N54 Piezo Fuel Injector Limits

[Twisted Tuning]

Higher pressure to a point is great, but beyond a point Flow does not increase with pressure, this is known of course. I did a lot of Injection start/stop angle tuning some years ago on another platform. It works, but it all boils down to exactly what jake said. You need an HPFP setup capable of keeping up with the additional window time. Otherwise, you're going to chase tails and be right back to square one.

The Knock sensor table thing, me and Jake spoke offline about. And while i think its good to have for those who need it. It not good to have for people who are not going to use it properly. Kinda of like when the current knock table availability was released, and everyone and their momma was desensitizing their knock sensors. That is not how that should be done. I can't remember seeing one person use the proper tools to reduce the knock sensors. Thats something that needs to be done by an individual basis. But hey, Darwins Theory, natural selection will weed people out of their engines, lol. Doesnt mean it should't be released though.

 

[kayzrx82]

The injectors don't sense anything. That functionality isn't being used on N54.

There's 2 wires into each injector for actuating the piezo stack there's no signal wire of any kind.

It wouldn't need a second set of wires or a dedicated signal wire. The wires that are across the piezo stack would generate a signal when they aren't energized. The injectors as is are fully capable of this. Its the circuitry within the msd ECU that is not set up to do so. You would also need a decent DSP to analyze that signal and detect a knock event . This signal can be seen on a decent oscilloscope just by reading the 2 leads that come off the injector. You could also use the spark plugs to detect knock with an ion sensing circuit. The S65 and s85 do this.

 

[rac]

personally my injectors maxed out at around ~2600psi, I didn't notice any significant flow increase after that and I tested them all the way to 3600+psi. there may have been but it would not have been much. for those that are interested, relative to max flow I found;

~93% flow @ 2200psi
~83% flow @ 1800psi
~68% flow @ 1400psi

although I know of another example with different numbers, probably from injector index 11+.
mine are old.

 

[Bnks334]

personally my injectors maxed out at around ~2600psi, I didn't notice any significant flow increase after that and I tested them all the way to 3600+psi. there may have been but it would not have been much. for those that are interested, relative to max flow I found;

~93% flow @ 2200psi
~83% flow @ 1800psi
~68% flow @ 1400psi

although I know of another example with different numbers, probably from injector index 11+.
mine are old.

To make the data relevant, what conditions did you "test" in. I was under the impression that in-cylinder conditions like pressure also affect flow.

And as stated, more pressure isn't just about squeezing more flow out of the injectors during the same injection cycle... it's also about the spray pattern/atomization of the fuel and improving combustion.

My ecodiesel can see 29,000psi rail pressure. 15.5:1 compression ratio. Makes just as much torque as my bmw and returns better gas mileage.

 

[SJ_1989]

I was under the impression that in-cylinder conditions like pressure also affect flow.

Correct. A back pressure (i.e. cylinder pressure) downstream of an orifice (i.e. injector) for a supply pressure (fuel rail) has a certain pressure drop available to push flow through the circuit. An increase in back pressure for the same supply pressure nets a lower lower pressure drop, thus lower flow rate. Google orifice area equation. People usually only talk about flow or pressure but they are directly related.

 

[Kommodore]

Using it properly meaning, your tuner listening with det cans and adjusting to the materials of your built motor?

Higher pressure to a point is great, but beyond a point Flow does not increase with pressure, this is known of course. I did a lot of Injection start/stop angle tuning some years ago on another platform. It works, but it all boils down to exactly what jake said. You need an HPFP setup capable of keeping up with the additional window time. Otherwise, you're going to chase tails and be right back to square one.

The Knock sensor table thing, me and Jake spoke offline about. And while i think its good to have for those who need it. It not good to have for people who are not going to use it properly. Kinda of like when the current knock table availability was released, and everyone and their momma was desensitizing their knock sensors. That is not how that should be done. I can't remember seeing one person use the proper tools to reduce the knock sensors. Thats something that needs to be done by an individual basis. But hey, Darwins Theory, natural selection will weed people out of their engines, lol. Doesnt mean it should't be released though.

 

[JBacon335]

The more cylinder pressure the more pressure you need at the other end of the injector to keep the pressure delta the same, it makes perfect sense

 

[rac]

To make the data relevant, what conditions did you "test" in. I was under the impression that in-cylinder conditions like pressure also affect flow.

And as stated, more pressure isn't just about squeezing more flow out of the injectors during the same injection cycle... it's also about the spray pattern/atomization of the fuel and improving combustion.

My ecodiesel can see 29,000psi rail pressure. 15.5:1 compression ratio. Makes just as much torque as my bmw and returns better gas mileage.

well the converse of those numbers is what I had to use to increase injection open time to get the same afr with the different rail pressures so I could modify a correction table on a stand alone (trial and error, no bench test). yes its going to change pending how late I am injecting on the compression stroke and boost because of the increasing differential, but only by a few hundred psi.

academically I'd be interested to know how much improvement in spray pattern / atomisation you can really obtain past choked flow. i'd be surprised if there was much improvement without injector design change to operate with the intended increase in pressures. I'd be more worried about what the spray pattern is doing when it hits the piston top late in the compression stroke - and I have seen some guys move away from the dished piston design on built engines which is there to help prevent fanning out of fuel and cylinder wall wetting.

 

[doublespaces]

It wouldn't need a second set of wires or a dedicated signal wire. The wires that are across the piezo stack would generate a signal when they aren't energized. The injectors as is are fully capable of this. Its the circuitry within the msd ECU that is not set up to do so. You would also need a decent DSP to analyze that signal and detect a knock event . This signal can be seen on a decent oscilloscope just by reading the 2 leads that come off the injector. You could also use the spark plugs to detect knock with an ion sensing circuit. The S65 and s85 do this.

Individual EGR probes on the exhaust manifold is not a new idea, but individual knock sensor capability would be. Is there a possibility of actually doing this?

 

[Hydra Performance]

Just my 0.02c,
Here is a page detailing the DI injection strategy for the Nissan VR30, can't be too different from ours I would imagine. It is interesting to note that, based on my experience with several different N54s , cranking up ignition timing causes the HPFP to crash. i.e. car is fine at say 9 deg, add race gas, target 15deg, and HPFP falls flat on its face. Target 9 deg on the spot and all is dandy once again. This would indicate to me that the injector fires late in the compression phase as well, shortly before TDC...

 

[Rob09msport]

As much as the n55 hpfp is lower in capacity it seems to run much higher pressures. Would it be beneficial to crank pressure up In the tune to allow more of a buffer for pressure drop? Also some have stated that the dme limits duty cycle is that limiting by use of return? Cause if it just limits output and not lobe position then it really won't change anything when pressure is crashing.

 

[Bnks334]

As much as the n55 hpfp is lower in capacity it seems to run much higher pressures. Would it be beneficial to crank pressure up In the tune to allow more of a buffer for pressure drop? Also some have stated that the dme limits duty cycle is that limiting by use of return? Cause if it just limits output and not lobe position then it really won't change anything when pressure is crashing.

I don't personally believe there is any benefit to a pressure "buffer." The fuel rail serves as a common rail but it hardly holds a large enough volume of high pressure to buffer any significant duration of pressure scavenging by the injectors. If it were a large enough chamber then we would never see a pressure dip at low rpms... Just about every tuner is requesting high pressure at low rpms as a "buffer" and we still see a dip on boost onset when HPFP output is exceeded.

I think the plunger style N55 cars (F-series/EWG/2012+) run higher rail pressure with the stock tune (2800psi) because of the aforementioned differences in injector holes and atomization. sacrifice a bit of pumping loss/economy for better fuel spray possibly for a net gain in performance.

I may be wrong but I interpreted the high pressure "volume control valve" as being a functioning component of the HPFP (per bmw docs). Apparently it's a little more complex than just a simple valve though. Here is an old write-up but I am not too sure the information is really even accurate: https://www.n54tech.com/forums/showthread.php?t=19447&page=5

Since we now know that there is a table which controls HPFP duty cycle (which has been limited supposedly) I would say that valve would be what is being controlled (it's the only thing that you plug in on the HPFP lol). And yes it would seem that it is bypassing excess volume from being passed to the rail. Meaning increasing the duty cycle would increase fueling. I think this has already been proven with logs to work and there is quite a bit of flow being left on the table with the stock HPFP that tuning can "unlock."

some of the other speculative data posted in this thread in regard ot timing and such can been updated with the factual data from known documents such as the N54 engine PDF hosted on this site and elsewhere: http://www.e90post.com/forums/attachment.php?attachmentid=50389&d=1165592709

 

[Jake@MHD]

I may be wrong but I interpreted the high pressure "volume control valve" as being integrated into the HPFP (per bmw docs).

You can take the two bolts out and actually remove the VCV.

 

[Bnks334]

You can take the two bolts out and actually remove the VCV.

Yeah i meant that it is part of the HPFP pump. Maybe should edit the word "integrated" lol.

Point being though that the control valve is managing both LPFP and HPFP pressure and bypassing excess. There are quite a few posts on Mini and diesel forums, using similar design HPFP, that actually point to o-rings on this solenoid as being the culprit behind HPFP failures lol. People have apparently replaced the o-rings with success in some cases.

And here is a quote from Sbrach he just posted in another thread (in regard to the HPFP duty cycle question):

I did this a year and a half ago and added the HPFP VCV control tables to increase duty cycle which is what allowed E50 DI only on N55 E series. The DME does not command 100% duty on a stock tune but 100% flow != 100% duty.

 

[SJ_1989]

Need to be careful about cranking up the fuel pressure. As @Bnks334 mentioned, the likely limit in my experience with hydraulic piston pumps and solenoid valve design is the solenoid valve. Typically it's due to the seals. Backup rings can be added though to protect the seal from blowing out under high pressure.

Honestly never looked into the HPFP design until the past few days. If anyone has a failed HPFP PM me. I'll pay shipping if you send it to me. I'm genuinely curious now lol.

 

[kayzrx82]

The solenoid on the side of the pump is simply a solenoid controlled bypass valve . It allows the high pressure side of the fuel pump to bleed into the low pressure side. It does not regulate the pressure on the low side of the pump. When fed a pwm signal it acts a variable flow valve adjusting the bleed rate back to the low side to control the flow of the pump which affects the pressure on the high side. It also allows the hpfp to be bypassed in the event it's is damaged so the car can run on the low side fuel pressure.

There is a mechanical pressure relief valve inside the pump that is not controlled by the dme. It just opens at 245 bar and limits the max pressure of the pump. You can spin the pump as fast as you want but you will never get higher pressure than what that valve cracks open at. Excessive bypassing of the fuel causes it to recirculate inside the pump. This just leads to heat generation and raises the temperature of the fuel leaving the pump significantly.

I'm curious if there is a table or scaller in the dme controlling the potential put across the injectors. Being that they are piezo injectors, they have an advantage over the solenoid injectors. Piezo injectors can vary the injector nozzle needle height off of the seat by adjusting the potential across the piezo stack. Piezo injectors can act as a variable flow rate injector where as the solenoid ones are either on or off. I would bet under limp home mode ,due to hpfp failure, the injectors are probabaly at their max height off their seat to allow enough fuel to flow at low side pressure. If that could be adjusted somehow there may be the possibility of getting more flow through the seat at higher pressures if the needle height currently is the flow restriction.

 

[kayzrx82]

Individual EGR probes on the exhaust manifold is not a new idea, but individual knock sensor capability would be. Is there a possibility of actually doing this?

It is possible. Would need a decently fast dsp , some coding development, and some work done on the circuit design to interface with that signal. I've worked with dsp and fpga's to analyze broadband signals. Figuring out the filtering shouldn't be to bad. If you just want to know when a certain threshold is reached that shouldnt be to difficult but I'm sure there will be some unexpected hurdles that will have to addressed. The ion sensing circuit on the plugs may be an easier route and can provide more info besides knock, but would take time characterizing the current signal. Would have to weight the pros and cons and cost to see if it's worth doing. I've messed around with the s65 coils as they have the ability for ion sensing. The s85 has a dedicated processor on the valve cover handling each bank. Certainly possible , at what cost I'm unsure.

 

[doublespaces]

It is possible. Would need a decently fast dsp , some coding development, and some work done on the circuit design to interface with that signal. I've worked with dsp and fpga's to analyze broadband signals. Figuring out the filtering shouldn't be to bad. If you just want to know when a certain threshold is reached that shouldnt be to difficult but I'm sire there will be some unexpected hurdles that will have to addressed. The ion sensing circuit on the plugs may be an easier route and can provide more info besides knock, but would take time characterizing the current signal. Would have to weight the pros and cons and cost to see if it's worth doing. I've messed around with the s65 coils as they have the ability for ion sensing. The s85 has a dedicated processor on the valve cover handling each bank. Certainly possible , at what cost I'm unsure.

Maybe it would be easier to adapt another solution and tune it for our engine?

 

[rac]

As much as the n55 hpfp is lower in capacity it seems to run much higher pressures. Would it be beneficial to crank pressure up In the tune to allow more of a buffer for pressure drop? Also some have stated that the dme limits duty cycle is that limiting by use of return? Cause if it just limits output and not lobe position then it really won't change anything when pressure is crashing.

definitely no value in a pressure buffer - the volumes are so small the additional pressure is a meaningless fuel volume.
for a 1/4 mile track car or other short bursts you could in theory make an accumulator tied into the fuel rail, not really practical and the solenoid will not be able to pull pressure down quickly going back down to idle.

 

[Jake@MHD]

but individual knock sensor capability would be.

DME already does detect knock per cyl just fine, no need to go do anything crazy.