DI injector and timing

Hey everyone, I'm working with a direct injection engine along with a second set of port injectors. I noticed that at around 6500 to 7000 i was getting a strange sensation and sound. When i checked the logs i noticed my direct injectors staying open well into the ignition event. My question is what happens if the direct injectors continue to spray while the ignition event occurs? Can it burn or melt the tip of the injector? I would assume its either causing a lot of heat and creating a flame front or washing the bore of the cylinder and extinguishing the combustion. both seem like a really bad idea. I'm wondering if I've created any issues by driving it like this.

Injector damage, pistons melt. It's extremely bad. You definitely need to leave a buffer between EOI and spark ignition.

As Mike said, it isn't a good thing to have happen.

What ECU are you using?

I’m using the stock MED 9. The tuner I used wasnt well versed in the tuning suite I have available. I’ve moved the injector timing to give a longer window and retarted the timing to separate the ignition event and injector time. I’m still seeing longer than I’d like injector times though. I’m working with eurodyne and there isn’t a lot of information out there.

I'm surprised that there isn't a protection strategy in that ECU that prevents the injection timing from crossing the ignition timing.

I would have thought so also but doing the math between ignition event and injector timing crossed over.

I imagine you've already done this, but how far have you pushed SOI and fuel pressure? Starting earlier and increasing pressure both aid in getting more fuel delivered before EOI becomes dangerously late.

long story mediocrely short:

Car: 2008 Audi A4 B7 2.0T FSI EA113

I paid to have my car tuned with a remote tuner. After about 10 revisions I kept getting hard cold starts and a lot of knock correction in each cylinder so i decided to move to a different tuner. The tuner I chose for remote was supposed to be a reputable tuner but after seeing that i decided not to continue in order to save my engine from failure. After a while of searching i was able to talk my local tuner into tuning the car. He's tuned my other vehicles and some friends so i know he is capable of it. However, the issue we ran into is I am using the Eurodyne maestro tuning suite and the tables are labels weird and the definitions (if any at all) are almost all incorrect. There is no references in the descriptions. I specifically purchased these courses for this car as i cant relying on the VAG community for tuning help.

example: ignition timing table 1 and ignition timing table 2 two tables for a 4 cylinder. When asking the company that creates the tuning suite they tell me they are not a tuner and to ask my tuner.

To answer your question my tuner was not well versed in the suite and did not adjust the injector timing or cam timing. Since then i noticed under full throttle i get a very strange sensation and it felt as if the car had sever loss of power at the high RPM. I did some math and research and the car was injecting starting at 410 and going for about 9.5 to 10ms. (well into the ignition event) since then I've gone in and adjust my cam timing to help with spool and low end power and I've made a lot of headway. As for the injector timing I've adjust as best as i can for the time being.

EVC: 352 / SOI: 320-351 / EOI: depending on request but I'm trying to keep it below 7.5ms

Ignition Event at 5500RPM is about 10* at the movement to keep the two events apart.

its about 45* per millisecond

First thing is, rather than running hot, the fuel will be cooling the injector tips, so you don't have to worry about that - it's also a big reason why, even if the fueling is supplimented by the port injectors, you keep plenty of fuel coming through the direct injection.

DI is outside my experience, but besides avoiding waste fuel passing through the exhaust on over lap, I understand that one big advantage is that it's possible to time the injection period to have a richer, and more easily ignited, mixture at the spark plug when it fires, which makes it easier to burn the wekar mixture in the rest of the chamber - basically a "stratified charge" strategy. Could this be a deliberate move by the OEM to have an initial ignition point that starts the fuel burn and then have the rest of the required fuel introduced to the burn, as required?

In theory it should have the benefit of reducing the pressure build up before TDC - reverse torque - and concentrate the temperature/pressure increase after TDC - positive torque. It could also allow higher compression ratios/boost to be used as the reduced, as there may be less 'free' fuel to be detonated?

It seems there are many variables, most of which don't seem to be fully understood, available to "tune", I'd suggest returning to the OEM (or as close as you can get it), figuring out EXACTLY what the various tables do - preferable with a suite that has support that can actually tell you what the tables do (if I read that correctly, and it wasn't you asking for values rather than identification), and when you understand what the OEM means on the tables, work from there.

Gord,

When you increase airflow significantly, or change to a fuel which has a lower stoich point, you can pretty quickly exceed the window of time where DI can function properly, which is what happened here.

Korey,

SOI 410 is really early so good catch. I've seen cases where massive exhaust pressure keeps fuel in the chamber, but to get things more reasonable it sounds like you should stay around 360 or less, and then you're building your gap between EOI and spark which is great. How are you doing this, by reducing airflow, targeting leaner lambda target, restricting max RPM, all of the above?

I take it you're already running fuel pressure as high as you can get it without causing drops?

Thanks, Mike, so more a case of lacking flow rate and so increasing duration into a possible problem area?

The solution would seem to be to use higher flow injectors - which may be where the trick lies?

Korey,

Can you provide us more details about your build. What are the mods, what is your goal etc. I am tuning European cars on a daily basis and might help you with it but I need more information.

Mike - to remedy this situation I've set the injector changed the injector timing to right before EVC so about 350ish. I've also set the Boost duty and PID to 0 so I'm running off of Wastegate (1bar) I'm still seeing too high of injector time at the high RPM. I've moved the RPM limit from 6500 to 7000 to help with the gearing. As for the timing I've limited it to 10* advanced above 5500 in order to keep injector time and ignition timing far apart until i can find a solution. Instead of running the tune i got from my tuner I've done the adjustments on a stage 2 off the shelf map and adjusted for my modifications. Fuel pressure is set to 150 bar request after 5500 RPM.

Gord- Thanks for you input. your knowledge far surpasses mine and i wish i could answer your questions about the ECU.

Valentine P -

The build:

2008 Audi a4 2.0T FSI, S3 injectors, IE Drop in Rods, IE Intake Manifold, stock 10:1 compressions, 150Bar fuel pressure return valve, Balance shaft dele, GT3076R, Port injection 1100cc injectors, TI450 in tank, HPFP upgraded internals, Intercooler.

I'm currently not running the secondary's in order to get the base map configured first.

I've attached some logs for reference.

You've got it Gord.

Aftermarket high flow gas DI injectors aren't readily available for many platforms, and those that are available often cause drivability losses while not increasing total flow potential that much. In the port injection world today the ID1050x in my first step and we have options far beyond that, but in terms of gas DI today, they're mostly 10% to 30% flow increase from stock, and at 20-30% flow increase, the injectors often come with significant precision of control penalties we aren't used to experiencing until we're using very large port injectors.

The upgraded high pressure pumps of today also often don't provide that much more opportunity for increased fuel delivery. For example if the stock system runs 3000 psi, and you can increase it via calibration to 3200-3300, is a $900 pump that gets you another 200 psi worth it? The % increase in fuel flow potential is small enough that I question the cost benefit proposition and loss of OE reliability in most situations. At a certain point if you increase pressure enough, you can also prevent injectors from opening. An external injector driver or potentially a reflashed stock unit may allow the injectors to open against higher pressure, but this is beyond the scope of many users and may risk the reliability of the injectors.

4 cylinder DI injectors and a single high pressure pump costs around $1800, so they are not cheap.

With what's available today, I believe it makes sense to work with the stock system to its limits. That involves maximizing fuel flow potential, and testing various combinations of airflow, ignition and cam timing, lambda targets to see what gets you the most reasonable benefit.

To go further, my preferred step is port injection plus a control system that integrates DI and PI so total fuel delivery is proper. You keep all the benefits of excellent control and reliability of OE DI components, but add the port injection benefits of pre charge cooling at high load, valve cleaning, and wide ranging options for fuel delivery increase to support almost any power level.

Korey,

Thank you for that additional info on the setup. It makes far more sense now why you're so far off your fueling requirement. I don't know your wastegate spring pressure, but with that turbo it sounds like you won't be able to rev the engine out without port injection assistance.

What are you using to control the port injectors? I'm guessing a piggyback?

If so, you're going to end up kludging the reflash as you go, limiting calculated airflow to avoid the DI system operating outside of a safe window, so I wouldn't put a ton of time into trying to make it ideal without PI. This method is inherently flawed, so do your best, keep your lambda targets what you want to achieve, and trimming will help you as best it can.

As you alter the reflash portion of the tune to reduce DI and keep it in a reasonable injection window, be careful as this skews estimated torque. Which transmission option do you have? I ask because auto or CVT will desperately need accurate torque reporting to function well.

The S3 injectors are OE "drop in". They are a meant for my application and have been proven to make around 400Whp with the correct spray pattern. (with correct tuning) The rail pressure release valve is out of a 2008 ish RS4 and the High pressure fuel pump is designed for this application. My goal is to try and find the limits of my setup before implementing the PI. Is tuning the Di to the limit then feathering in the port injection in before the limit a good tunning strategy? or should i tune both port and DI at the same time? Also does it matter if the port is sequential or batch?

I'm running a split second piggy back controller. It reads off the MAP sensor and spark timing. I'm also running a manual transmission. wastegate is set to 1Bar

Korey if you could tune the car for wastegate spring pressure DI only, I think that could be a worthwhile exercise, but it sounds like you've said you can't without kludging the ignition timing or reducing max RPM significantly.

I think there's value in making a limp home tune in that state in case the PI box fails, but if you can't tune the engine to run properly DI only, I'd work on your DI+PI tune.

At very high injector duty cycles there isn't much difference between sequential and batch because the injectors are basically open the whole time and you've lost the advantage of timing injection, but sequential works much better than batch when you aren't at high duty cycle. I always prefer Sequential.

I've used that controller and it's not ideal, but you're working with significant compromises no matter what since you have two independent fuel delivery systems that are not operating with common control.

Within the limitations of this type of control, I suggest bringing PI in well before it's needed since you want to allow for wall wetting to stabilize and avoid large transitions in PI fueling to give the closed loop fuel trim system a fighting chance of reasonable control. Generally I suggest brining PI in, then ramp PI and DI up together, until DI reaches your desired margin of safety to the available injection window. Make sure you not only leave room to avoid late EOI, but also allow for large fuel trim events which will be required in a piggyback DI PI situation.

Once airflow requires DI reach that margin of safety, from there up the port controller needs to provide the rest of the fuel, so the stock ECU is left with some range of lambda control.

thank you for that. So ill work on that and see if i can get it to do what i want. This has been a very helpful thread. thank you all and i hope i can bring some good news after.

Awesome, best of luck and please let us know how it goes.

Agree with Mike, start utilising the PI and limit the injection window of the DI. Use the fuel correction tables that are in 3D to reduce fuel amount supplied by the DI. Don’t forget to account for the dead time on the PI when tuning in part throttle. Wideband feedback is also a must but I believe your ECU should have one already. I would disable fuel trims when tuning.

I'm checking in to see how it's going, hopefully great!

Only jumping in here to say what a great thread, brilliant to see information freely shared on this platform and this topic is going to make me start visiting more to see if there is anywhere i can add value :)

Well done to all involved in helping this fella!

Regards,

Stu

Hey everyone sorry for the late reply. The older notifications went to my spam box and i missed them. I was able to figure out a little more info but I'm still working on this. The car is not my daily drivers so I work on it when i have the time. As for the injector timing there is a map for correction. The map for correction is only really used for small increments. As for the port injection i found another issue.

- the port injection was using my Boost reference sensor (MAP sensor but the map is located in the charge pipe not the intake manifold. While i was trying to blend the port injection the AFR's would go lean. After a while of trying to add fuel and getting the same results i couldn't get the car to rev past 6400RPM. I ran a data log on the secondary injection controller instead of the ECU and found that i was getting to Voltage drops from the MAP sensor. I didn't take the time to check to voltage on the MAP sensor to see why it would only effect my secondary injector controller. From what i can see my ECU is still reading voltage all the way through the RPM range. Instead of using the factory MAP sensor i moved to an external 4 bar map sensor and plumbed it into the intake manifold directly to avoid any issues.

I haven't had a chance to see if the new map sensor has fixed that issue or not. I will return with more information.

Well done, good diagnostic work.