Understanding Saturated Fuel Injectors

I'm looking to upgrade to 350cc (33lbs) high impedance injectors for my Merc SLK (OEM=295cc). Through some of the research I've been going, I've found out that the ECU is capable of a 15-20% compensation in fuel correction. My concern lies with the actual impedance of the replacement injectors.

Supposedly, within saturated injectors (12-17ohms), there are "activation points", or points where drivers are tripped-opening the injectors. Can anyone shed light on this? Are OEM ECUs not capable of ohm compensation, or are high impedance, saturated injectors so varying in ohms that OEM ECUs plan for replacement injectors to, well, vary?    

It's not so much that the resistance is different, as much as it is that different injectors have different dead-time characteristics.  If the computer is calibrated for injectors that have a .9ms latency and you install modern injectors that only have .5ms (for example), it's going to run even more rich than the size would have you believe, because the computer is adding .4ms too much time to the injection pulse.

It's a bit more complex than that,  but with injectors that small you probably don't have to worry about small-pulsewidth errors.

Do you have a way of retuning the computer for the larger injectors so it doesn't run massively rich at WOT?  Closed loop fuel control is generally only at part throttle.

Hey-who you calling small?!!! Haha, seriously, thanks Pete! This project could turn into a Challenge car, so my goal was to use some of the "tricks" historically used by the folks over in the Merc SLK forums- Big pulley, injectors, PSI up to 15 (ECU supports up to that), more efficient I.C. resulting in 280-300hp. Once the Challenge is over, I've got a standalone, bigger injectors and a turbo I'd like to add on!

Im thinking that maybe running rich won't be too bad of an issue- especially considering these things run lean (13.4 at WOT)!

High impedance injector drivers are generally pretty basic in practice; you can just drive then with a single transistor that the processor toggles off and on. As Pete noted, the main difference is going to be dead time, which may have its most pronounced effects at idle. There isn't really a good way the ECU could determine dead time electrically - it is possible to determine how long it took the injector to go from open to closed and closed to open with a current sensing circuit, but that won't tell you how much fuel got out when the injector was partially opened. So you won't see OEM ECUs try to measure this - instead, they will use a preset way to calculate injector dead times, which is programmed in at the factory. Exactly what that preset method is gets more and more complicated with newer ECUs.

In reply to MadScientistMatt :

Ah-I see! Thank you for the insight. Considering what's been presented, I'm happy to see that more fuel may flow, rather than less, as my concern lies with the computers' leaning of the A/F mixture at WOT. 

2GRX7 said:

In reply to MadScientistMatt :

Ah-I see! Thank you for the insight. Considering what's been presented, I'm happy to see that more fuel may flow, rather than less, as my concern lies with the computers' leaning of the A/F mixture at WOT. 

Computers don't so that on their own.  Also, being rich at idle due to injector problems does not translate to WOT- as the effect of opening and closing times are more muted when the duty cycle is large vs. small.  Which is to say that the rich problem will be a idle and it may not even be significant at WOT.

I would say in general the dead time differences that you will see with the % change of injector can be fixed through the fuel trims of an OE ecu. That not to say that the ecu will be totally happy, but it's possible that it is without retuning. 

MadScientistMatt said:

High impedance injector drivers are generally pretty basic in practice; you can just drive then with a single transistor that the processor toggles off and on. As Pete noted, the main difference is going to be dead time, which may have its most pronounced effects at idle. There isn't really a good way the ECU could determine dead time electrically - it is possible to determine how long it took the injector to go from open to closed and closed to open with a current sensing circuit, but that won't tell you how much fuel got out when the injector was partially opened. So you won't see OEM ECUs try to measure this - instead, they will use a preset way to calculate injector dead times, which is programmed in at the factory. Exactly what that preset method is gets more and more complicated with newer ECUs.

If you have a wideband sensor and gather enough data it's possible to compute the dead time by plotting it and looking for the X intercept.  Theoretically you could that in the ECU, although I kinda doubt anyone does.

In reply to alfadriver (Forum Supporter) :

Are you saying that the ECU won't correct on its' own? Not sure about your first sentence.

Paul_VR6 (Forum Supporter) said:

I would say in general the dead time differences that you will see with the % change of injector can be fixed through the fuel trims of an OE ecu. That not to say that the ecu will be totally happy, but it's possible that it is without retuning. 

Yes, according to other SLK owners who've dealt with tuning on the OEM ECU, trims adjust 15-20%.

codrus (Forum Supporter) said:

MadScientistMatt said:

High impedance injector drivers are generally pretty basic in practice; you can just drive then with a single transistor that the processor toggles off and on. As Pete noted, the main difference is going to be dead time, which may have its most pronounced effects at idle. There isn't really a good way the ECU could determine dead time electrically - it is possible to determine how long it took the injector to go from open to closed and closed to open with a current sensing circuit, but that won't tell you how much fuel got out when the injector was partially opened. So you won't see OEM ECUs try to measure this - instead, they will use a preset way to calculate injector dead times, which is programmed in at the factory. Exactly what that preset method is gets more and more complicated with newer ECUs.

If you have a wideband sensor and gather enough data it's possible to compute the dead time by plotting it and looking for the X intercept.  Theoretically you could that in the ECU, although I kinda doubt anyone does.

 

I will add a logging wideband to check A/F ratios before I close the book on the project. Worse case, I picked up a RusEFI that's keeping me under Challenge budget-I'll just have to limit the goodies I can add because of that $300 hit.

In reply to 2GRX7 :

rusEfi, eh? We'd love to have you over on the Slack! We have another guy running an SLK230 engine (in a W124) on rusEfi with pretty good results.

2GRX7 said:

In reply to alfadriver (Forum Supporter) :

Are you saying that the ECU won't correct on its' own? Not sure about your first sentence.

ECU's are only as good as their inputs.  Once set, they will always react the same.  If they go lean at WOT, that's either something wrong in the base code (which means you are using a bad ECU) or the user put wrong information in.  Computers are really good at doing the same thing over and over and over again.  

"Correct on it's own" means that it has a correction to even go to.  If you are using a WB sensor for feedback at WOT, then calibrator/tuner has a responsibility to put the right target in.   

Which is a long way of saying- ECU's don't go lean just for the sake of it.

edit- now I see you are using the original Mercedes ECU.  What year?  Depending on the age, the O2 sensor may be ignored at WOT, and it wont do any correction.  All it does is deal with the inputs that are there.  And I would not call 13.4 scary lean- we run cars at WOT at stoich for a long time- just to make sure they can.  It may or may not be a problem.  But even so, the longer open/close time is relatively nothing at WOT but huge at idle.  A 0.5 ms error us huge at 5ms pulse width, but tiny at 50.

In reply to Matthew Kennedy :

LOL- Actually, he's started helping me out on another forum (S.W.)!. He's been VERY helpful!  I've gotten on the forum, read through everything on the wiki, just haven't hit the Slack simply because until I know if I'll need to use RusEFI in the Challenge, or after, I'll just be lurking! Then again, if lurking's fine...

In reply to alfadriver (Forum Supporter) :

It's a 1999 5-speed, and if the developer says 13.4 isn't scary, i'll go with that!

In reply to 2GRX7 :

If you can find a way to go richer, 12.5:1 will make more power.  With boost, probably 12:1 is even better.  So I'm not saying don't do anything- just don't be terrified about it.  Change it for the performane benefit, if you can.  For a car like yours, I *think* the easiest way to do that is with a rising rate fuel pressure regulator on the rail.  Assuming you need to after all of the changes you are making....

If you do plan wot stoich make sure you have good fuel and the knock system works. You can tell pretty quickly if it doesn't. 

Where did the idea of running stoich under boost come from?

Putting ~12% larger injectors in is not going to make it run leaner.

In reply to Pete. (l33t FS) :

durability testing.