Subaru Outback Forums banner

1 - 15 of 15 Posts

·
Registered
2005 OBXT Limited, VF37, STI intake, 5MT
Joined
·
1,553 Posts
Discussion Starter #1
I'm looking for a little confirmation as I'm pretty confident I know the answer. Might also serve as a nice point-of-reference for future issue trackers.

Vehicle
2005 Outback XT 5MT
EJ207 Ver.8 swap
VF37
Catless uppipe
Factory twin-scroll downpipe
Factory cat-back
STI intercooler
06 WRX intake manifold

Issue
Under light load, car would "surge" rythmically. Similar to engine braking, then lightly accelerating.
Instantaneous MPG indicator would show "normal" gas mileage, then spike to +20mpg, in-sync with the "engine braking"
Did it when cold, operating temp, after every start/restart
No issue while decelerating or accelerating. Med-heavy accel, no issues.

Diagnosis
Using RomRaider, I monitored the following values to rule-out and underlying/hidden issues.

A/F Sensor #1 (AFR)
Engine Load (4-Byte)* (g/rev)
Engine Speed (rpm)
Feedback Knock Correction (4-byte)* (degrees)
Fine Learning Knock Correction (4-byte)* (degrees)
IAM (4-byte)* (multiplier)
Ignition Total Timing (degrees)
Knock Correction Advance (degrees)
Knock Sum* (count)
Manifold Relative Sea Level Pressure (4-byte)* (psi relative sea level)
Mass Airflow (g/s)
Primary Wastegate Duty Cycle (%)
Intake OCV Duty Left (%)
Intake OCV Duty Right (%)
Intake VVT Advance Angle Left (degrees)
Intake VVT Advance Angle Right (degrees)
Boost Error* (psi)

This list is a bit overkill, but I wanted a good picture (and my laptop/OBD2 cable can handle it without killing the sampling rate). I also had a decent idea where to look as I had this same issue about 6 months ago.

Analysis

Starting by isolating the engine load as that was the indicator of "cruising load". By comparing all the other values (OCVs, MAP, Knock, and other values that could effect engine operation) to the engine load, the only value that varied wildly was the AF Sensor #1. By watching the OCV values (for degrees and duty cycle), I could see they were steady (and not modifying valve timing). Manifold Pressure was steady, wastegate duty was steady, and no odd timing events such as Knock Retard fluctuating rapidly.

Verdict

The Upstream O2 Sensor (AF Sensor 1) has gone bad. The rapid oscillation as seen in the logs and snap-shot is the only varying value. I also confirmed this by disconnecting the AF Sensor and repeating the conditions. The issue was non-existent during "default" mode.

Where I'm a little skeptic is that I had installed a new Denso sensor about 6 months ago. While it's not impossible to sensors to fail, I will be watching closely and investigating why this sensor decided to go south.

@plain OM @traildogck Anything I missed?
 

Attachments

·
Registered
2005 OBXT Ltd Obsidian Black 5MT
Joined
·
1,079 Posts
So I am dealing with an almost exact same issue with mine.

I also replaced the Upstream sensor with a brand new Denso a few months back in an attempt to see if that would solve the issue, but it persists.

I've been focusing on the downstream sensor on mine. It's the only thing I haven't replaced yet. I've seen some strange things in the logs with it fluctuating without the AFR changing.
 

Attachments

·
Registered
2005 OBXT Limited, VF37, STI intake, 5MT
Joined
·
1,553 Posts
Discussion Starter #3
So I am dealing with an almost exact same issue with mine.

I also replaced the Upstream sensor with a brand new Denso a few months back in an attempt to see if that would solve the issue, but it persists.

I've been focusing on the downstream sensor on mine. It's the only thing I haven't replaced yet. I've seen some strange things in the logs with it fluctuating without the AFR changing.
While on the way to work this morning, it struck me to watch the Throttle Actuator system (Pedal position and throttle opening). I'd think the engine load would be all over the place if it were actual "kangaroo" action of the throttle, but maybe not.

As for the downstream, mine is disconnected and codes turned off, so I don't think that's playing into it...

Also, @Logie_OBXT what program is that for graphing? It looks easier than Excel.. lol

Edit: I did stall the engine (by accident) the other day and a day or two later, this appears.
 

·
Master Caster
2005 XT, Mildly Modified...2006 XT Limited, Highly Modifed
Joined
·
16,119 Posts
Can you post the csv. file? I am curious, what going on with a few parameters during the surging and how they relate to one another.

accelerator pedal angle
engine load
engine speed
fuel pump duty cycle
manifold pressure
throttle plate opening
 

·
Registered
2005 OBXT Ltd Obsidian Black 5MT
Joined
·
1,079 Posts
That's what I suspect as well.
And see, I was kinda thinking along vacuum leak for mine as well, seeing as it was doing a very elliptical idle. But I'm not seeing it in the logs. Unless I'm not looking for the right thing. I've attached my logs from a cold start and drive to work from the other morning for you to compare.

If the Denso AFR was faulty, then I'd see disparity between it and my Innovate wideband AFR, but they track almost identically.
 

Attachments

·
Master Caster
2005 XT, Mildly Modified...2006 XT Limited, Highly Modifed
Joined
·
16,119 Posts
@Logie_OBXT

I see fluctuations in engine load, manifold pressure, and throttle plate opening when the accelerator angle is constant. Minor fluctuations.
 

Attachments

·
Registered
2005 OBXT Ltd Obsidian Black 5MT
Joined
·
1,079 Posts
Here's a few logs, some without the throttle data, few without.
Yeah, you've definitely got a leak somewhere and it looks like it's after the turbo inlet. It's not making any boost.

And that seems to be causing the AFR to fluctuate like it is. I'm seeing similar on my rear O2 sensor, but not my AFR.

I've attached a log where I let my car warm up the other morning. It was super cold. It's somewhat choppy all the way through, but the rear O2 is like a sinewave, similar to how your front sensor is reading, but at a higher oscillation.
 

Attachments

·
Registered
2005 OBXT Limited, VF37, STI intake, 5MT
Joined
·
1,553 Posts
Discussion Starter #12
Yeah, you've definitely got a leak somewhere and it looks like it's after the turbo inlet. It's not making any boost.

And that seems to be causing the AFR to fluctuate like it is. I'm seeing similar on my rear O2 sensor, but not my AFR.

I've attached a log where I let my car warm up the other morning. It was super cold. It's somewhat choppy all the way through, but the rear O2 is like a sinewave, similar to how your front sensor is reading, but at a higher oscillation.
Most of the logs are little-to-no boost, so I'm not too surprised with that. At ~0.5 g/rev, and near 0 WGDC, I'd hope to not be in boost too bad.. lol
 

·
Registered
2005 OBXT Limited, VF37, STI intake, 5MT
Joined
·
1,553 Posts
Discussion Starter #13
Small update: Problem has been resolved, but I will gather a few logs to confirm.

TL;DR Upstream O2 sensor was giving erroneous readings, causing the surging issues.

After looking into replacement options (including the "correct" EJ20X O2 sensor with the short pigtail for twin scroll BP chassis), the best option was to replace with the chassis-correct O2 sensor (for the EJ255). This would match the original calibration in the ECU and give me ample pigtail to connect with the strut-tower connection point(as the EJ20X sensor pigtail would not; too short).

I had an 8 month old Denso unit in the car (matching to the EJ255-chassis), but I purchased a replacement Walker unit as the Walker unit appeared to have the 20X sensor snout (pencil shaped rather than multi-hole cylinder). Upon receiving the sensor, is was a standard multi-hole cylinder. I installed it as it was the only one on-hand.

But as for why the Denso unit failed, I suspect it was carbon loading from the rich tune that is currently loaded into the car. This was also confirmed by the increased ~2 mpg fuel mileage I experienced while running without the O2 sensor connected. This was partially expected (as the LVT also shows negative values for fuel trims in the midrange), but just confirms how rich this tune was. Time to reevaluate the tune (now that things are settling down again).

This is about 100 miles on the new sensor, and I plan on gathering data logs soon to confirm.
 

·
Master Caster
2005 XT, Mildly Modified...2006 XT Limited, Highly Modifed
Joined
·
16,119 Posts
This does make sense. Hopefully, as you stated, logs will confirm.
 

·
Registered
2005 OBXT Limited, VF37, STI intake, 5MT
Joined
·
1,553 Posts
Discussion Starter #15
I think was a bit of a case of "tail wagging the dog". One thing I wish I had logged was final fueling data vs. ignition data vs. O2 data. I would expect to see an inverse relation and slight time offset between final fueling and O2 readings. I still have the old one, may put it back in for testing.. haha
 
1 - 15 of 15 Posts
Top