Diagnosing rough idle / misfire

[kennedy]

No CELs, Stage 1 map from infamous that wasn't finished (rev01).. only running this because on the stock 522 map I found on scoobypedia I'm always maxing out the first bank @ 15 in LearningView.

 

IAM in LV drops to 87.5 after a week or so. I notice knock values will disappear and read 0 sometimes too even if IAM is 1 so i assume that it occasionally drops and clears all that, then makes its way back to 1.

 

Car idles funny, sounds like it stumbles and misses when warmed up and idling at stoplight/parking spot, etc.

 

Included a pic just to show the cyl roughness and anything else.. I'm not familiar with reading romraider data.

 

Thanks for any input and suggestions! Intake system pressurized to find leaks, none found - smoke machine yesterday found no leaks either.

 

Log around the city for ~10 min was too big, i hosted here:

http://www.ansresource.com/boyd/aroundthecity.csv

warmedupidle.csv

[kennedy]

OK - I decided to reflash back to stock and continue logging to eliminate variables. Since the 522n map was always showing 15 in 0-5.6 AF learning, I decided to try the 521n map.

 

With rom raider running, I observed 11 counts of knock within like 30 seconds of the car being started, entirely cold engine.

 

This only went up as I drove. I went about 25 miles with some mixed driving speeds, did a third gear pull, etc. Pulling min knock correction was -6, max 5. I parked and turned car off, checked learning view, and I saw this:

 

http://ansresource.com/boyd/lv320.gif

 

So, I turned car back on and re-opened romraider logger, pulled out and immediately did a 3rd gear run. Cruised rest of the way home back the same exact route, cruise control mostly, and got home. Pulled learning view again and saw this:

 

http://ansresource.com/boyd/lv1170.gif

 

So, my AF learning for 0-5.6 went from acceptable to unacceptable despite having done the exact same things as I did on the way to where I turned car off - half way point.

 

Again, pressurized intake system and heard no leaks a few weeks ago. Smoke test found no leaks yesterday.

 

Here's what romraider saw after my trek back from the half way point where I had turned it off and re-opened it.

 

http://ansresource.com/boyd/romraid.gif

I've observed roughness on all cylinders but #1 between this shot above and the one I included with my first post.

 

I've included a csv file of the third gear pull I did. I can see that I have knock correction occuring and it's baffling - at least it isn't showing up in learning view? As I was driving I was glancing over at romraider running and in 5th gear under slight acceleration from 65-70 I watched my knock sum go up by one, I was trying to rule out shift knock and I guess I did.

[EDIT] I think I included a different log that wasn't my 3rd gear pull, there's too much data and I know I pressed my defrost button again right after I hit 5000something rpm... what I included has a lot more data so it must not be it - I believe my third gear is this: http://www.ansresource.com/boyd/real3rdgear.csv

 

Once again, help greatly appreciated as I'm at wits end... not like I knew much to begin with, but stock maps aren't performing well either and I fear I've been doing and may be continuing to do damage to the engine.

 

Stock, short of hawgzaust.

 

Brandon

thirdgearpull.csv

[Max Capacity]

I'm no expert on reading your logs, but IMO if you haven't replaced the o-rings at the intake to TGV's. You might want to think about doing that.

 

When you see the old o-rings next to the new one's you'll understand.

[kennedy]

Thanks Max, I've started gathering information on this.

 

On another note, I went out to the car this morning and checked the MAF sensor again - gave it a little spray with electrical parts cleaner and let dry. Reset ECU and took for a spin..

 

Car died at like 4 stop lights before I made it to the open stretch of high way that I did last night. Same path, stops, etc... but was dying - it hadn't done that after a reset since about a month ago.

 

At my half way pt (25 mi) I checked learning view, had knock showing up (had none with this stock map yesterday at same half way point) and IAM was at .375

 

I popped hood and made sure maf sensor was connected properly since that was the only thing I had touched since last night. Had 5 counts of roughness in cyl 2, 7 in 3, and 3 in 4 - no roughness in cyl 1 again. I reset ECU, again, just to see if I could make it to IAM 1 by the time I got back from my half way point. Nope - was reading .5 with no knock.

 

I was at a gas station and used a 12oz bottle of techron fuel system cleaner and then filled up. First time using non-shell gas in a while - I had been using shell 91 exclusively from a few of the bigger shells around town since this started now like two months ago.

 

Took it out for a spin back about 15 miles, ZERO counts of roughness in any cylinder - idle no misfire felt or heard at idle which it was consistently doing. IAM was at 1, no knock logged in learning view.

 

I can't/don't really believe that a 12oz bottle could circulate and do it's job that quickly by any means, but I don't know what else to attribute it to. I find it hard to believe that I've had bad gas from several shell stations that see enough traffic that their tanks are routinely refilled.

 

Also, my gas mileage on this 15 mile last test was incredibly high! I did maybe 4 miles of 70mph, the other 11 miles were city, max 35mph and hit lots of stop lights. It was reading just below 27MPG! I'm always at about 22 MAX in city after a fill up and counter reset - and even when I've filled up at the shell 100 yards from interstate offramp, my MPG is generally 25 after driving 75 miles @ 75mph on cruise control.

 

If the fuel system cleaner is also a super octane booster and increased the efficiency of the fuel then I could believe the results - otherwise I'm not holding my breath, but very happy to see SOME sort of improvement. I hope to be swapping MAF sensors with another member for a few minutes in the near future to rule that out.

[bmx045]

vac leaks can cause idle misfires/roughness/stalls coming to stops, your idle vac looks low (~-14inHg, should be -19-21). this can indicate unmetered air entering the intake system (loss of suction due to a hose popped off, leaking gaskets) and cause a mis. perhaps you should revert to stock map and start with square 1 with mechanical issues.

 

subaru service manual states idle vac, in neutral, fully warmed up, a/c and vent system off, to be -19 - -21inHg, psi (romraider manifold relative pressure) to inHg is ROUGHLY x 2 so multiply them to get an estimate of inHg.

[kennedy]

Thanks BMX - In the last few weeks I've done a boost leak check by pressurizing intake (no hissing anywhere that we could hear) and smoke test - didn't show anything.

 

I swapped MAF sensors with another member (thanks Norman!) yesterday and problem persisted, still getting misfires but this time I was getting misfires on cyl 1 as well - up until now cyl 1 was the only one not misfiring.

 

I did a drive down to Denver two days ago and IAM stayed at 1 - then on the way back I noticed I had less power. Thought it was because I had more people in the car, but when I got home I saw my IAM was at .62. I reset ECU when I swapped MAF sensors last night and IAM made it back up to 1 and has stayed there since (~50 miles) - misfires in cyl 2,3 and 4. I included a 3rd gear pull I did about 20 minutes ago.

 

Ideas appreciated as always - going back to 5250 performance (mweiner) on Weds for much needed wheel bearing replacement and spark plugs - he'll do compression and leakdown test at this time and hopefully throw out some other ideas.

I've been running 521N map from scoobypedia, said to be stock. I've also tried the 522N but in trying to eliminate variables I loaded the slightly older one.

10-23_3rdgr.csv