NefMoto

So the only injector delay or lag I can find in ME9 FR is implimented via the map KLTVTSV. This map takes only one input parameter which is the pressure difference between the rail and the combustion chamber, and outputs a delay in milliseconds.

The table has an interesting form to it in my ME9.1 file. It starts off with a very small negative value for the largest differences in pressures, and tapers to a very small positive value for the smallest differences in pressures. For my file it runs from -0.05 up to 0.04, and the axis is supposed to be MPa, but my values run from 14.48 MPa down to 1.87 MPa.

So the question for those that understand this table's influence, how to I adjust to compensate for the increased injector lag of larger (RS4) injectors? What are the typical values one would expect to use. I have read that there is an B7 RS4 file around here with this defined in it, but for the life of me I cannot find it.

Any help understanding this is appreciated.

EPY

As it goes, after many (literally months) looking, I found what "might" be an 4.2 RS4 ori, but I can't tell because I don't have a damos to see if the maps looks good yet. So I started looking for KLTVTSV and I am not matching it up yet. Maybe we can start from here and if we can define the KLTVTSV in this file, we can compare.

Will take a look tonight!

Roger that DD. Thank you.
EPY

Well I started trying to compare and define the fuel angle maps, and I think I am close, but the axis look bad. Anybody want to take a look?

epy

http://www.google.com/patents/DE102009027290A1?cl=en

i think this has something todo with your topic.

Thank you. That is definitely about the topic, but I probably didn't state my confusion about this item very well. What I don't understand is what positive and negative numbers in this map mean, and what magnitude the numbers should be.

In particular, the values are on the order of 0.05, but lag for injectors is usually 10x larger than that (say 0.5 - 1.0 ms). So if this is truly a lag value, why is it so small?

Also, what negative and positive values? They should be all one sign, or this indicates that an injector can become predictive and actually start to fire before it gets a signal? I mean how else would the lag be negative? Or conversely, if negative values indicate lag before action, then how would it also be positive which would again imply the injector is firing before getting a signal to?

You must have forgotten one simple thing - in direct injection engines injection pressure is "somewhat" different if compared to traditional MPI. If you take it into consideration you will understand how long does complete injection event take and why values are so low...

I understand that there are 2 parts to the time for injection to take place:
1. time for the solenoid to open the injector
2. time for the fluid to flow from the seat to the holes

Greater pressure will make the flow from the seat to the holes faster once flow is established, but the time to open the injector should be similar.

For the first item, a shorter injection time lag could make sense for the values if we were talking a difference between a gas solenoid injector vs. a diesel piezo, then I could understand maybe a DI injector lag being 10x less in the piezo than a solenoid port injector. But both the port and DI injectors use solenoid technology, so that seems odd (at least to me) that the solenoid in a DI injector opens that much faster. I am not saying this is wrong, but do we have any papers, publications, or other reference on this?

Over the last 2 days I tried changing the values of the positive numbers by a factor of 2 and 10, and both times it started the same in cold conditions. ?

epy

isn't rs4 inj calibrations covered in one of the MED9 threads on here already? iirc it was a while ago. i dont recall TVUB needing to be adjusted, i def could be wrong though.

MED9 doesn't have TVUB, but does have a map that is "kinda" similar, and it is identified in the FR as for injector lag. This is the KLTVTSV map.

However, all the threads on the RS4 injectors don't provide any info beyond "go read the FR" or "you need to adjust angle and lag, now go read the FR". The only other thing ever mentioned was "the info of what needs to be changed is in the RS4 damos, so go read that FR and look at the RS4 damos". But nowhere is it indicated where the damos or map locations in either the MED9.1 file for the 2.0 or 4.2. So I am trying to get threads started that will actually give that info. As a start, I was hoping to get a 4.2 file with the lag map defined, and then also the 2.0 file with the lag map defined. That is what this thread is for. Next I would like to start looking at the angle maps.

Now if you do go read the FR, you see that AWEA has the einspritzen angle and lag maps, and I did track down potential maps locations for these and the kp is posted above. The values in the angle maps look OK to me (300-400 degrees crank angles), but the lag map makes no sense to me.

epy

ya, after searching there's no specific information i can find about exactly what to do and where to do it. this could def be a good source if anyone is willing to share any hard info. i think there's at least one non-pro here thats done this before, hopefully they will share

Well we have limited pro league support, because Taylor was kind enough to confirm the KLTVTSV map address, correct my definition of the axis address, and also provide some insight. The map I had tracked down in the file had the wrong axis address, and that made it more confusing, so correcting it helps a lot. Also, Taylor confirmed that this map does indeed have negative and positive values by looking at other files he has defined. I will post up the new map kp file at lunch.

The interesting thing is I am working on the new RS4 multi-hole injector, so there are additional considerations I am making that may or may not apply to the older single hole RS4. But the approach for tuning for them is still the same. Here is an interesting article that hints at some potential differences.

The first one that I need to solve is a hardware issue. The new multi-hole injectors operate on a 65 Volt booster signal, but I do not know if the MED9 ECU does this or not (it is discussed in the two articles below for the Delphi injector and the Bosch ECU).  Does anybody know what booster voltage the single-hole FSI fuel injectors work at (i.e., MED9)?

http://www.delphi.com/docs/default-source/sae-technical-papers/2014/Fuel-System-Pressure-Increase-for-Enhanced-Performance-of-GDi-Multi-Hole-Injection-Systems-pdf.pdf?sfvrsn=0
http://www.bosch-motorsport.com/media/catalog_resources/Manual_HPI5_2012_02pdf.pdf

And here is the same info for the new Bosch multi-hole injector saying it also runs on 65V like the Delphi. Notice it says that they use them for fuel consumption and performance in the 1L and V8 applications.

http://products.bosch-mobility-solutions.com/media/en/ubk_europe/db_application/downloads/pdf/antrieb/de_5/gs_datenblatt_magnet_hochdruck_einspritzventil_hdev5_de.pdf

I didn't even know they had a new design version! that's good info

They are like the TSI injectors but with a concave nozzle bowl instead of convex.


http://nefariousmotorsports.com/forum/index.php?;topic=7259.0

OK, I think I got a piece of the puzzle figured out for using RS4 injectors. At least for the cold start issue. This is my theory that I am going to test out.

If you look at the start condition, the engine uses split injection. The first injection is around 300 degrees BTDC in the intake stroke, and the second injection happens around 60 degrees in the compression stroke. Now if you think about the purpose of the two injections, you see the first injection gives a very lean homogeneous mixture that produces a small amount of soot compared with a traditional engine using overly rich conditions to start the engine. This won't fire by itself, at least not in a way that would provide smooth starts and a good idle. So you fire a second injection during the compression stroke that concentrates fuel around the spark plug to give it a little boost and idle smoothly with a combined lean mixture for emissions.

So when a person builds an engine with flat top pistons, lower compression, and larger RS4 injectors, you basically mess up this second shot condition. Now the RS4 injectors have more lag at the small injection times for split injection, so first you need a little more compensation during these small injection pulses. To do this you adjust KLTVTSV.

Also the second shot mixture isn't as efficient at being concentrated next to the spark plug. To fix this you can fire the second shot closer to TDC to help keep the charge next to the spark when it goes off. To do this you adjust the two split condition compression stroke maps in the low rpm/low load region. These are KFWEHK2K1 and KFWEHP2K1.

The end result should be a better concentration of fuel around the spark during cold start.

Thank you 99PWR for more information on the RS4 setup.

Now as for the setup on injector angle and lag. Here is the new complete mappack that I have defined so far for the 115G ECU. It has the updated KLTVTSV axis that Taylor fixed for me.

As for results so far. The info on the R8 has a very non-linear form (this is Taylor's, so I will not post it). However, I can say that it is everywhere with more negative values than the FSI values in KLTVTSV. I tried to put them directly into my car, and it did not run very well, so I think at least for the newer RS4 injectors, it is not the correct form. So next I tried to take 0.05 off of the entire KLTVTSV map, and I also advanced the injection timing on the compression stroke for starting rpm and loads. I have let the car sit for 5 hours to cool down, and this has resulted in a much better cold start. I will let it sit all night and see if a true cold start is also helped.

Here is my KLTVSTV

(http://i1073.photobucket.com/albums/w381/ericpaulyoung/Screen%20Shot%202015-06-05%20at%2011.35.00%20PM_zpsqas7xkpi.png)

Here is my KFWEHK2K1

(http://i1073.photobucket.com/albums/w381/ericpaulyoung/Screen%20Shot%202015-06-05%20at%2011.34.26%20PM_zpsrly7bxcx.png)

Here is my KFWEHP2K1

(http://i1073.photobucket.com/albums/w381/ericpaulyoung/Screen%20Shot%202015-06-05%20at%2011.34.41%20PM_zpsusxmzv7b.png)

And here are RS4 maps.

As you can see the new RS4 injectors (with multi-hole tips) do not behave like the old single-hole RS4 injectors.

For people tuning the old single-hole injectors, you should first try maps more like these.

(http://i1073.photobucket.com/albums/w381/ericpaulyoung/Screen%20Shot%202015-06-06%20at%2010.33.50%20AM_zpskkxht0ql.png)
(http://i1073.photobucket.com/albums/w381/ericpaulyoung/Screen%20Shot%202015-06-06%20at%2010.31.38%20AM_zpsumqwyr88.png)
(http://i1073.photobucket.com/albums/w381/ericpaulyoung/Screen%20Shot%202015-06-06%20at%2010.31.27%20AM_zpsluowhlv7.png)

Also, tried the morning cold start, and the maps above with an advanced injection timing for the compression stroke has helped noticeably. I will continue to refine it and post results for anybody else wishing to use these newer style.

This is really useful, is use 079906036D  which are injectors used in Audi and Lamborghini V10, I use them in a 2.0tfsi from an audi s3.

I copied the "lag" maps from Lamborghini, KRKATE I adapted to get  good LTFT.

I guess I need to workd on the inj. angle at cold start.

rgds

I use KRKATE of 0,02381 with the 079906036D injectors. LTFT is around -0,5%

Thank you KmosK04 for the MED17 info.

So the plot thickens, when you look at an MED17 file for a new 2.0 TSI with multi-hole injectors, it looks kinda like a cross between the MED9 RS4 information and the FSI 2.0 data. I think you can't just adjust the tables independently, and you have to consider the combination of how they work together. I will try the whole set of TSI maps in my FSI and see what happens - For science!


(http://i1073.photobucket.com/albums/w381/ericpaulyoung/Screen%20Shot%202015-06-07%20at%201.17.41%20PM_zpslk5tsojk.png)

here's what I use

Interesting. Can you post what you use for the angle maps also?
EPY

So I tried using all the angle maps and lag from the TSI MED17 to see how they match up with the new RS4 injectors. It doesn't work very well at all. I had to crank it for like 30 seconds to get it started.

EPY

I use standard from S3 (I think 70,5 and 60 if i remember right).

I tried today another version of what I found in a Lambo map (inj lag) but it didn't help my slow cold start.

rgds

Try to decrease the value of the compression stroke injection angle around the low rpm/load points.

epy

Does anybody have a damos for a 2011 or newer R8? That car comes with these injectors, so I would love to see what is in that file.

I tried your values fir inj end, no change, still the same poor cold start...

Back to school.

The injector does not close instantly either. The injector offset/dead time is the opening time minus the closing time.
If the opening time is faster than the closing time, you get a negative offset, simple.

Good point. I thought the definition for the "lag" in KLTVTSV was only for defining how much delay to take into account so that the ECU correctly sends a signal to start opening the injector at the correct time, but if it is a combination of both the opening and closing actions, then this makes total sense.

EPY

Oh, I get it. The "lag" is not a compensation as to when to start the injection, but a compensation to total pulse width used. Very helpful info PRJ, thank you. This makes much more sense now.

EPY

Well stated, prj. Best kind of explanation - short, simple.

Yes, "lag" is a misnomer.

"offset" is the correct term for this reason - it is a fixed time that has to be added/subtracted to the effective open time to yield the actual required signal on time.

"latency" is an entirely different animal (even though it is half of the "offset" story).

This matches what you would think physically. As pressure increased on the rail, the opening time should decrease and the closing time increase. So values for KLTVTSV should be least at the low pressure condition and greatest at the high pressure. In the case where there is a switch between the magnitude of the opening and closing times, this would have negative values at the low pressure points and positive at the high pressure points.

God, this map just snapped into focus now that is clear. I have been attacking it wrong, and trying to make changes consistently across the map, but in reality, the line could just "tip" more or less and pivot around the same zero point, or even shift the zero point with a change in slope, or translate consistently. There are many variations :/

Now we just need to find someone with this map defined for a newer RS4 or R8 and see what Audi specifies for these :) Please, I know you are out there somewhere. Help end my infinite flashing of different random map changes and bring peace to the realm.

EPY

I have 4.2 r8 and rs4 fully defined and disassembled, what do you need?

Such good info guys. Just wanted to say thanks for the info.

I need the supply maps, RKTI and AWEA

Is this injector code the Audi R8 4.2L and not the correct 5.2L?

After a lot of research I was able to get the RS4 B7 map pack, it's a gift for you.

thank you so much for this ols, really hard to find for me.

Problem is now bigger questions arise: why is that on the RS4 they use 0.0332 of KRKATE when people say to use 0.023 when fitted to 2000 TFSI???

because KRKATE is calculated for a GIVEN cylinder displacement, **MAYBE***???!!! what's your individual cylinder displacement on the 4.2 vs a 2.0?

being a V8 I would say it's 525cc vs 500cc on a 2.0 TFSI. only 5% different.

What?

can you be more specific in your question..