Can't Shake a Lean Condition - HELP

[mrbiglar]

So...have tried everything I can think of, but can't shake a lean condition on the car (at idle or WOT). Car actually runs great. Smooth idle, no error codes. Essentially after an ECU reset the car will max out the AF Correction, and then the AFR shows ~20 constantly.

Here is what I have done so far:
1) Replaced the Front O2 sensor with known good OEM sensor (was OEM before)
2) Cleaned MAF (Voltage looked good before, but did it anyway)
3) MAP reading looks good
3) Pre Turbo Leak Test - Ran a smoke test and didn't find anything
4) Post Turbo Leak Test - Same. One really small leak. Corrected, and no change
5) Exhaust Leak Test - Smoke test - No leaks found.
6) Swapped fuel injectors - Put in Cobb 1050's. Map was adjusted to account for change. Had Bosch 900's in it before

One thing...there are times when the car will not start...espcially after its been at temp for awhile. Turns over fine, but nothing. If I work with the throttle while cranking I can get it to start. Doesn't happen all the time, but some. Am assuming this is due to ECU thinking Lean condition (when maybe not actually lean)? Usually after I let it sit, and reset the ECU, it starts immediately. Just mentioning so I don't make it sound like there are no running issues.

Only additional thoughts I had:
1) Spark Plugs - Running NGK Irridium IX's vs OEM. Maybe a seal issue on the plug as well, but you would think smoke test would have shown that.
2) TGV Delete - Tune has compensated for this, although I left the servo's plugged in to avoid error codes. Any way either not truly deleted so ECU not adjusting, or vice veran sa with the servo's plugged in it thinks they are there?
3) Do a 02 Sensor delete and put in Wide Band? Might at least tell me what true AFR is and confirm if I have an issue with ECU\Sensors

So you have a sense for the set up:

Hybrid EJ257 (2.0 Heads) 91 Octane Map right now
2002 WRX ECU
TD05 20G Turbo
Perrin Fuel Rails
Walbro Fuel Pump
Fuel Regulator (Set at 36 pst at Idle w\ Vaccuum)
1050cc Cobb Fuel Injectors
NGK Irridium IX Plugs
PCV Delete
TGV Delete
Zero DB AWIC Kit
Tomei Expreme Exhaust, Up pipe
Mike Everson Stainless with Cat exhaust OEM Front and Rear 02 Sensors installed. No EGT
iWire Harness

[mac1]

Does it have a MAF still between filter and turbo and is it touching the hot coolant tank for the turbo? That's exactly what did it for me.

[mrbiglar]

Good call. I will check. Did yours show Lean immediately though? Mine gets there withing 10 seconds after an ECU reset.

[mrbiglar]

By touching do you mean the MAF sensor itself, or any of the housing? I have a COBB Short RAM. A small corner of the filter is touching but not the sensor itself.

[DMC7492]

Just to clarify, sorry if I missed, but,
Did you have a pro tune after installing into the 818,

Reason I ask ask, my installation with changing the intake injectors and exhaust and turbo my tune was so far off! It ran but threw lean codes. Ended up tuning the closed loop MAF tables using the romraider software

[mrbiglar]

Fair and good question. It has been tuned for the mods listed above. It hasn't been on the Dyno yet as the Tuner and I are trying to figure out what is causing the extreme lean condition in the AFR reading.

Car sounds great both at idle and at simple cruise, so we feel it is a sensor issue, but no sensors left to swap out other than ECU. TGV Delete is in, and tuner felt little to no chance that is causing the issue. Makes sense the more that I think about it.

Double checked the MAF proximity to the coolant tank per suggestion above. There is at least 3 inches of space between, so should be good there.

Going to try different spark plugs next unless others have ideas. Found a post that spoke to fouled plugs and a seal issue on #1 injector causing lean condition readings. Plugs are brand new, but running out of ideas.

[Sgt.Gator]

Maybe the front O2 sensor wiring has a short in it? Very odd that it's stuck at 20 constantly, that sounds like an electrical issue. Maybe a bad relay?

There is a list of diagnostic procedures to check for this in the FSM.

[longislandwrx]

Is your initial DAM still set to 8? Did your tuner change the advance tables at all?

[mrbiglar]

DAM is set to 16

Good call on the wiring check. Its a brand new iWire harness, but can't hurt to check.

[alpine227]

Idle partial throttle are closed loop, but wot should be open loop. If your getting lean in both open and closed loop there is a fueling calculation problem as wot wouldn't use 02 sensor for feed back. Do your air and coolant temps look correct (near ambient on first start) Do you have a stand alone wideband you can use to confirm AFR? Since it's both open and closed loop check that your injector scaling is correct. Odd to idle well running lean however.

[mrbiglar]

Air and Coolant Leaks look correct. Tuner doesn't feel table issue, and I have to agree with him from what he showed me

Next steps are:
1) Replace spark plugs (NGK Irridium IX's) with OEM's
2) Get an updated MAP for Rear O2 sensor delete
3) Install AEM Wideband and see what reading are. Will tell me if sensor issue or true lean condition.
4) Double check wiring harness to make sure no issues

Will let you know what I find

[mrbiglar]

Completed all of the above except the AEM UEGO wideband install (continued changes from original post). Unless someone can tell me I am thinking incorrectly, the only bung I have available is post cat, so won't get a correct reading? Also, if I weld a bung on pre cat, not sure I can make the 36" needed away from Turbo the manual calls for.

Either way, no change to lean. AF Learning 1 goes to 25 within 10 seconds. I did notice on a run though that it goes to Zero as I hit the throttle, maybe past 2500 rpm....goes back to 25 though when I go to idle.

Different direction...• I have never been able to shake the P1400 and P0447 codes. The Fuel tank pressure solenoid is deleted. I have the line running to a catch can and the intake connections are plugged. I have the PCV hooked up, and per normal tied into the T (tree) from the crank case, and then that ties into the lines off of each head. From there it goes to the AOS, and on to the bung on the Pre Turbo inlet. Do I have something wrong here, or just need the codes deleted? Wondering if this could cause my lean condition?

[AZPete]

Re P1400 and P0447, I had the same problem codes that I solved by adding resistors. These codes are not causing your lean condition but you can eliminate this annoyance. I got this info from guys here and do not pretend that I know anything about how a resistor works, or why, but these turned off the codes. I got the resistors from Radio Shack as they were selling off inventory so I just bought what they had, rather than the most correct resistor. BTW, electricity is just a magic passing fad!

P1400 Fuel Tank Pressure Control Solenoid Valve Circuit Low
I wired a resistor between the #1and #2 wires to R68, pressure control solenoid valve. I used two 100-ohm resistors with the ends twisted together (parallel) that give 50-ohm resistance, since the specs call for resistance 10 – 100 ohms.

P0447 Evaporative Emission Control System Vent Control Circuit Open
I wired a resistor between the #1and #2 wires to R69, drain valve. I used two 100-ohm resistors with the ends twisted together (parallel) that give 50-ohm resistance, since the specs call for resistance 10 – 100 ohms.

[mrbiglar]

Thanks AZ Pete. Agree on Lean, but am grasping at straws. Will put your fix in for codes, and hope someone has any additional ideas on what to chase for lean condition

[STiPWRD]

I just read everything from the beginning and will try to offer a few suggestions. But I did have some questions:

Car sounds great both at idle and at simple cruise, so we feel it is a sensor issue, but no sensors left to swap out other than ECU.

So the AFR during idle is 20 and yet the car sounds fine when it idles? Is that correct? This doesn't seem right - I'd follow up with the wideband to confirm. In general, with as many mods as you have, I'd consider a wideband essential. By the way, I'm running a very similar setup as you (hybrid, 02 wrx ecu, 18G, 824cc injectors, larger fuel rails, tgv deletes, etc...)

2) Get an updated MAP for Rear O2 sensor delete

I'm curious what sort of change was made to the ecu mapping here. Typically, the rear O2 doesn't affect the tune, it is there to sense the presence of the catalytic converter and cause a CEL if one is not present. It is there for emissions reasons. I've never heard of a tune needed when removing the rear O2 and have driven my wrx for years (with no rear O2 plugged in) with no issues. I was using the rear O2 bung for my wideband.

Unless someone can tell me I am thinking incorrectly, the only bung I have available is post cat, so won't get a correct reading? Also, if I weld a bung on pre cat, not sure I can make the 36" needed away from Turbo the manual calls for.

The wideband O2 sensor should work just fine in the available post cat bung.

Either way, no change to lean. AF Learning 1 goes to 25 within 10 seconds. I did notice on a run though that it goes to Zero as I hit the throttle, maybe past 2500 rpm....goes back to 25 though when I go to idle.

As others have mentioned, this sounds like an issue with your closed loop fuel mapping and/or sensors since the issue goes away when you are operating in open loop. I'd suggest logging CL/OL to confirm. Could you post one of your data logs at idle? Maybe there is something obvious we can see? What is the MAF voltage? Front O2 voltage? etc...

I have never been able to shake the P1400 and P0447 codes. The Fuel tank pressure solenoid is deleted. I have the line running to a catch can and the intake connections are plugged. I have the PCV hooked up, and per normal tied into the T (tree) from the crank case, and then that ties into the lines off of each head. From there it goes to the AOS, and on to the bung on the Pre Turbo inlet. Do I have something wrong here, or just need the codes deleted? Wondering if this could cause my lean condition?

I agree with Pete that these codes should not cause this sort of lean condition. You mentioned a "post turbo leak test" in the first post - did this include the intercooler, throttle body and intake manifold? Is it possible that air is getting into one of these three things and causing your lean condition?

Also, above you mentioned that you have the PCV hooked up but in the first post you say one of the mods is a PCV delete. Check the PCV hoses to make sure air is not getting into your intake manifold.

Another thought is you didn't properly scale your injectors. What was the value used in the ECU? For example, if using 1050cc injectors, a lower value must be put into the ECU otherwise it will run too lean.

[904svo]

What's your battery voltage with the engine running?

[alpine227]

Post cat o2 sensor will only be a little off from pre cat many tuners use a sensor mounted in tailpipe for tuning(not optimal but works when no other bung exist.) If running mods and not an off the shelf tune you should have a wideband for tuning. 02 sensors are relatively delicate sensors and prone to failure if hit. Mine failed stuck at 10.5:1 AFR. It would read okay during cool start than quickly go full rich. Engines don't idle great in a lean condition typically. Take your sensor out of the bung let it warm up in the open air it should read full lean. Place a lighter near the sensor head with the botton pressed but no flame so only the gas is flowing out should got to full rich after a few seconds to confirm your ecu is reading the sensor correctly and it is in fact working.

[alpine227]

Not sure how the tuning software your using works but usually the map doesn't require much change for bigger injectors just the injector scaling gets changed. Maybe some minor tweaking is need to the map but the primary change should be the scaling.

[mrbiglar]

Thanks all for suggestions. P1400 and P0447 were confirmed to be tune errors. He forgot to turn off in an updated rev he sent me.

AEM Wideband goes in this weekend, pre CAT. I welded a bung. just to be safe...although I agree with the many posts that Cat shouldn't affect too much. Depending on what that tells me, will move on to the other great suggestions listed above.

Thanks all for your ideas. Will let you know.

[mrbiglar]

AEM UEGO Wideband in, pre Cat....sure enough reads 13-14 max at idle. Doesn't exceed it at 4-5000 rpm either. As before, AP shows maxing out at 25 within first 2 minutes.

http://imgur.com/d54aaet
http://imgur.com/xlG9X4O

https://youtu.be/tL1u0vvWYIo

Looks like I have a sensor issue. Stumped, because I had already swapped out Front 02 sensor with OEM, cleaned the MAF, replaced with OEM MAF. Moving on to test 02 sensor and some of the other ideas above, unless someone has additional idea

[STiPWRD]

Looks like I have a sensor issue. Stumped, because I had already swapped out Front 02 sensor with OEM, cleaned the MAF, replaced with OEM MAF. Moving on to test 02 sensor and some of the other ideas above, unless someone has additional idea

It's either the sensor or the wiring from the sensor to the ECU, check both. If both are good, the issue could be the tune. You could try uploading a stock tune (with the exception of scaling the injectors) and seeing if the issue goes away. Sometimes a simple tuning mistake can be hard to track down - this would essentially reset everything. Otherwise, there are several other suggestions to try above. Good luck, you're on the right track.

[frankc5r]

What is your housing like pre MAF ? MAF sensors
are very fussy about different flow from the
air cleaner etc being not stock. It will cause
what you have seen. Can your tuner give you
a flash eliminating the maf?

[mrbiglar]

Housing is a Cobb Short Ram. MAF is OEM. The current one is 2nd OEM. Great idea on trying to eliminate MAF from tune and see what happens.

Next steps:
1) Clean MAF, and then replace once more, just to be sure.
2) Same on 02 Sensor Try 02 sensor test suggested above
3) Check wiring from ECU
4) Go back to my Cobb Stage 2 tune, to the idea listed above.

Will let you know. Thanks all for the help.

Forgot to include data log in last post

https://www.dropbox.com/s/gpgeo06d53...alog2.csv?dl=0

[mrbiglar]

Here is the latest.
1) Cleaned the MAF and replaced. No change. Voltage looks good
2) 02 sensor swap yielded no change. Ran the 2 sensor test suggested above. All looks good
3) Did ground check from the 02 sensor all the way back to the ECU. Rechecked ground for ECU. All good. MM check of voltage coming from sensor matches what is listed on the AP. Doesn't seem to be a wiring issue
4) Have not changed AP Map. Want to check with Tuner to make sure no risks given all the Mods.


In the end Wideband continues to show ~10-12 AF, while AF Correction on the AP continues to peg at 25 at idle, within a minute of start. AF Correction does go to 0 at closed loop

Out of ideas and frustrated. Going to try and put stock fuel pressure unit on, vs. aftermarket. Don't see how this could be issue, but since it is dumping so much fuel , doesn't seem to hurt to remove the variable.

[frankc5r]

Have u not tried removing the maf sensor like
I suggested? Easy for your tuner to give u a quick tuner
change to eliminate maf.

[mrbiglar]

I have not. Will go there next. Thanks for the reminder

Swapped out fuel regulator...no change.

[904svo]

reading your *.csv file it looks like the front O2sensor is open to the ecu. Check all the wires from the front O2 to the ecu and the power and grounds. You may have the sig and ground lead reversed

[mrbiglar]

checking sig and ground today. Thanks for the tip

[alpine227]

I haven't checked the wire diagram but am assuming the 02, maf, coolant and tps sensors are on a sensor ground that routes through the ecu first.

[mrbiglar]

Still waiting for the MAF delete tune. Reetraced the wires to the ECU to test for continuity and connected correctly. Here is what I have:

Reminder, I have a 2002 WRX ECU; Front O2 Sensor is DENSO OEM replacement for 2002 WRX. It is four wire and has Blue, White, and two Black wires. I have an iWire wiring harness.

When looking at male end of front 02 sensor plug, and assuming top left is pin 1 and bottom left is pin3, and top being where the release clip is:
Pin 1: White Wire....traces to D4 and D5 (B137) on ECU. Wiring Schematic says these two are Heater 1 and Heater 2. Assuming this means Heater 1 and Heater 2 are fused together?
Pin 2: Yellow Red Wire...12V Ignition Keyed on and to near ground point
Pin 3: White Wire...traces to D19 (B137) Schematic says this is Sensor + Shows ~3V to near ground point
Pin 4: Black Wire...Traces to D29 (B137) Schematic says this is Sensor - Shows ~3V to near ground point

A few questions:
1) Does the ground for the sensor come from the Sensor - connection? On the wiring Diagram, it looks like most of the sensors get their ground from B19 (B135), but doesn't list a ground to this pin or any other for front o2.
2) Confused...Every Front 02 Sensor wiring diagram I can find shows the Heater 1 and Heater 2 as separate connections vs. just going into Pin 1. You can wire the Heater 1 and 2 together?
3) How can I tell if my signal wires are backwards? I.e. Pin 3 and Pin 4 are backwards

Thanks

[DMC7492]

Hello here a link to the wiring schematic .
http://ken-gilbert.com/wrx/mans/7%20-%20WIRING.PDF

I used the ken Gilbert scematics for my 2005 wrx turbo manual. Study the drawings and you'll be amazed at their simplicity!
I dieted my complete harness and used only the ECU and the engine control and Antilock brake circuits. You'll be able to trace this circuits from the sensors to the ECU plugs. The colors and the pins matched 100%.
The O2 sensors are shown on page 72-74. Only the shields and the sensors ground through the ECU and then the ECU grounds are separate.
Look at the b18 , b19 plug on the bottom of the page and compare your wire colors to the drawing to determine if your looking at the plug end or the wire end of the connector. Looks like only power wire is common between both sensors. The I Wire harnesses are very nice and may be different than the subaru drawings, each is custom to your needs. Hope this helps!!

[alpine227]

Most sensors ground through the ECU that way there is no voltage drop.

[Bob_n_Cincy]

Still waiting for the MAF delete tune. Reetraced the wires to the ECU to test for continuity and connected correctly. Here is what I have:

Reminder, I have a 2002 WRX ECU; Front O2 Sensor is DENSO OEM replacement for 2002 WRX. It is four wire and has Blue, White, and two Black wires. I have an iWire wiring harness.

When looking at male end of front 02 sensor plug, and assuming top left is pin 1 and bottom left is pin3, and top being where the release clip is:
Pin 1: White Wire....traces to D4 and D5 (B137) on ECU. Wiring Schematic says these two are Heater 1 and Heater 2. Assuming this means Heater 1 and Heater 2 are fused together?
Pin 2: Yellow Red Wire...12V Ignition Keyed on and to near ground point
Pin 3: White Wire...traces to D19 (B137) Schematic says this is Sensor + Shows ~3V to near ground point
Pin 4: Black Wire...Traces to D29 (B137) Schematic says this is Sensor - Shows ~3V to near ground point

A few questions:
1) Does the ground for the sensor come from the Sensor - connection? On the wiring Diagram, it looks like most of the sensors get their ground from B19 (B135), but doesn't list a ground to this pin or any other for front o2.
2) Confused...Every Front 02 Sensor wiring diagram I can find shows the Heater 1 and Heater 2 as separate connections vs. just going into Pin 1. You can wire the Heater 1 and 2 together?
3) How can I tell if my signal wires are backwards? I.e. Pin 3 and Pin 4 are backwards

Thanks

Mrbiglar,
Per the schematic below your pin 3 & 4 are backwards. Black wire should be on pin 3 and white wire should be on pin 4 of the sensor.

Question #1: B19 out of your ECU is sensor ground (I call it sensor common) It goes to all the sensors. By having a separate common wire for the sensors, the ECU can measure the sensors properly even when high current and battery sag conditions.

Question #2: The Heater (resistor) in an o2 sensor is turned on with 12 volts. The Y/R wire brings in 12 volts for this purpose. The pin1 white wire is and open collector output (pulls to ground) from the ecu . When the ecu output pulls low, the heater is on. The reason for 2 pins on the ECU probably is the current rating of one pin was not enough.



2002_o2_wiring.jpg

[mrbiglar]

Ok. Looks like I have a direction.....Will fix the plug this weekend, and re-test. Thanks for the help

[mrbiglar]

Update: Sure enough.....The wires on Pin 3 and 4 were backwards on the harness. Swapped, and everything looks and runs great.

Thanks all for your help!

[mrbiglar]

So the wire swap solved my lean condition issue, but now I am having trouble getting the engine to start. It turns over fine, and feel I am getting spark because it will "cough" but not stay engaged. It will also end up starting once in awhile, but it is the exception.....i have to really work the throttle.

Is it possible I am not keeping enough fuel going during start? I ask as I saw a couple threads on this. I also was told I didn't need to reinstall the Subaru fuel dampener on the return side line. Wonder if that could be the issue?

[STiPWRD]

The fuel damper is not the issue - it's just there to smooth out fluctuations in fuel pressure. It does sound like you have a fuel delivery issue though. It could be anything from a bad fuel pump, bad injectors, bad fuel pressure regulator or even not enough gas in the tank. Try installing a fuel pressure gauge and troubleshooting whether the engine is getting enough fuel.

The two other things an engine needs to run is compression and spark so if you can eliminate those as the cause, I'd focus on the fuel.

[mrbiglar]

I have a fpr on the car. It reads ~39 after turn of key, and then the fuel pump disengages. After this, the fuel pump stops and then slowly drops to around ~32. During crank it stays around 32.


fuel pump is brand new wahlbro, injectors are brand new Cobb's, and fuel tank is 90% full.

Here's the kicker. If I can get it to start once.....Once the car is warm it starts immediately every time after. If I let the car cool down and come back the next day, process starts all over again.

[STiPWRD]

Possibly the ECU? It will typically make adjustments to the maps when the temp sensors are warmed up. Maybe some parameter is out of wack when things are cold?

[904svo]

Check that the MAF is installed correctly and that the wiring is correct to the ECU.