2.0 FSI Poor idle when cold

[saabman]

With the above in mind I just had another look at your sensor voltages and it appears that B1S1 is almost the inverse of the other 3 sensors. When 1 goes +Ve the others go -Ve.

I am under the impression that the first sensor is used for mixture control
And the others for checking cat effectiveness. If that is the case and it reports incorrectly rich the ecu will lean things off which seems to be what the othersensors are showing.

Interesting that the 1st sensors base line is about 1.5v and the others are around .5 v at engine start. I need to check but I seem to recall my first 2sensors at around 1.5v at engine start and the others at about .5v. Is this an issue or a function of the sensors in the manifold v the sensors further down the line?

[MEL744]

Yep my understanding that B1S1 is mixture control and the other 3 are cat effectiveness. The reading of 1.5V seems very strange indeed, and I can't find anything that points to a reading over 1.5V as normal. It would be good to see what the sensor 1s are reading on your engine. If both your sensors are reading 1.5V then I guess it is normal? (time for more research I think) If these were cheaper I'd be willing to take a gamble on buying a new one and seeing how it goes. Was going to consider disconnecting the sensor and taking some readings with a multimeter. However the only thing that could be checked is the operation of the resistance heater since we already have the voltage produced under operation. Could the 1.5V be a short of some sort?

I had come across one post on a forum where a cracked exhaust manifold was causing the rough start - might also be worth taking the heat shield off and having a look. Agreed it is absolutely fascinating to look at all the differences between what are on the surface very similar engines.

[MEL744]

Found a YouTube video about diagnosing the VW lambda sensors. This guy is getting a voltage of 1.5V on VCDS, which he says is not normal, and proceeds to check the voltage and resistance of the pins on both the ECU and sensor side. Will try this by the end of the week hopefully. This only applies to narrowband sensors, I'm not sure what these engines have - another thing to find out. Also surely the car would throw an error code with implausible signal if the reading was supposed to be between 0 and 1 volts yet it was showing 1.5v...

[saabman]

Yep everything I have seen on diagnosing o2 sensors says around 0.5V is normal but I noticed in the label file for the ALX engine which is what Im useing as a best guess specifies the pre cat sensors as between 0 -5V and the post cat sensors 0.1 - 0.8V so 1.48 looks like a plausible signal.

id love to see a log of sensor volatges from a known good engine to compare with

[MEL744]

Readings from a good car would be very helpful. I think that the pre-cat sensor is in fact a wideband one while the post cat are all narrowband. When the mixture is stoichiometric the current will be 0. Some cars report this as voltage, others as a lamda value. I have also heard its possible for wideband sensors to respond faster than narrowband, which could explain the spikes from the sensor. I'm not 100% convinced there's a problem with the sensor, really need some data from a car that is working properly.

[saabman]

i agree..

its frustrating that given how popular these cars are how little technical information is available, even an accurate parts list is like unicorn ****

I used to think the dealer mechanics were dumb arses but if they only have this crap elswin system they dont stand a chance

[saabman]

I just got elsawin working a bit better, It ddint work real well under win10 so I just created a winXP VM and installed it in that and its a lot more reliable, I can print and make pdfs as well which is really handy.

The BVY Exhaust sensors are
G39 - 5 wire Lambda probe in manifold in what appears to be connected to cylinders 1 and 4 Bank1 Probe 1
G130 - 4 wire Lambda probe after Bank 1 Cat. Bank 1 Probe 2
G131 - 4 wire Lambda probe after Bank 2 Cat. Bank 2 probe 1
G287 - 4 wire Lambda probe after second Cat. Bank 1 probe 3

The BLX exhaust sensors
G39 - 5 wire Lambda probe in bank 1 manifold for cylinders 1 and 4. Bank 1 Probe 1
G102 - 5 wire Lambda probe in bank2 manifold for cylinders 2 and 3. Bank 2 Probe 1
G130 - 4 wire Lambda probe after bank 1 Cat. Bank 1 Probe 2
G131 - 4 wire Lambda probe after bank 2 Cat. Bank 2 Probe 2
G235 - Exhaust temp sensor just before NOX storage cat
G295 - Nox Sensor Just after NOX storage cat

[MEL744]

5 wire will be a wideband, have seen some posts indicating that 1.5V represents stoichiometric A/F for VW o2 sensors. It appears you can run some Lambda aging tests, so might give those a go - other than that not sure what else to do really.

[saabman]

I had also read about the aging tests but have not been in a position to try it out.

I think getting to understand the significance of the fuel trim figures will lead us in the right direction. Haveing the 2 sensors in the manifold in the BLX engine will help as I can see a distinct difference between the two pairs of cylinders which to me indicates it is not something like the fuel pump or an intake leak.

Even though you only have the 1 manifold sensor you do have 2 further down stream all be it with a lower resolution there may still be enough to narrow the location down.

I’ve been working through a couple of YouTube videos about using fuel trim for diagnosis they are about 1/2hour each. I’m hoping I will learn enough to find the fault with out replacing everything in the engine bay until the fault goes away

Can I suggest logging all the lambda sensors and all the fuel trim readings for a complete journey.

A compression test or even better a leak down test would possibly be useful. I’m contemplating swapping cars with my daughter again for a few weeks so I can have another look and get some more data out it.

We have a couple of 2 liter tfsi caddy’s at work I’m trying to get my hands on one to grab some data out hopefully that will give us an idea of what a good engine looks like. My work caddy is a TDI so I don’t think it would gives us comparably figures.

[saabman]

Here’s the link to the fuel trim video I’ve been working through.
YouTube

[MEL744]

Do you know how to conduct the lambda aging tests? Haven't found anything that detailed on how to start them - have a day off tomorrow so might try and work that out. Will make a plot of all the lambda values and fuel trim values tomorrow too. If you could get your hands on one of those TFSI Caddys that would be great. Even though the sensor set up is not exactly the same it will give us a good benchmark for how sensors should react and values change with time.

[saabman]

couldnt get my hands on the TFSi caddy today but hopefully one day next week

have you got logs of Groups 32, 33 and 36 at idle?

I was looking at your graphs to day and it seems that

[saabman]

Not really sure about the aging tests. I believe you select the the appropriate mb and click the basic settings button then something like press the brake and accelerator.

[saabman]

Hmm I shouldnt post from my phone late at night

If you are able to get some logs maybe 32 33 and 36 at idle for about 30 seconds (after the engine has warmed up) and around 2000rpm for about 30 seconds.

It seems to take about 10seconds for the long term fuel adaption to change im hoping by doing at a fixed load just varying the rpm will reduce the noise from accelration etc.

Have you done a compression or leak down test?

I keep reading about the dirty intake as a posibility and from what I can gather it seesm to be caused by crud preventing the valves from closeing - if thats the case a leak down test will show a leaking valve - I didnt think to do it when I had nats car here

[MEL744]

Managed to get the data for 32, 33 and 36 at idle and holding the rpms relatively constantly at 2000rpm on a warm engine. Also worked out the aging test (you go to basic settings select lambda aging, press the on button then hold the brake and accelerator simultaneously until it completes) - this came back as B1S1 OK. I also went in to generic mode and had a look at the current through the B1S1 sensor as well as some other values when driving around (35 mins) you can view that here (includes cold start). Looks like long term fuel trim stays constant for both tests. Haven't conducted a leak down or compression test since I don't have the tools for it.

2000 rpm plots

Idle plots