Romraider misfire troubleshooting shenanigans

[awfulwaffle]

Hey guys,

 

I've been troubleshooting a misfire in my 05 wagon for a bit now, and some weirdness has me stumped. Hoping those of you with more experience with troubleshooting these cars might provide some insight. Situation is as follows:

 

Car has had what felt like a slight misfire at idle since I bought it a few weeks ago. About a week and a half ago, it finally popped a code (P0303, twice back to back). I started going through the regular checklist, and invested in a VAG cable so I can log some proper data.

 

Sure enough, I get elevated misfire counts in cylinder 3 at idle. Highest I've seen it is 41 (once) though usually it only got up to 10-20 during the polling period.

 

I've considered a few trends that made sense to me, and have only noted the following:

 

Occurs during warmup and hot idle, and rolls on to max count gradually while sitting at idle. Coldest coolant temp I've observed misfires at is 160, and have consistently seen the misfire counts drop down and pick back up timed with the thermostat opening and closing once the engine was fully warmed up.

 

Occurs during low end of injector pulse width range. I've seen sparse misfires as low as 1 ms and as high as 8 ms, but 90% of the misfires register during 2 ms injector pulses at ~800 rpm idle.

 

 

Before I looked at the injector pulse width dependence, I replaced plugs and swapped #1 and #3 coils to no avail. Plugs looked the same wear-wise. Once I saw the pulse width plots, though, I was sure I had tracked it to a faulty injector. Now that I think about it, this likely just shows that the misfire is happening at idle.

 

I swapped the #1 and #3 injectors this morning, and that's when I started seeing data that are throwing me off. Upon first start, I observed more cylinder 1 misfires than I ever had before (8, that's encouraging). Then, the cylinder 3 misfire counts spiked to 27 on warmup (aw, rats). Then, though, the misfires all but died down. Cylinder 1 would register a few on occasion, and cylinder 3 exhibited similar temperature dependence as before but never saw more than 5 misfire counts. I went for a spirited drive, and let the car idle while I logged data again. Never logged more than 1 count on any cylinder.

 

Outside temperature was about the same both days I logged data (50-60F)

 

Trying to figure out how swapping injectors could have possibly had this effect. Perhaps some underlying condition like bad valve clearance or bad valve in cylinder 3 that led to some degradation of the injector and compounded to produce the misfire I originally observed, and is now working on damaging the injector I installed in its place? I don't want to leave an underlying cause unaddressed just by calling it good.

 

At this point, I was going to do MAF cleaning, valve lash and compression test (in that order) to see if I can learn anything else, but I would love to hear some insights from those of you more experienced than me. I'd also appreciate any tips for other patterns/dependencies I might be missing.

 

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

TLDR

 

05 wagon with misfire at idle. Occurs both when warm and cold, and I've observed the following patterns so far:

 

Rolls on gradually after sitting at idle. Never below 160F coolant temp, and I've seen misfire count drop down to 0 and pick back up at the same time as the coolant temp oscillates due to thermostat open/close at warm idle.

 

90% of misfires noted at 2ms injector pulse width.

 

Changed plugs, looked the same wear-wise. Swapped #1 and #3 coils to no avail.

 

Swapped #1 and #3 injectors, and saw increased misfire count in cylinder #1 during warmup. Suddenly, cylinder #3 counts spiked, and dropped down. They still exhibited the coolant temperature dependence noted above, but were coming in at roughly half the counts as before. Went for a drive and re-logged hot idle, and observed no misfires on any cylinder.

 

Looking for any insight that might help explain what's going on, because I don't want to leave an underlying cause unaddressed. HALP

 

 

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

 

Many thanks in advance,

 

Alex

[awfulwaffle]

Update: Just relogged from cold start before a trip to the store, and there's no misfires as the engine warms up until a sudden spike in #3 misfires accompanied by a sharp drop in RPMs to 800 (from about 1000). Stumbles around a bit, and misses go away once at full temp. Zero misfires logged idling after the trip. At least no more hot misfires, but I be stumped.

[LPStandard]

I would start with a leakdown test to verify that nothing mechanically abnormal is going on in that cylinder.

 

How are your learning table values looking? If you had a bum injector, I would expect high fuel trims as another symptom.

[awfulwaffle]

Haven't thought to log them yet, unfortunately. I'll definitely include them in my logs when I get back from the work trip I'm on and get a chance to get some more data.

 

Leakdown doesn't sound like a bad idea, but I recently moved to another state and had to leave my air compressor behind because I live in an apartment. Was hoping to explore the free options before having to pay someone to do leakdown, or at least until I have a chance to drive back home and make use of the compressor. However, wouldn't a misfire due to poor valve sealing exhibit a more definite temperature dependence than what I've been seeing? I just can't reconcile the fact that switching two injectors made the hot misfire go away, and now there's a weird spike in misfires during a limited portion of the warm up idle.

[LPStandard]

Logging can be useful to see what the short trims are doing (AF1 Correction in Subaru language), but for the long terms/knock learning you can pull the learning table values all at once. When our LGT had weak injectors (we had two that were 15% lean) the long terms (AF1 Learning) were really high and there was lots of knock adaptation stored due to the pair of cylinders consistently lean.

[awfulwaffle]

I've added those to my logs. Just wondering what I'm looking for now. That is, what's the threshold that might indicate an issue?

 

I assume that I'm interested in A, since this is the lowest MAF range corresponding to when the car is at idle.

 

Also - am I understanding correctly that the AF1 learning value is what drives the long term stored value (A-D), and that the AF1 Correction value is the actual correction to AF ratio being calculated based on other realtime parameters?

 

Thanks!

[bugblatterbeast]

are you on the original o-rings between the intake manifold and TGV housing? there were a couple revisions to the part. the original design isn't tall enough and doesn't always seal . till I replaced the ones on my car, I had the same symptoms.

 

if you look carefully around each runner, you may find oily residue if there is a leak. another way to test is to spray some carb cleaner around each runner when the engine is idling and misfiring. if the behaviour changes, you've found the problem.

[awfulwaffle]

If I am to take the previous owner's maintenance record as complete, I'd have to say it is original. Were you misfiring on one cylinder?

 

Actually omw to home Depot to get parts for a boost leak check adapter, should tell me if there's a bad seal somewhere.

[awfulwaffle]

Well, just finally got a chance to actually try a boost leak check and I couldn't even get the pressure up to 10 psi. Hear it hissing out somewhere under the intake manifold. Not sure why a boost/vacuum leak would cause a miss in a single cylinder, but it's definitely something to address.

[awfulwaffle]

Wellp, smoked the intake and found the worst leak at the BOV return hose and put a hose clamp on it, but still couldn't get it to hold boost. Then, I decided to log TGV position at the time the misfires spike and it sure does look like bugblatterbeast was correct. Increased airflow when the TGV opsn exacerbates the leak, perhaps?

 

http://i92.photobucket.com/albums/l23/thesteakosaurus/misfire_tgv_zpsbwrmfg5f.png

 

 

Got the gaskets, we'll see what happens this weekend!

[bugblatterbeast]

good luck with the gasket install.

 

one thing to keep in mind, the mating surfaces may be in bad shape. mine were. I ended up having to do a lot of sanding an polishing of the surfaces before installing the new gaskets. took a damn long time. in my case, both the TGV to head and intake to TGV seals were bad. the former due to the surface going bad and the latter due to the crappy original o-rings.

[awfulwaffle]

Thanks.

 

The plan at this point is to pull of the entire manifold/TGV and clean everything up out of the car. Hopefully the fuel lines don't give me too much trouble.

 

 

I wonder what the odds of the valves themselves being bad are.

[awfulwaffle]

Here's a question for you guys:

 

Just read through a multitude of TGV threads, and in one the folks at Grimmspeed claim that they only operate for a few seconds at start. Can anyone confirm this? Mine sure look like they stay closed a while.

[iNfEk]

subscribing... any update on this?

 

what are you logging on romraider?

 

specifically for the misfire count pieces you mentioned in the beginning.

[solidxsnake]

Here's a question for you guys:

 

Just read through a multitude of TGV threads, and in one the folks at Grimmspeed claim that they only operate for a few seconds at start. Can anyone confirm this? Mine sure look like they stay closed a while.

 

For the '05-'06, I believe that is correct. For the '07+ LGTs, the TGVs will switch between open/closed depending on load/throttle/some-other-stuff.

 

You can check this for certain, just log TGV volts like you already are. When it's low (around 0.8V on my car), then they're closed, and when it's high (around 3.6V on my car), they're open.