2.0 FSI Poor idle when cold

[saabman]

Are there any values that would be helpful to log? Or just all of them like you mentioned?

Im using the scatter gun approach - I do a scan off all the blocks and then try and find some that I think MIGHT effect the cold start... I had been conecntarting on manifold pressure/temperature/egr/injector timing.

i wish it was possible to log everything while starting rather than been limited to about a dozen values.

Its pretty easy to check the gears in the egr you can pull the cover off with the egr valve still in the car, its only about 6 screws.

its also reasonable easy to remove the whole valve as well if you want to clean it out.

have you done a full vcds auto scan ??

if your engine usues the more common purge valve i have a good (new ) that doesnt fit my daughters car i ordered incorectly..

[MEL744]

Fair enough, I will try to log some more values over the next few days, fortunately I can use Matlab to quickly plot everything, maybe then something will jump out. Haven't done a full auto scan, just scanned the engine. Also today while changing the oil I noticed something interesting. When I started the car to move it into the driveway it did its usual high idle. Then about an hour or so later it did it again when it was time to go into the garage. This seems to suggest that there is another factor at play than simply the time (~= 6 hrs) as previously thought.

[MEL744]

Auto scan done - no issues other than door lock fault. I think I will adopt your strategy of testing as many blocks as possible, I'm currently writing up a quick script to automatically process the log files, plot them and export to PDF. Would be happy to process any log files you have too.

[saabman]

That would be good I can muddle my way through excel but thats about it.

Thats interesting that the issue showed up after such a short time - I FEEL like its a temperature sensor issue but so far nothing is showing up out of the ordinary.

Ive been comparing all my blockmap logs taken before starting the car - some will be when the car has run well and others aill be when it runs bad the only problem is I didnt make a note of which ones were good and bad

a couple of things I ve noticed that seem to be significant - Low manifold vacuum (770mmHg), severely retarded timing (-25deg), high throttle opening at idle (13deg)

[MEL744]

I FEEL like its a temperature sensor issue but so far nothing is showing up out of the ordinary.

I was doing some research on this recently, and found someone mentioned changing the pre-cat lambda sensor resolved the issue. Based on the age of the mk5 I would guess it has heated lambda sensors - I wonder if there is some strange condition with the high-idle warm up and the heating of the lambda sensors that causes incorrect readings? Keep in mind also found a million other suggestions for it - will have to see what the logs are showing - haven't plotted any yet.

[Idle]

Might be as simple as an injector not fully sealing.
Cheap piezo injectors in printers are famous for it.

[MEL744]

That seems reasonable, but I have read of other people with the same issue replacing injectors and nothing changes - which is why I am slightly hesitant to change them.

[saabman]

Yes the pre cat lambda sensor seems feasible. I wonder how we could see that I the data collected. If the ecu can see a problem the logs of data from the sensor should indicate it as well.

It’s obviously providing a reading that is valid but not correct for the rest of the engine perhaps if it’s say reading rich the ecu would lean out those cylinders this would increase knocking which would retard the timing ???

I remember seeing some knocking on cylinders 3 & 4 but put it down to the engines running crap of course it’s going to knock I’ll have to have another look at the data.

I’m expecting to see low injection times for cylinders 3 and 4 and longer times for 1 &2

[saabman]

Might be as simple as an injector not fully sealing.
Cheap piezo injectors in printers are famous for it.

An injector not sealing I would expect to show up as a drop in fuel pressure over time - I haven’t seen that yet ...

I’m using this fault as a good way to learn how to read the data from the car and am planning on not replacing anything unless the data shows a problem. We have amazing tools that give incredible insight into the running of the engine which I can only dream of on my Kombi. The trouble is interpreting the data in a meaningful way

[saabman]

I started doing a bit of research into this and come accross this page about reading lambda values in conjunction with other gas readings for engine diagnosis https://www.austincc.edu/wkibbe/lambda.htm

[MEL744]

Interesting read, definitely clears up a lot. I've been looking through the logs and I'm not convinced the label file is correct when it comes to values associated with the cat temp. In measuring block 43, the Catalytic Converter Bank 1 Temp varies from 12-150c, while in measuring block 46 it varies from 12-60c on the same start. Then in block 34 on a different start it varies from 12-714c. This makes me question whether the Lambda labels are correct as well. The supposed Lambda sensor values don't really seem to change with time. They remain constant at about 0.44V on Bank 1 and 0.435V for Bank 2, which would indicate a slightly lean mixture according to some data I have found on Lambda sensors. (0.45 - 0.5V apparently corresponds to stoichiometric air/fuel ratio).

[saabman]

The information in the Label files does seem to be a bit vauge and "best guess" with the amount of variations in VW engines Im not surprised no one can really hit it on the head ...

I was watching a youtube video about Fuel trim at lunch time today YouTube
and from what I could gather from that the voltages from the O2 sensor should remain pretty well steady ( should oscillate around 0.45V ) if the they sense a richer mixture the ecu should reduce the injector timing if the sense lean the ecu should increase injector time this would be seen as a variation of the fuel trim values. The ECU should be doing everything it can to keep the O2 sensor values constant. There are times it should vary like when accelerating or decelerating.

[MEL744]

Alright so from what VCDS is telling me, there is no oscillation whatsoever. It just stays on 0.44V or 0.435V, based on how the engine should be reacting there should be the oscillation in voltages from 0.1 to 0.9V. I am not sure if it is displaying a constant voltage because it is cold (can it even reach operating temperature in 30 seconds with heating?) or it is broken. Will try to make a log of the voltage when driving around today. Have also found that the second O2 sensor is downstream of the cat and used to compare O2 content after the exhaust has been through the cat, in order to check whether it is working properly.

[saabman]

I think your right about it not reacting because its cold. i dont have a log of the actual o2 sensor voltages but have Group 1 values which are the Lambda control values - Im assuming this is the calculated raw trim value to be applied. I have 2 lambda sensors 1 for cylinders 1 & 2 and 1 for cylinders 3 & 4 you can see in the graph there is nothing from them for 30seconds and then they start producing a signal that signal is then not applied to the fuel trim values for another 10 seconds

[saabman]

Given the current readings from the O2 sensors dont have an effect on the running of the car at start up Im wondering if there is something occuring when the engine is hot and that is pushing the trim values right out, thta figgure is stored and used as the start up value but is not right for the condition of the engine at that time -and the engine runs like crap untill the o2 sensor values kick in?

Apparently clearing engine fault codes resets the fuel trim values - I wonder if clearing the codes prior to a cold start makes a difference? Unfortunatley I cant try this out as my daughter lives 200km away so I can only work on the data I have - I might have to send her up my HEX-can cable (over my dead body LOL) I should see if the light version will do logging...