My car had been throwing this fault code:

16839 - EVAP System: Large Leak Detected
P0455 - 35-00 - -

So I changed the gas cap.

Then, the fuel filler metal flap disappeared (I just have the wire that it was attached to sticking down now). I've since looked for the metal flap in the gas tank and under the fuel pump, no sign of it anywhere. Maybe it is stuck in the fuel filler neck.

Then, I used STP fuel cleaner before giving the car several weeks to verify it didn't throw the EVAP code again. The next day, I got this fault code:

17535 - Fuel Trim; Bank 1 (Mult): System too Rich
P1127 - 35-10 - - - Intermittent

I have never seen the EVAP leak code again. This new code became my nemesis for the past 6 months. I changed the MAF, ECT, ECM, TCM, FPR, jet suction pump, small vacuum lines, EVAP purge valve, swapped O2 sensors between banks, etc. The Too Rich fault code eventually stopped happening (though I've now positively confirmed my original MAF is not bad).

However, I've only been getting 17mpg and my measur1ing block 32 values were recently like this:

02:41:47 Group 032: O2S control learning values (before Cat. Conv.) [B=Bank S=Sensor]
-2.3 % @idle B1S1 +/-10%
-15.6 % @part B1S1 +/-10%
-7.0 % @idle B2S1 +/-10%
-5.5 % @part B2S1 +/-10%

Now, here's the weird part that I'd like someone to try to explain. I have a fuel pressure gauge connected and fuel pressure seems to be running normally. Nevertheless, I recently did this really strange procedure: I disconnected the fuel return line from the fuel rail and used a hand pump to blow air into the gas tank (bubbles coming up from the surge housing were audible). The next day, fuel trims come down to within spec. for the first time ever in this debacle, as follows:

19:34:17 Group 032: O2S control learning values (before Cat. Conv.) [B=Bank S=Sensor]
-1.6 % @idle B1S1 +/-10%
-7.8 % @part B1S1 +/-10%
-7.0 % @idle B2S1 +/-10%
2.3 % @part B2S1 +/-10%

Ok, now a week or two later I get the oil changed, the guy overfills, the next day I get him to drain off the excess, but its too late, fuel trims go back to the abnormal and won't recover on their own:

02:41:47 Group 032: O2S control learning values (before Cat. Conv.) [B=Bank S=Sensor]
-2.3 % @idle B1S1 +/-10%
-15.6 % @part B1S1 +/-10%
-7.0 % @idle B2S1 +/-10%
-5.5 % @part B2S1 +/-10%

So, I do the weird thing again last night for the hell of it: Blow air down the fuel return line. The next day (today), fuel trims return to within spec.

20:02:43 Group 032: O2S control learning values (before Cat. Conv.) [B=Bank S=Sensor]
2.3 % @idle B1S1 +/-10%
-7.0 % @part B1S1 +/-10%
-1.6 % @idle B2S1 +/-10%
4.7 % @part B2S1 +/-10%

Okay, if I've still got your attention, what in the world could be going on? How can blowing spurts of air down the fuel return line (my version of a fuel return line plunger), how could that be improving my fuel trim values so tremendously, especially when fuel pressure was good before and after? Is there a tank venting valve or something in the EVAP system that is just out of wack, and the addition of fresh air (which includes adding oxygen into the tank, something I really don't like doing) just clears it up? Then how can a totally unrelated thing like an oil & filter change screw it up? I don't believe my MPG has improved at all despite the better fuel trim numbers, so maybe this is weird stuff is just a red herring, but what do you folks make of this?

 

I wonder if there is a tank venting issue that is causing poor fuel flow. What kind of cap did you get? I have had many a stant cap fail.

 

Though I thought tank venting would show up in fuel pressure, but it doesn't. So weird how something simple like blowing bubbles into the tank has such a dramatic and immediate change on fuel trims. Of course, it is also bizarre how the fuel level sender also reads TOO HIGH until almost the bitter end.

The cap is supposed to be an OEM cap, but yes, I didn't buy it directly from Audi so all bets are off... Again, never got the EVAP large leak code after changing the cap. The Too Rich codes didn't appear until immediately after the STP fuel cleaner was used.