sneaky98gt

Alright everyone. I need some help troubleshooting a problem on my car that arose today. I'm fresh out of ideas.

Car:
2001 Honda Civic LX with the D17A1, MT, 340k miles

Symptoms:
-Very noticeable miss at idle, both under the hood, in the car (vibration), and through the exhaust (psh-psh-psh-psh sound).
-The miss appears to clear up after about 1500 rpm.
-There is noticeable power lost from 1500 to about 4000 rpm, but still quite driveable.
-I can't perceive any power loss above 4000 rpm to redline. It actually runs quite fine up there.
-P0302 code, "Cylinder 2 Misfire Detected", is the ONLY code that has shown up, and it always comes very soon after I reset it. The CEL stays lit while driving, and flashes if I stop and idle for more than 15 seconds or so.
-The car went from 100% perfect when I left home this morning, to having all these symptoms by the time I completed my 50 mile drive to work. Somewhere along the way, whatever happened happened in a fairly short amount of time, and it happened on the highway at 70 mph with the cruise control set (i.e. I wasn't doing burnouts or anything).


Things I've Done to Troubleshoot So Far:
-Swapped the coils around. No change. Still have the P0302.
-Swapped the spark plugs around. No change. Still have the P0302. I also should note that after nearly 200 miles of driving like this, the spark plug looked identical to the other three. It wasn't fouled up or black or anything like that.
-Swapped the fuel injector for a spare one I have that I know is good. No change. Still have the P0302.
-Checked the valve lash. Well within spec. The valves also moved freely.
-Did a compression test across all 4 cylinders. Got 155, 110, 110, 155 (4, 3, 2, 1).

Before getting bent out of shape about the numbers being low, let me say this: I check the compression roughly every 10k miles when I change the oil, and those numbers are fairly similar to what I've seen for 50k miles now. I don't worry as much about exactly what the numbers are as much as I just look at the trends. #2 has always been the lowest, and #1 and #4 have always been about the same as the highest. But what's a little different here is that #3 is normally about halfway between #1/#4 and #2, but now it's showing up as about the same as #2. Significant? Not sure.

A little background: over a year and 50k miles ago, I had a burnt exhaust valve with one very significant similar symptom: the psh-psh-psh out of the exhaust. This sounds virtually identical to what it sounded like then. This immediately led me to believe that it was again a burnt valve this time.

BUT, that compression test says different. 110 psi is pretty good in my specific instance, and tells me pretty confidently that I don't have a burnt valve making that kind of compression, especially when that's the same 110 psi I've been getting before this problem arose up. Also, last year when I had the burnt valve, that cylinder had almost no compression at all, barely 20 psi on a hot motor. And there was a much bigger loss in power then, too, especially higher in the rpm.

However, #2 and #3 having the almost same numbers certainly points towards a head gasket issue. That being said, there is no oil in the water, no water in the oil, no overflowing coolant tanks, and no overheating. Plus, I would think that if they were bleeding together, I'd be getting misfires for both cylinders, and be able to audibly confirm that two cylinders were misfiring. That's certainly not the case here.

So, all that being said, tell me what you've got. Thoughts, ideas, where to go from here. Granted it's well past my bed time, but I'm seriously drawing a blank.

About the only thing I've got is there being some sort of electrical issue upstream of the fuel injector, something that's causing it to not fire, or to only fire higher in the rpm. I thought about the same scenario with the spark side of things, but if that were the case, the spark plug would certainly be fouled, and it's definitely not.

Other than that (which I have no idea how to test), I'm stumped. Thought about doing a leak down test, but I'm not sure how that would necessarily be helpful.

Fire away.

Thanks!

-Will

sneaky98gt

Something new to add this morning.

After I got on the highway on the way to work this morning, I cleared the code to see if it came back at higher rpm. It did not for the entire 50 mile interstate trip to work. But as soon as I stopped at the light at the end of the exit ramp, and idled the car, it was flashing within 15 seconds with the same P0302 code.

Is this significant? Heck if I know. I could still tell that it was down on power. I've no idea what the threshold is for it to throw the code, but apparently I'm below that once up in the rpm decently.

sneaky98gt

Ok, I'm on my phone now so I'll keep it short and post more details later.

We've narrowed it down to the coil not getting a signal. Unplugging the connector resulted in no change in the cars idle, indicating no spark on that cylinder. However, swapping coils and plugs around didn't change that it was #2 that was giving the problem. So it must somehow not be getting a signal.

The ground wire seems to have continuity with the ground, and the hot wire seems to be hot. But we don't know how to troubleshoot or check the signal wire.

Does anyone have any advice on this? Or have a wiring diagram we can use to trace it back?

Thanks.

sneaky98gt

Ok, back home after troubleshooting for several hours with some friends of mine. We think we have it nailed down what's wrong (although can't test for sure apparently).

Cliff notes version: we think it's a faulty ECU not sending the correct signal to the #2 coil pack.


Full version.

So, picking up where I left off last night, we did a leak-down test on it. Well...faux leak down test. The gauges weren't working properly, but I wasn't looking for that anyway. I just wanted to put air pressure in the chamber and see if we could hear any escaping past the valves (re: burnt valves) or into the adjacent cylinders (re: blown headgasket) or into the coolant via air bubbles in fill cap (re: blown headgasket). We got none of the above. With 75 psi air pressure in the #2 cylinder, we didn't get any peep at all through the exhaust, intake, or adjacent cylinders. No air bubbles in the coolant either. We could hear some air moving if we removed the oil fill cap, meaning it was leaking past the rings. But honestly, I knew this, and don't really care. The volume/loudness of the sound was indistinguishable from #2 with 110 psi on the compression test, and #1 with 155 psi on the compression test.

This is good news. There isn't major, mechanical, catastrophic damage inside the cylinder. That being said, we weren't any closer to actually figuring out what was wrong.

Next up was the wiring feeding the fuel injector. I picked up some noid lights from O'Reilly, and stuck one on the connector. It flashed perfectly. Was dimmer at idle and free-rev decel, and brighter at higher speed and free-rev accel. To confirm that it wasn't out of the ordinary, we did the same test on the #1 injector right next to it. Same results. On top of all this, we stuck a solid 3/8" drive extension down onto the injector, and put an ear up to the end of the extension. It could clearly be heard clicking. So power getting to the injector is not an issue.

Could it be a clogged injector? I reasoned that this was incredibly unlikely seeing as how I changed it out yesterday with one that I knew worked fine in the past. But, we wanted to be thorough and not overlook or underestimate the chances of something. So we cranked the car, and let it idle (with the miss plainly audible) for a good 5 minutes or so. Stopped the car, pulled the plug, and looked down in the cylinder onto the top of the piston. It was clearly wet with gas (it wasn't earlier when we first pulled the plug after I drove it to my friends' house), and sticking a small tube down into the cylinder, you could clearly smell gas. So there was no doubt that the injector is actually getting fuel into the cylinder.

This told us two things. The first, obviously, is that the spark is the problem. We confirmed this by unplugging the coil pack with the car running. Absolutely zero difference in how it ran. Plugged in, unplugged, plugged in, unplugged, etc. No difference whatsoever. Problem is with spark.

The second thing it told us, however, is where things start to fall in place. The fact that the piston top was dry when we first pulled the plug right after I drove it for 25 miles, but it was wet after letting it idle with the miss for 5 minutes, suggested that the spark was actually intermittent, and not just faulty altogether, all the time. This fits with my observation that under some conditions while driving it, it feels 100% perfectly fine, and in other conditions, it's got significant power loss.

So, we tried to somewhat replicate those different conditions sitting there in the driveway via free-revving it at different rpm. At idle and up to 1200 rpm or so, the miss was obvious, clearly audible everywhere. I said in the OP that it seemed to clear up by 1500 rpm. Well, when we stood at the tail pipe with someone else revving it to that rpm, you could still clearly hear the miss. It just wasn't really noticeable in the driver's seat. So revved to 2500 rpm, still could hear the miss at the tail pipe. Revved to 3500, still could hear it. Revved to 4500, still could hear it. And then, at 4500 rpm, just like that, it went away. Completely 100% smooth with absolutely no miss at all. Then, 10 seconds later, it came right back. Could hear it plain as day. Then, some time later, it'd go right back away. It sat there and did this as long as we held our foot on the gas, at seemingly random intervals.

Now we're getting somewhere. For some reason or another, that coil is turning on and off. (I think I mentioned it somewhere already, but we've swapped coils and plugs all around on the cylinders, and the problem has stayed with #2). So, we check the ground wire on the connector. Nice clean ground. We check the hot wire on the connector. Nice 13.5 volts. And for the signal wire.....yea, that's where my last post came from.

Thankfully, those friends I mentioned earlier, are friends with lots of resources. They were able to pull up a full, real-deal version of the Honda tech troubleshooting guide, the ones that the dealership uses. Pulled up the module for a misfire, and went through the whole thing. We literally had already done 98% of the steps, right up to the last two, which were exactly what we were looking for / asking for: how to test the blue/red signal wire on the #2 coil.

First up, check for continuity to chassis ground (i.e. see if it's shorted out). Checked it at the connector, and at the ECU. Nada. Not shorted out. And lastly, check for continuity between pin 3 at the connector, and pin 29 at the ECU (i.e. see if the wire is broken). All good there, good continuity through the wire.

And the last step in the troubleshooter: update the ECU with the most recent software, or swap the ECU for a known good unit.

And that's where we've left it. After all the other testing we've done, all of which indicated good parts, and combined with the 'works-at-some-conditions, but-doesn't-work-at-others" nature of this problem, a software bug in the ECU makes sense.

I've got to do some research to see what ECUs are compatible with my car. Once I get one, and get it swapped out, I'll definitely post up the results. The only other thing I wish we could do before I spend any money on anything, is to actually test if the spark plug is firing. I know that literally every sign points to the fact that it isn't, but I still wish there was a way to 100% verify that it wasn't before spending money on a computer.

So there's where we stand at the moment. Post up any comments or suggestions you may have. I'm not going to pretend to know a ton about these cars, or to have kept up with the forums over time or anything like that, but in the searching around I've done, I have never seen or heard of this symptom (single cylinder misfire) being caused by a faulty ECU. So I'll definitely make sure to keep this updated for any future readers who may have a similar problem.

ezone

Not sure I read all the story here......I picked this out of the list:

In this quote it seems you only found a lack of cylinder contribution at idle.
That alone does not verify spark or no spark, so do not assume.
I think you should verify spark visually, same way one would if it had plug wires and a distributor.
Pull the coil out of its hole and hook up the connector..... hold a screwdriver against a ground and get it close to the end of the coil....etc.






ECM/PCM won't randomly drop spark (coil control) over a software or programming issue.
If it's actually losing coil control, it would be an internal problem.


If anyone in your group has a labscope, you could watch coil control on a screen.

I wonder if a dwell meter would show anything, at least you could watch if it comes and goes --if it does read. (I've never tried it on a 3-wire coil system myself)

sneaky98gt

That's still my gut feeling, and why I mentioned at the bottom that I'd like to absolutely confirm, without a doubt, that it actually isn't sparking.

That being said, the combination of knowing we have fuel in the cylinder, knowing we have a hot wire to the coil, knowing we have a ground wire to the coil (verified those while it was running), and the problem staying at the same cylinder with different coils and spark plugs, doesn't leave us with anything else to think. I'm an ME guy, not an EE guy, so I'll admit I don't follow along with electrical stuff all that great, but I can't think of a reason why we'd have a 100% completely dead cylinder that has air, fuel, and sufficient compression unless there was no spark, and the only reason there'd be no spark is that it's not getting the signal to fire.

Random question that I think I already know the answer to: do these cars have a waste spark system?

We actually thought of that, but more than one of us had heard that doing that to a modern COP style coil can fairly easily damage it. Something about the current shooting way up and burning up the coil. Any insight there?

Maybe actually pull the plug out and see if it's firing?

There's the thing: I don't think it's random.

Here's a tidbit of data that I've left out that could be important: when driving the car, there is a CLEAR transition from running on significantly reduced power, to running 100% perfectly fine. For example, from a slow roll in 2nd gear, starting at maybe 1500 rpm. At first, it has an obvious problem, with it bucking somewhat and generally not really accelerating at all. Start creeping up on the throttle, and it only gets marginally better. Still way down on power. Continue going up on the throttle, 25%, 50%, now we're up near 2500 rpm, 75% throttle. And BOOM, it's like VTEC kicked in. Power comes on 100%, car starts pulling good, and runs out all the way to 7000 rpm with no problem.

Here's the important thing: this is repeatable. It acts like this EVERY time, every gear. So while I don't exactly know where that changeover point is, or what combination of conditions causes it (load, rpm, vehicle speed, etc.), it does behave about the same each time, signalling to me that it isn't random. When doing the free revs, it never once cleared up under 3500 rpm. Over that, I said random earlier, and it did seem that way to us, but a slight tweek in throttle position could have easily triggered it.

I don't have any idea how the tune works in this Honda, but I do have experience tuning Fords. And what happens a lot of time, is that a vehicle will operate on one table under a certain set of conditions, and will then jump to another table if certain conditions change. Open loop versus closed loop spark timing, based on throttle position, for example.

It seems a logical explanation to me that some table that the car operates off of at idle and low load has gotten corrupted, but the table at higher rpm and/or load is still good, and the spark command jumping between them is causing the issue to come and go. For example. I have no clue how it actually works.

I do indeed have a scope at work that I could borrow. Excuse my electrical ignorance again, but how exactly would you do that? What would you hook it up to? That blue/red signal wire?

ezone

Most electrical faults are mechanical in nature
(not electronic faults)


Valve leakage, for example, can cause enough compression loss to not fire the cylinder at low RPM, yet can fire at higher RPM. There is spark present under both conditions.

You DO have low compression on the problem cylinder.

A vacuum leak on a single runner of the intake manifold can disrupt enough to cause a miss at idle, yet not miss at higher RPM.
Same for a vacuum leak around an injector, it can disrupt fuel delivery to the cylinder.

No.
Waste spark would be set up as 2 plugs fired on each coil.

I create my own air gap to test available coil voltage regularly.
I expect to have 30kV+ available from a good coil, that's a roughly 1cm air gap to ground.
You've already swapped them around and it works on a different cylinder? Plug works.

The 0.44" gap of a plug is not sufficient to judge available coil voltage, it only takes about 5000 volts to ionize that gap.

It takes about 30kV for spark to jump an air gap of 1cm, according to basic rule of thumb (and internet charts).
I like to see how large the gap can be and still have spark jump across, most will reach about 3/4" or so.

If you confirm the coil has sufficient voltage available, then you may want to "watch" spark as the engine runs.
You could connect a standard spark plug wire to the coil and plug, connect an ignition timing light to it and watch the flashing (My preference because I have this and it's quick to do).
You could get an adjustable spark checker (commonly used to check small engines) and connect inline with the fabricated plug wire, set to a very small gap and watch for spark to quit.

The trigger wire from the PCM, at the coil or at the PCM whichever is convenient.
Not power, not ground. You already confirmed those, right?

I am too tired to finish and I have to work in the AM.

sneaky98gt

It does have low-ish compression, and my perception could be wrong, but I think 110 psi on a cold engine should be more than enough for combustion. I looked at my notes, and the last time I checked the compression was about 15k miles ago, and that cylinder had 125 psi. I could easily be wrong, but I just don't think a 10% reduction is enough to go from 100% working and good fire, to 100% not working and no fire. It probably varies by that much or more depending on external factors (warm vs. cold motor, weather, etc.). And the way it switched, pretty much instantly, from missing to not missing on the free rev, really has me convinced that it's something electrical, and not mechanical.

Not to mention the fact that the cylinder right next to it has the exact same compression, and is working 100%.

Then again, I feel like we're getting into some crazy territory now, so I suppose anything is possible, regardless of how unlikely.

I didn't think so. My Mustang actually has waste spark on a COP system, but that's only because I converted it to COP from a standard coil pack setup.

The thought crossed my mind that if it was, I could splice into the signal wire from #3 and get fire that way.

One of the machinists I work with is familiar with these spark gap testers. I'm going to run out to O'Reilly today at lunch and get one to try this afternoon.

Yes, power and ground at the plug are good.

We have a scope at work that I can use. If it turns out that I don't have a good spark from the gap tester, I'll probably put the scope on it and see exactly what we do have, compared to a cylinder next to it.

I feel ya. Was up till 1 Sunday and Monday working on the Mustang, and up till 1 Tuesday and last night working on this car. Waking up at 6 has been getting tougher all week long, lol.

But I do appreciate the help.

ezone

If I were using my equipment and method for the test and came up with the numbers you have, that would be a big problem.
I fully expect these engines to reach 180+ PSI when I do it.
Different testers may vary readings and different people do the test differently, but consistency is important.
You have far too much variation between cylinders. Misfire codes are expected.



Did you come up with a way to watch spark while running? Any results?

sneaky98gt

I agree about the exact results, but I'm not confident in exactly how accurate the numbers are. And honestly, not too concerned if it's not causing a problem. Prior to this, it was still running great and regularly returning 37+ mpg on average. So for a 340k mile daily that I paid less than $2k for over 4 years ago, it definitely doesn't owe me anything, and I don't intend to spend a ton of money on it going forward to keep it running.

I did get around to testing it, and it wasn't what I was hoping. Here's a video:


That's about a .750" gap, or almost 2 cm. According to the graduations on the tester, is about 30-35 kv. I went up to almost 1", and it would erratically gap it. Tried the coil next to it, and it was about the same, so I figured that wasn't abnormal.

I also pulled plug, stuck it in the coil, grounded it, and cranked the car. Wow, what a racket does that make! Haha. But I caught a quick look at the plug during this, and it was definitely firing.

I'm stumped at this point.

Here's a little more info/data as of yesterday.

When I pulled the codes last night, it now had misfire codes on all 4 cylinders and a random misfire code (P0300 I think). Cleared them, drove it a while, and it only gave back the #2 misfire code after about 25 miles of driving, so it seems it only throws the other ones after a good while of driving.

Filled up with gas yesterday afternoon, which is the first complete tank I've run through with this problem. It got a little over 34 mpg, which is a little less than average, but well within the range that I get regularly. And considering how much idling time it's had while troubleshooting it, and how much harder I've driven it to overcome the low rpm stumbling, I'd say that 34 mpg is right on the money. So it doesn't seem to have significantly impacted the fuel mileage.

And one more thing, which I still think is important. Yesterday afternoon, I was driving through the rolling hills of the NC Piedmont area, and I noticed something. Remember how I said we could tell it switching back and forth between missing and not missing during the free revving? I could also clearly feel it doing that on long, steady hill climbs in 5th gear with the cruise set at about 3000 rpm. You know how you can feel the power loss/gain when the A/C kicks on and off? It felt just like that. And almost everything was fairly constant when it was doing it. Vehicle speed and rpm were about constant because of the cruise control, and load was about constant from the long hill. And it was clearly cycling back and forth, although the timing of it seemed random to me.

This is where I still don't think it's mechanical. If it were not combusting from lack of compression, there's no way it would be such an instant, analog, on/off style of that cylinder cutting in and out. I'm out of ideas for what could be causing it, but I can't currently believe it's compression related.

Two more questions.

First, is there some freak-of-nature possibility that the spark plug could be shorting out to the head? And not just the head, but that specific hole in the head (since swapping the plugs didn't change anything). I drove over 250 miles between all the testing we did two nights ago, and last night. It could be purely coincidental, or it could have already been this way, but that #2 plug had a tan/brown discoloration on one side of the metal part of it between the hex and washer. I have no idea how it could happen in the same hole with different plugs, but the thought did cross my mind of the spark somehow shorting to the head versus actually sparking the gap. None of the other 3 had this mark, but of the other 3 that were run in the #2 cylinder, none were for more than maybe 75-100 miles. So, I put one without a mark back in #2, and I'll check it later this weekend.

And secondly, how does the ECU detect a misfire? I'm curious if there is a good chance of there actually being misfires in the other cylinders, or if the methodology isn't quite smart enough to always distinguish where it's coming from and it's just getting confused from the severe misfire in #2.

Thanks again.

ezone

No, a plug problem would move wherever you relocated the plug to.

A breach of secondary insulation, such as a pinhole in the side of the coil bottom boot or tube should cause misfire to occur as load increases and be very repeatable.

But that brings up this thought and pic, which can cause randomness in the firing:




The bad cylinder, does the plug show anything like this? DID the plug show this? If any did, the bottom boot(s) on the coil tube will have an identical track mark...and if both are not replaced at the same time, one can cause the other in a very short time.

How about a loose terminal in...the injector connector or coil connector? (or any connections along their entire circuits)

A female terminal that has lost its tension won't grip the male blade tightly and vibrations can cause it to intermittently make and break contact.

It could certainly explain the on/off effect of a single cylinder that quits working-- as you described, and also could explain the randomness of the events.

Google...'pin drag test'. You obtain a correctly sized male terminal, insert it into the female terminal and note the amount of drag as you slide it in and out. Compare to others of the same size that have no problems.

The PCM measures variations in crankshaft speed, every firing event, using the crankshaft position sensor (CKP).
There are criteria that must be reached before the PCM flags a code or flashes the CEL, like number of misfire events per number of crankshaft revolutions and certain operating conditions.
I'd consider it to be dead nuts accurate on the first single cylinder misfire code recorded (in the case of only a single cylinder dead, see below).

I have seen cases where other cylinder misfire codes can set due to driving around with the first one misfiring, even though they are fine. Hondas V6 engines will do this if driven around with a 'dead cylinder' misfire. I record findings, then erase the data and start over, a dead cylinder will reset a fault code pretty dang quick.

Wankenstein

This is some serious geek speak amongst real mechanics

ezone

Dude is an engineer.
Even if automotive isn't his area of expertise, I think he should be able to grasp the concepts or research a bit to figure out what I was talking about....

emich

You have an injector connector problem and bent valve/s

sneaky98gt

So I put a new plug in it, as well as a different coil that I had laying around. Also checked all the terminal connections, and they were all good. Swapped the #1 and #2 fuel injectors, too. None of that affected anything. Bummer.

Yea, I'd like to think I have a fairly good grasp of vehicle and ICE dynamics in general. And when I don't, Google is pretty darn helpful if you're willing to spend some time reading.

That being said, I think I've finally figured out what's going on, although I feel like an idiot for taking this long to do it.

In summary, ezone, I think you were right about it being a compression problem.

So all this time, I've been running under the assumption that 110 psi of compression is sufficient for good combustion. I still maintain that it is. The issue lies in how that 110 psi was obtained. When doing a compression test, I've always heard to open the throttle all the way, so as to allow it to breath unrestricted through the intake valve. And that's how I've always done it.

Well, there's the problem. When the car is idling, or at low load, the throttle is almost entirely closed and there's a large vacuum in the manifold. This means the cylinder isn't pulling nearly as much air in on the intake stroke. And if it isn't pulling as much air in, then, obviously, there's not as much air to compress. Therefore, the working compression at idle under vacuum is probably MUCH less than a compression test done with an open throttle would indicate. If there is indeed a bad exhaust valve, then this would be even more exaggerated at low rpm since the piston speed is slower, thus allowing the air/fuel mixture more time to escape as the piston compresses it.

This explains the symptoms perfectly. The actual working compression at idle and low speed / low load conditions is low enough that the air/fuel mixture simply will not ignite at all, thus explaining the completely dead cylinder at idle. Then, as the throttle is opened, it gets enough air in to build the compression needed to actually ignite the mixture. This explains the clear uptick in power that the car makes as I creep up on the throttle. And lastly, once we're at a load where it'll actually burn, the only power loss is through the little bit of gases that make it past the leak. Assuming it's an exhaust valve (pretty sure it is), if it's only a small-ish leak, I don't think the total power loss would be all that much as long as the mixture is still igniting. This explains the huge difference in perceived power from this go around compared to when it had a burnt valve over a year ago. That time, the valve was really bad, and was probably leaking enough by to not even allow ignition at higher loads and rpm, and thus power loss across the board.

This really does explain everything, except not hearing anything through the exhaust pipe on the leakdown test. Maybe it was bleeding into other cylinders with exhaust valves open at the same time? Idk. I wish I could do the compression test with the throttle closed, but I don't think I'd get meaningful results. I think it'd either 'steal' air from other cylinders, or if the really is a bad exhaust valve, it'd pull air in through that hole. So I don't really know any other way to test it.

Anyways, I'm off to the junkyard tomorrow to pull a head off a car that looks to be in decent shape. Probably try and get it swapped out sometime next week. I'll be sure to come back and let everyone know what winds up happening.

ezone

Well...The engine is just a large pump.. The compression test is to check if that pump is working right.

I'm pretty sure I mentioned what compression test numbers I expect from one of these engines somewhere in the thread.

A basic compression test IS supposed to be with the throttle blocked wide open, you do want a 'full gulp' of air for the testing, anything less causes incorrect test results.


If you have all of the plugs out for the test that should allow plenty of air through even if the throttle is closed, the readings will be the same as if the throttle were wide open (on an engine this small).
If you only removed plugs one at a time for the compression test---that would be the wrong way---and you would want the throttle blocked wide open.

For a leakdown test you want to be ready to reinstall unused plugs a few turns just to seal up those holes, as this can and will affect your results--any air coming out through those plug holes had to come from the pressurized cylinder, that's a leak that needs to be found.

For a leakdown test...I've gone so far as to strap blue gloves over the tailpipe to 'see' (inflate the glove) leakage that is too small to feel.
On the intake, I might leave throttle closed then remove a vacuum hose to check for air movement (smaller hole = velocity you can almost feel)



Running compression at low RPM is usually much lower than what you read during a cranking compression test. The pressures go high when the throttle is snapped open, and combustion pressures are much higher yet.


Good luck!

sneaky98gt

So, it's been almost a year since I posted this thread, and I actually got the car fixed not long after this thread, and it's even been sold since then. But I forgot to come back and update this thread until I was going through old pictures on my phone. Hopefully it'll help someone out one day.

Long story short, 2001 Honda Civic LX had a misfire on cylinder 2. Only occurred at low load / low rpm. Ran great under load and upper rpm. EVERYTHING we checked seemed ok. Spark was good, fuel was good, compression was good enough (based on previous readings), and leakdown didn't indicate anything wrong. We were completely stumped, and could only come to the conclusion of a very small leak through a slightly burned / tweaked exhaust valve. So the thread ended with me off to get a head from a junkyard.


Welllllll......the night before I was to head to the junkyard, we went ahead and pulled the head off, and just look at what we found:



Huge, gaping hole between 2 & 3 in the headgasket.

Ironically, in my very first post in the thread, I made this comment:

Apparently, this was exactly the case.

That being said, #3 never ever ever gave any misfires or issues. It was always #2. How that can be the case when both cylinders are exactly opposite in the combustion cycle, I have no idea. Still don't, to this day, nor does anyone I know.

As for it running well once under load or well up in the rpm range, I think my theory towards the bottom of the thread is what was going on:

At low load / rpm, the actual volume of air pulled in was fairly small, and the leak past the headgasket brought the compression well below the threshold for combustion, and voila: #2 miss. As the load / rpm goes up, more air is being brought in (plus everything is getting a little hotter, closing up the gap a little bit), and even if some is leaking past, there's still enough there for good compression and good combustion.

Though that still doesn't explain why #2 was 100% problematic and #3 was 100% perfect. Since the 2 cylinders are exactly opposite each other in the combustion cycle, I'd think both would suffer the same symptoms. For reasons still unknown, that obviously wasn't the case.


So, what was the end result? I checked the head, and it was crazy out-of-flatness tolerances. It was seriously bowed up in the middle. (The block was still fine.) We took it to a machine shop to have decked; they were able to do so, but had to take it almost to the maximum amount allowed by Honda. They wound up taking about .025" off. I did the math, and that would raise the compression by about a point. I figured I'd take the chance and see how it worked out. Worst case, I might have to run mid-grade or premium.

Did a warped head take out the headgasket? Or did the headgasket go out and the heat warp the head? Who knows. I ran it like this for several months and about 10k miles, so I definitely didn't do the head any favors, but it ran pretty darn good other than at idle. Testament to Honda engineering, there.

Put everything back together, and it ran like an absolute champ. The extra compression provided a very noticeable gain in power across the rpm band. Never once heard it ping on 87 octane, either. Another benefit was that it returned fantastic fuel mileage, the best I'd ever gotten in the 4+ years I had the car. After fixing this issue, I don't think I ever got less than 36-37 mpg, and that was driving the snot out of it around town. On my typical 90% highway tanks to/from work, I was averaging about 40 mpg. I got 42-43 on several all-highway tanks, and that was without using any hypermiling techniques. Just set the cruise on about 73 and ride. I was blown away at how well it did, and when considering the nice power bump, too, I think that's how they should have come from the factory. If anyone ever has the head off and has the spare $100 sitting around, have the machine shop take a good .020" off. You won't regret it.

After driving the car for about 30k miles after fixing it, I wound up selling it a few months back. No real reason, other than wanting a little bigger 4-door car. I replaced it with an '03 5-speed 2.4L Accord, so the apple didn't fall far from the tree. It was still running great the day it left.

So I think that about wraps it up. Hope all my pain and suffering troubleshooting this problem will help someone in the future.

-Will

sneaky98gt

If a mod wants to add a note at the top of the 1st post regarding the update I just posted, that'd be great. (So the regulars around here don't spend too much time reading through an almost year-old thread that they don't immediately realize is an almost year-old thread).

Thanks!

ezone

Thanks for the update!
Somewhere early on I agreed with the low compression issue (reply #7)....then somehow we went off on a different tangent....

I saw the new pics before reading your post and thought this must have been overheated to warp the head (they almost always lift in the middle during a severe overheat), then refilled with coolant and driven quite a while (with the warp and continued leakage) in order to 'flame-cut' the steel head gasket all the way through like that.

The usual scenario: An overheat warps the head. The head gasket may have developed its typical slight leakage that caused coolant to be slowly pushed out of the radiator (we kind of expect this failure on this engine) until it ran hot ,

that overheating event(s) (or any other overheat event from any other cause such as a fan or stat failure, even without prior typical slow leakage from a gasket failure) warped the head.

Continued driving with the warp and subsequent combustion leakage between the center cylinders allowed the flame cut to occur with every firing event just like a cutting torch.

sneaky98gt

Yea, and that seemed like a completely plausible cause of the issues. The only reason I was so stubborn in denying it is that it had exactly the same compression as #3, and #3 was running 100% perfectly. I still can not explain this.

As an aside, all 4 cylinders had great compression after I decked the head and replaced the headgasket. I forget exactly what the numbers were, but they were fairly even and IIRC in the 180 psi range.

Hmm, that sounds plausible. But to my knowledge, it was never overheated. If it did, it somehow did it without my knowing, and just with how I am regarding car stuff, I doubt that.

That being said, the head that was on it came from a junkyard a few months prior in an effort to replace some bad exhaust valves. I do not explicitly remember checking the flatness on it. So it is entirely possible that that head was already warped when I put it on the car, and over time it got to the exact point you mentioned above.

Either way, I learned some good stuff throughout the whole ordeal. Mainly to not underestimate how important good compression is, and 2nd, to make dang sure the heads are nice and flat before putting it back on. That's never a step I'll skimp on ever again.

Wildcat445

I came here to find out why I keep getting the P0302 code. I have had coils go bad--two on the Acura we had, and at least two others on this Civic. So with this misfire issue, I figured it was the coil...again...but after replacing it, the car would still misfire. I believe it also misfires only at idle. And like @sneaky98gt, the car has never overheated for us. It does go through a lot of oil, and was going to replace the oil pan gasket and valve cover gasket once the weather improved here.

I've been dreading a head gasket issue with this car and at 226k, it looks like it might be our turn now.

I don't have the equipment to do a compression or leakdown test. But I would think that's the next logical step. Along with checking the coolant for bubbles, and the usual head-gaskety things.

I'd replace the car at this point, but, we'd only get scrap value for it, and I just replaced (or retired, if you will) my '97 CR-V with a much newer one, so getting another car is not on the radar at the moment. My '97 is stuck in park (stuck in the tranny, not at the column), so I don't even know if I could press it into service as a temporary replacement.

ezone

Sounds like a compression test might be a first check.

Then a valve adjustment check.


Leaks or burning? Check plugs while doing the compression test, do they have whitish ash deposits?

At almost a quarter million miles some oil usage ought to be expected. Is it too rapid to keep up with?

Meh....see about compression first. Head gasket doesn't normally cause a loss of compression on this engine at first, but valve issues sure can.

The burned out head gasket pictured above in this thread was probably caused by a warped head installed from a used (overheated junkyard) engine.
Autozone rents stuff like that, give them a call? Surely other parts stores do too.


Cable problem, linkage under the trans?

Not an interlock problem?....a slightly open throttle (TPS voltage high) can prevent the shifter from coming out of park.

Wildcat445

The cable and interlock are fine--I had disconnected the cable from the trans and the shifter and cable work smoothly. The trans won't come out of park at idle when hot, cold, or even when the engine is off. Is there some sort of interlock in the trans when the idle is high (aka slightly open throttle)? It makes me wonder if that interlock, or whatever it is, is defective. At 289k miles, anything is possible.

I'll report back on the Civic once I can check compression. As for the oil, the oil pan is definitely leaking slowly, and I think there may have been a little oil residue inside the tube when I pulled the coil off the plug. I had already replaced the cam plug, which was leaking quite a bit. When I checked the plugs a year ago, there was no white residue.

The only odd thing I ever really noticed about this Civic was that there was this faint "diesel" smell to the exhaust ever since we bought it in 2014. It always ran well. Up until now, anyway.

ezone

At that point the cable was removed everything is strictly mechanical and manually operated within the trans (shifter lock and interlock are electrical), you should have been able to move the (lever or shaft) control for the trans by hand (or with a tool)..... that should have allowed the mechanical linkage for the park pawl to release manually, but if not it sounds like an internal trans issue.



Exhaust leak and snapping noise while cold? Cracked manifold casting is common at high mileage, noise and smell become less apparent after it gets warm

Wildcat445

I haven't looked at the service manual yet, but, isn't there a range switch on the transmission? Long shot, but I wondered if that might be binding somehow. Yet you can feel it in the whole vehicle when it gets yanked out of park (the times I don't remember to put it into neutral while parked) which makes me think it is internal. It is almost due for an ATF drain/fill but I doubt that will do anything to help. (The level is not low.) It works fine otherwise. I just don't want to put all that money into fixing it since it likely would cost more than it's worth.

Did you mean the gasket? I had read about that in these forums a year or so ago, and I already have the gasket on hand. I can feel the leak underneath the exhaust manifold where it meets the engine, when it's cold. Just haven't had time yet to work those rusted bolts off of the manifold. They are too rusted and rounded off to remove via conventional means...

ezone

Is this parked on an incline, or flat ground? An incline will make the park mechanism do this simply because it's keeping the car from rolling.
Manifold casting cracks are typical on Civic LX, we hear about these usually because it contains the cat and is expensive to fix.....an EX doesn't have a cast manifold (tube header), it's less prone to the issue....

Gasket and stud problems on either version are always possible, but usually much cheaper to fix.

In our shop I frequently see the cast ones cracked..... few gasket issues by comparison........but very few people are willing to pay to fix them either way.

Colin42

Are you sure the cable is good? I need to replace that on the '99 that I have, and you need to yank on the shifter with both hands to get the damn thing to come out of park. All the other gears are fine, just in and out of park. I got a cable from a junkyard for $20 that I tested before I pulled it.

Wildcat445

The driveway is flat, plus I put the parking brake on first, to keep load off of it. An old habit from 25 years ago when I owned a manual tranny. I also rocked the whole car back and forth to see if it might help, but...nothing. And it's fine when the engine is hot. Cold? Forget it.

Oh, wonderful. I was actually looking at that converter a while back, wondering why they would stick it in such an odd place. Now I see why the manifold is so pricey--it includes the converter. $250-ish for aftermarket. I don't think that is worth fixing unless the smell gets really bad inside the cabin. Some are made from stainless steel tubing vs. a casting. Could be better. Or could be worse, if the welds break.

Was your cable binding? When I disconnected the cable from the tranny, the cable moved very smoothly. Problem is, now, it is stretched a bit. I can adjust it again to accommodate, but I also think I am weakening it a bit when I've had to tug on it that hard. It has done this occasionally for the last few years, where it seemed a little bit stuck, but never to where it nearly refused to come out of park.

I find on mine that if I pull it very sharply, it can snap out of park. And that only happens when the engine is hot, like, up to its running temperature. If I put it into park temporarly, it will come right out. I leave it in neutral when it is parked now, but all it takes is me forgetting it once, and I'm stuck.

Colin42

I'm 95% sure its the cable on my car Doesn't matter hot or cold, engine on or off. Got the car for $500 and I'm working on getting it on the road for my girlfriend. There's been a slight pause while it's winter here, work should be resuming soon

Wildcat445

I have to get ahold of something to check compression yet, and some weather that's not in the teens or 20s. But until then, I cleared the codes again. This time it had three P0300 (misfire), P0302 and P0303 (misfire on 2 and 3), and it has been collecting codes for a week now.

Same symptoms as the first post, verified--the misfires appear strongly in the exhaust, and the misfiring stops somewhere between 1500-2000 RPM. No real noticeable power loss, and it runs smooth when at speed.

More to come...