The Dreaded Misfire Saga Continues

[subsince77]

And you should. Nothing wrong with it. I meant mine only as a complement.

Have a good day!

Yeah, I took it as agreement. All is well, except my car.

[subsince77]

OK, new CEL after injector swap - still #'s 3 and 4:mad:

 

So, let me just be sure I am thinking straight here before I move ahead. It seems that I am down to two possibilities. Either a vacuum problem likely caused by a valve (in which case I will have a warranty issue with CCR), or a problem somewhere in the fuel injector wiring harness. Is that right?

 

If so, my next step as discussed earlier is getting a vacuum gauge and figuring out how to use it, while I am driving, to see if it indicates a loss of vacuum when the CEL is flashing. Unless someone tells me otherwise that will be my next move.

 

If vacuum is eliminated, I don't really see any way to address the wiring harness (I've looked at the parts I can see and found no evidence of a break, or wear), and I'm not sure it would be worth it. Those wires run all over the place under the intake manifold. I don't feel capable of tearing that much engine apart to get at all the joints in that harness.

 

Anybody ever replace the injector harness, or is that a bit overboard?

 

Anything else I should do before the vacuum gauge?

[VSarge]

Consider yourself lucky. See http://www.ultimatesubaru.org/forum/showthread.php?t=89121 to find out why.

 

I assume the gas didn't seem to be under pressure (it shouldn't have been).

 

I second that. Just grenanded a 351W motor doing that.

 

Sarge

[hankosolder2]

OK, new CEL after injector swap - still #'s 3 and 4:mad:

 

So, let me just be sure I am thinking straight here before I move ahead. It seems that I am down to two possibilities. Either a vacuum problem likely caused by a valve (in which case I will have a warranty issue with CCR), or a problem somewhere in the fuel injector wiring harness. Is that right?

 

a vacuum gauge sounds like a good idea to check for valve issues. The intake manifold isn't as hard as it looks to remove. I think you could swap the FI wiring harness yourself, but I'm not convinced it's the problem. Really, the FI harness doesn't flex on deceleration at all...the other section of the harness between the connectors on the passenger side of the engine and the firewall would really be the only section which shifts as the engine moves on its mounts on decel. You could easily flex the wires and connectors by hand while it's running and if they were that bad, I'd guess you'd set a CEL at idle.

 

Your best bet might be to buy an OBD II scan gauge (or take it to a mech with one.)...I think some can capture real time data at the precise moment a CEL is set. You could then possibly find out which parameter is out of whack in the moment.

 

I admire your persistence on this one!

 

Nathan

[subsince77]

Your best bet might be to buy an OBD II scan gauge (or take it to a mech with one.)...I think some can capture real time data at the precise moment a CEL is set. You could then possibly find out which parameter is out of whack in the moment.

Nathan

 

I have one, that's how I'm reading the code and resetting, but its a pretty inexpensive one from Wally World. It says it has a freeze frame function, which I think does what you describe. But I haven't bee able to figure out how to see what is captured. I guess it's time to read the manual and see if will actually do this.

[subsince77]

Looks like it takes a better OBD2 scanner to read the freeze frame info from the ECM. A friend loaned me a vacuum tool today. I have to get some longer hose and tees before I can use it though.

[OB99W]

Looks like it takes a better OBD2 scanner to read the freeze frame info from the ECM.

Capturing freeze frame data is a good idea, if possible -- see post #23 of http://www.ultimatesubaru.org/forum/showthread.php?t=85794&page=3 .

 

 

A friend loaned me a vacuum tool today. I have to get some longer hose and tees before I can use it though.

Before you get longer hose, etc., just warm up the engine and do some basic testing using the gauge with the car standing still. Measure the vacuum at idle; record the reading and note how steady the needle is. Rev the engine, slowly and rapidly, and see how the vacuum changes (check online sites for how to use the gauge and interpret results). Let us know what you find.

 

I looked back at the threads (beginning late last year) covering your P0303/P0304 problems with both the previous engine and the current one. In rereading, I noticed that there seemed to be two "triggers" for the codes that you mentioned -- rough road surfaces, and deceleration after high speed running. Does a rough road still trigger the codes? If so, there may yet be an ignition-related connection issue.

 

While the cylinder #3/#4 combination of misfiring (due to the wasted spark ignition) often points to ignition problems, ignition trouble isn't always the cause. That's why it was good to swap the injectors, in order to eliminate them as a potential factor.

 

Deceleration causes high intake vacuum, and a leaner mixture. It's a condition when misfire can occur if the mixture is sufficently lean. A vacuum leak could make things worse, and that's one of the reasons checking intake vacuum is useful.

 

It seems you've eliminated most of the components relevant to P0303/P0304 that are ignition-related by replacement, and swapped the injectors, so those two areas are becoming less likely to be problematic. If vacuum checks don't turn anything up, consider that there's another system that can cause misfires -- the EGR. Because of the way it's tested under OBD, some EGR problems don't cause a code pointing directly to it. A slightly sticking/leaking EGR valve might only result in misfire under rare circumstances. Note that it's not just the valve itself that can cause problems. For some info, see http://endwrench.com/images/pdfs/Emissions.pdf .

[McDave]

I was wondering exactly when the cel came on too. Hitting bumps is a good clue.

 

I also was wondering if you actually feel the misfire when the cel comes on?

[subsince77]

The rough road thing was a working theory for a little while there, but I have to say that the vast majority of the time the CEL has come on on smooth pavement. I think the couple of rough road instances were coincidences. I can't pin the CEl down to any one specific thing exactly - that's part of the difficulty. I do think that most of the time it is happening on deceleration, but not always from high speed. i have really tried to watch rpms, time running, temperature, etc. none of it is totally consistent, but the deceleration seems at this time to be the most consistent factor. I tend to think that most of the time it happens around 3000 to 320 rpm, but can't swear to it. Certainly, the trip last weekend, with the light consistently flashing when topping a pass, was one of the clearest indicators and what got me thinking about vacuum.

 

As to feeling the misfire, NEVER EVER. Not a miss, not a nudge, not a whimper. Nothing to hear, nothing to feel. Even when the light is flashing for a time, nothing. That is why i drop this whole pursuit once in a while. If it weren't for the light I would never know there was a problem. I still think the ultimate solution might be a small piece of electrical tape that covers the cel.

[McDave]

The rough road thing was a working theory for a little while there, but I have to say that the vast majority of the time the CEL has come on on smooth pavement. I think the couple of rough road instances were coincidences. I can't pin the CEl down to any one specific thing exactly - that's part of the difficulty. I do think that most of the time it is happening on deceleration, but not always from high speed. i have really tried to watch rpms, time running, temperature, etc. none of it is totally consistent, but the deceleration seems at this time to be the most consistent factor. I tend to think that most of the time it happens around 3000 to 320 rpm, but can't swear to it. Certainly, the trip last weekend, with the light consistently flashing when topping a pass, was one of the clearest indicators and what got me thinking about vacuum.

 

As to feeling the misfire, NEVER EVER. Not a miss, not a nudge, not a whimper. Nothing to hear, nothing to feel. Even when the light is flashing for a time, nothing. That is why i drop this whole pursuit once in a while. If it weren't for the light I would never know there was a problem. I still think the ultimate solution might be a small piece of electrical tape that covers the cel.

You mentioned in another post that you had another ecu for it. Did you ever try it?

 

Also, this problem was there before the new motor, right?

 

(I know it sounds like we're going around in circles, but I'm going somewhere with this)

[subsince77]

Yes I am currently running the new ECU, and yes this problem predated the new engine. I was very surprised when the CEL came on about an hour into my drive home from Denver and CCR. I stopped, called them, they invited me to bring it back down, but all of us agreed at that time that the problem was likely in either the ignition parts or injectors that moved from the old to the new engine.

 

Right after that is when the information came that sometimes the neutral switch will cause this, so I replaced that. Then I replaced plugs and wires a few times, then the ECM, and now onto the injector phase of the story.

 

The fact that the problem seems to be exactly the same on both engines is why I have never really suspected the valves. I did suspect them on my old 196000 mile engine, and expected the problem to go away with the new engine. But here we are.

 

Today's experience is another good example of there being no good example of when this happens. I drove into and around town with a few starts and stops - say 25 minutes of driving - no cel.

 

The drove on the highway to Crested Butte, probably a 45 minute trip today with 10 minutes idling at a construction zone - no cel.

 

Parked for almost 2 hours of mountainbiking. Started to drive home. CEl came on on flat smooth road at about 55 mph. Steady driving, not hilly, my wife was driving but I don't think we were even decelerating, although it is possible that the light started on a slight downhill with light or no accelerator. She didn't know I was wanting to keep track of exactly when it comes on, so she wasn't sure on that.

[McDave]

Ok, here's what I think is happening. I may be all wet and those more familiar with how Subaru does things and acts up is welcome to correct me.

 

OBD II uses a correlation between the cam sensor and crank sensor speed readings to determine when a misfire occurs. A crank sensor failure usually causes a car to die or miss hard. A cam sensor failure doesn't. You generally can run without a cam sensor input.

 

Your model car has a history of cam and crank sensor problems according to the database I have access to. So much so that Subaru has released a kit that includes both for less that you can buy the two separately.

 

I'm betting the cam sensor is failing intermittently and is triggering the cel, but not causing a misfire, as they gererally don't anyway.

 

So, as much as I hate to tell people to throw their hard eaned money at a problem on just a hunch, I think it would be worthwhile to buy the pair (~$50 bucks I believe) and at least replace the cam sensor first, if not the crank sensor at the same time just to get it over with.

 

Ok guys... let the shooting begin. What do you think?

 

Edit: Wait a minute. I just saw your vehicle is listed as a 2.5? Scratch what I said, it applies to the 2.2s. The 2.5s exhaust valves tend to sink a bit (according to the database) and need to be adjusted. A vacuum gauge would likely show a flutter when it's acting up.

[subsince77]

You were going really good there for a minute though. If I wasn't so stubborn, I prefer to think of it as determined, to figure this thing out, I would just drop it. I don't think the car ever actually misses, the brain just thinks it does.

[McDave]

Yeah, there are two different trains of thought going on up there. If I could somehow talk you into putting the vacuum gauge on it already and taking a drive, we could narrow it down to one.

[hankosolder2]

Ok, here's what I think is happening. I may be all wet and those more familiar with how Subaru does things and acts up is welcome to correct me.

 

OBD II uses a correlation between the cam sensor and crank sensor speed readings to determine when a misfire occurs. A crank sensor failure usually causes a car to die or miss hard. A cam sensor failure doesn't. You generally can run without a cam sensor input.

 

Your model car has a history of cam and crank sensor problems according to the database I have access to. So much so that Subaru has released a kit that includes both for less that you can buy the two separately.

 

I'm betting the cam sensor is failing intermittently and is triggering the cel, but not causing a misfire, as they gererally don't anyway.

 

So, as much as I hate to tell people to throw their hard eaned money at a problem on just a hunch, I think it would be worthwhile to buy the pair (~$50 bucks I believe) and at least replace the cam sensor first, if not the crank sensor at the same time just to get it over with.

 

Ok guys... let the shooting begin. What do you think?

 

Edit: Wait a minute. I just saw your vehicle is listed as a 2.5? Scratch what I said, it applies to the 2.2s. The 2.5s exhaust valves tend to sink a bit (according to the database) and need to be adjusted. A vacuum gauge would likely show a flutter when it's acting up.

 

I thought the 2.2 and 2.5 cam & crank sensors were interchangable? Certianly, they are not failure-proof just because he happens to have a ej25. I don't think your two theories are mutually exclusive- he could have a bad cam sensor, or he could have a valve issue. I do wonder if it's possible for a bad cam sensor to glitch out long enough to set a misfire code, but not long enough to set a bad cam sensor code.

 

Nathan

[McDave]

I thought the 2.2 and 2.5 cam & crank sensors were interchangable? Certianly, they are not failure-proof just because he happens to have a ej25. I don't think your two theories are mutually exclusive- he could have a bad cam sensor, or he could have a valve issue. I do wonder if it's possible for a bad cam sensor to glitch out long enough to set a misfire code, but not long enough to set a bad cam sensor code.

 

Nathan

Now you're talking. A search with all three codes (inc. 341) at once showed only one instance though, and it turned out to be a cam off a tooth. So it sounds like it won't set with this mild of problem.

 

The pattern I'm seeing for just the 303 and 304 codes favors tight exhaust valves in the 2.5 motor. In fact it looks like there were warranty jobs on the early heads.

 

But on the 2.2 motor it looks like most needed a cam sensor and apparently the crank sensor for good measure since they offer(ed?) a deal on both. Maybe the early 2.5s need them too?

 

vacuum gauge vacuum gauge vacuum gauge...

[subsince77]

Yeah, there are two different trains of thought going on up there. If I could somehow talk you into putting the vacuum gauge on it already and taking a drive, we could narrow it down to one.

The gauge from my friend was actually not right. long story - he uses it to tune pianos.

 

I ordered one today, it should be in tomorrow. Getting time to hook it up and know what it's telling me - another story. But soon!

[McDave]

The gauge from my friend was actually not right. long story - he uses it to tune pianos.

"Oh, I thought you said tuning fork."

 

Sounds good. I can wait. No hurry.

 

When you get it, set it up so you can watch it easily while you drive. If it's a valvetrain problem it will start jumping/fluttering before the cel comes on, and could actually quit moving by the time the light comes on.

 

I usually slip it under the windshield wiper so it's right in front of me. Just don't turn on the wipers. Then take it out on a light traffic road and drive it at a steady 55mph like your wife was driving it while keeping an eye on the gauge and cel. Hopefully it will act up sooner than later.

[subsince77]

"Oh, I thought you said tuning fork."

 

Sounds good. I can wait. No hurry.

 

When you get it, set it up so you can watch it easily while you drive. If it's a valvetrain problem it will start jumping/fluttering before the cel comes on, and could actually quit moving by the time the light comes on.

 

I usually slip it under the windshield wiper so it's right in front of me. Just don't turn on the wipers. Then take it out on a light traffic road and drive it at a steady 55mph like your wife was driving it while keeping an eye on the gauge and cel. Hopefully it will act up sooner than later.

 

Apparently you use a vacuum gauge to tune player pianos, who knew?

 

I assume I will need to pick up a longer hose, and that there is a way to work it under the hood so the hose isn't pinched (I guess I could sawzall a window in the hood and leave it up). I haven't looked at any of that yet. I did see several places to connect to on the right side of the intake manifold when I was doing the injectors. The gauge should be here this morning. I even hit it on a one day 30% off tool sale at NAPA.

[McDave]

Apparently you use a vacuum gauge to tune player pianos, who knew?

 

I assume I will need to pick up a longer hose, and that there is a way to work it under the hood so the hose isn't pinched (I guess I could sawzall a window in the hood and leave it up). I haven't looked at any of that yet. I did see several places to connect to on the right side of the intake manifold when I was doing the injectors. The gauge should be here this morning. I even hit it on a one day 30% off tool sale at NAPA.

I remember player pianos. I had a great aunt and uncle that had one. Seems like I remember it broke once and my uncle explained to me how it worked. My aunt would have to keep pumping the foot pump until she got tired. I guess it developed a vacuum leak that got so bad they finally quit playing it. Loooong time ago.

 

Yeah, the vacuum hose is 3/16" or 5/32" and they can cut it as long as you want, and it's cheap. You might as well get enough to bring the gauge inside the car since it may take a few trips to get it to act up. Just route it out the back of the hood and through the drivers window (opened a crack).

 

Edit: The vacuum hose already on the gauge may be 1/4", but 3/16" or 5/32" is probably closer to the sizes used on your motor. Buy one of the smaller size hoses and step it down from 1/4" with a piece cut from one of the universal plastic tees.

[subsince77]

Ok, I picked up the gauge. I need a little clarification on hooking it up. It has one tee, and one cone shaped stick-it-in-the-end-of-a-hose type fitting. Since I am going to drive I obviously need to use the tee. Do you actually slice a hose, insert the tee, and leave it in there after the job is done? If so, I will need to get a plug for it before I do this. With just a quick look around, I didn't see anyplace where I could replace some kind of connector with the tee, so that leaves me thinking of cutting a hose - hate to do that.

 

I will look at my manual to try to see which hose would be a good one to use, but if anyone has a specific suggestion for the right vacuum, that would be great.

[OB99W]

Ok, I picked up the gauge. I need a little clarification on hooking it up. It has one tee, and one cone shaped stick-it-in-the-end-of-a-hose type fitting. Since I am going to drive I obviously need to use the tee. Do you actually slice a hose, insert the tee, and leave it in there after the job is done? [...]

Don't cut any engine vacuum hoses, just disconnect an end of one of them. Attach a small length of additional hose (it could be just a couple of inches or so) between the vacuum fitting you're going to use and one connection of the "T". Connect the gauge and the hose you pulled to the remaining "T" connections.

[McDave]

Cut a piece off the new hose you bought and plug it into one end of your tee.

 

Plug the gauge hose into the middle wing of the tee.

 

Pull a vacuum hose off the intake manifold that is about the same size as the tee and plug it into the last port on the tee.

 

Now plug the first hose that you put on the tee into the intake port you just unplugged. You are now tee-ed in and whatever thing you tapped into will continue to work as before.

 

Start it up. If the gauge doesn't move, the port you tapped into isn't manifold vacuum. Rev the engine a bit and if the gauge moves, you tapped into ported vacuum. If so, move your tee to a different port making sure it is downstream of the throttle blade.

 

It's all really much simpler than it sounds. We're talking analog here.

[OB99W]

I will look at my manual to try to see which hose would be a good one to use, but if anyone has a specific suggestion for the right vacuum, that would be great.

Do you see a fitting directly behind the ignition coil pack?

[subsince77]

Don't cut any engine vacuum hoses, just disconnect an end of one of them. Attach a small length of additional hose (it could be just a couple of inches or so) between the vacuum fitting you're going to use and one connection of the "T". Connect the gauge and the hose you pulled to the remaining "T" connections.

 

Man, you would really think that I could have thought of that little solution on my own. My current excuse is 8 hours without food - missed lunch. Now I'm fed and well.

 

OK, yes there are actually three connections right behind the coil, I tried the middle one, but it was a ported one. Nothing at idle, vacuum when I revved it. I'll try a couple until I find one. More later. Thanks for the patience.