NefMoto

Hey everyone,

I've recently sold my BT 20th GTI, but I'm still tuning it for the guy who purchased it from me. Im having an issue with the car losing o2 corrections under WOT, and I cant figure out why.

When I say it is losing o2 corrections, I mean B_lr=0 within a couple hundred RPM after going WOT. fr_w=1 at this time, and for the remainder of the pull, actual lambda reads full rich (.75), and target lambda hangs around .78-.79. This is on wastegate spring pressure ~10psi

Things that have been checked-
Fuel pressure is solid up to 35psi of boost (85psi rail pressure)
There are no intake leaks
There isnt a lean condition prior to the ECU dumping fuel (STFT is at ~.90-.95)
MAF output seems to be reasonable
O2 sensor had 3k miles on it when I noticed this, new owner replaced both sensors about a month ago
No exhaust leaks
All emissions deletes are resistored and coded out properly
Injector duty cycle is nowhere near max (USRT genesis II 1000cc and the car is probably putting down 250 wheel at this boost level)
Fuel trims are within 2%
All fueling maps are set to .85 to rule that out

Noteworthy information-
Car is a 2003 AWP 6MT
Original ecu was maestro tuned, something shorted out and fried the throttle drivers (possible o2 sensor)
ECU in the car currently is from a 2002 AWP 5AT jetta (032HF? cant remember) and is flashed with 06A 906 032 HS 0007 software (manual trans)
This issue did not present itself on my maestro ECU prior to it frying itself

I have been doing some reading, (okay a lot of reading) and there doesnt seem to be much info on the topic
I managed to dig up this thread where phila_dot mentions B_lr being inactive if lambda request doesnt fall in certain parameters, but I cant quite make out how that would apply in my situation. All other mentions of B_lr seem to be in reference to tuning partial load fueling via FKKVS/KFKHFM or other fuel correction maps.
http://nefariousmotorsports.com/forum/index.php?topic=2515.0     <-- (reply#8)

Does anybody have any insight into this issue? Any ideas as to what I can recommend as a solution? I cant help but wonder if its an issue with the o2 sensor wiring or the ECU being incompatible with the car for whatever reason. I may need somebody to school me on how and when ecu part numbers can be swapped around, and what software can be flashed onto any given ecu.

edit: Im having trouble attaching logs. heres a link to the log on my google drive. hope thats okay
https://drive.google.com/open?id=1fIgH9-kkWRC_lRwST72DpBIHwwfg4S4e

Logz

What percentage is lambda control at a full throttle pull? I'd start looking at your injector calibration. You need to dialed in kinda good with big injectors off 15% can be over fueling by 150cc's what's your long term fuel trim at?  Honestly you need to provide more relevant info. The sci should be fine any 01-05 me7.5 a chassis sci are the same at a hardware level. Although the software version yours using is for a 5 speed and yours is 6. Aside from not having crude in 6th gear and a few other annoyances it's fine.

Also that thread you linked is about narrowband lean burn. Not too much info relevant to your issue imo

Long term fuel trims were both very well sorted as calculated. Partial was -.1% and idle was 1.78% when I had him check last.

Log is attached below. Im still working with him on logging at relevant times and good rpm intervals (Im well aware that 1000rpm is hardly a good WOT pull) but he lives in the city so i have to work with what hes sending me. I can confirm that this issue was exactly the same when I owned the car, and B_lr would turn to 0 shortly after going WOT and remain 0 until I lifted. Logs essentially looked the same but had longer intervals where this was occurring due to me being WOT for longer

And yes I'm aware that it is about trying to run lean on narrowband cars, just saw that B_lr=0 when those conditions arent met and wasnt sure if that somehow applied to this case. The issue here after all is that the car is ignoring the o2 signal and just dumping fuel under WOT. Hopefully the logs posted above will be helpful to finding a solution

I still don't see any logs? Also have you tried another ecu and or another software version? I know from first ha d experience There is a broken version of the HS version that has o2/ fueling issues. Try another software version and just post over the basics maf, injectors, etc and see if younger the same result.

BL_R is supposed to go to 0 when WOT. Would you want a bad O2 controlling your fueling?

Start tuning your fuel to achieve your target.

pardon me, isn't WB me 7.5 supposed to work in closed loop even at WOT?

No. Not based on my experience or research for the reason I mentioned.

So when doing a wot log, watching block 001 shows no percentage correction?

Dont know why the log isnt attaching properly. ive tried 3 times now..

And I guess Im confused because this is absolutely not normal. If you can see the log you will all understand what I mean. And afaik, my car has always used o2 feedback on WOT. I guess I shouldnt say USED. It has always at least displayed a STFT. As of now, going WOT sets STFT=1 within 200rpm

To those saying to tune better, how exactly do you propose I do that? Fuel trims are spot on, and under WOT I have no idea if my STFT is lean or rich, its locked at 1. And AFR request follows something far richer than what is written in the tune..  The only clue I have as to if Im rich or lean is during tip in i have slight -5-10% correction for a split second. Im not going to blindly tune my fueling based on guesses. I have started from scratch 2 or 3 times with bare minimum tweaks to a stock base file (KRKTE, TVUB, MLHFM etc to account for different hardware) and this problem is always present.



edit. Im unable to attach anything for whatever reason. Log is posted on my google drive now. Hope thats okay

https://drive.google.com/open?id=1fIgH9-kkWRC_lRwST72DpBIHwwfg4S4e

Tune better is the correct answer. If you do, you will find that your correction will reactivate.
There is a limit to o2 reading and within what correction is active.

This is not preventing you from actually tuning in any way. If you are mega rich on boost with 1.00 fr_w, more than likely you have a boost leak or incorrect scaling.

As PRJ stated, you're flagging conditions that are locking the closed loop feedback.

Dig into the LRSEB function, and you'll find your answer ;)

Except there isnt a boost leak, and its rich because the ECU is requesting it. I dont have a request of .85 and actual of .75 or anything. I actually dont even know where the .75 request is coming from, only that it activates as soon as the throttle is wide open and B_lr switches to 0  ???




Im not claiming to be incapable of error, but can somebody explain this to me a little better. I like to get scientific about stuff like this and narrow things down methodically. Im clearly a little lost, hence the questions. Heres been my thought process while trying to deal with this:

-In stock trim, a wideband (not sure about narrowband) car should not see B_lr=0 on a WOT pull
-The intake has been leak tested up to 50psi (its a simple intake system. Brake booster, FPR, boost gauge, and BOV are the only accessories running from the manifold)
-The MAF seems to be working accurately.

SO, on an otherwise stock tune, with KRKTE and TVUB properly scaled for 1000cc injectors (KRKTE~.034), LTFT of less than +/-5%, boost set well under the MAP limit, and STFT active under normal driving, you might expect WOT fueling to at least be close enough to show high corrections in one direction or another, right?
Well that isnt the case. So I then scaled MLFHM down in incriments of 5% (yes I offset before and after scaling) down to a max of -30%. Still no STFT on WOT. Obviously I'm seeing large LTFT at this point so I reset the MAF curve and scaled down KRKTE instead, ending at .025ish and still didnt see any improvement.

I just fail to understand how this would be an issue of dumping too much fuel and causing lambda regulation to deacativate, or even conversely, not enough fuel causing lambda regulation to deactivate. Prior to this, calculating KRKTE and TVUB has always been very close and at least allowed me to see correction. And for anyone thats wondering, the plugs didnt have any indications of a rich mixture  :-\





As for digging through LRSEB, I'll do my best to attempt that, the german isnt too much of an issue, but I cant make heads or tails of any of the symbols in the function framework. Is there anywhere I can look to make tracing a variable through all these functions actually make sense? At this point I cant tell how any functions actually effect the value of B_lr or any other variable for that matter. Only what functions they belong to and where they seem to be passed on to afterward.

EDIT: Okay, I found the symbol defs at the beginning of the FR. Ill get this sorted hopefully...
 

You are focusing entirely on the wrong thing.

Why is fr_w 0.90?

Forget about request, fr_w goes VERY wrong long before request goes rich.

So what I'm seeing by looking at the function is that for B_lr to be 1, B_lrs must also be 1, B_lrms must be 0, B_dslsk must be 0, and lamsons_w must be >=LALIUSMN correct? Im not sure where to take it from there



Perhaps its that most of my tuning experience was in maestro, but wouldnt that indicate it is attempting to pull 10% fuel as soon as the car goes WOT? Cant the o2 sensor correct up to 25% which would correlate to .75 in fr_w.?Obviously thats not a number anyone wants to see, but am I wrong?

Either way, I will continue digging through the FR but at this point I feel like a third grader reading about string theory

If you have 0.75 request then obviously WBO2 is off, what do you expect?

Ah well, what a waste of time thread.

Thanks for motivating me to dig through the REAL FR and not the translated portion. I actually learned a lot from this experience  ;D


Yeah, thats true. My question wasnt so much "why is my request stuck at .75?" as it was "why is my lambda reg off?" because I kind of made the assumption that one was the product of the other. If you cant tell by my post count, I'm fairly new to all of this and learning as I go so my sincerest apologies for wasting everyones time with a question. I thought this was a place to for people to learn.

Anyways, I may have found the issue. Spoke with the owner more and I guess he somehow bricked his ECU last month and had a forum friend or someone he knows from a car meet make him a file to get the car started. MAF was scaled for 5v=2300kg/hr. Sent him a new file which will hopefully have o2 feedback at all times

I missed it. Was this a hardware (bad 02) or software issue?

It was a few things. Hardware is fine, MAF was overscaled. Now it seems like no matter how far I scale the MAF down, I can't get ps_w below pvdkds_w on WOT. It's a v8 MAF scaled to 1400kg/hr at 5v currently and I'm still seeing max ps_w in areas and at least a 300mbar difference through the pull. Digging through FR currently to see what else could cause ps_w to be too high.

On the plus side, o2 regulation is looking better. It's still turning off eventually, but not until at least 1000rpm into the pull as opposed to immediately

BGSRM ;)

Still digging through the FR but I'm new to this so some of the symbols are confusing the hell out of me. I think specifically, youre telling me to focus on BGSRM-BPS, which it looks like is the function where ps_w is calculated. Im unsure if thats correct, but if so, Im having trouble deciphering the functions. I dont quite understand how the accumulator works, and the symbols next to it are not explained from what I can tell (compute/10/syn)


Any insight on how that translates to a mathematical function that I can wrap my head around?