[MrDrunkPilot]

First of all s55 ecu is f chasis and has ptcan2 ptcan and flexray that doesn’t exist in e chasis and ecu won’t respond to old ptcan connection. Author of thread used somr kind of signal spliter or something similar, in my case is the only way to do something like he did, because mevd1726 on these pins which in f chasis controls second hpfp is used to control solenoid for wg.

Originally Posted by n55david

The original guy referenced a 135 which isn't even what we have here. Different chassis and different dme

Get an s55 dme too and make it simple

[thejeremyman9]

Originally Posted by MrDrunkPilot

So what? I just need to know what he used to run it even with som problems.

I just posted to warn you. Most new users do not read the date on threads. Hence, the chances of you getting a reply are much lower. Good luck

[n55david]

Originally Posted by MrDrunkPilot

First of all s55 ecu is f chasis and has ptcan2 ptcan and flexray that doesn’t exist in e chasis and ecu won’t respond to old ptcan connection. Author of thread used somr kind of signal spliter or something similar, in my case is the only way to do something like he did, because mevd1726 on these pins which in f chasis controls second hpfp is used to control solenoid for wg.

Yeah.. good luck buddy 😀

You may wanna try the s55 forums lol

[AmuroRay]

Originally Posted by MrDrunkPilot

First of all s55 ecu is f chasis and has ptcan2 ptcan and flexray that doesn’t exist in e chasis and ecu won’t respond to old ptcan connection. Author of thread used somr kind of signal spliter or something similar, in my case is the only way to do something like he did, because mevd1726 on these pins which in f chasis controls second hpfp is used to control solenoid for wg.

Sean is still active on the M2 side, you may be able to message him. I would also try to reach out to a tuner and see if they can assist you. Finally, there are other more technical forums than this one..

[lens]

Originally Posted by thejeremyman9

I just posted to warn you. Most new users do not read the date on threads. Hence, the chances of you getting a reply are much lower. Good luck

I remember calling out XDI on this one, stating that price and the work required to get this going would be absurd and would make zero sense when people were developing much better single HPFP upgrades that are headache free.

but yeah, I think Sean runs something else now (or new vehicle) so I think he won't remember everything.

[rlesko]

Necro'ing this thread because I was curious about the S55 dual hpfp retrofit (moreso just in theory than in actually doing it). Used part prices are at a point where this is worth considering over other off the shelf options.

MrDrunkPilot did you ever figure out your issue?

I don't believe "splitting the signal" will work. I read in the S55 technical documentation that the second pump is only activated above 3000 rpm. So it makes sense why the OP said the car wasn't running correctly. Its probably running way too rich at low RPM.

[Boostedphil698]

Originally Posted by rlesko

Necro'ing this thread because I was curious about the S55 dual hpfp retrofit (moreso just in theory than in actually doing it). Used part prices are at a point where this is worth considering over other off the shelf options.

MrDrunkPilot did you ever figure out your issue?

I don't believe "splitting the signal" will work. I read in the S55 technical documentation that the second pump is only activated above 3000 rpm. So it makes sense why the OP said the car wasn't running correctly. Its probably running way too rich at low RPM.

With how far DI upgrades have come, especially from Dorch, this retrofit just isn't worth the headache.

[rlesko]

Originally Posted by Boostedphil698

With how far DI upgrades have come, especially from Dorch, this retrofit just isn't worth the headache.

The challenge (what you call "headache") of making it work IS the fun part...

The easiest way of getting more fuel is modifying a B58 TU pump. Then next step just buy a dorch. But neither is as cool or satisfying as retrofitting genuine BMW parts

[wheela]

Originally Posted by rlesko

Necro'ing this thread because I was curious about the S55 dual hpfp retrofit (moreso just in theory than in actually doing it). Used part prices are at a point where this is worth considering over other off the shelf options.

MrDrunkPilot did you ever figure out your issue?

I don't believe "splitting the signal" will work. I read in the S55 technical documentation that the second pump is only activated above 3000 rpm. So it makes sense why the OP said the car wasn't running correctly. Its probably running way too rich at low RPM.

Per a conversation with AI, the n55 and dual s55 pumps are single piston, so there is pressure ripple in the output, as fuel is only getting sent to the rail on the output stroke. On s55, I believe the two hpfp's are physically 180 out of phase with eachother. That way, when the second pump kicks on, it's pump stroke is outputting fuel to the rail while the other pump is on its intake stroke. This reduces the magnitude of the pressure ripple effect on the rail, but doubles it's frequency.

The s55 DME is programmed to account for the different phase of the second pump, and the change in pressure ripple when the 2nd pump kicks on. Neither of these are accomplished by splitting the n55's single pump signal to the second pump, which explains the fueling issue when simply splitting the signal.

[LA1Z24]

There is a full cam lobe for each compression stroke. There is no such thing as turning on a pump out of phase. The regulator is actuated based on how far up the cam profile the piston has traveled. The PID is called something along the lines of DEGREES OF LOBE USE. It can not be out of phase with the engine or it would not work. As each cylinder arrives at TDC, a cam lobe is also passing the HPFP at TDC of the cam lobe.

[wheela]

Originally Posted by LA1Z24

There is a full cam lobe for each compression stroke. There is no such thing as turning on a pump out of phase. The regulator is actuated based on how far up the cam profile the piston has traveled. The PID is called something along the lines of DEGREES OF LOBE USE. It can not be out of phase with the engine or it would not work. As each cylinder arrives at TDC, a cam lobe is also passing the HPFP at TDC of the cam lobe.

I think you're misunderstanding what I'm saying. Yes, the regulator works by controlling how far up the piston is on its stroke when closing off the inlet (Volume Control Valve (VCV)) so the piston forces the fuel through the check valve to the fuel rail. And that's how both pumps are controlled on an s55 dual hpfp. When I say the two pumps are operating out-of-phase on s55, I'm referring the the mechanical phasing of the piston strokes. In-phase would be if the inlet and outlet strokes of both pumps were synchronized so they were happening at the same time (both pumps inletting, and outletting fuel to the rail at the same time). 180 degrees out-of-phase means that while 1 pump is on its inlet stroke (and not sending fuel to the rail), the other pump is on its outlet stroke (and is sending fuel to the rail).

To clarify, the pistons of both pumps are physically cyling at all times. By "kicking on" the second pump - this refers to control of the second pump. The DME keeps the VCV of the second pump fully open until the point the DME decides to "kick on" the second pump. At that point, it will begin controlling the VCV to close and force fuel through the check valve to the rail, closing the VCV based on the position of the piston on its compression stroke as you described.

[rlesko]

Originally Posted by LA1Z24

There is a full cam lobe for each compression stroke. There is no such thing as turning on a pump out of phase. The regulator is actuated based on how far up the cam profile the piston has traveled. The PID is called something along the lines of DEGREES OF LOBE USE. It can not be out of phase with the engine or it would not work. As each cylinder arrives at TDC, a cam lobe is also passing the HPFP at TDC of the cam lobe.

Lets talk this out so I can understand better. Here is my understanding of the system.

The HPFP works off the vacuum pump camshaft, which is driven by the engine but is not timed with the engine.

The camshaft follower in the HPFP rides a cam lobe in the vacuum pump. This is a pure mechanical part of the system. What varies is the solenoid which allows fuel from the low pressure side into the HPFP, which is controlled by the DME (possibly by PWM?). The solenoid opens variably, which controls how much fuel goes into the chamber, ultimately determining how much fuel goes into the rail. This is measured by the rail pressure only.

Adding a second HPFP into the mix, could the cam lobe that drives this HPFP be 180 degrees out of phase with the first, to deliver steadier rail pressure? It seems reasonable to me, but I am not a mechanical engineer or tuner

[wheela]

Originally Posted by rlesko

Lets talk this out so I can understand better. Here is my understanding of the system.

The HPFP works off the vacuum pump camshaft, which is driven by the engine but is not timed with the engine.

The camshaft follower in the HPFP rides a cam lobe in the vacuum pump. This is a pure mechanical part of the system. What varies is the solenoid which allows fuel from the low pressure side into the HPFP, which is controlled by the DME (possibly by PWM?). The solenoid opens variably, which controls how much fuel goes into the chamber, ultimately determining how much fuel goes into the rail. This is measured by the rail pressure only.

Adding a second HPFP into the mix, could the cam lobe that drives this HPFP be 180 degrees out of phase with the first, to deliver steadier rail pressure? It seems reasonable to me, but I am not a mechanical engineer or tuner

The hpfp is run off the vacuum pump, which is driven by the engine. So the speed of hpfp piston strokes does follow a relationship with with engine speed.

Since the pumps aren't supplying output on their inlet stroke, pressure is decreasing in the rail during injection cycles on the hpfp's inlet stroke. So those injections actually need longer open time to inject the same amount of fuel due to the dropping pressure. The pressure spikes (ripples) from the hpfp outlet strokes are happening too fast to see on logs, but AI says if you were to put an oscilloscope on the injector signal you'd be able to see the variation. The DME monitors rail pressure at a much higher frequency than our logging software, and can supposedly see these little pressure drops so it can compensate injection time real-time per each injection event based on actual rail pressure at the time of injection.

AI also said that since the hpfp piston strokes have a fixed relation to engine speed, thw DME actually knows when the pressure spikes from piston strokes will (should) happen for each injection event, and uses some feed forward features to anticipate how the injection time will need to vary based on where the hpfp piston is on its stroke. So the whole injector control methodology is a lot more complex than it appears based on log data.

So yes, by operating the second pump out-of-phase from the first pump, it reduces the magnitude of the rail pressure ripple from the hpfp outlet strokes. And it also doubles the frequency of the rail pressure ripple, as the rate of hpfp outlet strokes doubles once the DME signals the VCV of the second pump to start sending fuel to the rail.

[rlesko]

Originally Posted by wheela

The hpfp is run off the vacuum pump, which is driven by the engine. So the speed of hpfp piston strokes does follow a relationship with with engine speed. Since the pumps aren't supplying output on their inley stroke, pressure is decreasing in the rail during injection cycles on the hpfp's inlet stroke. So those injections actually need longer open time to inject the same amount of fuel due to the dropping pressure. The pressure spikes (ripples) from the hpfp outlet strokes are happening too fast to see on logs, but AI says if you were to put an oscilloscope on the injector signal you'd be able to see the variation. The DME monitors rail pressure at a much higher frequency than our logging software, and can supposedly see these little pressure drops so it can compensate injection time real-time per each injection event based on actual rail pressure at the time of injection
AI also said that since the hpfp piston strokes have a fixed relation to engine speed, it actually knows where the pressure spikes from piston strokes will (should) happen for each injection event, and uses some feed forward features to anticipate how the injection time will need to vary based on where the hpfp piston is on its stroke. So the whole injector control methodology is a lot more complex than it appears based on log data.

Yeah, engine speed determines the speed of the vacuum pump camshaft, but since its not timed with the engine how the DME knows when to open and close the hpfp control valves remains a bit of a question mark. On start up I guess it can sample rail pressure and figure it out.

[wheela]

Originally Posted by rlesko

Yeah, engine speed determines the speed of the vacuum pump camshaft, but since its not timed with the engine how the DME knows when to open and close the hpfp control valves remains a bit of a question mark. On start up I guess it can sample rail pressure and figure it out.

Good question, I don't know for sure, I didn't ask about that. But if the DME can see the rail pressure ripple like AI said, then I think your guess is very plausible that it can figure out the Offset on startup.

[LA1Z24]

The vacuum pump and HPFP are timed. They have a cam position sensor on the side of the vacuum pump housing. We have had threads on this forum this year with people that had mis-timed HPFPs

[rlesko]

Originally Posted by LA1Z24

The vacuum pump and HPFP are timed. They have a cam position sensor on the side of the vacuum pump housing. We have had threads on this forum this year with people that had mis-timed HPFPs

I have not been able to find evidence that these components are timed. In fact, quite the contrary. People install overdrive kits (smaller sprocket on the vacuum pump) which would not work if there was a timing restriction. Can you share why you think they are or the other thread in question?

[wheela]

Originally Posted by rlesko

I have not been able to find evidence that these components are timed. In fact, quite the contrary. People install overdrive kits (smaller sprocket on the vacuum pump) which would not work if there was a timing restriction. Can you share why you think they are or the other thread in question?

Those over drive kits are for the older n54 style hpfp. Instead of a single piston riding on a lobe like the newer n55 hpfp, the older n54 style works by rotating a wobble plate that sequentially cycles 3 pistons that ride on the wobble plate. Since there are 3 pistons pumping, there is always at least 1 piston producing flow while others are on their intake stroke, vs. newer n55 single piston style which doesn't output fuel while the piston is on its intake stroke. So the rotary n54 style produces much less pressure pulsating in the output, to the extent timing with the engine isn't required for sufficient control. On n54 style, the DME controls pressure by bleeding the output back to the inlet, vs. n55 single piston the DME needs to know where/when the piston's output stroke is happening so it can close the VCV valve at the right time to send the right amount of fuel to the rail to hit the pressure target.

Note: Some early e-chassis n55's came with the older n54 style hpfp. So you may may see n55 referenced regarding those over-drive kits. But it's just older n55's with n54 style hpfp - not the newer n55's with the single piston n55 style hpfp.

Per AI: