Fuel pressure Good...good. Thanks for checking.

How about the vacuum lines to the Fuel pressure reg. How many places through out the fuel delivery system was the fp checked. Meaning...either side of fpr...return pressure.

I bet vacuum hose leak...fpr or intake mainfold. $$$ on it. A "Lincoln" shall we say. Gentleman's wager..

 

Negatory. Replaced all 4 lines (In and out of purge control, FPR, and purge control into intake manifold) with no luck.

Did an HC test on the coolant, which thankfully was negative; don't really feel like doing head gaskets off the clock...

Pulled the #1 spark plug to check since it's brand new, nice and brown. #3 however was pure white across the ground, and had no burn indication on the electrode. Replaced that injector, and now I have a flashing CEL on startup for the #1 misfire.

I'm pretty sure it's arbitrary since return pressure is passive-ish, but I was seeing 12-15psi on the return line

So basically, no progress as of yet today.

 

Could a stuck valve be causing this.

 

Compression test was perfect and the valve train was totally spotless, so I highly doubt it's valve train related.

 

Had access to another fuel pump from an XT, dropped that in and no change. Cylinder 3 plug still white as a ghost, cylinder 1 misfire CEL, and no/very little change in idle when disconnecting #1 injector. #1 injector plug has 12+v power on as requested by the FSM.

I'm at my wit's end for the evening, and hoping one of you intuitive people have a suggestion for me in the morning. I'm too frustrated at this point to fiddle with it anymore.

 

Coil pack, crank sensor, cam sensor all changed with good used this morning; no change.

About time to throw the towel in I guess. Unless the ECU took a dump or the o2 is somehow causing a misfire with no coding.

 

Brainstorming . . .

Cylinder #3 seems to be working, but the white plug is interesting. #3 is the cylinder at the back of the engine on the right (passenger) side.

We are talking about this one and not the cylinder on the opposite side of the engine from #1, that is, at the left front of the engine, which is #2, right?. I ask because in the wasted-spark system cylinders 1 and 2 are paired, as are 3 and 4. If there's a problem in #1, it could affect #2, and vice-versa.

The plugs and wires are new, but the common element between #1 and #2 is the coil.

Does your 05 2.5 have the coil mounted on the top center of the intake manifold, or behind the intake manifold on the right side?

Also, did you check the intake manifold vacuum? It might be redundant, but it's not difficult to do and could reveal something that the compression test didn't.

I noted that the #1 injector has 12 V, but is the injector actually being triggered? There's two wires to it; one should have 12 V (battery voltage) when the key is at ON, the other is at the same voltage except when the ECM triggers the injector at which point that second wire goes down to near zero momentarily. I'm wondering if the injector isn't working, not because it's bad (you've tried a substitute), but because there's a fault in that second line. (Either there's a break/ground in the wiring, or the ECM has a problem.) A noid light could help identify if the electrical signal to the injector is working. Also, a mechanic's stethoscope could be used to listen at the injector for a noticeable click each time it's pulsed, and that can be compared to the sound of the other three injectors. An absent or "soft" click compared to the others could indicate insufficient current to rapidly and fully open the injector.

 

(4) NGK spark plugs
Set of new ignition wires
Good used injector in #1 cylinder
Again, good used injector in #1 cylinder
Fuel pressure tested according to FSM and has return pressure
Timing marks are on point

10-20-15

Tested with new coil pack and good used cam/crank sensors
Changed MAF
New denso oxygen sensor


Still runs rough at idle. Romraider data logging shows 100/100 misfire on cylinder 1 at idle, 0/100 with any throttle application. Flashing CEL for cylinder 1 misfire at idle, no CEL/misfire under any throttle application. Unplugging the MAF causes 0/100 misfire on cylinder 1, but still idles rough and obviously has it's own set of problems. Changing the map to increase idle does nothing for smoothness/misfire readings.

Injector "click" is identical in sound and rhythm across all cylinders

However, the car drives absolutely fine at any throttle input over idle. No hesitation, no roughness, no misfire pickups in any cylinder until I slow to idle...

At this point, I am so far beyond stumped and confused, not to mention far far past any reasonable amount of time/money spent chasing a problem that shouldn't technically exist, that I think it's about time to bag it and send it to another shop, I'm too shot to mess with it anymore.

 

I appreciate the frustration . . .

But the last post has some interesting, new, information; namely, the difference between idle and throttle application, and the ability to log with Romraider.

Have you logged engine parameters, such as the sensor signals, fuel trims, injector PW, timing, MAF g/s, MAP, engine load, etc, at idle, and separately, with the throttle advanced? (Does the car have to be driven, or just raising the rpm is sufficient to change from 100/100 to 0/100?)

There has to be something . . . it might be a matter of looking in the right place. Perhaps that might be the data that is available through Romraider logs.

 

It's any touch of the throttle that brings it out, park or drive, moving or not. I drove it out of necessity to pick the kids up, and there were a couple of stops where the car was much smoother at idle, but it still had a very distinct "running on 3" power level and behavior. There were also a couple of stops that the car was shaking very violently, and they were mixed it. Almost like an intermittent short.

I haven't actually logged anything outside of real time data as I drive, as I don't know enough about the correlation of parameter X and Y to do anything more than send myself further down the rabbit hole. If you have any input on the logs, I'll pull whatever parameters you can make heads or tails of.

I'm ashamed to admit, but I really only have the tactrix/romraider for deleting codes on WRXs for things like TGVs and air pump deletes, I really need to learn more about logging.

 

Upstream 02 sensor, had this problem today

 

Of course this morning it's decided to run much smoother with intermittent bouts of roughness every other stop and go. Here's the logs, hopefully you can make heads or tails of them, I didn't notice any drastic anomalies looking over real time data, but then again it was a lot of windows to keep track of.

Thanks in advance for the help.

 

Not that I intend to put much weight into this, but the car has began to run smoother and smoother as the day wears on (Using it to get parts) after running a can of Lucas fuel system stuff in the tank. Then the wife mentions that she got gas from a sketchy little place on the highway to go visit her mom. If this whole bout of automotive malarkey is due to bad quality gas, I will be less than pleased. It's about 1/2 a tank now, I'm going to run it nearly empty and run a tank of v power to see (Assuming it continues improving) if that smooths it out more. As of right now, it's only popping 3-7/100 misfires at a stop and 0/100 while driving.

 

The roughness monitors aren't reporting any misfires, at least within the log time-frames.

I haven't had a chance to spend much time yet looking at the logs and comparing to others I have, but what does stand out is the Rear O2 sensor voltage. In both logs, it's below 0.1 Volts (there's one short exception in the 1k rpm log), whereas it should be somewhere around 0.7 V most of the time. A low voltage is often seen with a lean fuel mix, but it could also be a faulty Rear O2 sensor.

The A/F Correction and A/F Learning fuel trims don't appear bad. The A/F Learning #3 is on the negative side and fixed at -0.31, whereas the chart suggests it should be on the positive side. Not yet sure about the significance but the combination of the fixed Rear O2 voltage and fixed A/F Correction #3 might be connected and due to a common fault. [In addition] the fixed low O2 voltage could indicate that fresh air is getting to the sensor. An exhaust leak in the area of the sensor, or even further downstream, could cause this.

That said, the news about that last fuel fill-up could be relevant.

If you have a chance, perhaps run another log, but this time drive the car from stops to cruise to stops etc for a minute or two. If the Rear O2 voltage doesn't vary, then it's probably not working. In the log, add vehicle speed and engine rpm. Remove the accelerator and sub accelerator voltage PIDs but keep accelerator opening angle. Same for the throttle, drop the voltage readings, keep the opening angle.

Incidentally, where are you getting the misfire counts (7/100 etc)?

 

Do you know if you carry an EGR?

 

I'm having similar problems with my 09 OBXT, but haven't swapped out a bunch of parts like you have. I'd be interesting if this problem was ever resolved, and if so, what fixed it.

@ cal_look_zero; can you please provide an update?

 

So ultimately, the rear oxygen sensor being bad was the culprit. Odd as it sounds to someone who works almost exclusively on turbo Subarus and can basically delete the rear o2 as it doesn't account into a/f ratios.

Anywho, problem solved.

 

Correct. I have an aftermarket sensor and gauge in that location. I get a readind from the sensor and it is down stream...so it registers what the car actually "finishes" after the ecu uses the data from the front.
But at least we know know that a bad rear O2 sensor can cause this

 

Interesting.

I looked at the list of Bank 1 Sensor 2 codes for my '07 and for the '05 2.5. The 2007 has a distinct code for O2 sensor "inactivity":

DTC P0140 O2 SENSOR CIRCUIT NO ACTIVITY DETECTED (BANK 1 SENSOR 2)
1. OUTLINE OF DIAGNOSIS
Detect abnormalities in the rear oxygen sensor output characteristics.
By referring to the intake air amount, engine coolant temperature, main feedback control, deceleration fuel
cut and other operating conditions, if the rear oxygen sensor voltage should be moving under these conditions
but is showing a low voltage, this is judged as a Low side NG. If the voltage is showing a high voltage,
it is judged as a High side NG.
When either Low side or High side is NG, judged as rear oxygen sensor property NG.

I could find no similar code in the 2005 FSM. I suspect that the low voltage signal shown in the Romraider logs would have triggered a CEL if there had been a P0140 (or similar) diagnostic.

I think this also demonstrates a link between the rear O2 sensor and actual engine operation in these engines, as opposed to some systems where it appears the rear O2 sensor only serves to monitor the cat converter performance.

 

Hmm.

I would have to talk with Ed to see if he has specifically disabled that code. If it exists on the '06 turbo. There was a way to spider the Innovate sensor and gauge harness into the stock sensor harness...But I didn't install that way. I just unplugged the stock harness and ran the new harness. I log the wide band from a separate data port. So I don't, can't log the stock rear O2 anymore...it isn't there. But is doesn't pop any codes related to O2. It pops the EGT code, that sensor is replaced also with an Innovate sensor. But Frank isn't happy with false readings. There some resistor fix for that...but that's off topic.