leaning out on warm idle and off accelerator pedal

[crazydriver81]

Hi to the Pro's on here,

I Need your help. :)

Now, that the winter is almost over and the Legnum is about to hit the roads again, I still have to deal with one issue which I did not fix last year. The car developed a leaning out issue. I also remember foxdie assuming that I might have a small boost leak while he was mapping my car. However, what I am looking for is some advice, on which points to look at.

The issue:
- car leans out to 19ish AFR's when warm idling and during driving when lift off accelerator pedal completely
- when pressing the accelerator pedal on idle (stationary) slightly, revs go up and AFR's go to 16ish before going back to 17ish to 19ish
- while driving when AFRs are lean car seems to take a "second" before AFR's are getting back to normal values
- when under boost and accelerating, car runs fine with stable 11.5 AFR's (as mapped by foxdie)

The car developed this issue on the way to the mapping Event (so before car was mapped!). The 600kms before after Manual gearbox swap it ran just fine.

So, what to you think guys?
- ICV to be cleaned and checked (plus install new base idle screw and seal)
- boost leak test to be done
- TPS Setting to be checked

Anything else I need to check for? Thanks for your input. :)

[Confused]

Right, lets tackle them one-by-one :)

When you come off the throttle completely, the ECU should disable the injectors until the engine speed drops to around 800rpm. You should see the Injector Duty Cycle drop to 0% and the Injector Pulse Width drop to 0.000 too, if you're logging via EvoScan. This is perfectly normal, and on a long enough zero-throttle coast, the AFR should lean out as far as your wideband will read. Well, it *should* be doing this, but foxdie did set the injectors to come back on at 2000rpm rather than 800rpm in his tuned ROMs - for pops/bangs on overrun - so this may affect what I said above (it certainly did on my Anglia!)

At idle, it *should* be aiming for 14.7, and it should "hunt" either side of this very slightly (14.0-15.4ish, as I've seen on my Anglia)

When you press the accelerator slightly, the engine ingests a sudden gulp of air - this causes the initial leaning out. When under no load, very little fuel is required to maintain that increased engine speed, so it is not surprising to see an AFR of 18.

There is also the response speed of your wideband - some will "smooth" the output over a few samples rather than giving an instantaneous reading - don't forget too that there's the delay in your gauge too - a gauge with a digital readout will always respond quicker than a gauge which has an arm that has to move.

Realistically - AFRs at high load high boost are most important - and 11.5 is absolutely spot on.


Checking for boost leaks is always a very useful diagnostic test - and if you're unsure, perform a boost leak test ASAP!
If the AFR isn't leaning right out on no-throttle coasting, then definitely do the TPS setup procedure to ensure that it's providing the Idle Position signal to the ECU (as a side note, I've never been able to get a value for the Idle Position Switch option in EvoScan, so make sure you use the feeler gauge/multimeter method to adjust it).

The ICV/screw should only be played with if it hunts badly at idle or can't catch the revs dropping away from high RPM (ie if it drops to about 500rpm before recovering to about 670 - it should catch it before it drops below 650).

[crazydriver81]

Right, lets tackle them one-by-one :)

When you come off the throttle completely, the ECU should disable the injectors until the engine speed drops to around 800rpm. You should see the Injector Duty Cycle drop to 0% and the Injector Pulse Width drop to 0.000 too, if you're logging via EvoScan. This is perfectly normal, and on a long enough zero-throttle coast, the AFR should lean out as far as your wideband will read. Well, it *should* be doing this, but foxdie did set the injectors to come back on at 2000rpm rather than 800rpm in his tuned ROMs - for pops/bangs on overrun - so this may affect what I said above (it certainly did on my Anglia!)

At idle, it *should* be aiming for 14.7, and it should "hunt" either side of this very slightly (14.0-15.4ish, as I've seen on my Anglia)

When you press the accelerator slightly, the engine ingests a sudden gulp of air - this causes the initial leaning out. When under no load, very little fuel is required to maintain that increased engine speed, so it is not surprising to see an AFR of 18.

There is also the response speed of your wideband - some will "smooth" the output over a few samples rather than giving an instantaneous reading - don't forget too that there's the delay in your gauge too - a gauge with a digital readout will always respond quicker than a gauge which has an arm that has to move.

Realistically - AFRs at high load high boost are most important - and 11.5 is absolutely spot on.


Checking for boost leaks is always a very useful diagnostic test - and if you're unsure, perform a boost leak test ASAP!
If the AFR isn't leaning right out on no-throttle coasting, then definitely do the TPS setup procedure to ensure that it's providing the Idle Position signal to the ECU (as a side note, I've never been able to get a value for the Idle Position Switch option in EvoScan, so make sure you use the feeler gauge/multimeter method to adjust it).

The ICV/screw should only be played with if it hunts badly at idle or can't catch the revs dropping away from high RPM (ie if it drops to about 500rpm before recovering to about 670 - it should catch it before it drops below 650).

Hi Garry,

thanks for all the explanations. I have a PLX wideband kit, so the readings are rather quick and precise. I know about the fully leaning out on coast.

We tried to setup pops-and-bangs on my car while mapping but it ended in massive overfuelling, so we disabled it.

I suspect the TPS as a cause, because when on coast and pressing the accelerator pedal slightly (e.g. just to keep a certain speed) it will also lean out. And this is not normal. Also when rolling to traffic lights with the clutch disengaged, it leans out to 19ish AFRs and almost stalls. AFR's are fine, if A/C compressor engages.

Anyhow, I just wanted to have some thoughts if I possibly miss something. So, I will work off the list above and Report back. :)

[Confused]

If you don't have it already, get yourself a cheap VAG KKL cable (about £5 on eBay, the one with the translucent blue plastic case works well), modify it (if you can't find it, I'll dig out a link when I'm on a computer) and get a copy of EvoScan ($25, and you'll need 2.7 for the VAG cable, not 2.9).

This will give you access to all the engine data, and has to be one of the cheapest and best diagnostic tools you can get!

[TAR]

I had an issue with car running lean a couple of years back. I have to admit I never got to the bottom of it. There is a lot of good info in it and worth a read.
https://www.clubvr4.com/forum/showthread.php?67335-Car-is-running-lean-on-light-throttle-any-ideas

Hope it helps. :happy:

[Confused]

To set the TPS correctly, put a 0.65mm feeler gauge between the stop and the spinny wheel part where the throttle cable attaches, then adjust the TPS so that pins 3&4 are just disconnected. They should then give continuity when you remove the feeler gauge.


To use the cheap VAG-COM KKL cable, you need to connect pins 1 & 4 whilst it is plugged in - I opened up my cable and connected them internally - you could also connect them externally.

[foxdie]

When you press the accelerator slightly, the engine ingests a sudden gulp of air - this causes the initial leaning out.

This can be addressed using the TPS delta tuning, it's a table for how quickly the throttle is pressed (distance between each "read" of the TPS, the delta) and enriches based on that :)

[foxdie]

Also hi Stefan, hope you're well :)
Confused has covered quite a lot here, I don't feel like I can add much more at this point, idling between 17-19 without any occasional swings towards 14.7 I would say sounds a lot like an old / faulty narrowband sensor, but as you have a PLX I'd expect better.

Your PLX is emulating narrowband output for the ECU yeah? Could be worth slightly offsetting the output voltage for that to steer it closer to 14.7 ?

[Nick Mann]

I'd check the voltage on the plx. If the narrow band emulation is swinging around 0.5v whilst the car is running 16afrs then you have found the culprit.
I would be suprised if that is the case.
Modifying the emulated narrow band output (assuming the output is correct) seems to be like putting a bandage round a leaking water pipe. It might work but it might just be disguising a bigger problem.

[crazydriver81]

Again, thanks for all the input guys.

I am not quiet sure if the PLX narrowband emulation can be tweaked. I just received the bung to test the IC pipework for leaks. I will work through all these points next week and will hopefully find the issue. Luckily I have access to the MUT-III tester and can check for correct base idle settings. I tend to believe that either the TPS is off or the ICV is faulty. But this will be discovered once I am in the workshop with the car.I'll also try to "calibrate" the PLX wideband which should be easy, now that the car stood in the garage over winter and exhaust is full of air.

What really bugs me with that are two facts:
- when coming to traffic lights with clutch disengaged, it will nearly stall and gets AFR's in the 19ish region (once engine is warm)
- when on cruise and try to keep speed and use the accelerator pedal slightly, it leans out

Anyhow, I'll report back. Keep the fingers crossed, I'll solve it.

[foxdie]

In both those situations you describe it's running in closed loop mode.

Stefan, just out of curiousity, can you measure the voltage on the o2 pin on the ECU when this happens?

Sent from my SM-G930F using Tapatalk

[Confused]

Davezj has found that his STFT (Short Term Fuel Trim) does not appear to kick into action properly until you give it a kick up the backside when idle.

[fassi1]

There could be few factors that cause it.
Most of the VR4s here on the club are modified. Basic modifications are higher flow air filter, decat pipe or 3" exhaust.
I suspect, that it helps engine to breathe easier and simply with the same throttle possition there is more air entering the engine.
I may be wrong but that's what I suspect.
Recently had chance to work on Nissan 350Z VQ35HR (convert to LPG) which had full custom exhaust system and 2 HKS panel filters.
First quick inspection of fuel trims and turns out idle is towards leaner side, idle fuel trim was around +12 .. .. +17 with max +25.
There was no air leaks (idle manifold intake pressure was correct) and no exhaust leaks.
Mid load fuel trims were perfect around 0% but high load and high rpm were stupidly overfueled with AFR 10 or less.
All these symptoms are similar to what most legnums suffer from.
To make sure that there is no air leak I would try to meassure idle manifold intake pressure. It should be roughly 0.28bar of vacuum with no extra load on the engine (aircon off).

[Nick Mann]

Back in the days of my hks boost controller my legnum used to idle at around 0.48 bar of vacuum. I don't monitor it so much now.

Also, do you have evoscan? That will show what the o2 sensor is doing.

[fassi1]

0.48 bar of vacuum is way too much air for idle. That will definitely cause lean idle and fuel trims out of working range.

[Nick Mann]

Sorry, it was 0.42 not 0.48. It would get to around 0.58 on over run. I am now wondering if I have remembered the units wrong! It was perfectly healthy.

I will check my current vr4 at the next opportunity.

[Piers1989]

I just checked mine as we were discussing the lean idle in another thread.

My gauge isn't aligned great but it looks to be around -0.6bar for my car!
It does always run a little lean on idle (15.7-16.5), but when its been left to heat soak for half an hour can go to 19-20 and struggles to run. As soon as you drive it returns to normal tho.
For reference this was the case before I had my AFR gauge (which does simulate the narrowband) also - the rough idling.

Did we discuss what could cause this?

[fassi1]

Correct idle manifold pressure without load is max 0.3bar or -0.7bar of vacuum with the engine warm. Anything above that is suspicious and 0.42bar or equivalent -0.58bar of vacuum isn't definitely right.

Sent from my SM-G901F using Tapatalk

[Davezj]

vacuum is normally measured in millimeters of mercury (mmHg) well my gauge does and in the plenum at idle it normally reads 500 to 600

This is the conversion i got from google

500 millimeter mercury (0 C) = 0.666612 bar
600 millimeter mercury (0 C) = 0.799934 bar

both of my VR4 have always been this value.

[Piers1989]

vacuum is normally measured in millimeters of mercury (mmHg) well my gauge does and in the plenum at idle it normally reads 500 to 600

This is the conversion i got from google

500 millimeter mercury (0 C) = 0.666612 bar
600 millimeter mercury (0 C) = 0.799934 bar

Both of my VR4 have always been this value.

I checked mine in more detail and its about -0.7 bar when i factor in the needle being misaligned. That sounds bang on what Dave's saying too.

[adaxo]

Correct me if im wrong but fassi1 and rest give different values as you measure it in different part of plenum? IE all readings via boost controllers are AFTER TB and actually vacuum should be measured before? it may be not the case al all, not sure

[fassi1]

Depemds on what do you understand by before or after throttle body. You know exactly Adam what meassurement I have in mind.
Meassurement has to be taken somewhere next to the brake servo connection to the plenum.

[Confused]

I saw around 18-20 inHg on my Anglia last night at idle. This is around 8.8-9.8 psi, 0.60-0.67 bar.

Taken from the usual place - on the plenum, between the throttlebody and the valves!

On deceleration this goes to around 23 inHg - about 12.3 psi, 0.84 bar.

[fassi1]

I saw around 18-20 inHg on my Anglia last night at idle. This is around 8.8-9.8 psi, 0.60-0.67 bar.

Taken from the usual place - on the plenum, between the throttlebody and the valves!

On deceleration this goes to around 23 inHg - about 12.3 psi, 0.84 bar.

On deceleration ( CUT OFF ) sounds about right but if idle was with no load at all, sounds too high to me.

[Davezj]

On deceleration ( CUT OFF ) sounds about right but if idle was with no load at all, sounds too high to me.

is this before or after you did all you inlet pipework modifications.

[fassi1]

is this before or after you did all you inlet pipework modifications.

That was after.
This is screen shot from LPG software live data which has its own MAP sensor. This is from almost fully warm engine.
MAP 0.28bar

[Davezj]

i understand what you are talking about with regard to the MAP sensor in the LPG system.

one of the sensors must not be operating correctly at the bottom end of its range. if my memory is correct my map sensor in the LPG system reads about the same as yours but my boost gauge reads the vacuum measurement i referenced in mmHg above, both being connected to the plenum chamber at the back.
It would be interesting to see what happens if i disconnect both of mine and put a vacuum gauge pump on each and measure the reading. or put both on the vacuum pump at the same time.

maybe this weekend.

[fassi1]

They must be out like you are saying Dave.
I do trust lpg MAP sensor. My Outback H6 has factory MAP sensor and its readings are identical to lpg MAP sensor.

[Davezj]

I will have to try it.

Sent from my MI 5s using Tapatalk

[crazydriver81]

So guys. I have the first results. I guess my issue is due to a faulty ICV. At least you cannot hear it moving with IGN on before actually turning over the engine.

Yesterday I cleaned the ICV and put it back in before actually starting the engine. I also had the PLX wideband heating up the O2 sensor until it detected air.

Since then, car runs and idles fine. The actual idle RPM's are at around 800-900 once warmed up and during warming up the RPMs are on the higher side as well - again, due to the ICV me thinks. I have to find the replacement one I keep in the shed and this will then show if it was that. AFR's stay close to or around 14.7 mark now. On idle it is going eventually to 17 but not leaner.

Other than that, I do not seem to have any boost leak. foxdie -> Now it actually goes up to close to 1 bar of boost when pushed. :) Remember your comments from the RR?

My vacuum readings are (taken from the intake manifold to stepmotor gauge):
- idle warm = approx 0.72 bar
- deceleration = up to 0.9 bar

I will post again, once ICV has been replaced. But my other conclusion is, that - even if not needed - recalibrating the O2 sensor from time to time even on the PLX cannot harm.

[fassi1]

Sounds good Stefan, glad you managed to find the reason of your problem.
Vacuum sounds ok. Idle rpm on warm sounds bit too high, it should be around 700rpm.