The Mysterious #3 & 4 mis-fire solved - Maybe

[Cougar]

You are way ahead of me since you have already replaced those two items. My next thing to try would be to run some Seafoam through the intake to see if there is some carbon buildup causing this.

 

I would keep the wires. I doubt very much they are the trouble unless they are not seated correctly.

[Gloyale]

XTs have a common misfire issue due to corrosion of a splice connection in the harness. specifically the wires for the injectors, since they are in pairs.

 

Is it possible that there is a splice somewhere in the harness that is corroded, and causing this?

 

Could someone post a Diagrahm for the ECU/ENGINE wiring to see if there are any splices in the Injector wires shown?

[subsince77]

You are way ahead of me since you have already replaced those two items. My next thing to try would be to run some Seafoam through the intake to see if there is some carbon buildup causing this.

 

I would keep the wires. I doubt very much they are the trouble unless they are not seated correctly.

 

Done it. Good stuff, but didn't solve this problem. I think that yesterday the #3 wire was not seated properly. It was really cold, the wires were very stiff, the sun was going down, and I think I didn't quite get it pressed in. I felt another little click when I pushed it again later.

 

Today, I rechecked both ends of all the wires, made sure everything was well seated. I also gave the injector connections a couple of on and off cycles to clean them up a little. I drove around for a while, and drove up the valley for about 30 minutes (tough deal driving from Gunnison to Crested Butte on a beautiful blue day with the West Elks all covered in snow). No codes so far, so I'm at least back to where I was. We have a longer trip planned in a couple of days, so we'll see how that goes. At least this problem is only irritating and not debilitating.

 

XTs have a common misfire issue due to corrosion of a splice connection in the harness. specifically the wires for the injectors, since they are in pairs. Is it possible that there is a splice somewhere in the harness that is corroded, and causing this?

 

Pffff... Way over this boys head. I don't touch scary splice connections in wiring harnesses. I could mangle something beyond repair. But that is great info to know for real mechanics. Along those lines, I hate all those little plastic wiring coonnectors these days. No two seem to work the same way. I'm always trying to figure out the secret code for pushing, and or pulling, exactly the right combination of tiny fragile little snaps to get them apart. I only actually break something about 20% of the time though.

[Cougar]

I think you should give yourself more credit. A lot of folks wouldn't have tried the things you have done already. Checking and cleaning connectors is a simple thing. They can be a little tricky getting them apart at times but usually not much of a problem. Hopefully reseating the plug wire will fix this trouble.

[subsince77]

I think you should give yourself more credit. A lot of folks wouldn't have tried the things you have done already. Checking and cleaning connectors is a simple thing. They can be a little tricky getting them apart at times but usually not much of a problem. Hopefully reseating the plug wire will fix this trouble.

Thanks. I am exaggerating just a bit, but those connectors get very brittle in the cold, as you may know. The only routine maintenance thing that I know of that I have not yet done on this car is the front O2 sensor. I've never gotten a code on it, and I doubt it's related to this issue, but I have a new one sitting on the kitchen table waiting for time and temperature to create an installation window.

[frag]

I hate all those little plastic wiring coonnectors these days. No two seem to work the same way. I'm always trying to figure out the secret code for pushing, and or pulling, exactly the right combination of tiny fragile little snaps to get them apart. I only actually break something about 20% of the time though.

 

Glad to know I'm not alone

[davebugs]

I wish the whole auto industry would standardize the style - and thus release method of ALL wiring clips.

 

Would make too much sense I know. But its almost always the most frustrating part of every project that I do.

 

If they were all the same even challanged folks like myself would catch on eventually.

 

BTW I swapped coil packs (I'm having this same issue on a 97 OB) and did 150 mile trip yesterday in the rain with no codes. Way too soon to say that it solved the problem. I did have trouble with the clip on the coil packs though.

 

Dave

[subsince77]

I did have trouble with the clip on the coil packs though.

Dave

 

Funny, last fall when I changed coil packs, I had no trouble with the clips. But yesterday, when i just wanted to "cycle" all these connections to clean them a little. I couldn't get that one loose. I finally left it alone for fear of breaking it in the cold.

 

I hope the new pack solves your problem.

[subsince77]

So the saga continues. I just returned from a couple of days away. We drove about 380 miles total. On the first day the car threw the #3 and #4 code within 1/2 hour on the highway. I just ignored it until we got to our destination, about 150 miles. Car ran perfectly, always does. If it wasn't for the light I would never know there was a problem. I did think of one soulution to this whole thing, a 3/4"X3/4" piece of electircal tape to cover the check engine light would eliminate this problem.

 

The next day I removed and cleaned all the injector connections, and the coil pack connector. I rechecked all the plugs and plug wires. The car ran around town for an hour or so with no CEL. Then we went out to a trailhead to ride. That meant 15 minutes of 75 MPH, then a short section of washboard dirt road. As soon as we hit the washboard I got the CEL. It blinked two separate times on a 5 minute stretch of rough road, again #3, and #4. I cleared the codes, drove back out the same road, no CEL. Drove 20 Minutes of highway, no CEL.

 

Last night drove 15 minutes of 75 MPH again and onto another dirt road to another trailhead, again CEL after about 1 minute of rough road. Now I'm thinking this isn't plugs or wires, but bad/loose grounds or something.

 

This morning, I cleared the codes, drove to the same trailhead, no CEL. After our Mtn. Bike ride, and before heading home, I removed and cleaned both the main engine ground and the grounds for the coil pack. Neither was filthy, but both had the light grey corrosion you get on any connection over time. I scraped them and used contact cleaner on them. I was not in a place to mess with the ECM, so I didn't touch that yet. We started home and the CEL came on in about 1 minute, this time it was #4 only, no #3.

 

We had to stop at a store about 15 minutes down the road, so at that point I cleared the code. Then we drove 150 miles home with no CEl at all. Not sure what to make of that last part, maybe we were just lucky. My plan is still to take the glove box out and clean all the connections on the ECM, and check any grounds. I don't know how the ECM is grounded, If anyone knows, that would be helpful.

 

Maybe what I already did made a difference. But given the tendency for the CEl to light as soon as I leave the pavement, I am thinking either connection problems or a problem in the ECM itself. I know this can be improper valve setting or a compression issue, but I have a hard time thinking it is either of those on a CCR rebuild with 1300 miles on it. BTW, the engine runs great! Any thoughts?

[WoodsWagon]

I'd be tempted to approach this problem from the opposite direction, which is how the ECM detects cylinder misfire. The ECM uses the crank sensor wave form to determine if the crank speed is responding as expected for each cylinder firing. Sudden load changes on the drivetrain can be seen as cylinder misfires, I believe GM had a problem with their cylinder misfire detection being too sensitive and setting codes while offroading because of this. With Subaru's, I believe that you have to have a certain # of consecutive misfires for it to set a code. If the crank sensor was getting marginal, or possibly corrosion was holding it from being fully seated in its hole, making a larger air gap, it could send a misleading signal to the ECM. The other part to consider is the reluctor wheel teeth that are on the timing belt pully on the crank. If these are bent/chipped/damaged, the crank sensor could be putting out a weak signal. Or, the crank sensor signal could be encountering resistance in the wiring between it and the ECM, in the form of a corroded connector or whatnot.

 

So there's some extra food for thought.

[Cougar]

I think you have a good idea about cleaning the ground connections. I think there has been others that had misfire codes and cleaning the grounds seemed to cure the trouble.

 

I would also clean the battery connections if you haven't done that already and the chassis ground connections near the battery and any others you see in the engine compartment using a brass wire brush. It may not help but it sure won't hurt at all to do this.

 

Some folks have added extra grounding leads and reported better perfomance with them. To me, just making sure the factory grounding system is in good shape should be enough but others may disagree with that.

[subsince77]

I would also clean the battery connections if you haven't done that already and the chassis ground connections near the battery and any others you see in the engine compartment using a brass wire brush. It may not help but it sure won't hurt at all to do this.

 

Battery Connections are clean. Chassis connections, I did the engine connection from the negative post to the engine. I'm trying to picture the ones you are talking about. Are they small braided wire type connection from the engine to the chassis or something? Seems I've seen that, but maybe that was on my truck.

[FSRBIKER]

I am trying to remember correctly but isn't the ECM under the passenger side carpet under a screwed on access door. I might be thinking of another car but I am pretty sure this is where mine is in my 97 Outback and there was no need to remove the glove box.

[subsince77]

I am trying to remember correctly but isn't the ECM under the passenger side carpet under a screwed on access door. I might be thinking of another car but I am pretty sure this is where mine is in my 97 Outback and there was no need to remove the glove box.

 

That's where I was thinking it was too. My Haynes said it is behind the glove compartment. Maybe I read it wrong, or maybe it is wrong, not the first time. I'll check it again. Anyone know for sure where the ECM is?

[FSRBIKER]

That's where I was thinking it was too. My Haynes said it is behind the glove compartment. Maybe I read it wrong, or maybe it is wrong, not the first time. I'll check it again. Anyone know for sure where the ECM is?

 

I have the Haynes book as well, that's where I remember reading about taking the glove box out first....which I did once. No need it's under the carpet.

[subsince77]

I have the Haynes book as well, that's where I remember reading about taking the glove box out first....which I did once. No need it's under the carpet.

 

Thanks, always good to save some frustration. I'm going to do that part, cleaning the ECM connections, then see what happens from there. Today's 150 miles with no problems was encouraging, I'm just not convinced. The Haynes actually suggests tapping sharply on the ECM while the engine is running to see if it throws a code or makes the engine stumble - now that is my level of wrenching!

[Cougar]

It's amazing how many times you can find trouble by tapping around on things. I like to use the handle of a rubber coated screwdriver as my tool of choice.

 

The braided ground wires should be on the chassis, bonding things to ground.

[subsince77]

I removed and cleaned the ECM connections - it is under the passenger floorboards. I also cleaned all the places where it bolts in because I guess those are the grounds. I ran the engine and tapped and wiggled the ECU for a while. I couldn't cause any problems that way. I see two chassis grounds toward the front of the engine compartment. I will get those in a couple of days.

 

I found out today that this engine did not come with NGK lasers as I thought. It came with some Nippon Denso plugs. I have ordered a set of NGK's that will be in tomorrow. When I pull everything out to do the plugs, it will be easy to get at those two chassis grounds as well. A bit bummed that I didn't get NGK's with the engine, I thought that was a given, but Emily says sometimes they use Densos. Unfortunately it is now supposed to snow for 5 days, so I won't get any of that done for a while unless I find a nice dry garage to work in. Life goes on.

[Cougar]

I think those are very good plugs also so I doubt it will change this problem, but perhaps it might. We can hope it will at least.

[subsince77]

I don't think CCR would use a bad plug, it just seems to be the consensus that this engine is very particular about both plugs and wires. If I were not having this problem, I'd run them until they died. But since I am, and I have been working from the premise that the plug and wire bases were covered, I'm going to change them.

[johnceggleston]

I don't think CCR would use a bad plug, it just seems to be the consensus that this engine is very particular about both plugs and wires. If I were not having this problem, I'd run them until they died. But since I am, and I have been working from the premise that the plug and wire bases were covered, I'm going to change them.

i commend you for your attitude, it sound like you are in this for the long haul. but save the old plugs for the next time you need them, i don't think they go bad like milk.

[subsince77]

OK, NGK lasers installed, 2 chassis grounds "cleaned." I couldn't get to them as well as I would like, but I scraped what I could and blasted with contact cleaner. In the process I found another main ground that runs from the neg. bat. post to the front of the engine compartment above the left headlight. Don't know how I missed that one before.

 

On putting the air filter box back together I discovered that I may have never really had that in right before. I didn't realize that the bottom of that cover has three plastic slots that fit over three plastic tabs to lock down the bottom side of the box. Oops - never saw them before. Probably why I only got 200k out of my first engine. Anyway, it will always be done right in the future.

 

Can't think of anything else to do on this, other than check the primary and secondary resistance on the coil pack. It's new so I haven't done that. It will probably be a week or two before another road trip to really see if I get another CEL. Here's hoping.

 

Oh, and yes I packed the Denso plugs up for a future need.

[davebugs]

FYI my 97 OBW threw the same codes today. I had swapped coil packs and didn't have the problem for a little while. Not a known good coil pack but the car is running. Seems to me if Diampond coil packs were experiencing this as a common problem we'd see it show up here at USMB.

 

It happened decelerating from probably 70 mph on the interstate - same as last time. I cleared the code, and later on a 2 lane road ran the RPM's up to 4k in 4th gear and decelerated to try and get the code again but didn't.

 

I haven't messed with the NSS or grounds yet. So I'm interested in following your progress. Either that or I can't go over about 70mph and allow the car to decellerate.

 

Dave

[subsince77]

FYI my 97 OBW threw the same codes today. I had swapped coil packs and didn't have the problem for a little while. Not a known good coil pack but the car is running. Seems to me if Diampond coil packs were experiencing this as a common problem we'd see it show up here at USMB.

 

It happened decelerating from probably 70 mph on the interstate - same as last time. I cleared the code, and later on a 2 lane road ran the RPM's up to 4k in 4th gear and decelerated to try and get the code again but didn't.

 

I haven't messed with the NSS or grounds yet. So I'm interested in following your progress. Either that or I can't go over about 70mph and allow the car to decellerate.

 

Dave

My coil pack is a diamond of the exact same model number as the original. I bought it at Napa, but I have no reason to think there should be a problem with it. Given the problems I am having, there was probably nothing wrong with my old coil pack. Too bad I tossed it.

 

I did measure the primary and secondary resistance across this pack, and it tests out fine. About 24 hours after my last post, new plugs etc., I got a CEL. It was #4 only, and it happened upon starting the car - that's a first. I have never been able to make the car throw the code by trying to duplicate the circumstances of the last time it happens - like your decelerating thing.

 

On the one hand, I want to find this problem and fix it, on the other hand, I don't feel that it makes the car unreliable in any way. I am about ready to throw a ECM at it, a used one, but I can't say I think that will fix it. It is possible I suppose that the valves were not set right, but I really doubt it. I think if that were true, the codes would be more consistent. As it is I can drive for a week around town, or 150 miles on the highway one time with no codes. Then it will just suddenly throw one. I still think putting electrical tape over the CEL may be the ultimate solution.

 

Seriously, next time it happens, I'm going to read it but not clear it. Then see how long it takes to clear itself. Last time it cleared itself within just a few starts.

 

If you find any new leads, please post them.

[aircraft engineer]

The sister's Imp threw a #4 code - it turned out that when the "old" plugs were replaced (gap was 080 - yeah, thats EIGHTY - about 2X recommended) the code went away and hasn't returned. Since 3 & 4 are connected across the same coil pack, I was surprised that it didn't show 3 AND 4, just 4.

 

Hers threw the code decelling at 35 down a long hill in 3rd (I think 3rd anyway)

 

I'm not familiar with these exact electronics but IF you have a bonding/grounding problem you will see it (clean it all and retighten the connections to ground - lots of flaky things happen when the system isn't grounded correctly)