3BG AWT Misfires, low power, lambda sensor error

Hello Diesel Mechanic Forum!

Okay, where do I even begin...

Best at the very beginning. In December 2016, I bought a 3BG Variant 1.8T as a replacement for my Golf 2 GTD. It had 306,000 km on the odometer, a Vialle LPG system (liquid injection), 1 year of TÜV (German technical inspection), and cost €2900.
The car drove well, but it consumed a bit more fuel than I would have liked, otherwise everything worked fine. When I purchased it, I already saw that the cooling system's expansion tank vent was broken (the pipe coming out of the block was crumbling).
At the beginning, I replaced all the belts, the water pump, all the control arms, the rear axle bearings, the brakes all around (the front brakes were upgraded to 312mm), the filters, and I changed the engine oil multiple times at very short intervals because the 1.8T engine is known to have a problem with sludge buildup in the oil pan/pump. The last oil change was with 5W30, and the service interval was also set to "longlife." Okay, now using a fixed interval and Mobil 1 5W50 oil.
So, I approached my first few days of vacation in April 2017 with a relatively clear conscience, knowing that I would be renewing the cooling system components. This included both valves, the hose, some tubing, clamps, a coolant pipe (which was heavily rusted inside), the intake manifold gasket, and the throttle body gasket. I also installed an S3 radiator shroud for better coolant flow and removed the fog lights to improve airflow. It should be noted that the completely open KGE (likely referring to a specific engine component or setting) did not reveal any weaknesses in the engine's performance, which I found quite surprising.
Later (summer 2017), there was a new mass airflow sensor (Pierburg) and a new N75 valve (VW) installed (autumn 2017). Slowly but surely, the car started to show some power! The maximum speed, according to... GPS 214 km/h – it didn't take that long - I considered the patient to be healthy.
Unfortunately, since my engine overhaul, it has always started very poorly after being idle for a longer period (3-5 seconds of cranking). I've learned to live with it because otherwise, I really couldn't complain. Everything else is fine. It's very reliable. Occasionally, I've noticed that the little four-cylinder engine seems to be very well decoupled and runs very smoothly at idle. You could tell the engine was running because a water bottle on the passenger seat was vibrating, but you could barely feel anything.
I drove several times for over 80 km continuously at 180 km/h on LPG, and I thought to myself, "This is exactly how I imagined it!" – Everything went smoothly.

Summer 2019 - I was forced to take a vacation trip to Denmark. Of course, I really wanted to go with my boyfriend! But I didn't trust the oil pan or its contents. So, I took the car to a friend's workshop. They cleaned the oil pan and replaced the oxygen sensor before the catalytic converter. I decided to replace the sensor because I hoped it would finally bring the fuel consumption down to what I expected from the new engine. Unfortunately, it didn't help.
There was sludge in the bathtub, but it wasn't worth mentioning. Onward to Denmark! Oh wait – I forgot the fuel filter! I had to replace it myself. The old one was completely clogged. Very bad.
But then - everything went perfectly.
For the entire trip to Denmark and the return journey, I used gasoline because, while repairing the cooling system, I had taken a look at the intake valves and they seemed quite dirty. I thought it might be a good idea to verschandeln them somewhat by running the engine on gasoline for a long distance. At least, that was my hope. When I fill up with gasoline, I only use Ultimate or V-Power. It also worked well in Denmark.
After these stages using gasoline, it also became clear why the car has such difficulty starting after being idle for a long time. If you drive the car on gasoline for a while, the problems disappear completely. Absolutely completely. So, I suspected that a gas injector might not be completely sealed and was slowly leaking gas, which made the engine run too rich when starting cold, forcing it to expel some of the rich mixture before it could run properly.

By the way, during the drive, it was also noticeable that the car, when running on gasoline, accelerated in a wave-like pattern at higher RPMs. It's probably not getting enough gasoline. Everything is fine when running on gas. Maybe the fuel pump is failing. I don't really mind, because I mostly drive on gas anyway. Before anything burns out, the lambda sensor will detect it. Down with it.

Two weeks after the trip, the check engine light came on. It was the first time since I owned the car. Error read: P0130 - Oxygen sensor electrical fault, circuit malfunction, bank 1, sensor 1.
I thought it was a manufacturing defect, so I took it back to the workshop, and they replaced the sensor. They removed the Pierburg sensor and installed an NGK sensor. Everything is fine now. The workshop foreman thought it was unlikely, but… Well, it helped!

Meanwhile, there were occasional periods of rough idling, and sometimes it just ran poorly. However, I couldn't pinpoint the cause. No particular driving condition seemed to trigger it, and it wasn't really disruptive, just not ideal. I thought it was different.
Sprayed the intake area with brake cleaner, including the electronic throttle body (ESV), the intake manifold towards the cylinder head, the intercooler hoses, the intake hose as much as possible, and the throttle body... Still no response! Topic postponed.

TEV, played, deburred, sprayed, no change.
Between the regular inspections (HU/AU) – the inspector said: very good readings, the system is perfectly adjusted!

In the new year 2020, I wanted to drive a bit faster on the highway and noticed that the car wasn't running as well as it should. The acceleration was sluggish. It wasn't easy to reach 180 km/h. I postponed it. It was only a short distance, I just wanted to speed up for a bit. Who knows.

Since replacing the EGR valve, a new SUV (VW) had been sitting around at my place. I eventually decided to trade it in. To my surprise, when I disconnected the SUV from the hose leading to the intake manifold, a large amount of water came out. Great. That much condensation? Funny.
But – it definitely works better. The noise when suddenly releasing the accelerator was no longer "wututututu" but "psch." Great. Then, on top of that, he drove around for two more years with a defective SUV.

About a month ago, something similar happened regarding the engine's performance. Driving home from Leipzig late at night, the highway was empty – the engine started making a strange noise. Again, very little power. Damn, the gas tank was almost empty too, so I needed to get gas. I was probably overdue for a fuel filter change anyway. I floored it all the way to Chemnitz, but I wasn't really happy with the performance. The speedometer hit 200... Well. I probably should have let that go.
Receipt tomorrow. Rough idle (making a lot of noise, but the tachometer needle barely moved), and the check engine light came on 1km after starting the car – P0130, oxygen sensor electrical circuit fault.

Damn! Why? Something's not right. I need to figure that out!

However, due to the intense activity the night before, I thought I had damaged the actual new probe and bought a replacement. Switched it out quickly, deleted the adaptation values, and moved on.
Idling roughly, but always performing poorly, and the check engine light comes on after a very short time – P0130… etc. Hmph!
Now I'm here... The problems are multiplying, and I can't necessarily identify a specific cause... To avoid always having to rely on my friendly local workshop (which is also 20km away), I bought VCDS myself.
I want to find the error and fix it!
In addition, I'm experiencing misfires on cylinder 1.

Compression on all cylinders (when warm) is between 9.5 and 10.5 bar, but cylinder 1 is the lowest.
Swapping the coils or spark plugs (independently of each other) does not cause the fault to move.
The wiring looks visually correct; I've checked the connections from the plug to the coil and everything seems fine. However, I haven't measured anything yet. The mass on the valve cover has been replaced.

The control unit box in the water tank compartment is completely dry, and the connectors look excellent.
Throttle learned – ADP is OK, and both potentiometers are working correctly and in opposite directions. In the center position, I'm getting a 50/50 reading.
LMM deducted, no improvement. Values are within acceptable limits (3.94 g/s).
Mass residue cleaned from the injection mold. Everything is OK.
"Boost pressure achieved (corresponds to the target value according to...)" MWB (while driving).
Variable valve timing operates according to... MWB + System in order will also be issued.
Both probes are working (111 and 110) and are not drifting too far apart. (Forgot to mention the MWB…)

All the values I checked were OK (MWB 31-37). However, the fuel mixture adaptation was consistently 0.0%. It is said that it will not be executed if there are errors.

"Readiness code is not generated, and the error message always indicates a problem with the secondary air system. Now here's the thing! I've never heard the secondary air system activate since I've owned the car! But since the last error occurred, it's been intermittently turning on during cold starts. Often, it happens after a drive, when there's still some residual heat everywhere, but the engine is still showing as cold according to the display. You can also hear the misfires in the exhaust from the air being forced through." Terrible.

Unfortunately, I'm not sure where to start. I generally like to keep my cars in good condition, but I'm hesitant to start with the "older" parts and replace things like the fuel pump and gas injectors (€€€), only to find out that a valve seat is burned out or something similar.
It has been working satisfactorily so far, and those two issues were more of a long-term concern.

It's also important to know that misfires are only noticeable at idle and are only recorded by VCDS at idle. While driving, all values are always at zero. The car also delivers torque at low RPMs and responds well to throttle input...

Please give me some tips on how to systematically achieve success. Unfortunately, the VW repair manuals have a small problem: when I search for a manual for Motronic and fuel injection, I am offered the manual for AWM, AUG, and ATW (Mexico) for download, which is not the AWT and also differs in some ways from it, just like the AUG and AWM.
The Audi manual for AWT is quite thin in comparison and offers less helpful information than the very comprehensive VW document.
So, I still bought both of them from ErWin, of course .

Slowly, I'm starting to believe there's a valve seat problem… Fortunately, the friendly workshop has an endoscope.

The car has now been driven 375,000 kilometers.

Regards,

Martin.

Hello,

If the misfires are consistently occurring in the same cylinder, I would try swapping two ignition coils and see if the misfires move along with them.

Best regards, Rainer.

Hello Rainer!

I swapped the spark plugs from cylinders 1 and 2 and let the engine idle for a while.

The error did not propagate. He used spark plug number 1, but with the coil from cylinder number 2.

Regards,

Martin.

Hello.

Not good, there's still the spark plug, fuel injector, and potential mechanical damage to the affected cylinder.

"Could you please provide additive and multiplicative adjustment factors for the air-fuel mixture?"
Okay, I'm ready. Please provide the German text you want me to translate.

Have you already swapped the spark plugs? What do the candle faces look like?

Best regards, Rainer.

Hello Rainer,

Thank you for your reply. In the, admittedly , very long initial post, I wrote that the candles and coils had been swapped. I tried that right away.

The error doesn't travel with it.

All the spark plugs look very good, literally a rich brown color, and there's no way to tell that anything is wrong with cylinder 1.

I've now learned how the misfire detection system works and why I haven't recorded any issues during my drives. It's clear that the misfires are too small to significantly slow down the crankshaft speed, even by half a revolution. That's more like idling.

The mixed adaptation is achieved through repeated error correction, starting from a 0.0% error rate, using both additive and multiplicative methods.
I read that the adjustment will not take place as long as there are errors in the memory. However, it's not in the repair manual, but somewhere on the internet.

Subject: Fuel injector...

The problem occurs with both gasoline and gas (LPG). It doesn't get better or worse, and I can't detect any change in the engine's running behavior with either fuel. The injection timings for gasoline and gas are correct. Gasoline timing is 1.81 ms, and then gas timing is 2.07 ms (these values are rough estimates from memory; I have separate software for the gas timing settings on this system).
I also consider that to be a highly unlikely error.

I'm going to use the endoscope to take a look inside the combustion chamber this week.

If there's ever a solution to this, I'd be happy to share it.

Regards,

Martin.

Okay, I'm curious .

I would also like to perform a pressure loss test.

Are you sure it starts on gas? Actually, that shouldn't be possible, because in case of a failure (if it doesn't start), gas could leak into the engine compartment. Normally, even when set to gas, the engine will run on gasoline for approximately 10 seconds before switching over, simply for safety reasons.

I once had a very similar problem with a Passat V6. In my case, it was due to the fuel pump (difficult starting) and the temperature sensor (low top speed). The mileage was also quite high, over 300,000 km, but it was only running on gasoline. There was no error, it just displayed the temperature about 25 degrees Celsius too high.


Misfires at idle with gasoline or gas engines, especially when the engine is warm and the ignition coils/spark plugs have been ruled out, suggest a compression problem.



Your all-out, reckless approach, based on the theory that the oxygen sensor will definitely prevent anything from burning, has a catch:
Only 25% of the exhaust gases come from each cylinder. If the total amount of gasoline to be burned is 100%, and you only want to allow a 5% shortfall (which is likely to be more in practice), then one cylinder would have to decrease from 25% to 20%, which is a 20% deviation for that single cylinder.

In terms of Lambda, full load for a turbo engine from that era was approximately 0.75-0.8. The cylinder in question would then be operating at just below Lambda 1. If everything else is okay, it would run and not have any misfires, but the knock sensor might retard the timing for that cylinder by a few degrees, and the exhaust temperatures on that cylinder would reach dizzying heights under full load.



The issue of the engine starting poorly after prolonged acceleration is likely due to the system not being properly tuned, or because the engine management system is overwhelmed.

Turbocharged gasoline engines typically have very large injectors due to their high cylinder output. Under full load, it becomes more pronounced (than a naturally aspirated engine).

This results in a relatively large difference between idle and full load conditions, with correspondingly short cycle times at idle.

While not inherently a major issue, this results in a reduced fuel injection amount (grams of fuel versus milliseconds of injection time) during acceleration, which needs to be compensated for by the gas system, effectively adding and subtracting fuel.

dleds wrote:

Are you sure it starts on gas? Actually, it shouldn't, because in case of a failure (doesn't start), gas could spread in the engine compartment. Normally, even when set to gas, the engine will run on gasoline for approximately 10 seconds before switching over, simply for safety reasons.

No, the car starts on gasoline, of course. After approximately 10 seconds, it switches to gas.

dleds wrote:

I once had a very similar problem with a Passat V6. In my case, it was due to the fuel pump (difficult starting) and the temperature sensor (low top speed). The mileage was also quite high, over 300,000 km, but it was only running on gasoline. There was no error, it just displayed the temperature about 25 degrees Celsius too high.

The coolant temperature sensor was replaced a while ago, based on suspicion, due to the cold start problem. It also displays meaningful values, but it didn't achieve anything.
I'm sticking with my theory about a leaky fuel injector to explain the starting behavior.

dleds wrote:

Misfires at idle with gasoline or gas engines, especially when the engine is warm and the ignition coils/spark plugs have been ruled out, suggest a compression problem.

Here it comes. "About two weeks ago, I drove to my friend's workshop to have them use an endoscope to inspect the engine and the catalytic converter."
We wanted to let the car cool down a bit, and we ended up chatting. So, nothing was checked .
Drove home. Parked the car and noticed that it was idling perfectly smoothly - very quietly. Dude!
Laptop out, VCDS in > Misfire 0.
Great. It's not uncommon for the LL (likely referring to a lighting system or similar) to fluctuate from very poor (with significant flickering - the MKL indicator is flashing) to slightly flickering, and even to being excellent.
There's always a problem with veterinary seating, and it never really gets better. Even if the valves rotate in their seats...

dleds wrote:

Your all-out, reckless approach, based on the theory that the oxygen sensor will definitely prevent anything from burning, has a catch: Only 25% of the exhaust gases come from each cylinder. If the total amount of gasoline to be burned is 100%, and you only want to allow a 5% shortfall (which is likely to be more in practice), then one cylinder would have to decrease from 25% to 20%, which is a 20% deviation for that single cylinder. In terms of Lambda, full load for a turbo engine from that era was approximately 0.75-0.8. The cylinder in question would then be operating at just below Lambda 1. If everything else is okay, it would run and not have any misfires, but the knock sensor might retard the timing for that cylinder by a few degrees, and the exhaust temperatures on that cylinder would reach dizzying heights under full load.

You're right! I thought about that some more.

dleds wrote:

The issue of the engine starting poorly after prolonged acceleration is likely due to the system not being properly tuned, or because the engine management system is overwhelmed.

It's a Vialle LPi system, and you can't really adjust anything on it. Like a Bosch Motronic, the injection system comes "pre-configured" for the engine. You just install it, connect it, get it approved by the TÜV (German technical inspection authority), and you're done. In rare cases, likely due to significant deviations from the standard configuration caused by variations or other factors, the engine control unit may need to be reflashed by Vialle. This requires a Vialle dealer to create a log using the appropriate diagnostic software (LPi Diagnostic Tool Level 2 or 3).

dleds wrote:

Turbocharged gasoline engines typically have very large injectors due to their high cylinder output. Under full load, it becomes more pronounced (than a naturally aspirated engine). This results in a relatively large difference between idle and full load conditions, with correspondingly short cycle times at idle.

If I understand correctly, you've installed the largest injectors available from Vialle in your car, and you've drawn the same conclusions from that as I have.

dleds wrote:

Generally not a major issue, but during the acceleration process, this results in a reduced fuel injection amount (grams of fuel versus milliseconds of injection time) that needs to be compensated for by the gas system, both being added and then subtracted.

In my opinion, that's also the case, but "unfortunately" in the opposite direction. The switching times of the gasoline direct injection system are about 0.3ms longer than those of the gasoline injection system.
I think that's acceptable for now, because the injection times were already within this range even before all the problems started.

I have some new information about that. Meanwhile, the vehicle is calculating values for the multiplicative mixture adaptation, but according to these values, the mixture is always too rich.
It includes everything from -9.3 to -12.5. Unfortunately, it was only measured at idle.

"Lambdasoll" and "Lambdaist" also work well together. Target = 1 - Actual = 0.965 - 1.03, fluctuating (as expected).

The heating is working (it cycles on and off, with the front sensor heating up first, then the rear sensor, and so on, alternating).

I want to measure the supply voltage coming from the vehicle today.

I find it a bit odd that the lambda regulation sometimes switches to "off." But unfortunately, this happens not when the engine is running at full throttle, but when it's idling.

It seems I still have some learning to do.

That's what happens when you switch from a Motronic 1.3 and VEP diesel engine to such a modern contraption .

Thank you for your feedback so far!

I would appreciate it if you could tell me which individual measurements I should include when creating a log.
The highways are so wonderfully empty.

I recently did a test, but the readings were too low. I measured boost pressure, lambda, and air mass. The air mass reading seemed a bit high. Good, 155g/sec at maximum speed.

I'm getting a pressure regulator on Monday, and then I'm going to bleed the intercooler system. Regarding the issue of power loss and the engine constantly running rich, I can still imagine that there might be a leak somewhere.

Regards,

Martin.

Hello, dear diesel mechanic forum!

In the meantime, quite a few things have happened...

First of all - the car is still running. He seems like a tough guy.

I had a nice logging trip and gave the data to my chip tuner, but he only works with diesel engines .
The log data doesn't look too bad, but it's noticeable that the boost pressure builds up relatively late (starting around 2100 RPM), even though the N75 valve has been at 100% since 1500 RPM.
However, the boost pressure remains consistently good.
Unfortunately, the measured air mass is too high at the upper end of the RPM range, at 160 g/s.

I've pressure-tested the turbocharger air system, and it's leak-free. LLK cleaned, hoses cleaned, clamps tightened.
Everything's fine. No change.

I replaced the engine control unit, but there was no change.

The wiring harness for the oxygen sensor has been tested and is working correctly. All connections have continuity. I wasn't able to measure a specific pin; it leads to a junction somewhere within the main circuit board, and it's a power line.

I have checked and cleaned the main power cables on the motor (e.g., the splash guard, motor mount). Everything looks good. No change.

A hose on the secondary air pump had ruptured, and I replaced it with a new sealing ring. Now everything is quiet again.

And now, here it comes...

It was time to get new summer tires, so I left the car at the friendly workshop in the evening. The next day, new shoes were supposed to arrive, and since the container would be cold at that point, I requested that someone use an endoscope to inspect the first cylinder and the catalytic converter.


A piece has broken off the valve seat of the intake valve on cylinder 1.
Furthermore, the seats (being a gas-powered car, of course...) are all pretty worn out.

Great.

I don't understand why the valve isn't shutting off. Perhaps it's because the missing area is so small compared to the total seating surface... (three intake valves...)

Okay, now it's time to figure out what to do. I've kept the car in excellent condition for years, and you wouldn't believe it has 380,000 kilometers on it. On the other hand, it's still an old car. But you can't make that decision for me.

Thank you for your support! It certainly won't be the last chance for that.

I'll attach the log file later; I don't have it here right now. Maybe someone else can still find something useful in it.

I just wanted to wrap up this topic for now. Resolutions are always the best part.

Greetings & Happy Father's Day.

Martin.

Thank you for the feedback!