EJ22 Idle and Intermittent misfire underload

[1980ea71Brat]

7 hours ago, lmdew said:

I have sets of Red Injectors.  Pay shipping and a Donation to the USMB and they are yours.

Larry

I sent you a message, thanks for reaching out.

[1980ea71Brat]

Well I wish I had better news today. I got off work early and rewired the 02 sensor, extended the lines about foot so they wouldn't rub on the axel. Didn't see anything wrong with it. Pending code never came back. 

I measured the new ECT vs the old ECT and they measured the same - or close enough to not make a difference. 

During a ~150 mile drive I had a P0505 code pop up. I tested it (brand new) and everything is within spec. Plugged it back in, cleared the code - it didn't pop back up.

Out of ideas! Unfortunately I think I am going to have to wait a while, save up some money and take it in to a dealer with the SSM code reader that can diagnose my problem more precisely. In the meantime I'm just going to drive it and keep looking for answers on the internet.

Kind of seems like it's probably going to be an issue with the wiring harness itself. Maybe a break on the circuit causing a fault? Corroded connectors? (I cleaned the main harness and it didn't look bad at all). Injectors?? I'm running injector cleaner, but I'm not a big believer in that stuff.

If it comes down to getting it diagnosed at the dealer, I will report back. Just going to take a while to save up some cash after I drained the envelope in parts and a new interior. 

[nvu]

Could your ecu still run in limp mode with the maf unplugged?  Just as a sanity check, drive the car and warm it up until it does the stumbling in idle.  Turn off the engine, unplug the maf, start it up again.  It should be in limp mode and running rich, but should be idling smooth.  If it's still rough, that could point to basic things like injectors, vacuum leaks, spark.  

[1980ea71Brat]

2 hours ago, nvu said:

Could your ecu still run in limp mode with the maf unplugged?  Just as a sanity check, drive the car and warm it up until it does the stumbling in idle.  Turn off the engine, unplug the maf, start it up again.  It should be in limp mode and running rich, but should be idling smooth.  If it's still rough, that could point to basic things like injectors, vacuum leaks, spark.  

That's a smart idea. Unfortunately, every time I have tried to unplug my MAF, it dies instantly. Won't even rough idle or anything. Not sure if this is specific to just EJ22s or ??

[nvu]

You'd have to turn off the ecu first.  With car completely off, unplug the maf, then start it.  The car will be stuck in limp mode until you reset the ecu by unplugging the battery though.  Then again, maybe the ej22 is different.  This method works on later ej20/ej25 ecus. 2002-2005ish at least

[1980ea71Brat]

I'm going to look into it. I don't know of a way to turn the ecu 'off'' while still having the engine run on the EJ22 but we will see.

Today I tried fully removing the throttle body and cleaning it, new gasket, etc. Seems to run slightly better? but didn't fix the problem.

[heartless]

10 hours ago, 1980ea71Brat said:

I'm going to look into it. I don't know of a way to turn the ecu 'off'' while still having the engine run on the EJ22 but we will see.

I think you need to re-read the instructions that nvu gave....

Quote

 .....drive the car and warm it up until it does the stumbling in idle.  Turn off the engine, unplug the maf, start it up again. 

 

[1980ea71Brat]

So here's what I did today: Got the car up to ~185 and shut it off. Unplugged the ECU, and started it up.

The car sat at closed loop idle of 1025 RPMS and didnt budge. No loping nothing. Stayed this way for 10 minutes, without any changes, was hardly drivable (limp mode as you described), It had almost no power.

Turn the car off, plug back in the MAF, restart the car - back to the way it was. 30s of closed loop idle, it auto switched to open loop idle where it surges from 500-1000 RPMS. After a minute of that it switches back to closed loop for 30s. and the cycle continues.

 

It was BUTTERY SMOOTH in limp mode. First time I have ever tried to do this - what can this tell me about the engine? No vacuum leaks probably? 

[1980ea71Brat]

It idled so well* in limp mode, it must mean I have a bad sensor causing havok? MAF? O2 sensor? IAC? - all those were replaced already, but it almost seems like the MAF is still at fault.

Does it running smooth in limp mode eliminate the possibility of having a wiring issue?

Edited August 8, 2022 by 1980ea71Brat

[nvu]

No, smooth limp mode mainly points to electrical.

Well you ruled out most mechanical issues.  So fuel pump, injectors, spark, are in good shape.  It doesn't rule out vacuum leaks though, in limp mode the car runs super rich and masks any leaks.  It's mainly sensor and electrical issues now. 

It's not exactly the maf though.  Unplugging the maf forces the ecu into limp mode, it ignores most sensors at this point.  If you really want to make sure, you'd need to throw your maf into another car or borrow one from a working car.

 

 

[el_freddo]

Was the donor vehicle an auto? 

I’m wondering if it is an auto ECU where the pin to tell the ECU it’s in front of a manual was not correctly fitted/removed/grounded - I can’t remember off the top of my head without looking up the wiring diagrams. 

If this engine ran perfect in the donor vehicle there must be an issue somewhere in the conversion and cutdown of the wiring loom. I’d be looking at any joints made in your loom to ensure they’re good and to rule them out. 

Cheers 

Bennie

[1980ea71Brat]

The donor car was a manual. And I should back track a bit - the original car had a 02 sensor code, and I found that the electrical connector was broken. I replaced the 02 sensor for good measure and rewired it from there. After that it ran perfect with no issues. 

 

I have been doing some internet sleuthing and found that its possible the ECU is going back and forth between closed/open loop because it's not able to correct the rich condition in closed loop. This also explains why if I apply even more than a light throttle to it will drop into open loop - drive mode.

 

I also observed that the I/M readiness function of my OBD2 reader continuously says 'incomplete' for the 02 sensor and the Cat. I have gotten the OBD2 to go green, and checkmark the 02 sensor, but I have no idea what triggers it. I drove for over an hour on sunday and the 02 sensor never got a green check mark. Luckily I don't have emissions in my state, but it almost seems to be pointing towards a problem with that sensor.

 

Might try ordering a $13 sensor off of ebay just to try? 

[nvu]

You did mention you had a 96 ecu?  Chasing the readiness monitor is near impossible with those as the programming was flawed.  If you turn off the engine, it loses the readiness state.  Even smog in CA has an exemption for 96 year subarus.

[1980ea71Brat]

Thanks for the heads up NVU - yeah it is a 1996. A lot of forums are suggesting that the 02 sensor really probably isn't the fault. It's just reading that the fuel is too rich, the computer is trying to correct it, and it's not able to.

 

Right now I am at a bit of an impasse - it runs and drives, and it's at a really good milestone right now. I may not find the answer or to this problem for a while, especially if I start to daily the car this winter. Thanks everyone for following up and helping. If I don't get this solved before winter, I will be sure to follow up in the future if I figure anything out.

 

Some other ideas I had:

Take the intake manifold off and reseal

Take the injectors off and clean them/ JY replacements (I really need to measure the resistance first)

Check all the manifold wiring with the intake off. 

Unless it is an 02 sensor failing, I am kind of out of options! 

 

Did yalls know ebay also has $25 brand new MAF sensors? No way those work right?

 

 

[heartless]

12 hours ago, 1980ea71Brat said:

Did yalls know ebay also has $25 brand new MAF sensors?

i would not trust those as far as a could throw one... and trust me, these days, it is not far.

one last thought... i dont recall seeing mention of trying to clean the MAF sensor?
there is a spray product specifically for cleaning MAF sensors - have you tried this?

[1980ea71Brat]

12 hours ago, heartless said:

i would not trust those as far as a could throw one... and trust me, these days, it is not far.

one last thought... i dont recall seeing mention of trying to clean the MAF sensor?
there is a spray product specifically for cleaning MAF sensors - have you tried this?

Yeah I cleaned it at first, then ultimately replaced it with a refurbished one from Napa. It fixed the intermittent misfire issue but not the idle.

 

I'm going to put together a TLDR of everything I have replaced and tested up to this point - I left it at work, so hopefully I'll remember to bring it home tomorrow.

 

[lmdew]

I believe the IAC are a little different from Auto to Manual Transmission.  Is it a match to you swap?

[1980ea71Brat]

1 hour ago, lmdew said:

I believe the IAC are a little different from Auto to Manual Transmission.  Is it a match to you swap?

Well it seemed to be a direct replacement to my old one. But I'm not sure. It was a pretty basic device, with two coolant ports and the main air duct.

[nvu]

On 7/30/2022 at 5:13 PM, 1980ea71Brat said:

The MAF was 100% causing the problem with the mid acceleration stumble. I used my code reader to chart the MAF performance, it turns out it was fluctuating too fast for the code reader to display, but was logging huge spikes every time the throttle cut out. It was definitely acting badly, and replacing it with a refurbished unit in fact fixed that problem. It will now accelerate from 1000 RPMs to 5000 RPMs no problem, no hesitation.

At this point I'm guessing and throwing out ideas since changing the maf made some improvement.  Check the connector and wiring itself to the maf.  

It might not apply here, on my turbo ej25 the fuel trims would drift too rich or lean over the years, I've always attributed that to dirty or failing maf and replaced a couple, it fixed the problem.  Randomly on one of the refresh builds, I broke the tab on the connector and it wouldn't lock.  Ordered a connector and found that one of the wires felt very stiff, it was the heater power for the maf.  I chased it maybe 1ft into the loom until it got flexible like the rest of the wires.  Then ordered a premade made pigtail and spliced it in. Haven't needed to replace any maf's since.  Then again the loom ran near the turbo area of the bay.

[1980ea71Brat]

3 hours ago, nvu said:

At this point I'm guessing and throwing out ideas since changing the maf made some improvement.  Check the connector and wiring itself to the maf.  

It might not apply here, on my turbo ej25 the fuel trims would drift too rich or lean over the years, I've always attributed that to dirty or failing maf and replaced a couple, it fixed the problem.  Randomly on one of the refresh builds, I broke the tab on the connector and it wouldn't lock.  Ordered a connector and found that one of the wires felt very stiff, it was the heater power for the maf.  I chased it maybe 1ft into the loom until it got flexible like the rest of the wires.  Then ordered a premade made pigtail and spliced it in. Haven't needed to replace any maf's since.  Then again the loom ran near the turbo area of the bay.

Yeah I think this is a good idea I can look into. I have looked at maf wiring, but I really didn't see anything out of the ordinary. Unfortunately if its truly a wiring problem, I will have to unravel my whole custom loom to find it unless its close to the connector.

I don't know if I'm ready to take it apart wire by wire back to the ECU yet lol. 

[1980ea71Brat]

I forgot my list at work of course, but here is a TL:DR of everything so far:

Replaced:

Fuel Tank, All Rubber Fuel Lines, Fuel Filter, IAC (New), MAF (Refurbished), Spark Plugs, Spark Plug Wires, Front 02 sensor, (I am probably missing something here I will add later)

Tested to Manufacturer Specs / Attempted Fixes:

Fuel Pressure Regulator (passed), IAC (Yes I tested the new one), Knock Sensor (560 kOhms), Removed and Cleaned Throttle Body (New Gasket), PCV valve, Cleaned Electrical Connectors, ECT Coolant temperature sensor, TPS (As seen on OBD2 scanner - 0% throttle at resting, 100% throttle at full press), Tested for vacuum leaks by spraying brake clean, Steady Peak Fuel Pressure (Confirmed no leaks), Fuel Pump Pressure is within spec, Multiple ECU reset attempts, was able to blow air into the fuel return line (indicating no blockage), (I am probably missing something here I will add later)

What's left to test?

Fuel Injectors (Ohms)

MAP Sensor?

Total Vacuum 

Neutral signal  

Purge Valve?

Bad Electrical connections under the Intake Manifold.

?????

What's the current problem? Note: I don't have any check engine lights, or pending codes.

Idle oscillates from 500-1000 RPMS in open loop, AFTER the car has idled for ~1 minute. 

Driving the car (while warm) with light throttle stays in Closed Loop. No throttle or slightly more than light throttle switches to Open Loop - Drive

Front 02 sensor reads .9 - .95 volts for most operations besides light throttle or deceleration, indicating rich mixture

Short Term fuel trims are maxed at -25%, indicating rich micture

Once fully warmed up and stopped to let idle, the car will switch between idling in Closed Loop at exactly 1025 RPMs, and Open Loop with a lope from 500-1000 RPMs. It will switch between the two every 30s-1 minute.

 

Edited August 12, 2022 by 1980ea71Brat

[el_freddo]

Can the front O2 sensor be replaced with a resistor to replicate one being there for testing purposes?  

Cheers 

Bennie

[1980ea71Brat]

16 hours ago, el_freddo said:

Can the front O2 sensor be replaced with a resistor to replicate one being there for testing purposes?  

Cheers 

Bennie

You know I'm not really sure I know how to do that, or what I could do with it. If I trick the computer into thinking the voltage at the 02 is .5 volts, I could see if the 2nd 02 is still showing rich? Rule out the front 02? But I've never heard of that before.

At this point I'm a starting to run out of time. I might consider paying the local subaru shop an hour of labor to plug in their SSM device and see if it tells them more than my OBD2 can.

I measured resistance at my injectors today - they all measured 12.4-12.6 ohms. They could potentially still be a problem? But not very likely... 

[1 Lucky Texan]

yeah, likely would need a resistor 'network' or a small potentiometer connected to created a volatge 'tap'.....?

 

but the ECU may be looking to actually detect a crossing as the sensor switches?

 

live data from an elm327 device or similar might help with this.

[1980ea71Brat]

Been a while. Just checking in to let people know I have decided to let some subaru experts take a look. Hoping 3 hours is plenty of diagnostic time. If they find a solution to the problem I will post it here. This car is 95% complete, just have to button up the interior after I get this idle problem fixed. Thanks all, wish me luck.