ECMlink (AWD) Lean AFRs when getting on throttle, levels off after (data log attached)

CluelessTuner · Dec 13, 2022

Issue:
My AFRs seem to be lean every time I get on throttle, but slowly adjust to a proper value afterwards. I am also going lean in high rpms, but I believe that is a different issue, possibly with the fuel pump.

Vehicle History/build list:
1995 Eagle Talon TSI awd
full 6 bolt swap (bottom end, top end, intake manifold)
MHI 16g rebuilt less than 2000 miles ago
ID1000 injectors
STM downpipe/catback
OEM o2 housing
ECM link v3 on a 95 ecu
Innovate LC2 wideband positioned in the o2 housing and using ECM link's own narrowband simulation (not using innovate's built in narrow band simulation)
CX racing FMIC
Speed density
Flex fuel (essentially only e85 no mixing)
Walbro 255

Problem history:
I've had a couple issues in the past with exhaust leaks pre-o2 sensor which caused similar problems, but I fixed it by reinstalling my O2 housing with a much nicer gasket, proper RTV, and properly torqing the bolts/nuts. I fixed this issue and got it dyno tuned without any issues for a while.

After commuting for a while and loving the car with the tune, it popped a CEL for maxed out fuel trims, and it was true. The fuel trims seemed to be maxed out and the issue of lean throttle tip in started showing up. I tried just adjusting global fuel upwards and it improved significantly, but still had the throttle tip in issue, it just returned to a proper AFR much faster. I know this isn't the right way to do it, but it gave me a hint that this was possible fuel pressure related (felt like my injectors lost like 300cc randomly). So I decided to splurge on an external FPR (wanted one either way) and decided to do a boost leak test on top.

I have a sort of weak compressor so a full boost leak test (turbo inlet) wasn't really possible. I know this is recommended, but since I'm speed density I decided to just give an only intake manifold test a try, might as well trying to find the most important boost leaks first before investing in a compressor. Well I did find some boost leaks, a vacuum T that I used for my MAP sensor and leaks from my TB. Replace the vacuum T with a brass one and rebuilt the throttle body with EVO milspec shaft seals and that solved my boost leaks problems. At the same time I installed my FPR and set my base pressure correctly (43 psi or so). Decided to take the car for a spin afterwards and still faced the same issues

.

At that point I started checking everything and discovered that I indeed had more pre-O2 sensor exhaust leaks. Starting the car cold and using soapy water before the exhaust got boiling hot, I found that the manifold to head and manifold to turbo gaskets were leaking, O2 sensor housing appeared to be fine though (after my original fix a while back). "Great", I thought to myself, this absolutely has to be the issue, and I've found and fixed other issues while I was at it, extra work but it wasn't wasted at least. While pulling my manifold off I realized some of the manifold to turbo bolts were loose, likely the cause of the leaks. I bought all new OEM exhaust manifold head studs and nuts, OEM exhaust manifold gasket, Copper RTV, and reused my STM SS manifold to turbo gasket. I additionally placed lock washers on every piece of hardware, torqing it a little higher than spec, and then retorqing it the next day to make sure. I've gotten really annoyed at these leak issues so I essentially attacked it with everything apart from getting the surfaces machined LOL

. I just drove the car again and it still had the same issues, so I'm asking if anyone can see an obvious issue just looking at the log because I'm scratching my head a little.

TLDR:
Long time ago had issue with lean AFR -> O2 housing leak -> fixed with STM SS gasket
Got tuned and was running great for a while, then CEL for maxed out fuel trims, this is when a lean on throttle tip in issue started.
Boost leak tests and fixes on intake manifold/TB, along with AEM fpr installed -> issue persisted
Discovered exhaust leaks between manifold->head and manifold->turbo -> fixed with new hardware, gaskets, and RTV -> issue persists

What I'm going to do next regardless:

I have a new innovate wideband MTX-L sensor with gauge that I'm going to install. Meaning that I will use a brand new O2 sensor and controller.
Going to give the O2 housing the same treatment as my other pieces (copper rtv and lock washers), thinking maybe there might be a leak around the back of it?

If that doesn't work, I think I might try:

full boost leak test
Fuel pump?
Injectors?
part out

About the log:
You can see the AFRs spike lean every time I get on throttle, multiple examples at 3.5, 10.5, and 41 seconds in (there are many more)
I did what I could in term a pull starting at about 62 seconds. In this example you can see that as I rolled into the throttle the AFRs spiked lean, but going down until 14.0 and then rising again as I do the pull. I think that later rise might be due to fuel pressure issues? Can't definitively prove it, but my fuel pump whine had always been louder than I feel is right. Maybe could be flowing less than it should.

What's interesting about the log is that I'm surprised that the car is not misfiring or behaving strange during the low RPM lean issues. Makes me feel like the AFR value isn't correct, meaning either more leaks are causing the O2 sensor to read wrong, or the O2 sensor itself is having trouble.

I've also included the current configuration that I'm using, you might notice that the global fuel is set to about -45%, while it should be -55% for 1000c injectors, but without riching it up a bit the car barely drives. So strange.

Let me know what you guys think, been struggling with this lately and have been missing driving my car around. Thank you

jpmxrider489 · Jan 1, 2023

Look like your global fuel is off. Its asking for a fuel. I believe the global should be -51 to -55 range. Then repeat idle process. I cant provide much input as far as the lean when letting off. But aside from the global being off, are you sure your narrowband sim switch point is correct. I dont have much experience with inovative. Most of mine has been with aem. I dont think a leak would cause this much variance unless it was large. Have you tried the front location? Link is saying you have it in the rear.

We're on Boost · Jan 1, 2023

CluelessTuner said:
you might notice that the global fuel is set to about -45%, while it should be -55% for 1000c injectors, but without riching it up a bit the car barely drives. So strange.

I notice in your settings, in AuxMaps, your "Fuel increase at E100" is set to 38%. That's probably way low. That might be why you are having to set global fuel richer than expected when running straight E85.
My "Fuel increase at E100" is set at 63%. That is also kind of the suggested setting given by ECMtuning on their "Ethanol sensor support page".
That value of 63% comes from dividing the fuel/air stoic ratio of ethanol by the fuel/air stoic ratio of gasoline. I'm saying fuel/air, the inverse of air/fuel.
So the inverse of 9:1 divided by the inverse of 14.65:1
That's 0.11111 divided by 0.06826
Which is 1.63
Which is 63% more.
Your 100% ethanol flow would need to be 63% more than your 100% gasoline flow.

On the ECMtuning support page it says that in some cases, with some injectors, this number might need to be even higher than 63% because some injectors actually flow less volume with increasing ethanol content.

I haven't looked at your running log yet.
Have you looked at the fuel pump? It might have a partially plugged sock. Goo on it, or whatever. Or your fuel filter.

We're on Boost · Jan 2, 2023

In your driving log I'm looking at the pull you mentioned at around 62 seconds.
Was your car running ok through that pull?
What gear were you in there?

It seems to me that from 63 seconds to 66.7 seconds the revs are coming up fast and smooth, the InjOn time is coming up smoothly, the Airflow is coming up smoothly, boost comes up fast and smooth, it goes into OpenLoop at a LoadFactor of 0.9, there's no KnockRet, the timing seems reasonable. Those things all seem good.

The only thing I see that looks way off is the LC1WB number.

At about 63.1 sec it all of a sudden jumps from 14.1 to 16.2 which seems ridiculous because the InjOn time goes up there like it should, not down. And the InjOn time keeps going up through the rest of the pull like it should.
At 64.1 seconds the LC1WB comes back down to around 15. But that is still wrong. The correct WB number through there should be in the elevens, around eleven.

I can compare your InjOn time to mine at an Airflow of 22 lbs/min where we both happen to be at 4,200 rpm. Same airflow, same rpm, and we have the same base fuel pressure. Your InjOn time there is almost exactly 2 times what mine is. (yours is 15.71ms, mine is 8.00ms) That's about right given that my injectors are 1650cc and my fuel is 45% ethanol, not 82% like yours. I can tell you that my A/F is in the elevens in that airflow range. Yours probably is too unless you are having actual bad running there. If you had clogged fuel sock or clogged injectors you'd probably have noticeably bad running.

If the car feels good through a pull like that and the wideband number is the only thing bad, then I think your wideband is bad. It might be burned out from being too close to the turbo. If it's in the O2 housing then I'm imagining it's about 5 inches from the turbo outlet, which is too close for a wideband. Or some other reason that your wideband is bad.

So, was your car running ok through that pull?
What gear were you in there?
And how far from the turbo is your WB sensor located?
I'm kind of assuming that your wideband gauge reads the same as your wideband logged numbers, but maybe not?

It would be better if you had a real OEM type narrow band sensor in the O2 housing to run the car, and use a wideband located farther back for logging and gauge only. The way it is now, you are simulating a narrowband from a wideband that is acting weird, so your closed loop operation is going to be off.

CluelessTuner · Jan 4, 2023

We're on Boost said:
In your driving log I'm looking at the pull you mentioned at around 62 seconds.
Was your car running ok through that pull?
What gear were you in there?

It seems to me that from 63 seconds to 66.7 seconds the revs are coming up fast and smooth, the InjOn time is coming up smoothly, the Airflow is coming up smoothly, boost comes up fast and smooth, it goes into OpenLoop at a LoadFactor of 0.9, there's no KnockRet, the timing seems reasonable. Those things all seem good.

The only thing I see that looks way off is the LC1WB number.

At about 63.1 sec it all of a sudden jumps from 14.1 to 16.2 which seems ridiculous because the InjOn time goes up there like it should, not down. And the InjOn time keeps going up through the rest of the pull like it should.
At 64.1 seconds the LC1WB comes back down to around 15. But that is still wrong. The correct WB number through there should be in the elevens, around eleven.

I can compare your InjOn time to mine at an Airflow of 22 lbs/min where we both happen to be at 4,200 rpm. Same airflow, same rpm, and we have the same base fuel pressure. Your InjOn time there is almost exactly 2 times what mine is. (yours is 15.71ms, mine is 8.00ms) That's about right given that my injectors are 1650cc and my fuel is 45% ethanol, not 82% like yours. I can tell you that my A/F is in the elevens in that airflow range. Yours probably is too unless you are having actual bad running there. If you had clogged fuel sock or clogged injectors you'd probably have noticeably bad running.

If the car feels good through a pull like that and the wideband number is the only thing bad, then I think your wideband is bad. It might be burned out from being too close to the turbo. If it's in the O2 housing then I'm imagining it's about 5 inches from the turbo outlet, which is too close for a wideband. Or some other reason that your wideband is bad.

So, was your car running ok through that pull?
What gear were you in there?
And how far from the turbo is your WB sensor located?
I'm kind of assuming that your wideband gauge reads the same as your wideband logged numbers, but maybe not?

It would be better if you had a real OEM type narrow band sensor in the O2 housing to run the car, and use a wideband located farther back for logging and gauge only. The way it is now, you are simulating a narrowband from a wideband that is acting weird, so your closed loop operation is going to be off.

So first of all thank you for the response, I really appreciate you spending the time trying to help me. I looked through a couple things you said trying to fix it.

So a while back I had issues with exhaust leaks causing false lean readings on the O2 sensor, and before those logs I resealed the exhaust manifold and manifold to turbo. I also just did my o2 housing again. I also suspected some issues with the O2 sensor, it is in the front position but wired into the rear o2 pin on the ecu. The old o2 seemed as if the lean condition was true, having a pale color to it. I put in a new o2 sensor and controller, an innovate mtx-l plus. All calibrated and it still had the issue. Front position might not be optimal but it worked for a while until the lean issue started showing. The values in DSMlink match the gauge so no issue with the sensor setup.

The additional fuel at E100 sounded interesting, I was so excited reading it hoping it was the issue, but again no luck.

During the pulls in the log I was in either 3 or 4th gear.

Interesting point on how the wack readings from the wideband mess up closed loop operation, but I drove the car around a little locked on open loop and it was very lean, I barely got it around my neighborhood. During the pulls in the log the car would buck at some point in the pull as it got lean, so at this point I'm pretty certain that the original wideband was getting a proper reading. I wanted to get a different gauge than what worked with the Lc2 anyways so I'm not angry about buying the new wideband kit.

So I have a FPR with a gauge in the engine bay and it never really had pressure drop. It might be the cause for the issue in the higher rpms going lean, but not sure about that lean spike on throttle. For a bit I just assumed the fuel pump was fine, but I got curious as the previous owner is the one that put in the walbro 255. I pulled the sending unit and realized the sock was absolutely disgusting. Like I've had a couple other projects with nasty socks, but this one took 1st place by a long shot. It was sooo dark and dirty that for a moment I thought that the sock was just a weird black color. I will swap it out with a new one and see how it goes. I swapped the fuel filter for a new one just a couple months ago so that should be good.

Hoping that will fix it but not having high hopes in terms of the lean tip in. I will give the factory narrowband a try and see how it goes.

Once again thank you for spending the time seriously looking through my tune and logs, got my DSM not too long ago and I absolutely love the community. I will report back with results.

diambo4life · Jan 4, 2023

Seems like a classic fuel delivery issue. Clean inline fuel filter? How clean is the fuel strainer? When were the injectors last flow checked and/or cleaned?

We're on Boost · Jan 4, 2023

CluelessTuner said:
I pulled the sending unit and realized the sock was absolutely disgusting. Like I've had a couple other projects with nasty socks, but this one took 1st place by a long shot. It was sooo dark and dirty that for a moment I thought that the sock was just a weird black color. I will swap it out with a new one and see how it goes.

Yeah good, the sock could be the whole problem or most of it. Although if the sock is bad it could also be that the problem has worked its way downstream to the fuel filter or the injectors.

CluelessTuner said:
I will give the factory narrowband a try and see how it goes.

A really cool thing about having the factory narrowband in there (or an NTK or Denso proper replacement part) is that it can give you a check on the wideband.
If your A/F is way lean, like over about 15, your narrow band will read 0 or close to 0.
If your A/F is rich, like less than about 14, your narrow band will read some voltage greater than half a volt.
That is after warmup though. So you have to be able to run the engine for a few minutes at least to see those things.

On a cold start, well, I can tell you what my narrowband does.
Basically the voltage starts at 0 and after about a minute the voltage will be up over half a volt. During cold starts the engine will always be rich because it has to be rich until it's warmed up. But during that first minute the narrowband is just warming up.

Today I cold started my car in 39 deg F ambient (outside) temps, the engine coolant temp was 37 deg F. After 53 seconds the O2 voltage was over half a volt and still climbing.
That was with only 21% ethanol in the tank. So I'll give you one from a few weeks ago when I had 45% ethanol in the tank.
On that one the outside ambient was in the high 40's F and the engine coolant temp was 48 deg F. After 52 seconds the O2 voltage was over half a volt and still climbing. Pretty darn consistent.

So that's just to let you know what normal cold start behavior is on the narrow band voltage. But after the engine is warmed up you can use the narrow band as a rough check on your WB.

If you have a good working narrowband in there and it reads 0.7 volts when your wideband shows a lean spike, then we can be pretty sure that your wideband is wrong. The NB will read 0 volts on a real lean spike.

CluelessTuner · Jan 4, 2023

We're on Boost said:
Yeah good, the sock could be the whole problem or most of it. Although if the sock is bad it could also be that the problem has worked its way downstream to the fuel filter or the injectors.

A really cool thing about having the factory narrowband in there (or an NTK or Denso proper replacement part) is that it can give you a check on the wideband.
If your A/F is way lean, like over about 15, your narrow band will read 0 or close to 0.
If your A/F is rich, like less than about 14, your narrow band will read some voltage greater than half a volt.
That is after warmup though. So you have to be able to run the engine for a few minutes at least to see those things.

On a cold start, well, I can tell you what my narrowband does.
Basically the voltage starts at 0 and after about a minute the voltage will be up over half a volt. During cold starts the engine will always be rich because it has to be rich until it's warmed up. But during that first minute the narrowband is just warming up.

Today I cold started my car in 39 deg F ambient (outside) temps, the engine coolant temp was 37 deg F. After 53 seconds the O2 voltage was over half a volt and still climbing.
That was with only 21% ethanol in the tank. So I'll give you one from a few weeks ago when I had 45% ethanol in the tank.
On that one the outside ambient was in the high 40's F and the engine coolant temp was 48 deg F. After 52 seconds the O2 voltage was over half a volt and still climbing. Pretty darn consistent.

So that's just to let you know what normal cold start behavior is on the narrow band voltage. But after the engine is warmed up you can use the narrow band as a rough check on your WB.

If you have a good working narrowband in there and it reads 0.7 volts when your wideband shows a lean spike, then we can be pretty sure that your wideband is wrong. The NB will read 0 volts on a real lean spike.

Alright, so swapped out the fuel pump sock and it made the pump waaay quieter. Still same lean tip in issue, but I bet it fixed the top end lean out issue that i was facing, though i didn't test it.

So I have the wideband wired into the rear o2 pin on the ecu and kept my narrow band harness. Threw the narrow band into the front position and while parked threw the wideband in the rear o2 position, also disabling nb sim in dsmlink while logging both. Interestingly the narrowband voltage did seem to go up as the wideband lean spiked. So sounds like an off wideband reading. You can see in the data log I gave it a couple revs. However despite me reseting fuel trims and only using the narrowband for closed loop, the car sounded like it stumbled a little bit on the rev. Not sure how well sealed my exhaust is down near the o2, but that could affect readings as well. Not sure if I can make any conclusions without driving the car. I will get it in a driving state again and get a log using only the narrowband.

Also AFRest seemed to stay at a good spot during the whole time. I'm not exactly sure how AFRest works, but just thought that was interesting.

Also by the looks of the log it seems like the alternator stopped charging. I'm feeling pretty cursed right now haha.

The Top DSM Community on the Web

ECMlink (AWD) Lean AFRs when getting on throttle, levels off after (data log attached)

CluelessTuner

Attachments

jpmxrider489

We're on Boost

We're on Boost

CluelessTuner

diambo4life

We're on Boost

CluelessTuner

Attachments

Similar threads

Latest posts

Build Thread Updates

Vendor Updates

Latest Classifieds