Hi, everyone. First time posting here although I've found myself here many times before now. For the better part of three months now, I've been trying to pinpoint the cause of an idle misfire that appeared shortly after I purchased my 2006 WRX (EJ255 w/ D25 heads @ 147k). No mods except a secondary air pump delete, Perrin 2.4" turbo inlet, and a catback exhaust. The initial symptom was an extremely rough warm-start idle that would decrease in severity over the course of 10-20 seconds--this presented with consistently negative fuel trim data, fluctuating AFR, cylinder roughness counts and associated codes for misfires on cylinder 1 and 3 via BtSsm. Since performing the repairs or replacements that I am about to list below, this particular symptom hasn't manifested itself. However, what remains is a non-cylinder specific misfire that occurs only in closed-loop during idle, which is accompanied by a negative fuel trim of -11% to -15%--learned values reflect that fuel is only being pulled at idle. FKC occasionally drops to a minimum of -2.71 for very short periods of time, which is within normal range as far as I can tell, and this is not accompanied by a change of FLKC--learned values indicate that timing is not being pulled due to knock. The misses are very easy to feel and hear, and they register more than 50% of the time as misfires via cylinder roughness count. They registered much more often when I was running oxygenated 91, but I have since switched to 93. Repairs, replacements, diagnostics: Compression test--140 psi ±5, cranks evenly; doesn't sound like low compression Leakdown test--did not personally perform but was told that numbers "came back good" Replaced spark plugs--no change, plugs are OEM and properly gapped Replaced front O2 sensor--seems to have eliminated the violent warm-start idle/misfire Cleaned and swapped OCVs from each bank--no change Cleaned MAF sensor--no change Checked FPR function/fuel pressure--fuel pressure normal and FPR functions with no leaks Cleaned injectors, swapped injectors, replaced injector O-rings--found a vacuum leak at #1 injector seal, most likely my fault for having previously removed injectors to clean them Cleaned throttle body, replaced throttle body gasket--no change Vacuum leak test of intake--no leaks found aside from above, replaced turbo inlet preventatively Inspected and cleaned all chassis and engine grounds--no change Swapped coil packs--no change Disconnected front O2 sensor and reset ECU to remove learned fuel trim values (effectively disabling closed-loop) after suspecting that it was the cause of my aberrant fuel trim data and AFR. My car ran perfectly during the week that it was disconnected (aside from limp mode limiting boost). Ordered a new front O2, installed it, and found that the rough idle in general, minus the violent 20-second shaking on warm starts, came back. I was unable to monitor AFR during the time that the front O2 was disconnected as I do not have a separate wideband sensor. However, the negative fuel trim data at idle preceding the replacement of the front O2 sensor in conjunction with the rich AFR (11-13) that I was seeing during open-loop operation makes me suspect that I was, in fact, running much richer than 14.7 during the time that I had the O2 sensor disconnected. However, I don't understand how running richer than stoich would alleviate my symptoms in any case. Conversely, my initial suspicion was that the front O2 sensor was inaccurately reporting a rich condition which the ECU then corrected for by cutting fuel, thereby creating a lean condition and the misfire at idle. Could this still be the case if the MAF sensor were reporting inaccurately? Is it possible for valve overlap to emulate a rich condition that the O2 sees and tries to correct for? Essentially, I'm not sure whether the rich condition is causing or alleviating my misfires. Again, the car runs perfectly without front O2 connected. The plan for now is to disable closed-loop altogether in RomRaider simply because it runs so much better in open-loop when fuel is not being pulled. I could simply disconnect my new front O2, but I lose the ability to monitor AFR and also risk fouling the sensor. I'm open to any and all suggestions or ideas as far as what might be causing these misfires. It's consumed far more of my time than I care to admit and has caused me a great deal of stress. Thanks in advance.
My best hypothesis: You have one rich (flowing more fuel for the same on time command) injector. When the computer subtracts fuel (-15%) for rich injector it makes the other cylinders lean causing lean random misfires. Thus when A/F sensor is disconnected and it is forced into open-loop, all of that doesn't happen.
Stock tune? And to follow-up on the injector cleaning, was this DIY, what did it entail, and have you flow tested them?
Yes, stock tune. The cleaning was DIY; a friend of mine and I removed all four injectors and ran carb cleaner through them with a 12v battery. Spray pattern appeared normal on all four injectors, but in my limited experience I am unsure if a spray pattern errant enough to cause a rich condition would be readily perceptible without proper flow testing, which was not done.
Alternatively to unplugging the O2 sensor, try unplugging the MAF. This will cause it to run in open loop as well, just to see if the problem persists. As to the possibility of a stuck injector, cleaning won't fix a worn shorted injector solenoid wires. If it has a short and gets "electronically" stuck open or the signal from the ecu is errant, no amount of cleaning will fix it. You might have to bite the bullet and bring it in to a professional, most people recommend JMAuto.
