Engineer/Geek Types, questions about PWM controlled Brushless and Voltage Boosters

[Angrey]

The great thing about the PCM logic is that it knows the fuel demand BEFORE the engine actually needs it. An external regulator can never know what's about to happen like the PCM does. The PCM knows the future airflow and future fuel flow before it happens because the PCM is controlling both. As such, it can send the appropriate voltage to the pump before it needs the extra flow. That's the genius of the Ford DBW system and the Feed-Forward controls. It already knows how much to open the throttle, where to send the cams, how long to leave the injectors open, what spark timing to send, etc because it is all modelled before the state-change. An external regulator has no predictive ability and can only react to changes after they happen.

It's just different strategies. The electronic regulator (with return style) doesn't need to anticipate because it has capacity in reserve. The whole point of variable control is to economize the fuel pump consumption/draw and reduce electric load and not put more heat into the fuel than necessary. The return style system accomplishes this in a slightly less than optimal way by sending a portion of the fuel back to the tank (in reserve). The PCM "approximation" has to anticipate any rapid changes in demand and the only way it can do that in large power applications is to run the pumps at higher output than is currently necessary, and even then, the lag from when it wakes up to when the additional flow is at the rails is what drives many tuners to just go return style.

Would be an interesting study to see overall which one manages the pumps/draw more economically. In any case, the part you didn't address is that I can mechanically/physically install my setup and away we go. Yours has to be tuned by someone with time, skill and experience and then revised/updated to account for any gaps or inaccuracies in the tables. You have to remember, not everyone is as smart or dedicated as you. For most people, simplifying this system is a win-win and fool proof (while eliminating the returnless lag issue, economizing the fuel pump loading, etc).

My setup simply rolls up the windows when it starts to rain. Your approach has to draw on schedules and information from weather data and pressures and light levels, etc to try to "infer" when it's going to rain. Can you get to the same outcome? Sure, but I walk up one block and you have to walk 5 blocks over 3 blocks up, 5 blocks back and 2 blocks rearward to arrive at the same place.

Sponsored

 

[engineermike]

The electronic regulator (with return style) doesn't need to anticipate because it has capacity in reserve.

An external regulator responds to demand changes by measuring the fuel pressure and closing if the pressure drops. But the pressure has to drop first before it can do anything in response. It's purely a feedback system with no feed-forward control. There has to be error in order for it to respond.

The whole point of variable control is to economize the fuel pump consumption/draw and reduce electric load and not put more heat into the fuel than necessary. The return style system accomplishes this in a slightly less than optimal way by sending a portion of the fuel back to the tank (in reserve).

With a return-style system, the pump is running at full voltage and amperage all the time. The flowrate through the pump is constant, as is the power consumption. The pump doesn't "know" that most of it's throughput is just getting dumped back into the tank.

The PCM "approximation" has to anticipate any rapid changes in demand

The PCM has a really good idea of what the changes in demand are because the PCM is controlling those changes. When you press the throttle, the engine calculates what all future parameters will be and controls them. When fuel demand increases from 2 lb/min to 15 lb/min, it's not like the PCM was surprised by this because the PCM created it. With a cable-actuated throttle, you would be correct because the fueling changes happen in response to airflow changes. But in a modern Ford, the computer is controlling all of it, so it already knows the future state before it even happens.

the only way it can do that in large power applications is to run the pumps at higher output than is currently necessary,

It runs the required voltage prior to the high demand.

Would be an interesting study to see overall which one manages the pumps/draw more economically.

I did some math for fun, and the result of a constant 13 voltage externally regulated system running at 30 amps would be adding 2 deg of heat to the fuel tank for every minute the pumps run. A modulated pump would be approximately 10% of that.

In any case, the part you didn't address is that I can mechanically/physically install my setup and away we go. Yours has to be tuned by someone with time, skill and experience and then revised/updated to account for any gaps or inaccuracies in the tables. ...For most people, simplifying this system is a win-win and fool proof.

I believe this is accurate. I see a common theme that tuners recommend customers to buy parts that make their jobs easier.

You have to remember, not everyone is as smart or dedicated as you.

Thanks you, but I am only as smart as those who have taken the time to guide me.

My setup simply rolls up the windows when it starts to rain. Your approach has to draw on schedules and information from weather data and pressures and light levels, etc to try to "infer" when it's going to rain. Can you get to the same outcome? Sure, but I walk up one block and you have to walk 5 blocks over 3 blocks up, 5 blocks back and 2 blocks rearward to arrive at the same place.

I would say the return regulator approach would be akin to holding your throttle wide open all the time and using the clutch to control torque to the wheels. Mine modulates the power of the engine.

 

[Angrey]

Mike,

You clearly do not understand how the regulator works. You need to read up on it.

First, no, the pressure does not drop for a response. The regulator monitors the return line diaphram and when it sense a drop in THAT pressure, it changes the PWM signal to the pump to increase output. The rails never see a change in pressure. Conversely, if the regulator sense an INCREASE in return line pressure, it sends a change PWM signal to ease the pumps and slow the flow to restore the return line pressure back to "base". The back side of the physical regulator remains constant pressure while the pump increases or decreases output to maintain a constant pressure all based upon return line flow.

That return line base pressure is a physical reverse anagram for the "flow" of the system. When the flow on the return line falls, the regulator knows the rails are consuming more and tells the pumps to send more. When the regulator senses a traffic jam and an increase in the return line flow, it knows the rails are consuming less and correspondingly eases the pumps to provide less flow.

The result is a variable rate system WITH FLOW/PRESSURE IN RESERVE, the best of both worlds. It doesn't have to wait for the pumps to spin up to maintain pressure on the rail side, it does that by simply operating naturally and returning less. That "base" return line pressure is analogous to how much reserve you want to carry. If ever a situation arises where you stomp on the throttle and the logs show a drop in rail pressure, you know that you're not carrying enough reserve to adequately maintain constant pressure under that condition. You can change the flow restrictor and increase it.

I'm not going to respond to the rest of your comments because they're all fruit of a poisonous tree where you start off with an improper understanding of how the Fuelab regulator actually works.

 

[Grimreaper]

Its measuring a drop in pressure and responding after.

Patent

The differential pressure created in the return reservoir and restrictor prior to the return line chamber is dependent on the amount of intake chamber pressure and then lift on the diaphragm and ball seat. In order for a change in pressure to occur in the return reservoir a change has to occur prior or after this reservoir. The patent shows some adjustments to the reservoir chamber restriction but that's fixed once setup from your input. So that leaves the intake chamber pressure being the driver of change in the reservoir chamber. In order for flow and pressure to exists in varying amounts in the reservoir chamber the ball seat has to open from intake chamber pressure on the diaphragm. Intake chamber pressure is rail pressure. This differential pressure is what is measured.

 

[bankyf]

@engineermike Do I remember correctly, that you had a hard time controlling a brushless pump from the OEM PWM signal, or did you get that worked out? If you did, would it still be able to effectively control a return style system, or do you think the return system is unnecessary and another product of tuners not wanting to do the additional work?

 

[Angrey]

Its measuring a drop in pressure and responding after.

Patent

The differential pressure created in the return reservoir and restrictor prior to the return line chamber is dependent on the amount of intake chamber pressure and then lift on the diaphragm and ball seat. In order for a change in pressure to occur in the return reservoir a change has to occur prior or after this reservoir. The patent shows some adjustments to the reservoir chamber restriction but that's fixed once setup from your input. So that leaves the intake chamber pressure being the driver of change in the reservoir chamber. In order for flow and pressure to exists in varying amounts in the reservoir chamber the ball seat has to open from intake chamber pressure on the diaphragm. Intake chamber pressure is rail pressure. This differential pressure is what is measured.

It's a drop in pressure on the RETURN side. The pressure on the rail side remains constant and the return side flow varies according to temporary increases/decreases associated with ramping in and out of pump duty.

Not sure why this is so difficult to understand. It keeps the desired pressure on the rail side by sensing a reduction in steady state pressure on the return side. Before the return side flow is reduced to zero (and a corresponding loss in rail side pressure), it electronically adjusts the PWM control of the pumps to provide more flow/output. Conversely, a pressure increase (on the RETURN side) indicates that the pumps are unnecessarily working resulting in excessive return flow and they adjust their signal to reduce pump duty accordingly.

It's a quite elegant and brilliant way of managing the fuel system.



In a typical automotive configuration, a returnless system can not respond instantly to spikes in rail consumption. You can automate pump increases associated with all sorts of inputs from load, throttle position, etc, but at power levels where the flow starts to become greater and greater, the felt "lag" becomes more problematic. You can obviously tune around this lag with adjustments to injector pulse, but the greater the issue, the more challenging and esoteric it becomes.

In the Fuelab setup, the rail side pressure remains constant to the physical set spring value. In the event that the return reserve is not large enough to accommodate maximum spikes in consumption, you can install a less restrictive fitting on the return side which requires more flow to reach the required steady state "back pressure" (for lack of a better term). While that increases the amount of flow available in reserve, it also further negates the desired reduction in pump duty for lower demand conditions.

The other option is to simply run a larger return line (reducing friction loss and pressure build up) but again, the pumps will simply operate at an elevated low level duty, which degrades the desired outcome.

 

[engineermike]

We understand it’s measuring pressure on the return line, but the return line pressure upstream of the regulator is virtually the same as the rail pressure. If fuel flow through the injectors suddenly increases, then in order to maintain the rail pressure the return regulator must initially close in order to restore rail pressure as quickly as possible. If you look at the cross section of that fuelab regulator, it still has a traditional spring and diaphragm to respond initially to changes in demand. The pump pwm change happens next. It’s a good design for sure, and about as good as I can imagine you could do short of a feed-forward system.

 

[engineermike]

@engineermike Do I remember correctly, that you had a hard time controlling a brushless pump from the OEM PWM signal, or did you get that worked out? If you did, would it still be able to effectively control a return style system, or do you think the return system is unnecessary and another product of tuners not wanting to do the additional work?

I was contemplating switching to a brushless dw440 but later concluded it wasn’t a big enough improvement to be worth the headache. If I go that route I’ll go with the e5lm pump instead.

I do think these double and triple pump full-on systems are the easy-button for engine tuners. They get to sell the parts and it makes the tuning easier. It’s a win-win for them, similar to selling e85 tunes.

 

[Angrey]

We understand it’s measuring pressure on the return line, but the return line pressure upstream of the regulator is virtually the same as the rail pressure. If fuel flow through the injectors suddenly increases, then in order to maintain the rail pressure the return regulator must initially close in order to restore rail pressure as quickly as possible. If you look at the cross section of that fuelab regulator, it still has a traditional spring and diaphragm to respond initially to changes in demand. The pump pwm change happens next. It’s a good design for sure, and about as good as I can imagine you could do short of a feed-forward system.

No it's not. The regulator spring will maintain and hold pressure on the rail side up to nearly no flow out the return. It behaves similar to many pressure relief valves.

I've observed the pressure gauge (on the regulator) and the logs and the whole thing works as advertised. I've listened to the pump duty, although I haven't gone as far as to use the wire output to log the pump duty (or controller signal).

As soon as the little actuator on the return side reservoir moves it changes the flow, the rail side never sees or realizes what is going on.

The one complexity is that you have to get the sizing of the return size correct (or the total head loss) because if it's too free flowing, the pumps will find a "base" that's too high and if it's too free flowing the regulator will start to go herky jerky with pump stall and surge.

The regulator is AFTER the rails, so the feed side never loses pressure (when it's operating correctly within proper setup).

 

[Angrey]

I was contemplating switching to a brushless dw440 but later concluded it wasn’t a big enough improvement to be worth the headache. If I go that route I’ll go with the e5lm pump instead.

I do think these double and triple pump full-on systems are the easy-button for engine tuners. They get to sell the parts and it makes the tuning easier. It’s a win-win for them, similar to selling e85 tunes.

The "Bugatti" pump is very long and it's a bit troublesome to fit in tank. Fore (himself) told me it couldn't be done. The Radium hat that is able to accommodate it didn't come out until after I had bought most of the components for my system.

I would have used Fuelab pumps (and avoided the stress of whether the regulator would work with the DW controllers) but Fuelab at the time didn't have in tank option. They (and Radium) made/make surge tank options to fit their bigger pumps, but I didn't want to go that route.

The 440's were the only pumps I could find (at the time) to fit in an in tank hat. Now everyone is on the train. Fuelab had developed a system for the S197 but at the time they were resistant to a similar system for the S550 because they were just too early and spent money on the R/D and didn't really sell a lot of them (even though they were an in tank 1200 rwhp variable system). They were figuratively 10 years too early.

The improvement is significant when you consider DW rates their brushless at 18V continuous and 22V intermittent (and with a brushless I believe them). So with a booster, I have pretty much a 1600 hp capable system out of 2 brushless pumps (in tank).

Now aeromotive has fielded theirs (after telling me they weren't interested) and Injector Dynamics is now offering their controller setup to enthusiasts for a more reasonable price. None of that was available when I started down this path.

The automotive (CI) industry is 20 years behind the small tools industry that went to brushless motors decades ago. Now only cheap chinesium drills and cordless tools have brush style motors.

 

[engineermike]

@Angrey ask yourself this question….”how does the electronic regulator know it needs to respond?” The only link between the regulator and the rail is the pressure. The device is 100% feedback and 0% feed forward. Without a reduction in pressure, there is nothing to respond to. It may respond fast and it may have a nice secondary function to control the fuel pump, but it still has to respond to a drop in pressure first. Look at the cross-section of the regulator. It still has a traditional spring and diaphragm which responds to pressure changes in the return line which are caused by pressure changes in the rail.

 

[Angrey]

@Angrey ask yourself this question….”how does the electronic regulator know it needs to respond?” The only link between the regulator and the rail is the pressure. The device is 100% feedback and 0% feed forward. Without a reduction in pressure, there is nothing to respond to. It may respond fast and it may have a nice secondary function to control the fuel pump, but it still has to respond to a drop in pressure first. Look at the cross-section of the regulator. It still has a traditional spring and diaphragm which responds to pressure changes in the return line which are caused by pressure changes in the rail.

Again, you haven't read or studied it. It has TWO reservoirs. Look at the diagrams in the patent link he posted. On the regulator side of the spring diaphram is the rail pressure. It maintains that rail pressure by releasing excess flow into the return side. On the return side it has a reservoir with a another actuator, one that is on a simple rocker. When the return side reservoir increases in pressure, it changes the voltage and the controller corresponds by telling the pumps to slow down. When the return side chamber drops in pressure, the swing arm moves, telling the controller that the regulator is heading toward a deficiency. As a result, the controller tells the pumps to pick up the pace. All the while, the rail side of the regulator stays constant. The only situation where it could run into trouble is if the pressure in the return side chamber drops so rapidly, that the regulator can not maintain pressure fully closed.

Think of it this way. As the rail side fills with fluid and pressure. The diaphram spring remains shut, all the way up to the desired/set point/pressure. Once it reaches that pressure, it relieves some flow to maintain that pressure. The resulting flow has it's OWN flow and pressure created in a secondary reservoir/chamber. THAT is where the electronic regulator gets it's baseline. If the baseline return flow slows, it tells the pumps to send more fuel. If the baseline pressure INCREASES (i.e. there's too much return flow) it slows them down.

If I want to have more reserve flow, I simply adjust the fitting on the return side to open it up more. It then needs MORE flow to create the back pressure needed on the return side chamber. The result is, the pumps operate at a higher baseline speed. But that's more baseline current draw and less efficient.

I can't explain it any clearer than that. My rail side pressure doesn't change (relative to boost index) and all the while, full load or idle, the pumps vary, all based upon the return line flow and pressure.

This will be my last attempt to explain it. You should really read and understand it. Because I think you're continuing this whole conversation based off a complete misunderstanding of how the regulator actually works and controls the pumps. Furthermore, if there were fluctuations in the rails side, I'd have been able to both observe it visually on the external gauge AND see it in the logs. The regulator works perfectly as advertised (once I sorted out the trigger source, but that's a whole other discussion). Whether I'm idling at a stop light or mashing on it, it gives me 55 psi + boost constantly, while the pumps go from a walk to a run to a sprint and back to a walk as necessary to keep up.

 

[engineermike]

It maintains that rail pressure by releasing excess flow into the return side.

As the rail side fills with fluid and pressure. The diaphram spring remains shut, all the way up to the desired/set point/pressure. Once it reaches that pressure, it relieves some flow to maintain that pressure.

Forget about the electronic part for a minute and focus on the above statements. The diaphragm and spring mechanically respond to setpoint error. This is the mechanical version of a pid feedback control, except without the d. Without error, the diaphragm does not move. Therefore there is error, specifically during dynamic conditions.

Conversely, a feedforward system attempts to correct error before error exists by modeling the future state and commanding what it knows will be needed before it’s needed.

 

[Angrey]

Forget about the electronic part for a minute and focus on the above statements. The diaphragm and spring mechanically respond to setpoint error. This is the mechanical version of a pid feedback control, except without the d. Without error, the diaphragm does not move. Therefore there is error, specifically during dynamic conditions.

Conversely, a feedforward system attempts to correct error before error exists by modeling the future state and commanding what it knows will be needed before it’s needed.

IT IS A SEPARATE MECHANISM FROM THE REGULATOR DIAPHRAM. IT IS AFTER THE REGULATOR SPRING/DIAPHRAM.

I'm not sure which part you're misunderstanding.

The regulator (just like any regulator) has a spring with a set screw. It can make 55 psi with a very small flow out the return or it can make 55 psi with a very large flow out the return.

When the pumps are at full operation, obviously it's going to result in high flow in the return line.

There is a point at which the spring constant/extension suffers ramp and so it isn't a binary on/off, it has to have a trickle, but it can still maintain 55 psi down to small flow levels.

Now imagine for a moment on the RETURN SIDE of the diaphram, there's another small reservoir with a pivoting arm to an electronic sensor. When the pressure in that RETURN SIDE reservoir increases, it moves the arm. When the pressure in that RETURN SIDE RESERVOIR (again BOTH ON THE RETURN SIDE OF THE DIAPHRAM) lessens, it moves it in the opposite direction. That arm LITERALLY HAS A SEPARATE SPRING ON IT. IT IS NOT, IT IS NOT, IT IS NOT THE REGULATOR DIAPHRAM SPRING.

Jeezus, if you'd look at the schematics, we wouldn't be spending 3 pages of tennis back and forth. The physical diaphram HAS NOTHING TO DO WITH THE ELECTRONIC MANAGEMENT OF THE SYSTEM. WHETHER YOU SET IT AT 100 PSI OR 40 PSI IT HAS NO BEARING ON THE CONTROLLER. THE DIAPHRAM IN THE REGULATOR OPERATES JUST LIKE EVERY OTHER PRESSURE REGULATOR.

As you would imagine in such a system, the sizing of the return line is crucial to getting the controller to baseline in the desired range. If the return line size flows too well, it doesn't provide enough pressure backup and the pumps then run full duty because the sensor thinks the motor is drinking up all the resources. CONVERSELY, if you were to put a coffee straw as the return line, the pumps will never get to full operation because the sensor thinks you're creating too much return flow and wasted effort.

For the 10th time. IT HAS NOTHING TO DO WITH THE REGULATOR SPRING AND DIAPHRAM. It IS A SEPARATE SENSOR AND SPRING MECHANISM ON THE RETURN SIDE OF THE REGULATOR. THEY ARE BOTH CONTAINED IN THE SAME HOUSING FOR SIMPLICITY AND ECONOMY.

 

[Grimreaper]

The second spring and ball/seat(component 169) is a bypass for throttle lift to avoid destroying their sensor, per their patent. Is this the swing arm you referenced? The only moving part is the main diaphragm otherwise (once setup).

The reservoir chamber allows the differential pressure to exist and fluctuate based on how the main diaphragm is behaving. The main diaphragm behaves as it does based entirely off of pressures. Feedback is now available to show how much more or how much less the pumps are over/under producing and going into bypass essentially. The reservoir chamber is dependent on the diaphragm/intake chamber. Not independent.

Sponsored