G55 W463 Kompressor rough idle / lumpy

[borat52]

Had a G55 for 3 months now, engine pulls nicely on WOT, but on idle its a little bit rough, almost as though its missing every now and then. It also bogs down a little on part throttle and the gearbox doesn't seem to know what to do on part throttle.

Fuel consumption is also terrible (getting about 9-11mpg in town, 12-13 on cruise).

Had it plugged into to a local mercedes garage and there are no fault codes present. It's also been serviced and had new plugs.

The one point of interest is that during the warm up cycle when the Secondary air pump is in action its nice and smooth, and when the air pump cuts out it starts misbehaving (does the ECU disregard Lambda / MAF readings during this phase?)

Any idea's on where to start? I'm thinking its likely not coils or wires due to the fact its smooth throughout the revs on WOT and also becuase its smooth when the Secondary air pump is running.

 

[jaymanek]

Sounds like a MAF issue then... MAF wont always bring up a fault.

 

[borat52]

Many thanks for the advice, I'll change the MAF and report back either way. Car is a 2005 with 70,000 miles for reference.

 

[BlackC55]

Fault codes are only half the info. The live data needs to be checked.

There is no maf on a supercharged 55.

 

[borat52]

No progress with this, had the G55 into several places to get it looked over, tried new plugs, leads but the problem persists.

Car is not throwing fault codes.

As before when secondary air pump is active, the idle is smooth as silk, as soon as it kicks out there's a roughness to it every now and then.

I've also identified that if you hold the revs at 1500, the car will hunt within a 500rpm rang up and down with a constant amount of accelerator pedal applied.

Any suggestions? My list to try next is EGR, throttle body clean and then lambda/cats.

Worth compression testing?

The fact that it is fine on WOT and fine when secondary air pump kicks in suggests to me that it is closed loop related but it doesn't appear to be badly enough behaved to throw a CEL or to give any diagnostic feedback to the techs through the live data.

 

[borat52]

Plugged my own scanner in, STFT look like this at idle, both banks very similar. Looks too crazy to me, and the roughness in the engine is in time with the turning points on the graph, so it's chocking on lean and rich and the lambda's are constantly fighting a battle with it.

I guess that rules out Lambda/cats as it's unlikely both sides would be cooked.

Any ideas's? List now is:
Vacuum leak
Valve cover gasket leak
Dirty throttle body
EGR
Definately no MAF on these?

Anywhere else to look?