I have a misfire in cylinder #2 and I can't make it go away.
** UPDATE ** Long story short, my #2 piston has a small crack.
As my story has been progressing I've been updating this first post with relevant details. My desire is to give back to the LegacyGT.com community, so as long as someone finds this info helpful I'll have succeeded. Oh yeah, and I REALLY want to get my car running properly again.
At about 97k miles the pesky P0302 code showed up (misfire in cylinder 2), accompanied by a sometimes rough and low idle (around 400-550 rpm) once the car reaches operating temperature. Sometimes the CEL comes on, other times it doesn't. When coming to a complete stop the RPM occasionally dips really low and the engine stalls. The engine doesn't always idle super-low, but the misfire in cylinder #2 is always present and noticeable.
The first time I noticed there was a problem: It was a very cold morning and I was letting my car warm up in my driveway before my drive to work. When I came back out of the house after a few minutes my car wasn't running. My guess is that after it reached operating temp it started idling badly and then stalled.
Worth mentioning: Every once in a while I'll see a misfire or two in one of the other cylinders (1, 3, or 4). I'm choosing to ignore those because of how few and far between they are. I'm betting that the inconsistent misfires will go away after my root issue is resolved.
Here's what I've done so far in attempts to scare away the misfire:
- 2005 Legacy GT wagon, 5mt. ≈102k miles on the clock (97k when I first noticed the problem).
- I bought this car back in June from someone here on the forum. The car's excellent appearance, info from the previous owner, the pre-delivery inspection I had done @ Dunning Subaru (Ann Arbor, MI), Subaru's records, and the Carfax report lead me to believe that the car has been well maintained since birth.
- Original turbo, timing belt, ignition coils, injectors, OEM flux capacitor, 2-yr old NGK spark plugs, no engine mods, or tuning, Bosch battery (fall '14), WRX clutch kit + LWFW (Jan '15).
- I use Peak Global coolant and Castrol synthetic 5w-30 oil. The engine consumes basically no oil between OCIs.
- I have a Tactrix OpenPort 2.0 that I use with BtSsm on my Android and RomRaider on my laptop.
I have a bluetooth OBDII adapter that I use with Torque Pro on my android. Torque works great, but it doesn't go deep enough into the ECU for what I need.
Between each change on 1/3/15 I put the car back together and let it warm up and run for 10-15 minutes each time. I've been very methodical about doing my best to only change one variable at a time.
If the #2 injector isn't the problem what's next?
- 1/1/15: Replaced the intake manifold o-ring gaskets per this thread.
- 1/3/15: Tested the compression of each cylinder and got 59/60/61/60. I think my numbers were so low because I tested it cold (~65°F). If needed I'll test everything again later this week with the engine warmed up.
- 1/3/15: Cleaned all spark plugs and then re-gapped them to .035" ± .001".
- 1/3/15: Swapped spark plugs #1 and #2.
- 1/3/15: Swapped ignition coils #1 and #2.
- ** None of the above things caused the misfire to stop or move to a different cylinder.
- Tomorrow evening I'm going to swap the injectors for cylinders #1 and #2 and see what happens (per this thread.)
Do I perform another compression test but with a warm motor? Should I borescope the #2 cylinder and pray it's shiny and clean? Maybe I should toss a few washers in the turbo inlet and see if that fixes it. I'm getting really frustrated and I'm starting to get that achy feeling deep down in my stomach like a trip to the stealership is approaching. I can't afford to do that and I really don't want to pay someone else to fix a problem that I feel like I'm very close to solving.
On a more positive note this has been a really great opportunity to learn my way around the engine bay (its my first Subaru).
All help will be greatly appreciated!
↓ ********** ↓ UPDATES ↓ ********** ↓
***** Jan 5 (2015)
I took out the #1 and #2 injectors and lightly cleaned them with carb cleaner, and then put them back in the engine but swapped cylinders (#1 into #2, vice versa). The #2 injector had a little bit of gunk on it, but it didn't look too bad. The surface that actually sprays fuel looked no different than injector #1. The o-rings were pretty clean and didn't show any signs of cracking. I also lightly cleaned the surfaces that the o-rings seal on. My goal here was mostly just to swap injectors, not perform a major cleaning.
right off the bat I saw the P0302 code (no CEL), but I didn't clear it and it went away as quickly as it took Torque to rescan for codes. I let the engine warm up while I shoveled my driveway and threw down some salt... still no P0302 or misfire sound effects. I drove a couple miles near my house (a few stop signs and a couple gentle 0-40 pulls)... still no P0302 or signs of a misfiring cylinder. The rough idle that accompanied the misfire is also gone. Nothing else that I've tried so far has generated results this encouraging.
I'm going to drive it to work tomorrow and see what happens...
***** Jan 6
On my drive to work I had the usual misfires and low idle.
On my drive home I had no noticeable misfires and the idle was as it should be.
No engine codes.
***** Jan 7
Drive to work = no noticeable misfires and the idle was as it should be.
Drive home = no noticeable misfires and the idle was pleasant.
No engine codes.
***** Jan 8-14
Maybe one or two misfires (per minute) when the car is idling, but nothing you'd notice if it wasn't being looked for.
Idling is as smooth as butter and I didn't see any codes that Torque can pick up.
***** Jan 15
I've been driving the car every day for the past couple of weeks (it's not my usual DD) and the misfire still hasn't gone away completely.
Today I hit the jackpot and Torque finally picked up an engine code... P0301. The misfire followed the injector to cylinder 1.
Now that I'm 99% certain the injector is the problem, I'm going to thoroughly inspect and DIY clean all four of them this weekend and see what happens next week. For the love of doing things right I'll "professionally DIY" rebuild all four injectors next weekend using the $30 the kit from MrInjector
***** Feb 23
- The misfire is still obnoxiously present and frustrating me every time I drive the car. It's only thrown a CEL a time or two in the last month, but the Torque Pro app on my android picks up the misfire pretty reliably.
- The o-ring kit I bought from MrInjector (via ebay) didn't have the right sized o-rings or pintle caps... so I don't know if that will solve my problem or not. I haven't heard back from MrInjector yet on whether he'll let me exchange the o-rings and caps for the right ones.
- About a week ago I soaked the problematic injector overnight in carb cleaner and then ran it through my Harbor Freight Racing ultrasonic cleaner (soaked in Lucas fuel treatment) for about 20 minutes. All of that didn't help at all.
- This weekend I'm going to move the malfunctioning injector to cylinder #3 or #4 just to be extra sure the injector is really the problem.
***** Feb 28
Mr. Injector to the rescue!
I sent them a message through ebay on the 24th, and a new set of o-rings and caps were in my mailbox on the 27th. The first set I ordered arrived at my house in within a couple of days, and the replacement set came just as quickly. Wow, customer service like that is rare. If I EVER need something that Mr. Injector
sells I'll absolutely buy it from them.
I'll put on the new o-rings/cap tomorrow evening to see if that fixes my problem.
***** March 8
If the P0302 code was a quarter I'd be a wealthy man.
The P0301 code I got on Jan 15 must have been a fluke. Other than that one time it's been a P0302 code every time. I guess I just got too excited thinking I had the problem narrowed down to my injector that I didn't realize the code I kept having to clear was for cylinder #2.
- Moved the injector from cyl#2 back into cyl#1 (where it first started).
- Replaced the injector in cyl#2 with the new (rebuilt) one i got on ebay.
***** March 18
My idle was especially rough on this day, and once after 30 seconds or so of ~5mph driving my CEL came on and code P0030 showed up (HO2S heater control circuit, bank 1 sensor 1). The computer also picked up code P0301 (misfire cyl#1) at the same time.
One time way back in November I saw the code P0141 (O2 sensor heater circuit, bank 1 sensor 1), but since then this was the first time I saw anything relating to the O2 sensor.
Maybe my front (bank 1) O2 sensor is going bad, but due to the high replacement cost I'm going to exhaust all of my other options first.
I'm choosing to ignore the cyl#1 misfire code as an isolated occurrence that I only briefly saw because swamp gas from a weather balloon trapped in a thermal pocket and reflected the light from Venus.
***** March 8 - April 2
Every time I drive my car the misfire is noticable. The P0302 code isn't always present, and my CEL rarely comes on.
Here's what's really strange...
on several occasions my drive to
a destination (grocery store) as been rough, and my drive BACK
was virtually flawless. One particular evening my wife and I stopped by her friend's apartment to pick up some kids clothes. Our drive heading into the apartment complex was miserable, like the car was trying to stall on me whenever the clutch was disengaged (pedal in) for long enough to let the engine slow to idle. While my wife ran inside to chat about whatever girls like to chat about, I stayed out in the car. While I sat the car idled rough, but it didn't stall. Our slow drive out of her friend's apartment complex was perfect... I'm talking about a buttery smooth idle at stop signs. The engine was fully warmed up for both 'legs' of our errand, so what could cause the engine to apparently fix itself after just a few minutes of idling? Letting the car idle in my driveway has never had a positive effect like it did here.
***** April 3
My problems with this car have become a topic of frequent discussion with the other car guys at my work, and earlier this week we collectively decided that my vacuum lines should be the next thing I look at. Coincidentally, I stumbled on this thread
earlier today, so this weekend (between Easter parties) I'm definitely going to inspect all of my vacuum lines (see here
) and replace anything that looks suspicious.
***** April 8
Last weekend I sprayed carb cleaner all over my engine bay and the idle didn't noticeably change.
I'm going to attempt a DIY smoke test this upcoming weekend.
Voltage readings from my O2 sensor indicate cyl#2 is running lean, so I think my my next rabbit hole will be looking for a wiring issue.
***** April 18-22
I smoke tested my engine at about 5psi and found a leak coming from somewhere under the intake manifold (the two black arrows in the attached image). The smoke rose above/around the intake manifold in two places but it looked to be leaking in only one place. I didn't have time to take off the intercooler or manifold to investigate further, so that'll happen maybe next weekend when I do a leak-down test. I haven't found any wiring problems yet.
I also installed a $15 harbor freight boost/vacuum gauge so I can get a better grasp of what's going on whenever my idle acts up. The 7 LED colors are pretty obnoxious at night, but for $15 it does a decent job as a cheap diagnostic tool. I wouldn't recommend it for everyday permanent use.
***** May 17
Crapozone doesn't loan (or sell) leakdown testers, nor do any of the other auto part stores near me... other than Harbor Freight. I know, I know, I shouldn't have, but all things considered it was my best option. It only cost $30 (after 20% off) and it leaves me with a little bit of money left in this month's budget for whatever comes up next.
- U.S. General (Harbor Freight) cylinder leakage tester.
- Compressor set at 75 psi, no measurements recorded while compressor was running.
- Prior to measuring each cylinder I ran the engine until the coolant reached 180°. I did this for each cylinder, not just once.
- I measured each cylinder five times, carefully resetting the tester to 0 and rotating the crank 360° (22mm socket) between each measurement.
- Cyl 1: 13.0% (median 13)
- Cyl 2: 22.4% (median 21.5)
- Cyl 3: 13.6% (median 13.5)
- Cyl 4: 20.6% (median 22)
- All leakage noise for all four cylinders came from the oil filler cap hole. I didn't see any bubbles at the radiator, or notice any noise at the intake or exhaust.
- All four plugs appeared to be the same in color; A little bit of black carbon build-up as expected, but no orange or white.
For shiggles I also swapped the coil packs between cyl #2 and #4.
This is the first leakdown test I've done, but I took my sweet time getting familiar with the tester and I feel very confident that my numbers are accurate within the limitations of the HF tester.
Should I be concerned with cylinders #2 and #4?
It seems like most everyone on here loves to talk about ringland problems... I'm praying I don't turn in to one of them.
***** June 6
Current mileage: 101k and some change.
Early this morning warmed the engine and I ran about ½ a can of seafoam through the intake. I 'injected' the liquid into a vacuum hose using a children's medicine syringe (while engine running), let it sit for 30 min, then ran the motor at about 1500 rpm until I ran out of Batman's smokescreen.
Results: no change.
***** June 10
I swapped the two camshaft position sensors (bank 1 <--> bank 2) and the misfire remained in cylinder #2.
Wait... why is my engine smoking? Oh awesome another thing to work on, the bank 1 sensor is sitting in a puddle of oil.
***** June 28
Checked valve clearances on cylinders #2 & #4 (bank 2, driver's side).
Everything is within spec except for the two cyl #2 exhaust valves. The clearance between the cam lobes and the buckets only measures .008" (.0085" won't go). This seems to be pretty consistent with everybody ragging on the stock tune for burning exhaust valves.
***** July 3-5
This weekend I adjusted my two cyl #2 exhaust valves to get the clearances within the FSM (factory service manual) specified range.
The clearance between the buckets and cam lobes measured .008", so I ground .007" off each exhaust bucket to get my new desired clearance of .015" (on the high end of the acceptable range).
Results: the engine runs, maybe
even a little better than before, but the misfire is still obviously present.
Data logging via RomRaider indicates that my misfire is almost only happening at lower RPMs, and my most resent LV snapshot indicates that I may have a vacuum leak. I'm dreading taking off my intake manifold to pin down the leak I found back in April, but it looks like that may be my only remaining sub-$5k possible solution.
Reference: Adjusting valves per the "Heiche budget method" here
Installed new: valve cover/spark plug gaskets, idler pulley, both accessory belts.
Reinstalled existing: cam sprocket bolts (no problems), timing belt.
Cleaned: valve cover, TB covers.
***** July 19
Borescope'd the pistons, didn't have enough room inside the cylinder to rotate the camera to see the cylinder walls.
(First posted on page 10, post #145.)
***** August 3-4
About 750 miles since previous update, the idle's been buttery smooth for the last 150 miles and no more CEL's... wah wah waaaah... until today. P0302 and P0304.
The data shows that my cyl#4 is healthy, and I believe the P0304 code was only thrown because cyls#2 & 4 are adjacent and share camshafts.
From BtSsm log 08/03 prior to the CEL (15,267 data points):
Cyl#1 misfiring 0.000% of the time (0/15,267)
Cyl#2 misfiring 9.792% (1,495/15,267)
Cyl#3 misfiring 1.827% (279/15,267)
Cyl#4 misfiring 0.210% (32/15,267)
From BtSsm log 08/04 after ECU reset (22,155 data points):
Cyl#1 misfiring 0.000% of the time (0/22,155)
Cyl#2 misfiring 4.694% (1,040/22,155)
Cyl#3 misfiring 1.620% (359/22,155)
Cyl#4 misfiring 0.000% (0/22,115)
***** August 5
Smoke tested (again) and found my air leak at the seam between the intercooler and the throttle body hose.
Solution: remove flappy soft rubber end pieces from the TB hose, clean, reinstall.
Results: misfire still present.
***** August 17
I installed a new digital pressure regulator on my air compressor, so I did another leakdown test:
Results from leakdown test #2:
Cyl 1: 27%
Cyl 2: 23%
Cyl 3: 24%
Cyl 4: 27%
***** Sept 25
I bought a WIX oil analysis kit off Amazon and sent it to ALS Tribology in Ohio for analysis.
Results are here
***** Oct 3, '15
Replaced OEM turbo inlet with a new silicone Mishimoto inlet (p/n MMHOSE-SUB-08IHBL).
Misfire results: no change.
***** Nov 5
Replaced: Timing belt, tensioner, pulleys, water pump, thermostat.
Fun: Painted timing belt cover.
***** Dec 2-13
Mileage: needs updated
- Removed intake manifold and TGV assemblies to inspect and replace the TGV-to-head gaskets. After removal there were no signs of the gasket leaking. I'm replacing the OEM TGV-to-head gaskets with Grimmspeed p/n 32001, and I'm reusing the IM-to-TGV o-rings that I replaced about a year ago.
- Completely gutted TGV assemblies, painted red (for speed, duh). Disabled associated codes in my current ROM, reflashed ECU.
- Accidentally broke a small piece of plastic off one of the fuel injector connectors. It still clipped on and felt pretty secure, but I used Q-bond (awesome stuff) to reattach the broken plastic because OCD.
REFERENCES (in no particular order):
Misfires & Rough Idling
Pistons / Ringlands:
Compression / Leak-down Testing / Vacuum:
(mostly for carb'd cars)
Misscellaneous (pun intended):
Logging / Learning View / Tuning:
ER HC error code + solutions:
(tons of links)
(solution: poor ground connection on DS TGV)
(could Er Hc be caused by my leaky gas issue?)