Fault Code 17538, running lean...
 

I've got a fault code which keeps popping up.

As above, it's 17538, fuel trim, bank 2, system running too lean.

With the way the system works with faults like this, it has to happen a certain number of times before it bring on the engine lamp on the dash.

By the time it's done this, the corresponding fault code 17536 is also there but it's always the other bank the is first with the fault code.

I've been researching, checking lots and the obvious ones to look at are air leaks, air flow meter, lambda sensors etc.

I've cleaned the air flow meter, checked the readings which register around 20g/s at tickover and 60-80g/s when you boot it. There is a spike when you initially boot off tickover which shows around 160g/s.

Also, readings are similar when driving along. At 2,000rpm it's reading 18g/s and 60g/s when you boot it with the spike as above.

I've checked all over the inlet manifold and found a small plug on the front which was very perished and splitting so I've replaced that.

I've been running it on cheap fuel so this tankful, I've put super unleaded in. No change.

The lambda sensors are all reading similar voltages, the cats are at similar temps etc.

The one thing I've been watching is measuring block 32 which is 02S Adaption Maxims. From what I can see, Bank 2 running leaner than bank 1. See photo below.

http://i690.photobucket.com/albums/v...IMAG0677-1.jpg

I'm still investigating but if one bank is running leaner than the other, that would suggest that the air flow meter is not the problem...

Any ideas?

MG

running lean
 

I feel for you...a problem like this is annoyingly difficult to resolve as there are so many possibilities.

Could be as simple as a dodge fuel injector by the sounds of it....but I would try swapping things from bank1 to bank 2 a bit a time and see if there is any change in the fuel trim....

Lambdas are a common weak point, and on your AVP there are pre and post cat lambdas, doubling your chances of a tired one I guess.....

Swap em between banks like Ranj says and see if you problem moves.... ?

Ok. I guess 185K miles is enough for any injector to be not at it's best.

I'll swap them over bank to bank and see if the problem follows the injectors.

MG

Blocked injector is an option but that would most likely throw up misfire codes as well unless its all injectors slightly blocked on that side and not just one. When you swap sides you will see if its that or something else. Try also swapping lambdas.
There is quite a difference between both sides which could mean there is breathing problem one bank getting more than it's fair share of air if is the same after swapping injectors and lambdas.

You can also test lambdas on a bench while off the car with blowtorch and voltmeter.

Well the cause has been identified.

This last week the readings in measuring block 32 have been at their highest and then this evening, the engine died on the way home.

It's the fuel pump that's died so it'll be running lean due to lack of fuel pressure.

MG

Recheck measuring blocks after changing pump as the difference is a bit high between banks -08 .. +10.9% giving total 11.7%. IIRC absolute value must be within -12.5..+12.5% but difference between banks no more then 6% on A8 and 4% on S8 engines