NefMoto

Hi!

ive been tuning my R32 turbo Project recently, and i think i am about 99% done with the software.. fueling(maf, injectors, afr) is 100% perfect, timing tables made in dynapack and so on.

now ive run into a problem with the fueltrims over time, which seems to be a known issue.
after doing some reading, it seems to be related to both widebands beeing fitted into the same pipe, which makes the trims go crazy between the banks.. and so i would like to remove the b2s1 probe and utilize just b1s1.

i read a thread where Phila_dot did a quick fix for this and made the b2s1 read the b1s1 input, by swapping adresses at 3 locations.. is this a working solution? (injectors, MAF etc are fully calibrated, and intake is flowed, so no compensation is needed between the banks)

any help, ideas, suggestions is much appreciated  :)

I think there is only a quick fix for narrow band sensors.
But im pretty sure there is a finished code out here, maybe someone will share with you :)

I have the same problem I need a tune with mono Lambda :(  Please tell how you solved it!

Hi all,

I've been reading a lot here and on s4wiki.com and found that the more you read the more you realize is still to learn...

Anyway - I have turboed an Audi A3 3.2 with DSG and got it mapped by a good tuner in Stockholm, Sweden.
We have searched and tried all leads and suggestions we have found but not succeeded getting the system to act as one bank.

Has anyone found a working solution our problem?

// Björn

Im still looking... some S4 tuners use single turbo application. I also think Fotis @ revlimit can fix this.

A simple code patch is what you need. The information about this can be found in atleast 2 threads here on nefmoto.
Which sthlm tuner are you working with?

Hardly a good tuner then if can't do such basics.

ooh, another A3 3.2t in Sweden.. 8L..?

are you running the 032CN software?

contact me in pm.

The number of tuners who can do code patches you can count on one hand.

"good" generally means "somebody who won't fuck things up when only altering maps" IMO.

In any case, it would be good if the patch information were put in a single location somewhere on Nef (or the s4wiki)... because otherwise it is pointless to complain if somebody asks.

The whole point of a FAQ is to have a centralized location to point noobs to.

Bitching is ONLY valid if the FAQ is reliable.

Count me in.. im in same position..  Lets Coop on this..  PM me you mail adresses so we can cooperate.. I have come far with this tune my self and found some problems in my tune when i was logging.. But we have diffrent ECU.. But we are in same boat with the O2 sensor.

Yes, and those are good tuners. Everyone else are box shifters not worthy of the name.

That would be a good idea..

I find it hard to see how this kind of comment could be considered useful.

"If you are not going to say something nice, don't say it at all."

As difficult as it might be, you have to filter out prj's useless comments from his very useful ones.

There have been discussions about mono Lambda here

http://nefariousmotorsports.com/forum/index.php?topic=2112.msg26863#msg26863

Never forget that prj is not just an asshole, but he is an asshole who doesn't understand the difference  between tuning or calibrating maps within software and writing code or programming. Or so it would seem...

Meh. He just thinks that if you can't do custom code in ASM you aren't a tuner.

Meaning that there really only are a half a dozen ME/MED tuners on the planet.

This forum has made all of us better at what we do either with our personal effort or from info we found here..All of us have contributed some way and giving names to members that have contributed is going to brake a certain peace that exists on this forum.....

So in my humble opinion leave anyone be as long as he is not violating any forum rules and contributes or has contributed on this forum....Leave the bad names out of this...This is the only forum that is left for real tuners helping each other.

Dont forget its fun to.. Everyone wants to bee an "Jedi" but training and patient is a must.

I was on an interview at a large "automotive company" couple weeks ago. They need new engine calibrations engineers. They said, well  almost: Heres a bucket and a sponge... You can start cleaning the floors..
And then i looked around.. testing rooms, engines and engineers running around everywhere in hyper stressed situations... I thought to my self.. no way.. Ill stay where I am and play with my car in my own garage instead :P  lol...  And thanks to this forum i cn play with my cars :P Many ppl in this forum are way better on engine control than many of the engineers on the car companies. (Well they work for emissions though soo probably not fair to compare.)

You guys are saying tuner? Its called calibrations engineer ??

Everybody has different definitions.

At a car company, the calibration engineer does not write ASM.

Indeed they don't. But they are not "tuners".
If someone can not write custom code, they can not do a proper R32 turbo software for example. Only a bodge.

There are a ton of people who can shift some maps, there are few who understand the ECU well and know the internals.
Knowing code also helps you to take any software and find maps or reverse engineer stuff.
The real tuners started a long time ago and without any documents, everything was worked out through reverse engineering.

These days there are a lot of kids who buy FR and Damos and think they are god, well it is not so.

Hate all you like - but there are very very few "good tuners" on this forum and in the world. Most of the better ones actually visit here.

Two main things...

One is that you yourself seemed to advance rather quick. I will state I make the assumption you payed for a lot of help too. You talk huge for a guy asking basic stuff in 2011. My other assumption is that you are someone who was good with code and then you became tuner god after discovering this stuff called cars. As someone who was born into cars and did them as a life before getting savy on the digital side, I can understand your negativity in a opposite direction however I do not condone it. You talk to everyone like they are a caveman hack. On the flipside, where are the big hp prj built cars that look nice and inspire other builds?  

Anyone who seeks out to make good files on their own as a tuner or calibrator can seek out and pay for custom code for their customers needs. Focusing on your skill sets and knowing when to get help are part of running a good business. The end product to a customer is all the same.

I think if you had better people skills you could go pretty far.  

I've surely gotten quite far when you seem to be so obsessed about me. It's amusing really.
Keep going.

None of your assumptions about me are correct besides the fact that I've been around code for decades.

you sound like you're 60 years old prj :D :D :D

I actually came to the same impression after re-reading what I wrote Mihnea.

You are far less of a daily thing for me than you would hope. I have never seen or heard of you outside this forum and I try to visit here a few times a week. More often than not I see you typing some dick head posts to people so you get my attention at times.

I should forward you the random pms I get everytime we bump heads. You rub a lot of people the wrong way and it seems to really make their day when I tell you what's on my mind. I even had a fellow American say he wanted to beat you up, which I thought was a bit much but it still made me chuckle. You sure do leave a lasting impression.

Job well done then. Couldn't care less what anyone on here thinks about me :)

So why are we having this conversation then?

Did you ever tell this guy the locations of what variables he needs to change? I am sure for someone like you it would be a 5 minute thing.

He needs code mods so both sensors read the same signal.
It is pretty straightforward to add code mods to copy over the readings and fuel trims from bank 1 to bank 2 using the information freely available on this forum.

If one does not want to do this themselves, they can pay someone to do it.
The offsets and code are different for every file, because RAM locations are different and code locations are different. I've done CN before, but it takes longer than 5 minutes.

Been trying to fix this on mine aswell, And as PRJ says its a code job.. So the ASM files is needed i think to reconnect som signals..
Found soomething in FR but im uncertain..  Theres 2 signals connected to two functionblocks  "uulsu_w"..  Note sure if thats correct signal from LSU Its maybe the Heat circuit out something not sure?.. But since i dont have the code i cant do anything anyway.. its a bit frustrating..  also found some signal LAMSONI_W and LAMSONI2_W..  My german is not so good.

From wiki.. But every ECU has its own adress for these and the functions can be diffrent aswell on each ECU.. "I think"....

Delete B1
0x3C8BE
0x3C8D0
0x3E832

Original value will be 1A91, to replace with B2 input change to 1891.

Delete B2
0x3C8F0
0x3C902
0x3EA90

Original value will be 1891, to replace with B1 input change to 1A91.[47]

There's alot more to it on wideband ECU's

Nobody "has the code"

They use a disassembler and hand code the patch.

Exactly. And this is done per file revision, because the offsets are different, so you can't just copy paste.
The code is a bit different on different ECU's too.
Narrowband and Wideband was already mentioned, but there are differences on Wideband ECU's depending on whether it is C167, ST10 etc.

What needs to be done is not a secret at all - copy Lambda sensor readings and all the fuel trims from one bank to another.
But it takes some effort, and no one is going to sit down and spend that time to do your specific binary for free.

If you can't do it yourself, pay someone to do it.

jim_coupe: you running the 032CN software?
someone who called me a few weeks back mentioned you.. i think.. if youre from sweden that is..?
shoot me a pm, i might be able to help you out  ;D

I have a newer version mine is from BUB engine 8J 2007:  ECU 022906032H

I also got a BUB 8J and still now can't find a way to convert it to mono lambda.

I got the neccessary ram adresses and memory mapping out of a etas a2l project and think they should be correct. Some Variables are stored in ram adresses like 0xFXXXX i cant find any of this in the ida project as this adress space is empty! And then there are the other half of the variables (mostly uninteressting ones) wich are stored in 0X38XXXX wich i find all in the ida project. Sadly all lambda related variables are in the 0xFXXXX area. I tried many different methods to get it to mono. I searched for the lambda regulator variable , for the injection time, all no success as they all lay in the 0xFxxx area wich give me 0 results in ida not one pointer/jump insctruction points in this memory area.
On the older CR167 Ecus i got no problem to find everything in the ida project. But on those ecus the variables are all in the same space not divided in 2 seperate memory locations.

My project constists of the ST10 processor code and the external 29f800 flash code. Maybe i miss something? In the datasheet of the ST10 they mention a very small xram area maybe its stored inside this and i cannot read it?  Did anyone successfully find a solution for the ST10 R32 Ecus and can help me out?

This is my memory map  :

    RESERVED FLASH INTERN 0x0 0x8000
    RESERVED FLASH INTERN 0x18000 0x8000
    RESERVED FLASH INTERN 0x20000 0xB0000
    CODE EPROM EXTERN 0x800000 0xE0000
    DATA EPROM EXTERN 0x8E0000 0x20000
    VARIABLES RAM EXTERN 0x380000 0x6000
    VARIABLES RAM EXTERN 0x386000 0x2000
    VARIABLES RAM EXTERN 0x388000 0x200
 
    As you can see the Variables from 0x38XXXX  are existent but the ones with 0XF000 are not mapped. Maybe i need to convert those Adresses with a offset ?

Regards

Intressting work there... I would like to help and work to find out more about these ECU´s but I have been to busy lateley.. sucks*

I'm building a 8J BUB turbo ATM. When get the hardware done i let you know if i Can get it working.


Verzonden vanaf mijn iPhone met Tapatalk

With which ECU software version?

022906032HJ 384450

Very late reply, but better late than never, huh. ;)
Your memory map lacks the XRAM which is located exactly in the 0xFxxxx region. Look up ST10 documentation or spend some extra time with that A2L. Also you need to set up default DPPs differently for the ST10. Search for DPP writes in the code to figure it out.

Haha this is a good thread, I feel PRJ is the wizard of oz. I think he is behind
so much of whats out there is not even funny.

The problem is that the st10 documentation doesnt fit to the a2l memory map. Or atleast i dont find the link between them.

A hardware solution where you mod the ecu to send the same analog signal to both wb a/d inputs would do the trick as well. Although the sw mod is nicer but the result would be the same.

NO, that will be not be a reliable solution.

IF, changing the ref to stored AD-values so that the code ref the true value of just ONE sensor as suggested earlier in this thread really is a solution to the problem, then hardwire the AD-input of WB sensor2 to AD-input of WB sensor1 would effectively acomplish the same thing.

I'm not saying that you should just connect the ECU input pins together for both WBs, what i mean is that the input on the CPUs AD-channels should be modified as described. (Soldering is indeed needed)

Or have i missed something here?
 

There is only one ADC, they will not be sampled at the same time.

It won't, the fuel trims will still go -25% +25%. I don't think you understand why this happens, so your solution doesn't work either.

Of course, you are correct prj.  :)
I was blinded by the c16cx hw architecture facts and those are not really worth being mentioned compared to how the actual function works, hardware would have had minor effect compared to the code.

Yes thats right it doesnt work to feed the wb Sensor 1 signal to the AD from Sensor 2. That mod can be easily done with some jump wires and some cuts on the two wb02 IC's but it doesnt work it went straight to -25% +25%. I even tried to set any parameter wich was bank selective to equal but it doesnt help.
The only hw mod wich would work would be to use the final injection signal from injector 1+2+3 and feed this delayed to 4+5+6 but this will be a quite big circuit to add to the ecu and i don't trust it. A SW solution wich feed the injection time to 4+5+6 would be more effective. Or maybe it would be possible to buffer the wbo2 result and double it to supply the ADC with the same value.

The SW solution is to force bank 2 trims to be equal with bank 1 trims.

Basically copy over these variables:
fr_w
frm_w
fra_w
frau_w
frao_w
rkaz_w
rkat_w

Think that was it from memory.

Yes , on older cr167 even only cloning final Ti bank1 to Ti bank2 worked well. But on st10 without correct memory map and correct dpp's i will never find anything :D Wich is why i'm stuck.

You are not alone im stuck on this to :)

Just be like me and build a new manifold with seperate lambdas, one days worth of work beats a couple of months worth of work lol lol

What do you do about the post cat/turbo sensors?

Code them out?

Depending on your local emission test requirements (if you have any) you won't be able to pass that way.

Oh sheeeeeet, ya sorry forgot I live in like a 20th world country lol

Obd and emission tests passed with 2nd lambdas unblugged.

Not in New Jersey to my knowledge. Like I said. Local. Every state in the US has different standards.

haha Man :)

Nm