Car is running lean on light throttle - any ideas?

[TAR]

thanks for keeping this thread alive Dave.... and enjoy the Guinness

Not had time to look at anything on the car, but please keep the ideas coming.

[fassi1]

Unless injector scaling address needs to be changed aswell as injector scaling figure.

[lateshow]

10306 is the right address for injector scaling!

[fassi1]

At xml definition injector scaling address 10318,
ROM injector scaling address 10318
Hmmm both are the same.
Do both need to be changed to 10306?

[fassi1]

Big thanks to @lateshow and @foxdie for helpful advices.
Finally managed to sort out occasional lean idle.
Having injector size scaling DSM 409 and injector size scaling address 10318 in xml definition file vr4base.xml
changed injector size scaling DSM to 376 (that gives EVO 344) and injector size scaling address to 10306 in vr4base.xml file, injector size scaling address has been changed automaticly to 10306 in rom file.
As @merlin stated eddress value is an important air flow parameter.
http://www.clubvr4.com/forum/showthr...jector+scaling
Quick rom upload and straightaway noticed that around 10% been knocked off of the STFT which gave me 5-7%, LTFT remained 4.6875%.
Other thing I noticed is richer open loop which is now very close to stock fuel maps.
Borat would say "Great success"
Dave that bell was reliable

[TAR]

LOL Chris... glad you got yours sorted, just wish I could get mine working as well.

[Davezj]

Chris @fassi1 @lateshow does this mean that the injector scaling address is wrong in avert bodies def file.
Are you using Kenneth's rom and def file, or you using a home grown one of you own.
Does Kenneth's rom need to be updated.

[fassi1]

Im using home xml definition which had injector scaling address 10318. I edited it and changed address to 10306.
ROM I'm using is the one i read from my ECU and the only thing I changed is injector size scaling to 376.
I didn't use Kenneth's ROM.

[lateshow]

Chris @fassi1 @lateshow does this mean that the injector scaling address is wrong in avert bodies def file.
Are you using Kenneth's rom and def file, or you using a home grown one of you own.
Does Kenneth's rom need to be updated.

yes the original vr4base has the wrong value, atleast in 20030013 has 10306 and many others. Kenneth's definition has this one right as it is 20030013 based mod.

[Davezj]

cheers chris, thanks for the info.

[fassi1]

Just gone through warm up process. Cold start and left it idling for about half an hour.
LTFT remained 4.6875% like before rescaling which is maximum but STFT started from around -2% and was getting higher with engine getting warmer reaching the highest around 9%. That means scaling still could be lower than 376.
Looks like most of this cars suffer from high positive fuel trims, remember pezza's VR4 had high positive LTFT and STFT but still within the range to keep it in closed loop
with AFR 14.7.

[disco-ian]

Hi all,
OK when I first bought my car I had similar issues (obviously I only knew about them when I fitted the AFR) I was dipping to 18's and 19's when at low rpm cruise.

Turned out to be the air flow meter. Not sure if you have a spare there to try but might be worth ago.

I also had the car cut out on me at low rpm a couple of times when we were in the extreme summer heat we got a little of last year and again changing the air flow meter fixed this.

[TAR]

I finally had some time to try a few things out but it gave me even more grief!!

To begin with I replumbed the vacuum line to the FPR as it was looking a bit suspect - no change
Reset the FPR to ensure it was correct - no change
Changed the MAF - no change
Tested the TPS and reset the IPS - no change
Changed the stock narrowband lambda sensor - no change

Then, all of a sudden, it started to run 17.7 to 18 all the time, except on boost when t dropped to 14's - WTF! the car was running like a dog, so I rechecked everything I had done and reset my wideband just in case, but nothing made any difference.

By this point I was very frustrated.

I have now spent the weekend trying to see what's going on. Looking at Evoscan, I could only see 0.03V to 0.05V for the O2 sensor, this didn't fluctuate much either. I checked the wiring and also wired in the narrowband emulation from the wideband but it still didn't make any difference. I was scratching my head now and getting very hot.

I then checked what would happen to the lambda if I disconnected the MAF as I know it should over fuel like mad..... and it did, but the O2 reading in Evoscan still didn't fluctuate very much, it certainly didn't hit 1.1V at any point. The fact it now ran rich proved the fuel delivery and the ECU's ability to fire the injectors sufficiently.

I connected everything back together and tried revving the nuts off it... don't know why but it seemed like a good idea at the time.

Disgruntled, I locked the car up and walked away.

A few hours later and back in the car, strat it up and BINGO!! its running just as it should without leaning out and idling at 14.5 to 15.1.

I have no idea what has happened but I'm really hoping it lasts more than 5 minutes. It's now running with the narrowband emulation, which I'm not really happy with as I'd rather run the stock O2 sensor but I'm too scared to swap it back!!

If anyone has any theories why this has happened... you know what to do

but very hot and now very sunburned.

[Davezj]

Good round up, glad you remembered to update this thread.

Could you have had a sticking valve or blocked injector, o2 sensor or some thing like that, that has now cleared itself.

I personally do an Italian tune up (5000rpm in 2nd or 3rd for a couple of miles done the road) every so often to clear out the pipes, I don't do many miles and the car is hardly up to temp when I get to work in the morning.

Just a thought.

[wintertidenz]

I had really weird problems with my Innovate wideband - it liked to lose calibration, which would make the car run super rich one time, and then another time way too lean.
I figured out after I removed it that the turbo timer was causing a power loss for a fraction of a second, and upsetting the calibration - it may be worth checking if this is the case.

[adaxo]

I finally had some time to try a few things out but it gave me even more grief!!

To begin with I replumbed the vacuum line to the FPR as it was looking a bit suspect - no change
Reset the FPR to ensure it was correct - no change
Changed the MAF - no change
Tested the TPS and reset the IPS - no change
Changed the stock narrowband lambda sensor - no change

Then, all of a sudden, it started to run 17.7 to 18 all the time, except on boost when t dropped to 14's - WTF! the car was running like a dog, so I rechecked everything I had done and reset my wideband just in case, but nothing made any difference.

By this point I was very frustrated.

I have now spent the weekend trying to see what's going on. Looking at Evoscan, I could only see 0.03V to 0.05V for the O2 sensor, this didn't fluctuate much either. I checked the wiring and also wired in the narrowband emulation from the wideband but it still didn't make any difference. I was scratching my head now and getting very hot.

I then checked what would happen to the lambda if I disconnected the MAF as I know it should over fuel like mad..... and it did, but the O2 reading in Evoscan still didn't fluctuate very much, it certainly didn't hit 1.1V at any point. The fact it now ran rich proved the fuel delivery and the ECU's ability to fire the injectors sufficiently.

I connected everything back together and tried revving the nuts off it... don't know why but it seemed like a good idea at the time.

Disgruntled, I locked the car up and walked away.

A few hours later and back in the car, strat it up and BINGO!! its running just as it should without leaning out and idling at 14.5 to 15.1.

I have no idea what has happened but I'm really hoping it lasts more than 5 minutes. It's now running with the narrowband emulation, which I'm not really happy with as I'd rather run the stock O2 sensor but I'm too scared to swap it back!!

If anyone has any theories why this has happened... you know what to do

but very hot and now very sunburned.

@TAR, did you ever manage to find what was causing this problems as it behave very similar to my car atm

To avoid starting new thread I join here, if I may

Its used to happen occasionally that my car run lean on light/part throttle, but since I don't use it daily, rather once a week now, this issue become more common, to this degree that it's bad most of the time now. On idle afrs are mostly nice and steady (14ish) but sometimes it start leaning out to 18ish and car trying to stall then its some kind of a 'click' and all back to normal.

To be more precise, when cruising around on light load afr go up to 16-18, I assume only in closed loop as when I put my foot down its behave as it should, ie afrs drop to 11-12s.

So far I did this

swap MAF- no change
swap bov from forge to stock- no change
recalibrate wideband- no change, well it was normal for first few miles then start leaning out, then back to normal and again and again
check all vacuum/turbo pipes, all seems to be ok

Im getting annoyed and quite desperate to sort this out as I really want to go on RR day this sat, so any clues will help.

On the side note, I'm running @Kenneth rom and something ring the bell in the back of my head that leaning issue was mentioned b4 by someone on here but cant find actual thread about, may @foxdie mention this, I'm not sure.

Plan for next step is to flash my ecu with my original rom and see what's happen, is that make sense?

btw I've got evoscan logs but they about 6MB and are to big to upload here so if any one wants to see it I can send via email.

I try attach googledrive links let me know if they work, 12.11.08 is filr with issue present 13.17.34 is log when all was ok
https://drive.google.com/file/d/0Bwm...it?usp=sharing
https://drive.google.com/file/d/0Bwm...it?usp=sharing

[foxdie]

Hi folks, coming into this quite late and trying to remember everything. My memory isn't what it once was

Here are a my thoughts in bullet points;

* I believe the MTX-Ls need calibrating every 6 months, calibration needs to be done with a completely clean exhaust, leaving it standing overnight with no cat should suffice, if you have a cat you may need to try something more radical (calibrating it whilst on a long downhill stretch of road, low gear, engine revs steady above 3000 RPM with your foot off the accelerator so pure air is being pumped though)

* VR-4s do like to run a little lean but not that much. Raul used to regularly cycle between 13-18:1 AFR cruising under closed loop (hunting)

* Oxygen sensors typically have a 100,000 mile / 10 year life expectancy - don't trust it to be accurate if its anywhere near this age

* Faulty oxygen sensors typically result in richer running, not leaner running.

* Raul briefly used to run stupid lean (anything up to 20:1) when starting on a hot engine, always put this down to engine temp vs enrichment tables in the ECU not being set up properly, but as it sorted itself out after 30 seconds it wasn't a priority to fix it

* Starting him on cold would be around 11-12:1 until it warmed up, running rich on cold is only really programmed in to warm up the cat as soon as possible for emissions targets - the engines should be fine idling at stoich (14.7:1) from the get-go

* The software on the VR-4 ECU (and indeed most EFI cars) doesn't use the signal from the oxygen sensor for at least 20-30 seconds and then starts to dial it in slowly as both the sensor and engine temperatures approach working temp - ie. they run open loop for 20-30 seconds when the cars started, HOWEVER they do use the engine temp sensor for feedback and adjustment of fuelling

* Give or take a few other variables such as engine temp etc, the VR-4 ECUs will run closed loop until either the throttle is depressed beyond 50% or the engine load goes above 50% (whichever comes first), at which point it then switches to open loop

* A narrowband oxygen sensor has a sharp swing between 0.9 volts (+/- 0.1v) for about 14.5:1 and under, and about 0.1 volt (+/- 0.1v) for 14.9:1 and over. The ECU tries to meet stoich by repeatedly running rich and lean, crossing over that "line" on purpose to hunt for 14.7.

With the above bullet point in mind, logically one of three scenarios is happening:

1. Either the ECU is running open loop and the cell for that RPM vs LOAD value is out and needs adjusting (or like mentioned previously, injector scaling is out, however this is unlikely if you're running a stock ECU)
2. The narrowband output of the MTX-L is set too smoothly, resulting in slower response times for fuelling (again, doesn't sound likely)
3. If you're still using the stock narrowband for fuelling, its age could be resulting in slow reactions, making the ECU temporarily think the car is running richer than it is

These are just thinking out aloud, probably wrong on a few points, take it all with a pinch of salt

[adaxo]

Hi folks, coming into this quite late and trying to remember everything. My memory isn't what it once was

Here are a my thoughts in bullet points;

* I believe the MTX-Ls need calibrating every 6 months, calibration needs to be done with a completely clean exhaust, leaving it standing overnight with no cat should suffice, if you have a cat you may need to try something more radical (calibrating it whilst on a long downhill stretch of road, low gear, engine revs steady above 3000 RPM with your foot off the accelerator so pure air is being pumped though)- correct, manual says it should be done every 10000 miles, I did it by leaving car over night, no cat present

* VR-4s do like to run a little lean but not that much. Raul used to regularly cycle between 13-18:1 AFR cruising under closed loop (hunting)- it was ok b4, usually keep 14-15 all the time on light load

* Oxygen sensors typically have a 100,000 mile / 10 year life expectancy - don't trust it to be accurate if its anywhere near this age- i use only wideband which is about two years (~10000 mileas) old, emulate narrowband to ecu

* Faulty oxygen sensors typically result in richer running, not leaner running.- correct, when my narrow band connection fail car run richer than normal as ecu don't get lambda signal

* Raul briefly used to run stupid lean (anything up to 20:1) when starting on a hot engine, always put this down to engine temp vs enrichment tables in the ECU not being set up properly, but as it sorted itself out after 30 seconds it wasn't a priority to fix it- my afrs looks normal on hot/cold start, just leaning out after minute or so but not always

* Starting him on cold would be around 11-12:1 until it warmed up, running rich on cold is only really programmed in to warm up the cat as soon as possible for emissions targets - the engines should be fine idling at stoich (14.7:1) from the get-go- this how its now, start cold afrs are about 12-13 and then when warm up going to 14-15s

* The software on the VR-4 ECU (and indeed most EFI cars) doesn't use the signal from the oxygen sensor for at least 20-30 seconds and then starts to dial it in slowly as both the sensor and engine temperatures approach working temp - ie. they run open loop for 20-30 seconds when the cars started, HOWEVER they do use the engine temp sensor for feedback and adjustment of fuelling

* Give or take a few other variables such as engine temp etc, the VR-4 ECUs will run closed loop until either the throttle is depressed beyond 50% or the engine load goes above 50% (whichever comes first), at which point it then switches to open loop

* A narrowband oxygen sensor has a sharp swing between 0.9 volts (+/- 0.1v) for about 14.5:1 and under, and about 0.1 volt (+/- 0.1v) for 14.9:1 and over. The ECU tries to meet stoich by repeatedly running rich and lean, crossing over that "line" on purpose to hunt for 14.7.

With the above bullet point in mind, logically one of three scenarios is happening:

1. Either the ECU is running open loop and the cell for that RPM vs LOAD value is out and needs adjusting (or like mentioned previously, injector scaling is out, however this is unlikely if you're running a stock ECU)- this is what I also consider, one thing is, that it was ok for a while, well since got it flash on japfest2, it is possible that maps, somehow 'adjust' itself and causing all mess?
2. The narrowband output of the MTX-L is set too smoothly, resulting in slower response times for fuelling (again, doesn't sound likely)- all is set as it used to be, just to be sure I reset all to 'factory settings' after last calibration yesterday.
3. If you're still using the stock narrowband for fuelling, its age could be resulting in slow reactions, making the ECU temporarily think the car is running richer than it is- as said above, im on wideband only

These are just thinking out aloud, probably wrong on a few points, take it all with a pinch of salt

THANKS

[swinks]

Its used to happen occasionally that my car run lean on light/part throttle, but since I don't use it daily, rather once a week now, this issue become more common, to this degree that it's bad most of the time now. On idle afrs are mostly nice and steady (14ish) but sometimes it start leaning out to 18ish and car trying to stall then its some kind of a 'click' and all back to normal.

To be more precise, when cruising around on light load afr go up to 16-18, I assume only in closed loop as when I put my foot down its behave as it should, ie afrs drop to 11-12s.

So far I did this

swap MAF- no change
swap bov from forge to stock- no change
recalibrate wideband- no change, well it was normal for first few miles then start leaning out, then back to normal and again and again
check all vacuum/turbo pipes, all seems to be ok

Are you running adjustable fpr?
If so it's likely it's gone bye-bye.
Had identical issue few years back when my 1st fpr refused to work.
And if you hooking up new adjustable fpr, then get vacuum directly from manifold, just by-pass solenoid.

[adaxo]

Are you running adjustable fpr?
If so it's likely it's gone bye-bye.
Had identical issue few years back when my 1st fpr refused to work.
And if you hooking up new adjustable fpr, then get vacuum directly from manifold, just by-pass solenoid.

No FPR, petrol side is untouched.