Kingtal0n

I guess for me, 100hp isn't "alot" of horsepower, it isn't worth what you are asking in this venue IMO. It depends on the cost of the engine, and ease of engine swap. "I'm unable to replace the pump in question for some personal reason, so lets boost the input amplitude at risk of burning or confusing electronics and technical instrumentation for very little power increase".

really if there are people willing to risk using a higher than expect voltage, then this was pretty much covered:
I can buy the 18v input could be reliable if it was only short bursts and not constant. But, the controller is the issue,
And I will love to elaborate on that. When i said:
In this case 'cpu' is the controller and the issue is communication in closed loop with the driver. The controller is predicting engine demand so the fuel is available to the DI mech. pump at the exact moment it needs it, rather than met by the delay caused by slower control methods (apparently return regulation is very slow). For the controller's code, there are two issues, transients and WOT. It is a transient issue at low power outputs when the pump capacity is more than adequate. If the person driving puts the pedal down suddenly the controller is both monitoring engine rpm and fuel pressure, and communication to the pump-driver with an expected response in terms of fuel pressure. In other words, the controller doesn't care about the pump-driver or fuel pump. All it cares about is the fuel pressure. If the input magnitude suddenly changes without being expected (the input voltage in this case causes more torque in the fuel pump which causes an unexpected rise in fuel pressure) then the controller will have over-estimated the required amount of driver excitation to 'hold' the fuel pressure at some exact number especially during such a sudden, high demand(the whole point of the controller is to do that, put the right amount of fuel instantly in the line).
Now the controller must learn and remember that overshoot has occurred and make accommodations. It might take a couple runs or something. For it is actually 'un-doing' the higher voltage's effects on fuel pressure for situations below max output, lowering controller output once it realizes that the same outputs give more fuel pressure signal it has to reduce its output some for the new voltage. That in this case is giving it more headroom to fuel the engine, as the controller knows it can only command 99.9% or some set number maximum it has to work within those bounds. I doubt the controller is ever programmed to 'just give 100%' like in a traditional sense we run our return pumps 100% and what you get is what you get. In other words, I don't think the PWM factory in-tank pump can ever just be 'let loose' it would never work with a 100% output in DI because max flow demand only occurs for one specific instant at the very height of engine power output or energy use. Because the DI mechanical pump and factory computer are expecting the low side fuel pressure to remain at some number (say, 60psi) this can only be accomplished through the use of less than 100% duty cycle at the pump, via the fuel controller which only watches fuel pressure. And since that controller is only able to operate from 0-99% output, that range is pre-set by the factory based on expected fuel pressure behavior.
For example if somebody bought the car and never went over 50% throttle, ever. Even still, The fuel computer will still 'know in advance' how much fuel pump PWM duty to give the pump for the first time it went WOT, in order to achieve a desired low side fuel pressure result. The same way the ECU knows how much fuel injector pulse to give the first time an engine goes WOT.

But unlike WOT injector duty, the WOT fuel pump duty is not open loop control. Feedback from fuel pressure sensor tells controller 'how well it did' during a first run taking place, and it seeks to modify it's output in order to achieve a desired result in fuel pressure behavior. If it suddenly starts 'doing very poorly' due to a spike up or down in voltage, then the first run at the new voltage might have given it more fuel pressure or less, than the second run will due to the closed loop control having a memory on the second run. Finally third, fourth etc... In the end, if closed loop is able, it will return to the factory specific fuel pressure result for all power demands, there will be no difference in fuel pump output at 14v or 18v or 22v because closed loop corrected for it at low power outputs (it won't be rich anymore if it started out rich at low outputs). If the controller consistently depends on the higher voltage it may raise headroom for commanding fuel duty so there is that one single possible benefit allowing more power to be made from the same fuel pump than before. But there is still the issue of unwanted, excess stress. For example if the car is running out of pump at 500rwhp. It is fine up to 500. Technically if we could run the pump to 500rwhp and then at that exact moment switch the voltage to 18v or 21v to simply squeeze the last bit of flow from the pump at that instance, would be the ideal way to minimize pump exposure to higher voltage. Again however, the issue is the controller. The controller is expecting and watching fuel pressure, and has no idea you are going to spike the voltage at some instant, and correcting for it might not happen properly (unstable disasters, limp mode etc). So the point at which voltage spikes is a critical entity of this operation. If the voltage were spiking "properly" at the exact moment it was needed (501+rwhp) then the damage is minimal, but the controller isn't able to deal with that situation. So instead, it sounds like (correct me if I am wrong) voltage booster devices are using WOT throttle position instead, to spike the voltage, and the pump is enduring the voltage spike the entire time instead of for just the time it is really wanted/needed(501+), and this is done to allow the controller's closed loop to 'remove output duty cycle to the pump through the driver' thereby possibly creating a little headroom at the very top where it was just short before 50-100hp. They are assuming that the controller is able to move the entire map (it would be like a maf sensor being able to change it's frequency->fuel injector duty correlation) which correlates output to fuel pressure, and that the closed loop memory is going to raise the headroom for the fuel pump output by lowering its input command to the driver whenever the voltage is spiked. In any case, there are sensitive electronics already stressed to the max that voltage boosters are trying to squeeze even harder for small gains. Such an elaborate, self correcting system that some of us would love to have, being put at risk for a couple hps

atljar

I've stayed out of this recently because this is just a circular argument with vast amounts of bad information from all directions.

But with that said, I'd love to see what evidence you have to support your theory that the fuel pump control module is a learning system.

All the GM documentation I have read makes no such mention. There is a request for a specific fuel pressure and the pump is run at increasing duty cycle until target pressure is met. If pressure is too high, duty cycle scales back. With a booster you just will run less duty cycle.

Yes you can induce pressure spikes, which is why you want a system that has a ramp in algorithm. This is also true with the low systems, some of them have a ramp in rate to help alleviate those spikes.

Proof of concept is now out there. This works for additional low side help. If you're on the edge, it's a good idea. If you have bigger plans like e85 or significant power increases, then not your ticket

HessViper

The good news is, all we have to do is throw one of these on a car and data log the fuel pressure.

If it does any damage, it can be removed, and the warranty will cover the pump, as it can be tested on the stock tune first, and then bolt ons added like pulleys and ported blowers.

In the end, all this typing becomes pointless, throw it on a car and lets go. And it seems like some have already, without pump failure (yet)

Rflow, how many miles have you put on your car with the booster in use?

Warp Factor

KingtalOn, I think you are imagining this as being way more complicated than it needs to be, rather than realizing the advantages of the simplicity.

Kingtal0n

The definition of a controller in this context implies closed loop. Otherwise how could it maintain a target fuel pressure during a wide variety of changing conditions.

You said it yourself, it 'meets the requirement'. Thus, it makes a change, checks to see how good it did, then makes another change based on the past information 'learning' as it goes, and continues doing this forever.

If the voltage increases or decreases suddenly, that is the same thing as any other sudden disturbance, it will 'learn' about the new conditions. The idea of closed loop isn't just for higher or lower voltages; it also accommodates for pump wear and tear like aging, fuel system changes to quality such as clogging fuel filters or different specific gravity fuels, or even a small enough leak somewhere could be made up for.

Kingtal0n

As to the 'complexity', on one hand, it isn't very complex. The control system in place here is a very common theme and simple compared to most higher machines. Autos just get the tiny little bit of control for the most basic applications, this is perhaps one of the most simple systems. That said, it is far more complex than most people realize, a staggering amount of information, processing, circuitry, communication, timing, shapes of waves, its really not that simple.

atljar

Closed loop feed back yes. Learning no. Its like a car with only 1 cell in the fuel trims, and short term only, no LTFT.

Kingtal0n

This is good, HessViper has good things to say, his original post about testing and this one is good routine/manners: 'test it properly', get the necessary data, and that is about all we can do here.
The only issue I see is interpretation of the data. If this device isn't tested properly it could easily lead to false results. For example if the system is installed and tested right away, it will yield incorrect results because the controller maybe hasn't had enough time to learn yet. I think testing should be done immediately, and then again after a couple runs, and then again even more time after that, and finally the device should be removed to show what the controller has 'learned' by taking away the voltage we should all see a drop in fuel pressure on first couple of runs which would help us see how 'fast' the controller is able to compensate and remember. It might remember for 5 seconds or 5 days.

Kingtal0n

neither one of us knows if the fuel controller has a memory or not. Lets assume the controller does not have any "LTFT" style of memory as you suggest.
This just means the controller is intended to be 'fast'. It still runs a closed loop for target fuel pressure so it will still dial back the pressure due to any unexpected voltage increase.
You wind up back where you started at the original voltage in seconds, rather than over the course of days.

i dont mind adding:
Systems that use more past data are likely to stay stable as long as all that data stays relevant. Really what would be ideal is a controller that correctly predicts the voltage spike (past information to be true) and corrects for it the moment before it happens. Like tagging the IAC position before engaging an A/C compressor, predict the 'load'. send 18v or 20v to the fuel pump at the exact moment you needed the extra flow (for that 0.2 seconds at the end of a run). Instead of trying to teach a closed loop system for a second in advance and spiking the pump the entire time with unnecessary 'insane' overclocking values. The longer it is exposed to 18v the more likely it will fry. I think voltage boosting may have an application here but you want a controller to do it properly, replace the fuel computer (yeah right). And even then there is still risk, etc... and nobody would go through all of that for 100hp, instead just tax the oem computer and let it learn in advance for a second or two, and hopefully it figures out at that instant in those seconds that it's got some additional headroom right before it hits the old max flow rate. As long as you are planning/willing to change the pump anyways. wouldn't use this device without having a spare a pump to change in if I got stranded and a way to change it quickly 30min~ unless a warranty fix is the selling point on this.

It looks tempting to do anything to a fuel pump, some cars have annoyingly complex fuel systems though, even just inside the tank can be a pain to deal with.

Warp Factor

Suppose that a voltage booster did cause a momentary spike in fuel pressure, when it first kicked in at a certain throttle position or boost pressure, before the pulse width was reduced to compensate? So what?

Suppose that it did reduce pump life, from the voltage increase when the car was operating over 7 pounds boost? So what? What percentage of their running time do most people spend over 7 pounds boost?

atljar

I know Ive read all GM manuals on the system, and there is no reference to a learning system. Its a correcting system, but not learning.

Brilliant isnt it? Now you get the simplicity of the system. The ECM still controls target fuel pressure. If the voltage booster is "on", and the fuel pressure is over the request amount, then the pulse width is simply reduced to less duty cycle to keep pump pressure in line with request

No, we wind back up to the original requested pressure, instantly, but at less duty cycle at the pump.

Kingtal0n

it is not instant. The controller has to learn about the new pressure. This will take some time. It could take seconds or milliseconds, or it could be done completely wrong and throw the system into an unstable oscillation at that exact moment due the way it was programming. It might do whatever it wants, activating the pump/driver when they are not supposed to or causing some kind of chatter, etc... When the controller is exposed to new sudden changes out of the blue like that, all bets are off unless it was a pre-designated programming function or somehow managed to fall within the category of stable operation by thoughtful programming. The PWM pulse is a pump signal input, is reduced by the controller through the driver once the controller senses the new fuel pressure resulting from the current information. It then must use that as past information to send a proper new signal to the driver which then reduces that PWM pump signal in a sloping, changing manner to drive to desired pressure without moving too slow or fast. And then the controller checks again as fast as it can but also must account for the rate of change of pressure, because the controller can sample the new pressure before the pump has even responded to the new PWM signal from the driver. So output values to the driver are based on predictions AND past information, and past information needs to be generated first which includes the higher voltage's effects, and it can't over or under predict responses to inputs. So there is definitely going to be a nice delay, mayb 700ms 1.2 seconds, it would surprise me less than 500ms but I could see getting it down past that if you knew in advance and programmed it in advance for this situation. What I keep bringing up is that the magnitude of the output changes when you turn up voltage, and magnitude there is expressed in dB, so just 3dB is a LOT of amplification and the controller could easily over or under predict and make all manner of terrible mess to the pumps behavior.

Warp Factor

"The sky is falling".

HessViper

rflow306 , how many miles have you put on the vehicle with the voltage booster added?

NicD

This is like a dream come true, kingdingaling from ls1tech against the mike/hess/whoever... I'm just watching the massive **** show of disinformation but figured I would throw out there that people have been using BAPs since 2009 in a PWM controlled environment on the GMs and the Ford guys have been doing it far longer. Oh yeah and if anybody saw ANY change in your final A/F by running a BAP then you were maxing out injectors before the BAP and your high side was dropping since these cars adjust pulse width based off of the injector pressure vs manifold pressure on the high side not low side. The FPCM certainly has trims that it remembers and uses for it's pre-programmed open loop DC table but it's irrelevant since it sees pressure changes damn near immediately and compensates accordingly.

atljar

I welcome some knowledgeable insight. Is that a certainty that PWM of low side is controlled by rail pressure vs map? I know most PWM port setups are done this way, but then I would ask what role does the low pressure sensor have? My understanding was rail pressure - map = injector tip pressure was used to simplify the system and eliminate the need for a low side sensor on port setups.

As for the FPCM learning? What's your source ? No mention in any service manual. I just went through all applicable modules with a GDS2 and there is no correction factor I can find anywhere, only a PWM command and low Side pressure reading

RoxyCarter

Rail pressure at the engine is dictated mostly by the high pressure pump.

the low side pump just needs to give the high side pump some fuel to work with and "grab" from.

computers and pressure sensors work in real.time to maintain a specific low side fuel line pressure. It's a very simple process. Learning is not required to monitor and maintain low side fuel line pressure, just a simple, real time feedback loop.

might there be learning on the high side? Perhaps. But it is irrelevant to this discussion when you actually stop and think about it.

the discussion is simple... We need to give the high side pump more fuel since there are times the stock low side pump with stock voltage are failing to provide that fuel in certain WOT situations and with certain air flowing upgrades to the engine, or in extreme cold and dense o2... so We need to have the low side pump move larger volume of fuel in certain circumstances. Higher voltage to the low side pump, accomplishes this.

anytime a pump has to pump harder, it's lifespan is decreased. The difference, again, is when the stock low side pump fails, it could easily be covered under warranty, for free... thus the voltage booster is the most economic in the short term, as well as long term.

the post above is classic nicD. He offers no explanations where our team was wrong, but insults us and all others in the thread.

while I do not agree with everyone's point of view in this thread, I am glad it is being discussed so that misconceptions or false priorities can be squashed.

that's the point of a forum... create posts that teach others.

Nick walks in with insults for all of us... and rest assured, if anyone insults him back, he will be complaining behind the scenes to moderators.

there are certain people on this forum who are poison. They may know a lot about vettes... but on this forum, they are poison.

he is in second place behind the other fellow who was already banned this past month. Their strategy is one in the same. Provoke others, and complain.

and the sad part is, nicd refused to believe that our team could get a stock vehicle to religiously go into limp mode in the cold, after all, he is an expert on this since he is living the hot deserts of the southwest, over a thousand feet above sea level, and he told us we were wrong about the cold weather/ sea level problem (he is an expert in all things And at all times)... (Yet, we have all the service tickets on multiple vehicles where Mike and others have shown they can do it, there will be a YouTube channel about it soon I've heard, where nicd and some other's posts may be highlighted for the comedy portion of those episodes)

and now nicd is saying that our team is wrong about this voltage booster idea and calling us names again.

we say it's a good idea and may need bugs ironed out perhaps.

if he wants to take the opposite stance as ours... then let the record show that nicd does not believe the voltage booster will provide additional fueling for our vehicles.

and if he wants to state otherwise, and further clarify his position, then so be it, but that would require an informative post instead of a short one where he insults us all, and then runs away, hoping we insult him back, so he can tell on us...

as he brings no useful knowledge to a discussion other than his opinion that "We are all wrong and have no idea"

which MUST mean he disagrees and has the opposite stance as we all do, otherwise insults are not what he should have typed out. Let me know where I'm missing the boat on forum etiquette 101

NicD

No, low side is simply controlled by the commanded flow tables so when the computer is looking at low flow conditions it commands a lower fuel pressure, in high flow situations it commands 72.5 psi and has a cap that lowers as flow rate increases.

The low pressure sensor is for the closed feedback loop to command higher duty cycles to the pump when it needs it and lower when it doesn't to maintain it's commanded fuel pressure since it has set tables for various flow conditions. Same thing for the high side control but directly references the rail pressure/map as it's control for final pulse delivery. Low side doesn't come into the equation for final pulsewidth, just needs to supply the high side with enough fuel to hit it's commanded rail pressures.

This is how they all function, same reason the main ECMs has trims as well. I believe it was in GDS2 where you can monitor the trims and watch how they respond, HPT doesn't have it in their scanner for the Gen5 stuff yet. It's the same exact controller logic for low side control as all of the Gen4 stuff that actually had a PWM controller for the fuel pump and it functions identically. Ford uses a lot of the same strategies with their fuel pump control modules as well.

atljar

Thanks for the follow up. I misunderstood what you were saying in your first reply when you said "adjust pulse width based off of the injector pressure vs manifold pressure on the high side not low side" I thought you were referencing pulse width of the low side pump, not of the injectors. Everything else said is more or less my understanding of the system as well
​​

Still can't find any trims in GDS2. Any memory of where you access them?

HessViper

exactly, never did we say that they are a new product. We realize that newer versions of the voltage booster may be required as the computer and circuitry on the low side pumps are changed, manufacturer to manufacturer, year to year, model to model.

The drawbacks to the voltage boosters used to be that it had a more direct affect on the injectors, and their flow rates.

Since the High side High pressure pump is now in the middle of the mix, the previous drawbacks that may have existed on conventional fuel injectors, no longer exist in this DI system on the c7's, when the volt booster is pushing the low side pump harder than in the stock configuration.

We appreciate the OP for bringing the spotlight back onto voltage boosters for the low side pump, not because we did not know what they were, but because we had never fully examined their potential on the DI systems until he created this thread.

We waited for people with real world experience with DI c7's to chime in, but from our point of view, we believed this would be a very useful product on the DI c7's as soon as I came across this thread.

And then, of course NicD insults us all in his first post. And then in his second post about us, he says that we believe this is a new product.

He sure does think highly of people's understanding, prior to putting words in our mouths, we have noticed