N54 Piezo Fuel Injector Limits

fmorelli

Lieutenant General
Staff member
Aug 11, 2017
3,748
3,592
0
57
Virginia
Ride
E89 Z4 35i, F10 535d
Injection window is the big limiter to the stock di setup. At some point, that window closes and the only way to inject more fuel without fully wetting the cylinder walls comes by way of pi.
Can you explain that further? I'm not sure I understand what that actually means.

Filippo
 

doublespaces

Administrator
Oct 18, 2016
9,303
4,331
0
AZ
Ride
2009 E93 335i
To bring an old but popular subject to this forum, this is a copy paste of some information that is at least 5-7 years old is my guess, but its valuable so here it is. Credits to the N54tech.com member for originally sourcing this information.
======

Today i had my follow up conversation with the tech guy at continental. Here are the questions and answers. Enjoy !

Are the pistons forged pistons ?
- Don’t know

Does the N54B30 use homogenous or stratified charge under high load ?
- The engine was developed to run in homogenous mode. The N53 engine was to run in a stratified mode. Although the engine is capable of running in stratified mode, it currently does not.
It uses an outward opening injector.

Have you experience with this engine and E85 fuel or do you have any thought on it ? Can e.g. the injectors handle alcohol ?
We have a lot of experience with E85. Currently, we have an issue that if the injector is hotter than 100 Celsius, we are seeing melting of the fuel filter in the injectors. As long as the temperature of the injector stays under the 100 Celsius, this engine is perferctly capable to run E85. As a matter of fact, it will run E100 without any problem under cold conditions. There is no need to preheat the fuel.
We are investigating the filter issue, and we hope to have a solution next year. Whether or not we will create injectors for this engine, depends on the market (read: BMW).
It is not possible to replace this filter yourself; you cannot open the injector to get to that part.

Is the HPFP pressure of 90 bar constant or is it controlled by the MSD module ?
It is constant but indeed controlled by the module. The engine has been tested to up to 200 bar. If you keep the HPFP pump below that number, you could inject a lot more fuel. The relation between the amount of fuel sprayed and the pressure is not a linear relation but a Bernulli equation.

In our earlier conversion, you mentioned that under high load the TI is about 2 ms ? Correct ?
Yes maximum is 2 ms. Notice that the engine runs under homogenous mode and that means that the injection start in the intake phase. With this engine, it is no problem to keep injection into the compression phase, but not too long; it would be bad for emission. However, if the pressure of the HPFP is increased, this may not be needed.

The injectors have a maximum of 40 mg/ms spray at 200 bar. This is an increadible amount of fuel, much much more than any solenoid injector. It is this high, because when we designed the application, BMW insisted that the engine must be able - in limp mode - to only use the LPFP pressure, which is 5 bar, and still drive the vehicle at 170 km/h.


If correct, do you agree then that considering the ti in the compression phase, we can get it up to 4 ms for 7000 rpm ? Can we use the full compression phase ?

You would have to calculate it, but yes, there is extra time at mentioned before.
The injector can inject a lot of fuel into the cylinder. When increasing the injection time, and also when injecting in the compression phase, look out for cylinder wall wetting. This is dangerous. It would wash away the oil on the walls, causing wear. One way to check for wall wetting is to check the engine oil if it has fuel in it. Another way would be to check if the vapor of the exhaust gas recirculation (EGR) and see if there's fuel vapor.
The risk of wall wetting may be reduced by running higher boost, as it affects air flow inside the cylinder.
Running the HPFP at higher pressure may increase the risk again.

You said we should tune the TI duration of the ignition pulse. Can we also tune with the MSD via a reflesh when the ignition time starts ?
There is indeed a calculated ignition start time in the MSD. It is not a map, but i'm not sure.

How does the ECU measure a knocking problem ? Does it use voltage ? Does it do it for all 6 cylinders individual ?
Each piezo injector has a knocking sensor. It measures vibration. By looking at certain frequencies, the knocking signal is sent.

EDIT: seems that the engine also has the standard knocking sensors in place.

The community is desperate to lean more about the MSD80. We would love to have information that can help us reflesh the unit to tune to engine for 600 PS. Can you help ? Can you direct us to a colleague (do you have a name ?) This is vital for my own project too !
I don't know about the MSD80. I'm an injection / fuel system engineer. I will talk to some colleagues and see if i can help you with anything. We have to make sure we're not infringing any copyrights and i think BMW owns the rights to the information on the MSD.

Can we buy this information ? Who should I talk to to ?
there may be a legal way to buy this information, i will talk to some colleagues and mail you back.

================


60sec/6000rpm gives you about 10ms to inject fuel on an intake stroke at 6k rpm. The problem has always been the fuel injectors take time to open and close. Thus on port fuel injection engines, 75-80% duty cycle was the max fuel flow.

BTW, I found the Continental info pdf on the fuel injection systems again (finally!): http://www.conti-online.com/www/dow...ain/common/pow_powertrain_products_pdf_en.pdf

These are the specs on the injector:
System pressure 50 up to 200 bar
Working flow range > 50
Static flow > 35 g/s
Dynamic flow 14.5 mg/ injection 10% at 0.4 ms PW
Min. dynamic flow < 2 mg/stroke
SMD size ~ 15 µm
Opening/Closing time = 150 µs
Spray angle 94°/98°± 4°
Temperature range – 30 up to +140°C

and on the HPFP:
Max. fuel pressure up to 200 bar
Supply pressure 4.5 up to 6 bar
Temperature range – 40 up to +130°C
Control flow controlled
Drive end of camshaft or belt driven
Maximum flow up to 240 l/h (at 3500 rpm, depending on the # of pistons 2 up to 6)

So, assuming these numbers are accurate and keeping mind they are static flow numbers, using some rough calculations:
35 gm/sec = 277lbs/hr
277lbs/hr * .5lb/hr for 1HP = 554HP from a single fuel injector.

There is probably more to the story restricting how long the injector can actually open because BMW put about 3000HP worth of fuel injectors in a 300HP engine.
smile.gif


The HPFP is 240LPH so, 1.64lbs/L * 240L = 393lbs/hr = 785HP (@ .5lb/h per HP).

Pretty rough numbers (BSFC is wrong) but the general point is there's supposedly a lot of fuel available via the injector and HPFP if their dynamic supply numbers are close to say 80% of their static flow numbers.

This doesn't seem to match up to the real world though so other factors are undoubtedly influencing fuel delivery (lines and fittings, fuel rail, etc).

====
 

Rob@RBTurbo

Lieutenant
Dec 7, 2016
626
401
0
St. Louis, MO USA
www.rbturbo.com
Ride
'08 335i, '14 M6, '15 Tundra
There is 2-3 long discussions sprinkled across the internet dating back to when that one person actually managed to contact the engineer who worked on the fuel system and obtained those cc numbers. Probably a good discussion for another thread, although I think one of those threads is here somewhere.

100% this can turn into a long long discussion and truth be told I'd need a revisit it myself to describe it perfectly. However in simplest form comes down to the length of time a Direct Injector has to actually pump fuel into the cylinder and how its relation to a Port Injector differs (being external to the cylinder combustion process)... when comparing the cc potential of the injector.

Rob
 

iminhell1

Sergeant
Jun 17, 2018
419
207
0
Injection window is the big limiter to the stock di setup. At some point, that window closes and the only way to inject more fuel without fully wetting the cylinder walls comes by way of pi.

This is where the different injection strategies come into play. They really are something that time should be spent on seeing if they can be enabled and tuned.
I see zero benefit from a community knock table.
Port is 100% limited to valve open time (actually a fraction of that). But direct has no cam limitations.



100% this can turn into a long long discussion and truth be told I'd need a revisit it myself to describe it perfectly. However in simplest form comes down to the length of time a Direct Injector has to actually pump fuel into the cylinder and how its relation to a Port Injector differs (being external to the cylinder combustion process)... when comparing the cc potential of the injector.

Rob

Direct has infinite time.
Port is very limited.

The hangup is in the controlling of DI. It's software limitations and not hardware.




I got into this platform because of the technology and power potential I saw. I don't want to travel 40 years into the past (PI) to make power.
 

gmagnus7

Corporal
Dec 3, 2018
174
73
0
Ride
135i N54
Wait... each fuel injection has it's own knock sensors? ON TOP of the normal knock sensors in the block? That's pretty rad!
 

The Convert

Captain
Jun 4, 2017
1,487
1,052
0
Ride
335
Wait... each fuel injection has it's own knock sensors? ON TOP of the normal knock sensors in the block? That's pretty rad!
That is not at all what he was saying. He’s saying a community funded knock table is a waste of money. Period.
 

gmagnus7

Corporal
Dec 3, 2018
174
73
0
Ride
135i N54
That is not at all what he was saying. He’s saying a community funded knock table is a waste of money. Period.
Confused and unsure if sarcasm. Also never followed that knock table stuff - not that hardcore.

"How does the ECU measure a knocking problem ? Does it use voltage ? Does it do it for all 6 cylinders individual ?
Each piezo injector has a knocking sensor. It measures vibration. By looking at certain frequencies, the knocking signal is sent. "
 

kayzrx82

Corporal
Apr 4, 2018
110
88
0
Ride
n54 335i
Confused and unsure if sarcasm. Also never followed that knock table stuff - not that hardcore.

"How does the ECU measure a knocking problem ? Does it use voltage ? Does it do it for all 6 cylinders individual ?
Each piezo injector has a knocking sensor. It measures vibration. By looking at certain frequencies, the knocking signal is sent. "
Piezo crystals can act as a sensor . They produce a voltage when a force is applied. By looking at the voltage across the piezo crystal it is possible to read noise signals from the cylinder. Each piezo injector essentially can be used as a microphone. That being said I do not think the msd80 is looking at or running algorithms to convert that signal into a usable knock signal.
 
  • Informative
Reactions: Torgus and fmorelli

rac

Sergeant
Nov 14, 2016
341
204
0
Australia
Ride
135i ST
Note we don't have a full intake and compression stroke because of exhaust valve timing, ignition advance, and finishing the injection close to ignition timing results in poor combustion. DI start and end angles have a significant impact on power - realistically there is likely something like ~300-320 degrees of optimal crank angle which is only ~7.4ms @ 7k, 8.6ms @ 6k.

My rough calcs would suggest ~800whp or something of that nature on E85 and there is various posts floating around of injector reliability / failures at lower figures.
 
  • Informative
Reactions: fmorelli

Bnks334

Lieutenant
Dec 1, 2016
524
340
0
New York
Piezo crystals can act as a sensor . They produce a voltage when a force is applied. By looking at the voltage across the piezo crystal it is possible to read noise signals from the cylinder. Each piezo injector essentially can be used as a microphone. That being said I do not think the msd80 is looking at or running algorithms to convert that signal into a usable knock signal.

Some diagnostics, like the inpa/ista "rough running" procedure, are most likely to be utilizing the feedback being referenced.
 

Jake@MHD

Major
Platinum Vendor
Nov 7, 2016
1,593
2,058
0
Philly
The DME does not use any feedback from the piezo injectors for any knock detection. It is all done by the 2 knock sensors attached to the block and fed through a DSP processor, etc. The knock fund was not a waste of time / money, as it is a very real problem for basically all built motors on this platform. The only way to properly fix it is to tune the signal processing correctly.

Injection window can be expanded, but you need to have the rail pressure to support lengthening the injection window, or else you will just drop even more pressure. I am working with vendors on the software side to get more HPFP solutions to market. If any of them end up hitting a wall due to the injection window, it will be addressed at that time.
 

veer90

Lieutenant
Nov 16, 2016
1,000
774
0
West Nyack, NY
Ride
e90 335i 6MT
How does the ECU measure a knocking problem ? Does it use voltage ? Does it do it for all 6 cylinders individual ?
Each piezo injector has a knocking sensor. It measures vibration. By looking at certain frequencies, the knocking signal is sent.

The injectors are capable of measuring knock. However the DME is not using this functionality to my knowledge, only the standard knock sensors.

This is why you can put 2 trim pots in the knock sensor wires and all your timing corrections go away.
 

GreyNBlueE92

Sergeant
Oct 3, 2018
266
92
0
27
OH
Ride
2007 E92 335i
1k hp is cool and all but how many have a built motor to actually use the knock findings? Using them on a stock motor is probably not going to end well. Injection timing on the other hand...
 

Rob09msport

Major
Oct 28, 2017
1,929
664
0
Monroe CT
Ride
09 335i msport le mans 18 x5
All your false knock goes away. The injectors would sense actual pre detonation
 
Last edited:

veer90

Lieutenant
Nov 16, 2016
1,000
774
0
West Nyack, NY
Ride
e90 335i 6MT
All your false knock goes away. The injectors would sense pre detonation

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.
 

Jake@MHD

Major
Platinum Vendor
Nov 7, 2016
1,593
2,058
0
Philly
1k hp is cool and all but how many have a built motor to actually use the knock findings? Using them on a stock motor is probably not going to end well. Injection timing on the other hand...

All built motors need the tuning changes regardless if you are making 500hp or 1000hp. If someone makes the changes on a stock motor, that is their problem. You can also run 17:1 afr and 30psi stock motor, should we take away the AFR tables and WGDC tables?

Again, SOI / EOI tables aren't going to do you any good unless you are maxing the current windows AND have rail pressure to spare. I have not seen any evidence of anyone reaching this point yet, as there aren't enough 2x HPFP setups out there.

All your false knock goes away. The injectors would sense pre detonation

The injectors aren't sensing anything.
 
  • Agree
Reactions: GreyNBlueE92

Rob09msport

Major
Oct 28, 2017
1,929
664
0
Monroe CT
Ride
09 335i msport le mans 18 x5
Im curious if they could be source of post shift timing pull

Edit read jayoma answer so never mind
Wouldnt you not need a 3rd wire to sense voltage generated by a crystal
 

GreyNBlueE92

Sergeant
Oct 3, 2018
266
92
0
27
OH
Ride
2007 E92 335i
Not sure that forcing as much fuel through the injectors in as little time as possible is going to benefit as much as letting them stay open longer. But if the rail pressure is not there it cannot be done. I understand that.
 

Jake@MHD

Major
Platinum Vendor
Nov 7, 2016
1,593
2,058
0
Philly
Not sure that forcing as much fuel through the injectors in as little time as possible is going to benefit as much as letting them stay open longer. But if the rail pressure is not there it cannot be done. I understand that.

If the pressure is within the operating spec of the injectors, I see no issues. It is more beneficial to inject at a higher pressure as you get better atomization and a better burn, which allows you to run a bit leaner and make more power too. All the OEM's are going even higher with their next gen HPFP, up to 350bar which is 5000psi. Granted, it's partly for emissions, but it does help performance too :)