NefMoto

Dear Forum Members,

I'm trying to tune audi s3 8l.
It is the 2.0 stroker engine, fully forged with 100mm MAF, 1050cc injectors, 78mm throttle.
At the moment I'm trying to tune idle. I have been reading a lot on this forum about how to tune ME7.5, learned bunch of things and understood two main points.

1. ps_w should be lower than pvdks_w.
2. mlhfm_w and msdk_w should be as close as possible. :) 

I have calculated KUMRSL, KRKTE, MAF linearization, rised idle RPM but the car idles still very bad.

I have spent many hours and tried different things to understand what could be wrong but no success.

PLEASE HELP !

I'm attaching the log(pressure variables are doubled in value that is because I was preparing the .ECU for 5120 hack), original and the modified file.

Thank you in Advance,

IMHO you should do yourself a favour and do things one by one as it would be really hard to properly dial in higher engine capacity with custom MAF and non-stock throttle body. For now all models within ECU are screwed up and you started to prepare for 5120 hack?

First of all Thank you for your reply.
No I just edited the .ECU file to understand things.
What do you mean one by one?  The car was build like that. How should I start from stock injectors and throttle body?
I thought that injectors and MAF is something I could calculate. I could start from stock TB theoretically.

What do you suggest??

I would suggest starting from as much stock hardware as it is possible because you still have more than 10% more displacement which should be tuned first. Next could be anything but 1 thing per iteration - it does not matter if it would be MAF, throttle body, injectors or anything else.
Now you do not know which parameter is good and which not so you could compare 2 bad parameters thinking everything is good because they are close to each other.

Thank you again! :)

Let's assume I have stock hardware like MAF, and TB in place. What is important beside KUMSRL and KISRM for more displacement??

You forgot to mention fuel trims.

Hm.. is this some kind of hint to my problem? :D

Yes because with all those hardware changes you have you need to be sure everything is in check as you may screw ECU models to a level where everything checks but still you are running bad fuelling and ignition ;D

Hello experts,

New day new luck!

I was told that the values in KFMSNWDK/KFWDKMSN are describing the TB and are not vehicle specific.
I have looked at KFMSNWDK/KFWDKMSN from 1.8t stock files. As far as I know the TB are the same on 1.8t ME7.5 vehicles so logically the maps for TB should be same but surprise surprise ! They are different for example in 225 BAM and 180 AUQ.

Could somebody please explain why this is like that.
If this maps are describing the TB and are not vehicle specific they should match.

Thanks in advance,

I can help if you like. looks like you have most of it done.

Hello,

Yesterday I have done the following:

1. Fitted stock MAF.
2. copied TB maps from the ECU where it was used.(it turned out that this TB is from VR6 TUAREG so ME7)
3. adjusted idle torque reserve because engine load requested was only 21 and the car run unstable.(I tought it was not a bad idea to adjust this because light flywheel is installed.)
4. calculated KRKTE@50PSI = 0.031619962.
5. Copied TVUB@50PSI from injector datasheet(attached).

This gave the result that the Idle was very lean and I have had to adjust TVUB up(which is not right I think, do I need to adjust VE model instead?) in order to meet lambda request?
Even now idle is more stable ignition is still oscillating.
I have tried to track down where this ignition oscilation comes from. My german is perfect so started to study FR further.

zwist follows zwsol.
zwsol brought me to etazws.
etazws to mizsol_w.

miist_w follows mizsol_w

mizsol_w to mizsolv_w
mizsolv_w to misolp_w + dmllr_w
misolp_w = miopt_w*etazwb
miopt_w  = mioptl1_w * ETALAM
mioptl1_w = is the output from KFMIOP
etazwb = output ETADZW
ETADZW uses variable dzwb = zwoptl1 + dzwotm + dzwoag + dzwol - zwbas
zwoptl1 is the output of KFZWOP/KFZWOP2.

I don't think that ETALAM, DZWOLA and ETADZW are the maps to change.

Does my current problem comes from KFMIOP/ KFZWOP or there is something hidden deeper that I could dive?!

To summarize I have three main questions:

1. do I need to adjust VE model in order fix AFR at idle? If not why are the specified TVUB off by that much?
2. How to scale maf correctly? I have tried to scale it manualy: substract offset 200 scale by increase in area. If I scale from Stock s3 maf to 98 mm housing I get different result than when I scale from RS4 83 mm maf. Scalling up from 83mm is more acurate I think. Anyway It is still off because it reads more at idle than stock configuration.
3. do I need to tune KFMIOP/KFZWOP to have better idle timing?

I'm attaching injector datasheet,  log and updated bin.

As you all can see I'm not asking anybody to serve me a ready solution so it would be nice if somebody could give me the right direction.

Thank you.

Hello,

Nobody replies so I have decided to share my findings.
Beside MAF linearization my problem was the IDLE CONTROL. In particular module LLRRMM.
It is controlling IDLE by controlling torque via ignition angle.
dmllr_w is the varialbe that affects mizsol_w.

I could not imagine that IDLE PID was the problem. Because the engine is heavily modified LLRRMM need to manage less torque to maintain steady IDLE. The stock idle PID was overreacting because the angular momentum and produced torque of the modified engine differs from the stock. It looks like the IDLE PID was constantly increasing and decreasing torque by decreasing and increasing igniton via dmllr_w and the increase/decrease steps where to big . That is the whole story.

I have recalibrated the IDLE PID and now all is fine again. Ignition angle does not oscillates and the RPM is steady.

Still I have the problem with MAF linearization. As I wrote in my previous post I have used already all kinds of formulas to linearize my 98mm maf housing. No matter how I calculate it, it is still not right. Is there a specific approach to linearize custom maf housings?

Thank you


 

Flow bench or datasheet.

Thank you for your reply. Flow bench is not a option for me. Could you please tell what datasheet do you mean?

Datasheet for the MAF sensor you are using. If you are using a stock MAF sensor in a large housing, make sure the output is scaled correctly for each voltage input compared to stock, including MLOFS.

With non-zero MLOFS its a bit trickier but you should able to use matlab or rudimentary algebra to check your results.

Well as far as I know, I have to scale stock MAF linearization by increase in area. I don't know anything about scaling MLOFS. Could you please tell what should happen to MLOFS?

Thank you again

Can you specify what you did to the PID?

I would really like to know as well, my car have the exact same problem.