything here to ensure the swirl tank remains full. ie, could there be a warning system, or some system in place to ensure it IS always kept full, or warning issues should the level drop for whatever reason
The "inertia" input could also simply be used for the engine ECU to "enable the pump run" (or disable it when engine stops etc) or for an immobilizer input etc
CAN doesn't add a huge cost (something like ~£10 worth of components, and would be an easy "delete" option if people didn't want that option, but enables datalogging and remote control by more advanced systems etc
The system doesn't require a pressure regulator if properly calibrated, the Pressure drop Valve allows better pressure control on transients, and the ability to self-bleed (useful if car is ever accidentally run dry etc)
I thought about a swirl pot "full" sensor, but it adds significant cost (either IR "refraction" sensor, capacitive or a pressure sensor (that would require to be added to swirl pot) As the system knows exactly how much fuel is being delivered, i would just run the lift pump to suit the outgoing fuel mass (+ a bit for safety etc) Overall it would be an easy add, but again the costs go up.
Re 2 pumps.
Surely people looking for such a controller, would be those pushing big power, and few would want to push a single pump to its limits.
So for those seeking say 6-700+, there are few single pumps available ( other than big yank stuff ) that can cut it.
2 pumps is almost a must.
If they are making less power, that a single pump can easily supply....I can see little reason they would need an expensive controller in the first place.
Surely people looking for such a controller, would be those pushing big power, and few would want to push a single pump to its limits.
So for those seeking say 6-700+, there are few single pumps available ( other than big yank stuff ) that can cut it.
2 pumps is almost a must.
If they are making less power, that a single pump can easily supply....I can see little reason they would need an expensive controller in the first place.
Good point about the two pumps!
Assuming there were two high power outputs, would you want the pumps to come into play sequentially or in parallel?? (i.e. pump1 starts low gets to 100%, then pump2 cuts in low and ramps to 100%, OR both pumps start low and ramp to 100% ??)
It would be easy in a sequential system to add a bit of code to ensure that the system always started with a different pump to ensure both were used evenly (on things like turbo bentleys where they have two pumps, the problem is that they hardly ever actually use the 2nd pump, which can cause it to fail, and then not work when required etc!)(less likely to be such an issue on a more performance car however)
Assuming there were two high power outputs, would you want the pumps to come into play sequentially or in parallel?? (i.e. pump1 starts low gets to 100%, then pump2 cuts in low and ramps to 100%, OR both pumps start low and ramp to 100% ??)
It would be easy in a sequential system to add a bit of code to ensure that the system always started with a different pump to ensure both were used evenly (on things like turbo bentleys where they have two pumps, the problem is that they hardly ever actually use the 2nd pump, which can cause it to fail, and then not work when required etc!)(less likely to be such an issue on a more performance car however)
If you could current monitor the 2 HP outputs, then running the 2 pumps in parallel at all times makes the most sense.
No pump ever sits dormant, but more importantly current draw should really be the same.
If for whatever reason current draw varies by say 1A, clearly something is up. So perhaps a warning monitor could be assigned using that method ?
It would cover short circuit, overload, open circuit. Pretty much anything really.
having the 2 mirrored means that one almost monitors the other for performance
No pump ever sits dormant, but more importantly current draw should really be the same.
If for whatever reason current draw varies by say 1A, clearly something is up. So perhaps a warning monitor could be assigned using that method ?
It would cover short circuit, overload, open circuit. Pretty much anything really.
having the 2 mirrored means that one almost monitors the other for performance
I was indeed planning to monitor all the output currents, enabling detection/protection of failed pumps / wiring, and also, by measuring the pump current (proportional to pump output pressure) and knowing the pwm percent, the back emf of the pump can be calculated, which is proportional to pump speed, and hence fuel delivery quantity.
Max_Torque said:
I was indeed planning to monitor all the output currents, enabling detection/protection of failed pumps / wiring, and also, by measuring the pump current (proportional to pump output pressure) and knowing the pwm percent, the back emf of the pump can be calculated, which is proportional to pump speed, and hence fuel delivery quantity.
But as you are electronically controlling output and then electronically regulating pressure...will you be able to tell if the pump is working using pressure ?Wouldnt that require a specific set of data to compare to ? TBH, its going over my head....my 2 pump idea was simple lol.
It's not too complicated luckily, and all revolves around that old favourite V=IR (and its close friends. P = VI & P = IsquaredR)
Take a Battery voltage of say 14V, and a PWM value of say 50%. Measure the pump current at say 3Amps. The only thing left to know is the pumps series resistance, typically in the order of 0.8Ohm(either measure this with a multimeter, or system can measure it automatically when the pump isn't running by injecting a small voltage and measure in the current)(needs to be small to prevent pump actually turning)
So for:
BattV = 14V
Pump R = 0.8 Ohm
PWM = 50%
Current = 3.0A
The pumps effective supply voltage = (PWM% / 100) * battV, in this case 50% of 14V = 7V ( with a 50% PWM, the voltage is effectively OFF for half the time, hence the applied voltage is also half the DC supply voltage)
The pumps "forward" voltage (that which actually drives the current through the motor to make torque, (V= IR)) V = (3.0A * 0.8Ohm) = 2.4V (most of the supply voltage is simply used to conteract the Back EMF (see below), if the pump is stalled and there is no back emf, you can see why the motor will quickly fail from over current (7V supplied, through 0.8 Ohm = 8.8A, or at 100% PWM = 17.5A !!)
The difference between the effective "supply" voltage and the forward voltage must be the pumps' back EMF (the reverse voltage generated by the spinning magnets) so, back EMF = (7.0 - 2.4) = 4.6V
And as the back EMF is generated by the pump spinning, we know that this voltage is proportional to the speed of the pump (and hence its flow rate), where the max speed is clearly where the Back EMF = the supply voltage (in reality this can never quite occur due to frictional and copper losses in the rotor/windings) in this case the pump is going around about 33% of it's maximum speed (and the fuel mass supplied can easily be worked out from that speed)
We also know that the electrical input power (in watts) of the pump is (P = VI), = 7v x 3A = 21 Watts
The useful work done by pump is proportional to the Back EMF x current (just like IC engine power = rpm x torque) = 4.6 x 3 = 13.8W
That leaves us with a difference of 21 - 13.8 = 7.2W of loses (and conviently, the copper losses of the pump (P = IsquaredR) agree (3 squared x 0.8 = 7.2 W ;-) )These losses are transfered to heat in the copper wires of the rotor, and ultimately are transfered to the fuel (motor is submersed in fuel for cooling etc)
From all that we can see that at that particular operating point, the pump motor has an electrical efficiency of 65.7% ((13.8W/21W) *100).
Hence by measuring / knowing those 4 inital conditions, we can calculate the exact status and condition of the pump !!! Simples ;-)
Take a Battery voltage of say 14V, and a PWM value of say 50%. Measure the pump current at say 3Amps. The only thing left to know is the pumps series resistance, typically in the order of 0.8Ohm(either measure this with a multimeter, or system can measure it automatically when the pump isn't running by injecting a small voltage and measure in the current)(needs to be small to prevent pump actually turning)
So for:
BattV = 14V
Pump R = 0.8 Ohm
PWM = 50%
Current = 3.0A
The pumps effective supply voltage = (PWM% / 100) * battV, in this case 50% of 14V = 7V ( with a 50% PWM, the voltage is effectively OFF for half the time, hence the applied voltage is also half the DC supply voltage)
The pumps "forward" voltage (that which actually drives the current through the motor to make torque, (V= IR)) V = (3.0A * 0.8Ohm) = 2.4V (most of the supply voltage is simply used to conteract the Back EMF (see below), if the pump is stalled and there is no back emf, you can see why the motor will quickly fail from over current (7V supplied, through 0.8 Ohm = 8.8A, or at 100% PWM = 17.5A !!)
The difference between the effective "supply" voltage and the forward voltage must be the pumps' back EMF (the reverse voltage generated by the spinning magnets) so, back EMF = (7.0 - 2.4) = 4.6V
And as the back EMF is generated by the pump spinning, we know that this voltage is proportional to the speed of the pump (and hence its flow rate), where the max speed is clearly where the Back EMF = the supply voltage (in reality this can never quite occur due to frictional and copper losses in the rotor/windings) in this case the pump is going around about 33% of it's maximum speed (and the fuel mass supplied can easily be worked out from that speed)
We also know that the electrical input power (in watts) of the pump is (P = VI), = 7v x 3A = 21 Watts
The useful work done by pump is proportional to the Back EMF x current (just like IC engine power = rpm x torque) = 4.6 x 3 = 13.8W
That leaves us with a difference of 21 - 13.8 = 7.2W of loses (and conviently, the copper losses of the pump (P = IsquaredR) agree (3 squared x 0.8 = 7.2 W ;-) )These losses are transfered to heat in the copper wires of the rotor, and ultimately are transfered to the fuel (motor is submersed in fuel for cooling etc)
From all that we can see that at that particular operating point, the pump motor has an electrical efficiency of 65.7% ((13.8W/21W) *100).
Hence by measuring / knowing those 4 inital conditions, we can calculate the exact status and condition of the pump !!! Simples ;-)
I am interested in a fuel pump speed controller and I know at least one other person seeking one as well.
Now my knowledge of these is limited to the research that I have done recently so excuse me if I start to speak complete rubbish but here is what I would want.
1. needs to pulse rather than voltage controlled as I understand lowing the voltage to control speed will destroy a fuel pump.
2. needs to hold 20amps and have a peak start voltage of 40 amps.
3. ecu controlled is preferred (i'm using a Motec)
4. reliable (unlike the Motec controler)
5. cost effective.
6. easy to setup (as Im a dork when it comes to electronics lol)
Im not sure how this fits in with your gradings as I didnt entirely understand them all hence putting my list up, Im definatelly not running CAN for the reason in number 6
Now my knowledge of these is limited to the research that I have done recently so excuse me if I start to speak complete rubbish but here is what I would want.
1. needs to pulse rather than voltage controlled as I understand lowing the voltage to control speed will destroy a fuel pump.
2. needs to hold 20amps and have a peak start voltage of 40 amps.
3. ecu controlled is preferred (i'm using a Motec)
4. reliable (unlike the Motec controler)
5. cost effective.
6. easy to setup (as Im a dork when it comes to electronics lol)
Im not sure how this fits in with your gradings as I didnt entirely understand them all hence putting my list up, Im definatelly not running CAN for the reason in number 6
(0 = don't bother, 10 = must have!)
1) CAN enabled (so you can control / data log via CAN
0
2) speed control for a seperate(additional to main pump control) lift pump (10A max)
5
3) Voltage output (0-5v) for current "fuel pressure" (saves having to have another sensor for your EMS to read)
?? Im fitting a sensor on my fuel rail for the motec to read, am I missing a trick?
4) Crash switch input (inertia switch signal to shut off pump(s) in event of crash etc)
8 great idea, not sure my car has anything like this currently
5) pump fault output signal (to dash lamp or even relay to cut engine etc)
8 good idea again, although my ecu reads pressure and has can to my stack dash so not sure i need it??
6) Control of a "pressure drop valve" to allow easy system bleeding (for returnless systems) and also more accurate pressure control on negative throttle conditions(otherwise you have to wait for the engine to use the fuel to drop the pressure)
8 as above not sure I would need it but great idea.
7) MAP sensor voltage input to allow "load dependant" pressure setpoint (and Fuel injector deltaP output signal for ems etc)
10
8) Main pump voltage boosting (approx 20V max output) to allow main pump to "overrange" as required
does this include startup peak?
9) MIL spec connectors (some more reliability, MUCH more cost!)
no idea
tried to add my interpretation of an inteligent response?
The great thing about having a pwm controlled pump is you totally avoid the "Inrush" current spike by being able to "softstart" the pump(s). (you' don't just hit them with the full voltage when they are at zero speed)
I am currently ering towards 3 x 20A pwm outputs, with configurable scheduling, so each output can either run a high pressure pump, or a lift pump (so either 1 x lift and 2 x pressure, or 2 x lift (for odd shapped / multiple tanks) and 1 x pressure)
There are really 2 different levels of "inteligence" for the control stragegy:
1) the simplest: Using a standard fuel system with press reg and return line, and just doing basic speed control of the pumps, controlled with a calibratable(inside pump controller) "speed" vs "load" table (directly from a 0-5v MAP signal) or just following a 0-100% pwm value (for example, using Motec to output a low current pwm that the pump controller "amplifies" to drive the pumps, (requires a table within motec set to aux pwm output with engine rpm and load as it's axis) For either case it would require the used to calibrate the pwm table to ensure that the pumps output voltage (0-100%) is sufficient, vs load and speed, to generate enough pressure to just start to "spill" a very small amount of fuel back to the tank (and hence leave passive pressure reg in control of fuel rail pressure)
2) Pump controller runs in "Pressure control" mode, closed loop presure control (with or without return fuel line / pressure regulator). Pump controller either sets its pressure target with a calibratable table in the controller, with again axis of rpm and load(MAP), OR sets its pressure target proportional to a pwm signal from say Motec, OR just has a fixed pressure target at all loads/speeds. In these cases the pumps speed (and hence it's flowrate) will be set via a closed loop controller using presure error as it's main parameter. In this case the user would have to calibrate some basic PID values to "optimise" the control response to their particular system.
In all cases the system will monitor directly the following:
1) battery (supply) voltage
2) pump current (for each pump used)
3) controller internal temperature
4) an INHIBIT input (either crash, immo, or other "shutdown" signal)(disable-able in s/w)
and indirectly (requiring additonal or pigybacked sensors)
5) engine MAP (0-5v MAP sensor or via CAN)
6) engine rpm (0-12v coil "trigger" hardwired (same as a std "tacho") or via CAN
7) ignition line voltage (KL15) hardwired "ign on" signal to let system bootup
8) fuel rail pressure (for "pressure control version only") (0-5v FP sensor) or via CAN
I am currently ering towards 3 x 20A pwm outputs, with configurable scheduling, so each output can either run a high pressure pump, or a lift pump (so either 1 x lift and 2 x pressure, or 2 x lift (for odd shapped / multiple tanks) and 1 x pressure)
There are really 2 different levels of "inteligence" for the control stragegy:
1) the simplest: Using a standard fuel system with press reg and return line, and just doing basic speed control of the pumps, controlled with a calibratable(inside pump controller) "speed" vs "load" table (directly from a 0-5v MAP signal) or just following a 0-100% pwm value (for example, using Motec to output a low current pwm that the pump controller "amplifies" to drive the pumps, (requires a table within motec set to aux pwm output with engine rpm and load as it's axis) For either case it would require the used to calibrate the pwm table to ensure that the pumps output voltage (0-100%) is sufficient, vs load and speed, to generate enough pressure to just start to "spill" a very small amount of fuel back to the tank (and hence leave passive pressure reg in control of fuel rail pressure)
2) Pump controller runs in "Pressure control" mode, closed loop presure control (with or without return fuel line / pressure regulator). Pump controller either sets its pressure target with a calibratable table in the controller, with again axis of rpm and load(MAP), OR sets its pressure target proportional to a pwm signal from say Motec, OR just has a fixed pressure target at all loads/speeds. In these cases the pumps speed (and hence it's flowrate) will be set via a closed loop controller using presure error as it's main parameter. In this case the user would have to calibrate some basic PID values to "optimise" the control response to their particular system.
In all cases the system will monitor directly the following:
1) battery (supply) voltage
2) pump current (for each pump used)
3) controller internal temperature
4) an INHIBIT input (either crash, immo, or other "shutdown" signal)(disable-able in s/w)
and indirectly (requiring additonal or pigybacked sensors)
5) engine MAP (0-5v MAP sensor or via CAN)
6) engine rpm (0-12v coil "trigger" hardwired (same as a std "tacho") or via CAN
7) ignition line voltage (KL15) hardwired "ign on" signal to let system bootup
8) fuel rail pressure (for "pressure control version only") (0-5v FP sensor) or via CAN
From a simplicity point of view, retaining a mechanical FPR would probably suit most people, and I suspect make the cost of the controller itself cheaper ?
As for crash, virtually every ecu already has fuel pump control built in, in terms of shutting it down.
I'd imagine the ecu FP output will still be the primary trigger to enable the pump.
So is a secondary kill needed ?
But in saying the above....given the number of modern cars running dead end systems, there could be a market to allow these guys to run a lot of fuel pump, and retain the dead end for simplicity sake. SO fuel pressure control there may make sense.
Other than that, a simple 0-5v or 0-100% Duty for it's control signal might be the simple option ?
I guess there are many possibilities. But either way, the unit needs to come in at a sensible price. Doesnt matter if the spec is amazing, if the price is too high, nobody is going to buy it
As for crash, virtually every ecu already has fuel pump control built in, in terms of shutting it down.
I'd imagine the ecu FP output will still be the primary trigger to enable the pump.
So is a secondary kill needed ?
But in saying the above....given the number of modern cars running dead end systems, there could be a market to allow these guys to run a lot of fuel pump, and retain the dead end for simplicity sake. SO fuel pressure control there may make sense.
Other than that, a simple 0-5v or 0-100% Duty for it's control signal might be the simple option ?
I guess there are many possibilities. But either way, the unit needs to come in at a sensible price. Doesnt matter if the spec is amazing, if the price is too high, nobody is going to buy it
Max_Torque said:
andygtt said:
1. needs to pulse rather than voltage controlled as I understand lowing the voltage to control speed will destroy a fuel pump.
I'm not really sure what this means btw??andygtt said:
Max_Torque said:
andygtt said:
1. needs to pulse rather than voltage controlled as I understand lowing the voltage to control speed will destroy a fuel pump.
I'm not really sure what this means btw??For the case of a fuel pump, (brushed DC motor, approx 100mH inductance) a PWM frequency above approx 500Hz will be sufficient to limit excessive ripple currents.
The primary advantage of using a high frequency "square wave switched" supply is that the semiconductor switches (the MOSFETS) spend the majority of their time either fully off or fully on. This is the state at which they have the least resistance, and hence the least losses.
If you try to control the output voltage with a "Linear" device, the losses are enormous because you effectively have to continuously "throttle" the applied voltage within some resistance (as opposed to reacting it against a virtually lossless inductance).
Take a pump motor with a 3 ohm resistance for example, if you supply it with the full 14.7 battery voltage (for full speed running) it will consume 4.9Amps (V=IR). Now if you wanted to run it at half speed, you need to only supply it with 7.35V (half the battery Volts). This means you need to "drop" half the supply voltage across a resistor.
In this case we will need a 3ohm resistor in series with the pump (so that the supply voltage is evenly split between the 3ohm pump and the 3ohm resistor).
The problem is now that we are still supplying 14.7V from the battery, but effectively wasting half of it over some resistance that does no "useful" work. (in fact both pump and resistor each use 18watts)
This produces a very inefficient system (and a lot of heat to deal with)
No damage caused to a DC motor if it is stalled with a limited current running through its windings.
(Stalled motor's are worst case because the multiple windings are not sharing the current between them (the rotor is not turing, so the brushes just keep feeding a single set of windings). if you stall a motor and supply the full voltage to it, it's low resistance will quickly cause a huge current flow and overheat the pump to failure.
In actual fact, with a fuel pump in proper working order (i.e. not jammed up by a foreign object etc) the pump never completely stalls, even at a current that is insufficent to generate enough torque to overcome the pressure head, because the roller or vane pump has enough "leakage" paths to allow the pump rotor to slowly turn anyway.
ok so your saying that regardless of method of reducing pump speed no additional wear will occur unless its completelly stalled and restarted during the process which cant really happen?
However if the voltage is simply reduced for example to a flat 7V to reduce pump speed, then we have 8V being dumped which produces heat in both the resistors or whatever used to drop the voltage and the pump.
Therefore the correct way to reduce pump speed is to pulse the full 14v from the battery to the pump... which in effect actually reduces the voltage the pump sees as long as its done fast enough?
So the speed controller you propose to make pulses the full voltage to the pump to slow it down?
Also you are intending to 'soft' start the pump/s therefore my huge single A1000 will not require 40amps or so to start up?
BTW I like the idea of being able to use the controler to manage a lifter pump if I decide to fit one later.
As I said Im not great on electronics etc lol
However if the voltage is simply reduced for example to a flat 7V to reduce pump speed, then we have 8V being dumped which produces heat in both the resistors or whatever used to drop the voltage and the pump.
Therefore the correct way to reduce pump speed is to pulse the full 14v from the battery to the pump... which in effect actually reduces the voltage the pump sees as long as its done fast enough?
So the speed controller you propose to make pulses the full voltage to the pump to slow it down?
Also you are intending to 'soft' start the pump/s therefore my huge single A1000 will not require 40amps or so to start up?
BTW I like the idea of being able to use the controler to manage a lifter pump if I decide to fit one later.
As I said Im not great on electronics etc lol
andygtt said:
ok so your saying that regardless of method of reducing pump speed no additional wear will occur unless its completelly stalled and restarted during the process which cant really happen?
However if the voltage is simply reduced for example to a flat 7V to reduce pump speed, then we have 8V being dumped which produces heat in both the resistors or whatever used to drop the voltage and the pump.
Therefore the correct way to reduce pump speed is to pulse the full 14v from the battery to the pump... which in effect actually reduces the voltage the pump sees as long as its done fast enough?
So the speed controller you propose to make pulses the full voltage to the pump to slow it down?
Also you are intending to 'soft' start the pump/s therefore my huge single A1000 will not require 40amps or so to start up?
Yup, correct However if the voltage is simply reduced for example to a flat 7V to reduce pump speed, then we have 8V being dumped which produces heat in both the resistors or whatever used to drop the voltage and the pump.
Therefore the correct way to reduce pump speed is to pulse the full 14v from the battery to the pump... which in effect actually reduces the voltage the pump sees as long as its done fast enough?
So the speed controller you propose to make pulses the full voltage to the pump to slow it down?
Also you are intending to 'soft' start the pump/s therefore my huge single A1000 will not require 40amps or so to start up?
Electronics is actually pretty simple really ! I like to think of the "water" equivalent model, where Voltage = water pressure, Current = water flow rate, and resistance = a small hole that water is forced through, and inductance = a large damping volume (an inertia)
The key to PWM is just ensuring that you pulse the voltage on and off fast enough so the thing you are powering doesnt have time to react. (the things "inductance" being the electrical equivalent (more or less ;-) of mechanical inertia
Here's a "motor" (in fact a resistance and an inductance combined to sim a motor) being driven by 12 volt supply, switched at 50% On-Off ratio at two different frequencies:
500Hz:
and 125Hz
The green line is the applied voltage (switching between 0 and 12v) and the blue line the motors current. In both cases the eventual "mean" current is just under 5amps, but with the slower frequency, the inductance of the motor does not "damp" the current swings as much, so although the mean current is the same, the current ripple is much larger. In reality this current ripple would generate a torque ripple, and the motor would (depending upon its mechanical inertia) try to accel/deccel to follow it. At some frequency of supplied pwm, the motor's electrical inductance, and its mechanical inertia result in a sufficient damping to effectively "ignore" the current ripply, and the motors torque will just be that of the mean current flow.
As far as the motor is concerned, the faster frequency the better, but the semiconductor switches that control this frequency take a certain power to switch, (and generate more heat during these switches) so chossing the pwm frequency becomes a compromise between torque ripple and power electronics efficiency.
Correct, you just pulse the power on and off fast (well fast is mechanical terms) and the On to Off ratio controls the delivered voltage (50% on 50% Off = 50% of supply voltage, 10% on 90% Off = 10% of supply voltage)
A small DC brushed motor (like that inside a fuel pump) has a really quite high inductance and mech inertia, and because it's only moving a bit of fuel around you don't care too much about the ripple current/torque.
For something more critical like say a brushless traction motor for an EV, you'd be using a pwm frequency of approx 20kHz (20 thousand cycles per sec) (and in fact measuring both motor rotor position and phase current and doing the requisite FOC calcs on every single pwm cycle (50uS per cycle)
It took me a while to get around just how fast modern processors are, even at the basic level, where somthing like £2.50 buys you a chipset that delivers 20million instructions per second (20MIPS) !!!
A small DC brushed motor (like that inside a fuel pump) has a really quite high inductance and mech inertia, and because it's only moving a bit of fuel around you don't care too much about the ripple current/torque.
For something more critical like say a brushless traction motor for an EV, you'd be using a pwm frequency of approx 20kHz (20 thousand cycles per sec) (and in fact measuring both motor rotor position and phase current and doing the requisite FOC calcs on every single pwm cycle (50uS per cycle)
It took me a while to get around just how fast modern processors are, even at the basic level, where somthing like £2.50 buys you a chipset that delivers 20million instructions per second (20MIPS) !!!
Gassing Station | Engines & Drivetrain | Top of Page | What's New | My Stuff