O2 sensor leans out after key on

[Chris77]

Hello

This is what happens when amateurs mess with things

 

Short version , afr guage goes lean a few minutes after start , second afr system does the same but comes back at mild rpm/mild load .

I've wired an atom g4x into a off road vehicle running a rover v8 . The carb intake has been replaced with a later model injection intake with factory injectors , custom trigger wheel in the dizzy and wasted spark coils . AEM o2 guage via analogue input . Much to my surprise the stupid thing fired up as I was trying to test the trigger pickup , go figure . Injector dead times are unknown , if anyone knows what 94 rangerover injector figures are or where to find them I appreciate a head up .

Anyway thats about as far as I got as a couple of minutes after start the afr started going lean . My first thought was the cold start was ramping off too fast so I started to chase it with master fuel (this was the first start etc ). It quickly became clear this was a false reading as the engine bogged and belched smoke . Tuning it to max vacuum at idle put me back to the original master fuel . The afr guage dashed out , however the engine seems to be idling just fine , smooth with no misses etc . Wondered if I have just cooked the sensor .

Day two and I swapped on my o2 sensor system from a working setup and it did the same thing , interesting . However lightly driving the thing and the sensor comes back after about 10 to 15 seconds or so , after returning to idle it eventually rolls lean again . During this time the fueling does not appear to change significantly , the motor sounds and runs smooth . So at this stage I think its guage wiring or placement issue . Hence the suspected heat/location issue .

As above , amateur , so bugger all experience with failed o2 setups . First guess is its lack or heat , key on heats the probe so life is good , however the probe is too far down the exhaust so cools off and shows lean .  Drive it a little and the sensor sees more hot gas and warms up thus reads again. The sensor is in the end of the 4-2-1 collector , about the base of the firewall , so further than I would like but not miles away ( I thought ) .

Second guess , well dont really have one . Maybe an air leak up stream , but why is it good for the first couple of minutes and why does it come right being driven . I have had a quick google and one post has the same fault but does not give any solution , another from this forum says after replacing the faulty o2 sensor he found the master fuel was way to high . I do not believe the sensor is faulty as two systems show similar results , different o2 sensors and wideband controllers .

 

So what are the common o2 sensor issues ?

I'll google some more but do the controllers heat the sensor constantly or only at power up , this will debunk my heat theory straight off . How far is too far down the pipe ?

If the sensor is not in the pipe enough can this cause issues , I would have though it would have to be way wrong to give this sort of an issue . The sensor is located at horizontally so not collecting moisture etc . 

 

Sorry for the novel but interested in any responses , cant progress with this until this issue is resolved . Need some confidence in the afr reading . Oh and if it matters (as it was mentioned in the thread with similar issue) master fuel is currently 17 and the fuel table is about 48 (too high to me) . Change of master fuel from 16 to 17 moves the afr (when it seems to be reading ok) from 13.5 to high 14's . Bigger jump than I expected , but if it was working I'd run the higher master fuel and bring the table down .

 

Cheers

Chris

[Chris77]

Ok , well thought about this for a bit and realized its easy enough to answer one of my questions . Hooked my o2 sensor system up to my power supply and it cycles from 1.5 amps to 0.05 amps every few seconds and it did this for 10 minutes , I got bored at this point . So its safe to say the sensor heats continuously , so heat is not my issue . AEM manual says 18 inch from the head and we would be about that from what I've seen . It may not be the 10 deg up per the manual , I did not install it so I'd have to check that . Not that this makes it correct but my sensor a few deg down and it has survived , I know tow wrongs dont make a right .

 

Left field idea , could it have a blown head gasket and be spitting water down the pipe and its buggering the reading . When under light load there is enough airflow over come this therefore it reads . It reads ok for the first few minutes from start as the pipe is dry .

 

Yeh as crazy as my first idea .

 

Cheers

Chris

[Adamw]

I would say more likely a small exhaust leak, vacuum leak or slight misfire on one cylinder.  You only need a tiny bit of extra oxygen in the exhaust to completely mess the lambda measurement up.  Also sources like leaking brake booster, pcv system etc need to be checked.  

If you attach the tune and a log showing the behaviour I can take a look at tune related stuff to confirm, but if it seems to be running ok with the gauge pegged full lean then I would suspect the measurement is off rather than the tune.  Most engines wont idle any leaner than about 1.1lambda.

[Chris77]

Thanks for the reply , even over the festive season . The person who assembled the intake and did the mechanical bits wondered about an exhaust leak , I can see this with the o2 seemingly coming better with more air flow (revs/load) but have not worked out why its fine for the first few minutes after start . maybe its just the cold start after all .

 

I will get the file and logs off the other laptop an post up in the next few days . The lambda sure is a mess in the log . The engine seems to run smooth and does not seem to fluctuate along with the o2 signal .

 

Cheers

Chris

[koracing]

I had a 3" downpipe to exhaust system flange come loose, but not separate and this was 3" after the wideband o2.  Idle o2 readings went crazy lean.  Tightened it all back up and good to go again.

[ac123]

I had something similar to this happen and it ended up being a bad sensor. I guess when the sensor goes bad it reads lean so you run rich so it's a failsafe.