vvt solenoid

 

I thought of that but the data shows them to be working correctly on both sides. And both solenoids going bad at the same time? Good thought though.

 

Interesting . . .

No answers, but a few comments/questions that might help.

Problem is cylinder 3 & 4 misfires when letting off the throttle. . . .

Complete release of the accelerator pedal, or just backing off a bit?

During the throttle release, does the CEL flash (signalling misfires) as well as P0303 and P0304 codes being set?

Torque converter maybe shuddering causing the knock sensor to misinterpret a miss?

Misfires are determined from the timing of the pulses from the crankshaft position sensor. Knock sensor isn't involved.

I have combed through hours of data

Are you recording data using the Subaru Select Monitor logger, or equivalent, accessing the ssm data stream, or a tool that accesses the generic OBD data stream? If you have .csv logger files, can they be shared?

Fuel trims do go a bit lean when letting off throttle but not leaner than another car of same model.

Not sure if the 2011 2.5 is programmed this way, but from what I've seen in other drive-by-wire Subarus, when the accelerator is released while cruising, the fuel system goes into "fuel cut", AFR goes to max (~20:1) or equivalence ratio (Lambda) goes to 1.4. (Also, typically, the fuel system goes into Open Loop, and A/F Correction -- STFT -- is at zero.) When that happens the OBD would not be detecting a misfire as that would be inconsistent (can't have a misfire when no fuel is being injected). I'm wondering if there's something here.

 

Interesting . . .

No answers, but a few comments/questions that might help.



Complete release of the accelerator pedal, or just backing off a bit?

During the throttle release, does the CEL flash (signalling misfires) as well as P0303 and P0304 codes being set?



Misfires are determined from the timing of the pulses from the crankshaft position sensor. Knock sensor isn't involved.



Are you recording data using the Subaru Select Monitor logger, or equivalent,accessing the ssm data stream , or a tool that accesses the generic OBD data stream? If you have .csv logger files, can they be shared?


Not sure if the 2011 2.5 is programmed this way, but from what I've seen in other drive-by-wire Subarus, when the accelerator is released while cruising, the fuel system goes into "fuel cut", AFR goes to max (~20:1) or equivalence ratio (Lambda) goes to 1.4. (Also, typically, the fuel system goes into Open Loop, and A/F Correction -- STFT -- is at zero.) When that happens the OBD would not be detecting a misfire as that would be inconsistent (can't have a misfire when no fuel is being injected). I'm wondering if there's something here.

I agree very much with the shouldn’t misfire under decel because of the no fuel so no misfire but I can’t for the life of me see it. Yes I am using the Subaru SSM. And no the check engine light doesn’t flash it just comes on. The misfire happens when you let off both completely and mostly off. I know there is something in the data but I’m just not seeing it.

 

So the misfires occur both when you let completely off the throttle and when you partially let out of the throttle. If you slowly back off the throttle and then get back on it then lift again the misfire counts climb and climb but only when the throttle is let off. It will not misfire under load. Full throttle acceleration power braking climbing a steep hill reving in neutral no misfire. Like if you are pulling a slight hill when you get to the top you back out of the throttle to reduce speed that’s when it misfires. The car also doesn’t feel like it’s misfiring. The engine doesn’t run poorly it’s just the roughness counts. Every time you let off the throttle it will build more and more counts until the light sets but it doesn’t flash. I will see if I can figure out how to upload the data when I get back to my computer on Monday. The timing components have been replaced with Subaru oem parts. I’m struggling with interpreting the data as I am learning this as I go. My knowledge base on this type of thing is only what I’ve been able to teach myself so I greatly appreciate your help. Computer diagnostic is not my strong suit for sure. And I am logging every piece of data the Subaru SSM will let me log.

 

Ramping up of the roughness counts for the two cylinders during reduced throttle, and then, at some point, the CEL turns on with the two codes.

If files can be uploaded, that would at least get some more eyes on them, and perhaps new ideas.

 

I will try and figure that out on Monday. Again I appreciate it. At this point I’m grasping at straws like with the transmission. I have fixed a ton of misfires a bunch of different ways and I’ve seen all kinds of things cause them but man o man I’m stumped on this one.

 

I would have to see data from the car at the time the misfires occur in order to narrow it down to a specific thing or at least to a specific system.

RPM
TPS
AF Sensor Lambda
AF Correction
ECT
IGN Timing
Knock Correction
Roughness on all cylinders
Rear O2
AVLS both banks
Batt

Here's my theories in the mean time.

This generation still maintains a light throttle fueling. So there is a difference between OFF the accelerator and barely on it. Any slight movement of the APP tells the ECM you are on the throttle. And that brings to mind the APP feedback and range performance.

If the timing belt is stretched, or the tensioner is weak, when off throttle and during engine braking the timing will be off. Think about the connection to the converter and a locked TCC. Like letting off the throttle with a manual, the crankshaft gets a brake affect and is slowed quickly, (this is one of the reasons that belt guard is over the crank gear on manuals). With a loose belt the cams will be late to follow suit and the piston movement against the valving changes the engine's breathing and combustion chamber pressure which can cause imbalance on the crankshaft. But this doesn't flow with the fuel cut off throttle because the engine is an air pump and the firing spark plugs are just firing without really burning anything. So something is causing an imbalance.

If timing is the case, then when the crankshaft "kicks" off throttle from the engine braking affect, that would give the ECM an erratic signal and it would think it's misfiring off throttle. On the same token, if the crank and/or rod bearings have wear, that excess movement would create the same type result, an imbalanced engine.

Leaking fuel injector comes to mind, but since you've swapped them, that may be ruled out. Unless the fuel from the leaking injector is traveling to the other cylinders via the manifold runners and off throttle the cylinders continue to get fuel but just in differing amounts this would create a scenario for a misfire with one or two firing while the others are leaned out. This is where Lambda measures come in to play to see if the engine is really in a fuel cut mode, or it's still getting fuel, or it's not cutting fuel for some reason related to a sensor feedback issue. OFF the throttle, as @plain OM posted, the Lambda goes to 1.4 indicating extreme lean out and this is normal. If you are watching AF Stoich, then it would be 20-21 AFR. If it's not, the cylinders are getting fuel either due to injector(s) issues, or sensor(s) feedback.

Keep in mind, a single injector can fail and the others be okay. On the flipside, the injectors have been operated equally throughout the life of the engine. They get the same fuel, the same maintenance, same operation. You could have a full set that have issues and your moving them around isn't necessarily proof to discount one or all of the injectors.

Now, the most important thing on the car has not been checked. At least I didn't see it in you posts.

Battery and Grounding

Do you have a Midtronics battery conductance tester and do you know how to use it to check the cables and various ground points? The main ground is good with no more than a 50 amp loss between the negative battery post and the lug at the end of the ground. You do not want any more than a 75 amp loss between the negative battery post and the alternator case.

Plenty of times there have been posts on this forum where the grounding was causing issues. The grounds are the way the computers communicate and operate their outputs.

Lastly, what codes are stored on the other modules on the car? All of them. Did you run a full scan of all the computers?

 

This issue sounds very similar to what I have going on with my 2012 and I have attempted many of the same fixes to no avail. Did you ever find a solution?

 

Posts like this are tough without stored codes. @texas0377 what is your code(s)?

 

I am getting a cylinder 3 and cylinder 4 misfire code. Occasionally cylinder 2 as well but that one is less consistent.

Like the discussion starter, these codes only pop up when decelerating - most notably when going down a long hill and utilizing the engine brakes. I don't get any codes when idling or cruising along the highway.

Here's what's been done so far:
New plugs and coils
New upstream and downstream O2 sensors
New fuel injectors
New PCV valve
Visually checked all vacuum hoses for signs of leakage
Swapped camshaft and crankshaft position sensors out with known good ones
New timing belt and timing belt tensioner
Valve job
New MAF and MAP sensor
Removed and cleaned VVT solenoids, swapped the left and right side solenoids to see if any change would be noted.
Tested exhaust back pressure - tested at 1psi
Tested fuel pressure at idle and 2500 rpm - pressure held steady at 50psi

The car is a 2012 Outback with 70k. Everything in the engine compartment is very clean and it runs perfectly; just keeps throwing codes.

Any advice or thoughts would be greatly appreciated!

 

@texas0377
Is the MIL flashing when you are off throttle?
Battery?
Grounds?
Knock sensor?
Has anyone looked at the data stream when it happens?

 

The MIL does not flash but will illuminate and stay on for several days or until I clear the codes.

The battery is just over a year old and the terminals are perfectly clean.

I double checked all the grounds and spent a while yesterday cleaning all the electrical terminals with CRC electronic cleaner.

I'll try a new knock sensor this weekend and report back.

2 separate mechanics have looked at it but I'm unsure how deep they dove into the data. I've been recording the freeze frame data from multiple events in a spreadsheet to try and find a pattern but my scan tool is basic and my understanding of how to interpret the data is basic as well. If you think that would be helpful to see I'd be happy to share that spreadsheet.

Thanks again for your help!

 

Sure. Put it in a file on a remote storage site and link it here.

Have you checked fuel flow and pressure?

 

This should get you to the file that contains freeze frame data for when codes were generated:

Dropbox

www.dropbox.com

I checked fuel pressure at idle and while holding the RPM's at 2500. It held steady at 50psi. I did not test the fuel pressure under actual driving conditions but maybe that would be insightful?

 

Sorry, was busy over the weekend with the new house and other things.

The freeze frame shows that the car is running about the same every time the code sets but doesn't give a hint as to where to go from there. It's also just a millisecond of data and does not show what led up to the fault. This is where the datastream logging comes in so a pattern can be determined. The AF sensor is showing rich at the times the code set so this does indicate raw fuel in the exhaust.

Cylinders 3 & 4 being opposing cylinders, it may be an issue with one of the cylinders and that issue is causing the other to react to it. An engine imbalance.

Fuel, Air, Fire and electronic management of 2 of the 3 combine to get the combustion process correct. Fuel metering by the AF sensor, fuel supplied by the pump, fire provided by the ignition coils via ECM management and signals that affect it like the cam and crank sensors. And on the last 2, has the timing been double checked?

If you have a windows laptop, Rom Raider is free and a Vag-Com KKL is about $10. You will be able to log and record data, post it up on DropBox and link it here.

 

Thanks for your help @cardoc

The timing has been checked and double checked multiple times and I feel quite confident that's not the issue. A while back I tried swapping out the timing belt and tensioner from my 04 Impreza and the same codes came back.

I've looked into Rom Raider before but never used it. I'll get on that and start playing around with it.

I'm traveling this week but will try a few more things and ping you again with anything new next week.

 

On another thought, where did the coils come from?

 

^^ Good point. Cheap knockoff Chinese-made coils can cause all sorts of issues on all sorts of vehicles. I would stick with either official Denso or official NGK coils.

 

Diamond makes the coils for Subaru.

 

I ddunno if they do anymore as my NGK coils came with the Subaru logo on them and the diamond coil that was in#5 did not have a Subaru logo on it.

 

From Application Data

Ignition coil

Type .................... FK0368
Ignition system .................... Independent ignition coil
Manufacturer .................... Diamond Electric

 

Just joined this community and after reading this thread it appears that it describes the issue I'm having on my wife's 2012 outback 2.5i. MIL comes on in about 30-50 mi. Changed bad alternator, fuel pump, cleaned EGR valve, crank and camshaft position sensors replaced, oil control valves replaced, new spark plugs, coils (denso), O2 sensors (up & downstream), MAF over the past yr and still get the misfires codes (P300-304). They all show as misfiring with the initial cylinder that caused the fault varying between all the cylinders. So not sure if the initial cylinder that fails causes all others to also misfire? Also, sometimes get a P0700 (TCM). Which leads me to another question. If replacing the TCM with a used one does the used TCM need to be programmed? What about the ECM? does it need reprogramming if replaced w/used unit? Going to attempt to log the failure on laptop when it occurs. The freeze data was similar to what someone discussed in an earlier post, so not much help there. Any insight would be great.

 

I put this up every once in a while on this forum, so it's in a lot of places. I'm going to add it on this thread because all this throwing parts at a car is just a waste of time and expense. Always start with the most basic thing on the car that is also a common denominator to everything on the car. The battery and grounds.

This is an LL Bean with the EZ30D H6. Same concept applies to every single car built in the world.

 

I put this up every once in a while on this forum, so it's in a lot of places. I'm going to add it on this thread because all this throwing parts at a car is just a waste of time and expense. Always start with the most basic thing on the car that is also a common denominator to everything on the car. The battery and grounds.

This is an LL Bean with the EZ30D H6. Same concept applies to every single car built in the world.

Cardoc, thanks for the llink. However, the problem still occurs. Is there a way to affect the fuel/air mixture settings in the ECM? I wondering if I have to tweak the settings. At this point don't know what else to try or investigate.

 

Ok, thanks Cardoc. I have a similar OBD cable for my laptop that I can use w/OBDWiz diag app. What params should I consider logging to narrow this issue down?

 

AF Correction (STFT)
AF Learning (LTFT)
Battery Voltage
RPM
TPS
Knock Correction
MAF
ECT (Coolant Temp)

Start there and record about 2 minutes cruising at different throttle applications.

 

mass airflow sensor died on me...misfired like mad... unplugged it... car ran OK on manifold absolute pressure data.... got new Denso sensor on order

 

Just a thought...have a look at the PCV valve especially the hose on the PCV valve for cracks, and a tight fit. When throttle is closed car is pulling full vacuum and a vacuum leak will cause cylinders to go lean and misfire.

 

Nice suggestion. I'll take a look.

 

Captured some data. Now need to examine it.

 

Cardoc, have several *.graph files where I captured all the params you suggested 'cept for the knock values as there was no setting for that in my app. Can you possibly review if attached? Thx.

 

I use shell and my engines run better on it.

 

I use 2 gas brands. One is Valero, the other is a Texas grocery chain. HEB. Surprisingly, the VDC loves HEB gasoline. I can tell from the way it reacts when I put it in. The butt test hits high marks. Haven't tried it in the BAJA yet.

 

Cardoc, hopefully you can access the following: It's the hispeed graph capture. If you can access then I'll post link to the low and mid speed graphs. Thanks. Meet Google Drive – One place for all your files

 

Cardoc, hopefully you can access the following: It's the hispeed graph capture. If you can access then I'll post link to the low and mid speed graphs. Thanks. Meet Google Drive – One place for all your files

Go back to Drive and give permission to view to "anyone with a link."