Help sorting out misfire CEL

[phil527]

Need some help/advice of where to look next for a misfire CEL I've been getting.

 

The background:98 Legacy Outback, 2.5, 122K, Tbelt changed at 65k. Earlier this spring I replaced a number of parts to correct some low mileage and hard starting. I replace the front O2, the CTS, as well the usual fuel and air filters, PCV valve, plugs and wires. That worked great- car runs superb smooth idle good power and great mileage but I still get a misfire CEL whenever I go on long trips. I recently put in a new coil pack, and cleared the code but it came back on a trip this weekend.

 

Here is the interesting thing, the car runs fine and won't throw a code with around town driving or on short (slower speed )trips. Only when cruising at 75-80 on the highway after about 1-1.5 hrs will it come on. When it does you would not even know anything was off - no hesitation or change in how the car is running.

 

So I am thinking now that it probably is not ignition related. Should I start looking at fuel? I unfortunately can't remember which cylinder the misfire code came from earlier in the spring but the last couple came from cylinder 3. Could this be a bad injector? Or maybe low fuel pressure on the fuel rail?

I am not sure what the engine computer looks for with a misfire, high HC in the exhaust?

 

Any advice would be greatly appreciated!

[idosubaru]

let me guess, they are not Subaru ignition wires? this is a phenomenon of the EJ series engine (in your car) only, not all Subaru engines. if they aren't Subaru then you need to replace those first and i bet your problem goes away. if you're new to this, then i know it sounds weird but it's a "feature" of all EJ engines, they are highly sensitive to ignition wires. those of us that are very familiar with Subaru's have seen this quite often. you should also use stock NGK plugs on this motor as well, but that's probably not your problem.

 

 

the one suitable aftermarket option is Magnecor and I highly recommend them, they have put considerable time and effort into designing an EJ specific wire, which is why they only offer one size for the EJ engine instead of the multiple products for every other car/engine out there. i use them exclusively, they last the life of the vehicle, i've used one set on two different vehicles before.

[phil527]

No, they are Genuine Subaru wires. The coil pack is not from Subaru but I seem to think it is probably not that. This problem started this past winter, but has been more noticable with all the long trips I've been taking this summer.

[idosubaru]

have you pulled the plug wires to see if everything looks okay? i'd probably read the code, note which cylinder it is and then swap wires. if the code "moves" with the wire, then you know the wire is bad. if it's cylinder 3...then you swap with cylinder 2 and the code is now on cylinder 2, then you know it's that wire that is bad. it's an easy test that's the only reason i suggest it.

 

if the problem didn't change at all when you swapped coils then i agree, it's not likely to be a coil problem.

or if you have an extra set of wires, swap them over and see if that does anything.

i'm not sure how the ECU sense cylinder misfires, that's a good question.

 

what kind of plugs did you use?

did it do this before you changed wires and plugs last year?

[nipper]

You need to use OE ignition wires, though magnicore seems to have a good rep. This is a typical aftermarket wire failure.

 

Aftermarket suplliers use one common wire for almost all applications, just varying the lengths and boots as needed. Cost being the bottom line they go with the lowest comon denominator. Some use the best wire that will work with all apllications, but they cost as much as OE.

 

How wires are tested:

They spend 300 hours in an oven at 400 degrees (though it may be 500 now). They have 100,000 volts running through the wires, with the inside of the oven being the ground. If they arc any time before the 300 hours they fail.

 

This is the bare minimal accepted test method. So as you can see they do not reflect real world conditions. This is also a set and forget test. This was fine when you only had a distributor. Now with waste spark this is no longer acceptable. The more expensive wires do more extensive testing.

 

nipper

[phil527]

Just to clarify these are Subaru wires, NGK plugs. The wires and plugs have probabley less than 5K on them. The car ran better with the new plugs and wires but the misfire code still came back.

 

Only the coil is aftermarket, I replaced it last minute before I headed out on a trip -but it did not fix the problem.

 

That is why I am thinking it must be something else. Small vacuum leak?

Fuel supply? If the fuel rail was running a little low on pressure causing a leaner mixture by starving the injectors- would that give a misfire code? Or is a misfire detected by unburnt fuel in the exhaust?

 

The thing is it only shows up when the engine is spinning above 3500 continuously for an extended period.

[nipper]

Just to clarify these are Subaru wires, NGK plugs. The wires and plugs have probabley less than 5K on them. The car ran better with the new plugs and wires but the misfire code still came back.

 

Only the coil is aftermarket, I replaced it last minute before I headed out on a trip -but it did not fix the problem.

 

That is why I am thinking it must be something else. Small vacuum leak?

Fuel supply? If the fuel rail was running a little low on pressure causing a leaner mixture by starving the injectors- would that give a misfire code? Or is a misfire detected by unburnt fuel in the exhaust?

 

The thing is it only shows up when the engine is spinning above 3500 continuously for an extended period.

 

Why did you replace the coil?

 

If you replaced everything at the same time, reinstall the old coil.

 

What brand is the coil?

 

Missfire is detected by the voltage spike from the coil . These computers are really smart, much more powerful then we give them credit for, but at other times we forget that they are only tools, and tools are only as good as the person operating them.

 

What you are seeing is a voltage breakdown as the coil heats up to a steady state temperiture, at least thats my theory.

 

COil gets a lot hotter on the highway then just idleing around town.

 

 

nipper

[phil527]

I replaced the coil (from NAPA) because I was thinking along the lines that you mentioned. I thought maybe it was getting hot on a long trip and arcing out, also since I had just changed the plugs and wires this seemed like the only thing left. The resistances were all in spec on the old coil.

 

Found this on the web http://www.cleanairforce.info/us/obd_dtc/p0300_engine_misfire_detected.htm

 

They are indicating that the misfire is determined by monitoring the crank and cam speed. The flashing CEL indicating a high misfire rate is something that I have seen. My CEL came on then started flashing for a short bit,then stayed on steady. If this is true, based on the way it was flashing I think I should have felt the enging stumble, which I didn't - so I am thinking it might be a sensor issue.

 

Would a flaky crank or cam sensor give a misfire code? I have not gotten any crank or cam sensor fault codes. Or do these components usually fail completely?

[nipper]

Crank sensor controls spark, cam controlls injection.

 

I dont trust no brand replacement parts with a waste spark ignition system.

 

Crank sensor can thorw a code, when it fails the car wont start.

 

Thats a generic article, so i wouldnt automatically apply it to subaru, as subarus can be a little "odd". look how finicky they are with ignition parts.

 

What specific code are you getting.

 

There are generic codes, and codes specific to individual mfg's and engines. The codes are standardized, but not everyone uses the same methods to control emissions, ie some cars have a knock sensor or O2 for each bank of cylinders.

 

nipper

[phil527]

Yeah, I realize that is a generic article, but I was just looking for a good explanation as to how the ECU comes up with the code. After the last trip I was keeping a close eye on the CEL and I could not feel any change in engine performance when it came on.

 

That is why I am starting to think it might be a sensor problem thinking there is a misfire instead of actually being one.

 

Maybe I am wrong in thinking I would feel a sputter of hesitation when the misfire occured. The specific code was P0303 the last couple of times and unfortunately I did not write down the preveious ones - I thought my memory would hold:-\. They definitely were misfire codes though, 0304 seems familiar.

[WAWalker]

Subrau ECM's do NOT monitor ignition Kv's nor does any other engine control system that I know of. That would be nice though. Would save having to spend money on a scope for that.

The ECM is looking at crankshaft speed via the crank sensor when monitoring for misfires.

 

With Subaru engines a misfire on accelaration is 99% of the time an ignition miss. Worn and or carbon tracked spark plugs and wires. Under heavy to moderate load is where this shows up, and you can really feel it. Most of the time you will not get a illuminated CEL, even though you can feel the misfire.

 

A misfire at highway speed (cruising) is 99% of the time a mechanical problem in the valve train. Tight valves. Valves hanging up momentairly in the valve guides. This is hardly ever noticable, but will illuninate the CEL even though you can't feel a misfire.

 

Go figure huh! But this has been my experiance working on these cars for a living.

[phil527]

So sticking valve could be a cause and could make sense as to why I might only be seeing it on cylinder 3. What are your thoughts on usefullness of oil additives to "free sticky valves"?

 

Would a valve lash adjustment also create potential for misfire? I believe it is recommended at around 100K, I have not done it yet. I was figuring I would do it when I needed to do some more major involved work.

[WAWalker]

Yes, tight valve lash adjustment will cause misfires at hight RPM.

[nipper]

Time to get out ye old vacume gauge. The vacume gauge is perfect to tell if there are sticking valves or an ignition issue.

 

I'm sticking with ignition myself.

 

Put on a vaccum gauge and tell us exactly what it is reading and what the needle is doing.

 

nipper

[WAWalker]

Good luck with a vaccum gage when it only happens very intermittently at 70-80 MPH on the freeway.

 

Oh, and make sure you have a helper, someone is going to have to have their eyes glued to the gauge while the other drives.

 

Please let us know how this works out if you try it.

[OB99W]

[...]The ECM is looking at crankshaft speed via the crank sensor when monitoring for misfires.[...]

Yep. When the ECU senses that the engine speed should be relatively constant (no significant changes from the throttle position sensor, etc.), it "assumes" that each pulse from the sensor will occur at a specific time. Because a misfire will slightly slow the engine for an instant, the ECU determines that one has occured if a pulse from the sensor is "late". A characteristic of this monitoring approach is that any problem resulting in momentary slowing of the engine will be interpreted as a misfire; that means ignition, fuel, or valving problems all look the same to the ECU.

 

One of the difficulties in trying to diagnose the problem is that "the car runs fine and won't throw a code with around town driving or on short (slower speed )trips. Only when cruising at 75-80 on the highway after about 1-1.5 hrs will it come on.". The ECU needs sufficient time with a warmed engine running at a constant speed/throttle to determine misfire, especially if the trouble is intermittent. So it may be that if the "around town" trips are very short, don't allow for full warmup, rarely include constant speed (lots of stop-and-go), etc., an intermittent problem may actually be occuring during local driving, but not be detected. Highway cruising, on the other hand, is ideal for the ECU to "see" such a problem.

 

If we assume the misfire is indeed only occuring at highway speeds, then certain possibilities come to mind. I don't have the time to get into that right now, but I do have a couple of questions:

 

1) Does it actually take "cruising at 75-80" (not at 60, for example) for the CEL to be set?

2) Just how great is the "great mileage"; in particular, what's the local versus highway mileage you're getting, if you can separate the two?

[phil527]

To clarify "around town" and "highway speed":

 

I changed the coil around labor day and cleared the codes. My wife headed back to Niagara Falls the following weekend - she went the back road route so she would have gotten up to 60-65 top speed in stretches but a little under 3 hrs of driving should have been plenty of time to get good and warm. Doubt the engine RPM got above 3500. Trip there and back no CEL codes, local around town no problems.

 

Recent trip to DC this past weekend, 45-60 minute drive to Binghamton (also 2 lane back roads) no CEL. Get south of Binghamton on 81 set the CC to 75 and a short while later CEL comes on. It came on for a half a minute-minute, started flashing for 10-15 secs then stayed on the rest of the way there and back. When it came on I could not feel anything different in the way it was driving.

 

As far as mileage I got a little under 26 for three tankfulls. This might be skewed down as there was a fair amount of stop and go in Harrisburg and construction north of Scranton. When we got the car I would get a bit over 27 on trips (this is a 5 speed). So I am happy with where the mileage is and how it feels when it runs.

 

Nippers thinking that it is ignition sounded good at first, but after changing the plugs, wires and coil that would seem to be ruled out. Also the fact that is seems to only happen when this engine is spinning above 3500 continuously for a while may point to the sticky valve theory that WAWalker suggested.

 

I think I am heading back to Buffalo this weekend, so I might pull the vac gauge from my truck and find a spot to set it up in the car for the trip. I am due for and oil change so maybe I will do a Seafoam treatment and throw some Marvel oil in with the change.

 

My only other passing thought is oil pressure. I have seen posts about resealing the oil pump during Tbelt work. If that has not been done would I get a reduced oil pressure - maybe not keeping up at the high RPM? I don't get an oil pressure indicator light though.

[WAWalker]

Oil pressure is not the problem.

 

The intake valves can be cleaned up running Sea Foam or othe top engines cleaners through. The exhaust valves are a different story. I have droped the exhaust on a few of these cars in order to soak the exhaust valves with Sea Foam while turning the engine over buy hand to make sure I'm getting the valve stems good. Seems to have helped, no news is good news.

 

Have also seen this type of misfire problem on Rebuilt 2.2L engines untill they were good and broken it. I've got to believe it is the valves hanging in tight new valve guides. Since I can't spend days driving customers cars for 100's of miles to try and duplicate the problem, all I can do it speculate. But I believe this is what is happening.

 

Your last resort is going to be to check valve lash clearance. I know it is labor intesive, and expensive if you have to pay someone to do it...............Good Luck

[OB99W]

To clarify "around town" and "highway speed":

 

I changed the coil around labor day and cleared the codes. My wife headed back to Niagara Falls the following weekend - she went the back road route so she would have gotten up to 60-65 top speed in stretches but a little under 3 hrs of driving should have been plenty of time to get good and warm. Doubt the engine RPM got above 3500. Trip there and back no CEL codes, local around town no problems.

 

Recent trip to DC this past weekend, 45-60 minute drive to Binghamton (also 2 lane back roads) no CEL. Get south of Binghamton on 81 set the CC to 75 and a short while later CEL comes on. It came on for a half a minute-minute, started flashing for 10-15 secs then stayed on the rest of the way there and back. When it came on I could not feel anything different in the way it was driving.

 

As far as mileage I got a little under 26 for three tankfulls. This might be skewed down as there was a fair amount of stop and go in Harrisburg and construction north of Scranton. When we got the car I would get a bit over 27 on trips (this is a 5 speed). So I am happy with where the mileage is and how it feels when it runs.[...]

Okay, so 65 and below doesn't reveal a problem, but 75 or so does, and your gas mileage is good, but slightly down...

 

The problem is obviously very subtle. The things that are typically different with a change from 65 to 75 MPH cruising are additional demand for fuel/air/spark, and an increase in heat and exhaust gases that the engine has to rid itself of.

 

You previously mentioned the possibility of low fuel pressure; that, or ability to deliver sufficient volume could be an explanation, and cylinder #3 may just be more sensitive. Since you've already replaced the filter, the pump itself could be suspect.

 

As WAWalker has mentioned, you may indeed have some valve issues.

 

Another possibility (and I admit this is going a bit "out on a limb") is a slight exhaust restriction. That can cause high-speed-only miss and some gas mileage reduction, but usually wouldn't be limited to just one cylinder.

 

Two more questions:

1) Do you see any difference in the temperature gauge reading when cruising at 75 versus 65?

2) Could you try loading the engine more on a trip at 65 MPH (turn on headlights, etc.), and see if that will turn on the CEL at the lower speed?