CEL at idle and fuel smell in cabin?

[nmlss2006]

My 90 325iS has developed a nice set of symptoms:
1) Upon the first start in a while, it will not idle correctly: it tries to idle at ~450 rpm... then goes up to 1200 or so... then drops back down... occasionally can't catch itself and dies. After 3-4 miles of driving this disappears.
2) If I sit at idle at a light, more often than not I'll get a CEL. It goes away as soon as RPM go >1000.
3) there is a fuel smell *in the cabin*, not outside, which makes itself known mainly at stoplights/when going slowly. It seems to be more prevalent when the tank is >1/2 full.

All these started toghether after sitting in a 2hr backup on I-76 some 10 days ago. Any ideas? The car has ~139kmi on it and most of the injection system has been replaced in the last 10kmi /2 yrs to diagnose a hot-start issue (which, effectively, remains, though changing the FPR ameliorated it).

[jlevie]

As to (1) & (2)...

Idle or hard starting problems are most commonly caused by intake leaks and/or
a sticky or defective Idle Control Valve (ICV). The only reliable method of
locating intake leaks is to have a smoke test run on the intake and crank case
and to test the brake booster with a gage and vacuum pump. The complete list
of possible causes of an intake leak is:

Intake boot
Throttle body gasket
ICV hoses & connections
Brake booster, hoses, and connections
Crank case breather hose
Evaporative control hoses, valve, and expansion tank
Fuel pressure regulator & hose
Injector seals
Valve cover gaskets & bungs
Oil filler cap
Dip stick o-rings
Oil return tube o-rings

While leaks in some of those can be found by inspection or by spraying carb
cleaner on suspect areas, not finding leaks that way doesn't eliminate the
possibility. Only a smoke test will really work.

Once the possibility of intake leaks is eliminated, the ICV needs to be
removed and cleaned with carb cleaner until the vane inside moves freely. When
the ignition is switched on you should be able to feel vibration from the
ICV. If no vibration the ICV is bad, there's a problem with its wiring or
connector, there's a problem with the TPS, or the DME (or Idle Control Module
(ICM) on an ETA car) is faulty.

For the DME (or ICM) to control idle, the idle switch in the TPS must work
correctly. The switch should close when the throttle stop is 0.030-0.060" off
the idle stop screw.

The fuel system should be tested via the suite of tests in the Bentley manual
as invalid rail pressure can be a contributor to idle and starting problems. A
simple injector check is to pull the injectors, jumper the fuel relay to run
the pump, and see if the injectors are leaking. You can also point the
injectors into a towel, remove the coil wire, and crank the engine to see if
all of the injectors appear to be spraying in a similar fashion. The best
approach to possible injector problems is to have the injectors cleaned and
flow tested. Since raw fuel can or will be released in these tests, have a
fire extinguisher handy.

While a bad check valve in the high pressure pump can result in longer than
normal cranking, if the fuel system is working as it is supposed to the rail
will reach normal pressure in a few turns of the engine. A weak pump, clogged
filter or leaking FPR in conjunction with a failed check valve can result in
longer cranking.

The O2 sensor can be a contributor to idle problems. The O2 sensor is a
scheduled maintenance item with a useful life of no more than 100k. If the
sensor has that mileage or more (or is of unknown age), replace it.

The AFM can be a contributor. If the vane doesn't move freely or the
resistance track is worn the DME may be receiving invalid data from the
AFM. And if someone has fiddled with the bypass air adjustment the DME may be
unable to stabilize idle. The bypass air adjustment should only be adjusted
per the procedure in the Bentley and with an exhaust gas analyzer. And even
then everything else associated with engine management has to first be
operating properly. If the AFM becomes a suspect, replacement with a good used
unit is the best approach.

Improperly adjusted or malfunctioning valves will affect idle and starting. As
can compression issues from ring or cylinder wear. A valve adjustment is
called for every 15k. A useful diagnostic is to run compression (dry and wet)
and leak down tests on the engine. Aged ignition wires, plugs, distributor
cap, or rotor can cause problems. Insulation does break down with time and
heat. And since the youngest E30 is going on 20 years old, if the ignition
system is original or the plugs are old replacement is indicated.

Although not usually a problem, a bad DME temp sensor is a possibility. That
generally won't cause an unstable idle, but can cause hard cold or hard hot
starts and/or a rough idle. As can problems with the timing reference
sensors. Although not commonly encountered, a bad harmonic balancer on an
M20B25 or M30 engine will cause problems.

When all other possibilities have been eliminated and idle or starting
problems persist, replacement of the DME, or if applicable the ICM, is
indicated.

(3) soundslike it could be a fuel leaks, either in the engine bay or at the tank under the rear seat.

[nmlss2006]

Jim,
thanks for the advice. Given the fact that everything you mentioned was diagnosed and replaced multiple times between 06/06 and 07/09, I was hoping to find something more specific that tied all three things toghether. I'll go through the laundry list and report back.

[bimmernutz]

ill say the fuel smell is the fuel tank seal is leaking or rusted up which is simply to ck out. Just pull the bottom of the rear seat up take the cover off...it shoud be obvious if leaking...splashing around up there is why its doing it with alot of full in it...fuel is to expensive to be leaking it.. you may even have a fuel line leaking back there and this loss in pressure would cause the car to throw the cel.

its like being at mcdonalds and your straw splits your not getting the same volume of drink and your having to suck even harder than before..

[nmlss2006]

bimmernutz, I checked: There is no obvious leak and as to the leaking fuel line, I'd expect that to be more of a problem at higher RPM than at idle. I sort of put this on the back shelf since my 318 came in, I'll get back to it later.