Burnt valve causing idle misfire?

[slow ride]

I've been chasing a idle misfire on #4 for a few years and basically have checked everything possible except for a running compression test only due to having the fueling off the car and not wanting to reinstall it to test. At first it was not so bad with stockturbos/fueling/tune and adding load would make it stop like turning on the AC, etc. Last summer I added turbos/298 injectors/tune, etc and the problem was still on #4 but seems just a bit worse which I attribute to the new larger injectors. The car runs great under load and makes good power and no misfires at 23-25psi while we were tuning on pump gas. After more and more tuning pulls the problem was getting worse than in the past so I decided to dig in one more time as I will be pushing the stock block to the limits on e85 later on.

I have checked all the wiring for the coils/ignitors and have spares to try as well. I checked for pressure leaks up to 25 psi, etc. I Flow tested the new injectors and they are perfect, so no issues or leaks with any of those parts. Compression test is 157, 153, 154, 140, 149, 148 cold in order so you can see #4 is down a bit. Leakdown was 4-6% cold for most, but #4 still showed the most at ~8-9% and most of the air is past the exhaust valve/s vs any of the other cylinders, but again not that bad. LTFT's are also richer by 5-7% on the drivers side so that aligns with the lower compression IMHO. Cam lobes look great for 150k miles and the springs all feel the same when pressing down on the exhaust valves on the drivers side. I was sort of hoping it would be a broken spring or worn cam, but it appears to be a slightly burnt valve/s on #4 or a pumped up lifter holding the valve/s open slightly, but I guess it's so mild that the compression is decent. I'd hate to pull the head and not find an issue, but at this point it almost has to be either one or both valves on #4 or a bad lifter not adjusting correctly and holding the valve off the seat a tiny bit.

I have had the car since 2002 so quite some time and maintain it well. The engine was re-built at the dealer due to water pump failure at 30k back in 2002 also. I attribute this issue somewhat to poor machining on the head by the dealer or some other fluke, but it's hard to say since all the other cylinders seems to be ok and not have a misfire issue. I've probably diagnosed this thing to death and just need to pull the head, but I'm kicking myself for not doing it when I had it out for the turbos. At this point I'm also tempted to swap the rods out also, but then it's pulling it out again. I just wish the problem showed up more pronounced in lower compression or higher leakdown test numbers. Anyone else fought one like this? Burnt valves don't seem very common on this platform so that is why I ask for more examples.

Ed

[spacecadet]

I'd imagine you would have much more leak down/compression loss than that with a burnt valve.

[blitz2190]

I've been chasing a idle misfire on #4 for a few years and basically have checked everything possible except for a running compression test only due to having the fueling off the car and not wanting to reinstall it to test. At first it was not so bad with stockturbos/fueling/tune and adding load would make it stop like turning on the AC, etc. Last summer I added turbos/298 injectors/tune, etc and the problem was still on #4 but seems just a bit worse which I attribute to the new larger injectors. The car runs great under load and makes good power and no misfires at 23-25psi while we were tuning on pump gas. After more and more tuning pulls the problem was getting worse than in the past so I decided to dig in one more time as I will be pushing the stock block to the limits on e85 later on.

I have checked all the wiring for the coils/ignitors and have spares to try as well. I checked for pressure leaks up to 25 psi, etc. I Flow tested the new injectors and they are perfect, so no issues or leaks with any of those parts. Compression test is 157, 153, 154, 140, 149, 148 cold in order so you can see #4 is down a bit. Leakdown was 4-6% cold for most, but #4 still showed the most at ~8-9% and most of the air is past the exhaust valve/s vs any of the other cylinders, but again not that bad. LTFT's are also richer by 5-7% on the drivers side so that aligns with the lower compression IMHO. Cam lobes look great for 150k miles and the springs all feel the same when pressing down on the exhaust valves on the drivers side. I was sort of hoping it would be a broken spring or worn cam, but it appears to be a slightly burnt valve/s on #4 or a pumped up lifter holding the valve/s open slightly, but I guess it's so mild that the compression is decent. I'd hate to pull the head and not find an issue, but at this point it almost has to be either one or both valves on #4 or a bad lifter not adjusting correctly and holding the valve off the seat a tiny bit.

I have had the car since 2002 so quite some time and maintain it well. The engine was re-built at the dealer due to water pump failure at 30k back in 2002 also. I attribute this issue somewhat to poor machining on the head by the dealer or some other fluke, but it's hard to say since all the other cylinders seems to be ok and not have a misfire issue. I've probably diagnosed this thing to death and just need to pull the head, but I'm kicking myself for not doing it when I had it out for the turbos. At this point I'm also tempted to swap the rods out also, but then it's pulling it out again. I just wish the problem showed up more pronounced in lower compression or higher leakdown test numbers. Anyone else fought one like this? Burnt valves don't seem very common on this platform so that is why I ask for more examples.

Ed

Since you've ruled pretty much everything else out and your leakdown and compression show low numbers for 4 my bet is on a bent valve or possibly just carbon buildup, the dealer may have left a bad valve in there when rebuilt thinking it wasn't bent(may have been very little) and that could have caused a burnt valve or extra carbon to build up. But regardless its all speculation and sounds like its time for a head pull.

[ElementR]

Only other thing that comes to my mind is slow coolant leak into combustion chamber/cylinder. Any signs of slow coolant loss?

[christianb5s4]

I had a slight miss at idle that was caused by a leaky valve stem, so if everything else has been ruled out that's certainly a potential culprit.

[slow ride]

I agree on the speculation and I know better from all the other setups I've worked on or had a part in building, but the older I get the less enthused I become to work on these cars. Doing a valve job when it was out would have been ideal lol.

No signs of coolant loss at all, engine is spotless on the bottom side and I just did the rear main when it was out. Also took the car on a 2000 mile road trip from Iowa to Colorado 2 years ago with no issues (other than a ICM in the mountains, thank god for spares).

I will pressurize the intake side and look at valve stems on that side to check for a leak before pulling too much. I'll check the leakage on #4 also when the cam/lifters are removed to be sure they are not in play. Such a pain in the ass when it could have been done last summer on a stand dammit. I should also say that it isn't a dead miss and probably counts up to 20-40 before refreshing, much less with load. I can drive around all day with no codes, but allow it to sit and idle too long and sure enough you will get a #4 misfire code. Clearly something is not happy on that side since the LTFT's are -5-7% off (rich) on the drivers bank.

[Zba]

Weak or broken valve springs can do this too.

[slow ride]

Weak or broken valve springs can do this too.

I agree, when I pull the exhaust cam and re-check leakdown I will be looking at the springs more. I did see the old post on here where Volker had a broken exhaust spring, but he also noted that he could push down the lifter easier due to seat pressure reduction. I did check how hard the lifters pushed down with a plastic panel popper and they seem tight or the same as the other exhaust valves. Not scientific, but just a quick check.

[Zba]

If it's not pulled apart yet you can do a dynamic compression test. That will help pinpoint where your problem lies.

[slow ride]

That was the only thing I mentioned not doing, but only because injectors/rail are out due to testing injectors. I suppose it's not that much work to reinstall.

[Zba]

Not too bad to reinstall to get a little more info

[wbrunner23]

Can you please explain dynamic compression testing procedure/purpose to me?

[Zba]

While a static or regular compression test shows you the ability of the engine to build compression, it's really not all that useful. A leak down test is best in that regard since if you have leakage, it will show up and you can now pinpoint where the compressions learning at.

A dynamic or running compression test which gives you a better idea of how the valve train and bottom end are working and provides more useful information regarding engine health. For example; a weak have spring will not show up on a regular compression test because the cylinder will still build compression. A leak down won't show a weak spring either because the air in the cylinder will force the valve closed. A demise test on the other hand will show air escaping the cylinder.

A vacuum gauge can also be used to determine what is going on n your engine but that obviously won't narrow down to a cylinder.

Check out this link since it can describe it way better than I can.
http://support.alldata.com/article/t...pression-tests

Vacuum gauge readings.
https://actron.com/content/engine-pe...g-vacuum-gauge

[slow ride]

Well I'm digging at the car to take the head off in the car. I did not do the running compression as I got stubborn and kept plowing through things and didn't want to reassemble to do the test. Man what a pain in the ass, but it's about out now. I've determined the exhaust valve/s in 4 to have a decent amount of leakage which after consulting with other mechanic/engineering friends seems totally logical to be causing the misfire IMHO. The intakes have about zero leakage so the rest is blowby which the car doesn't really have much when running at all and should not really cause a misfire like this with such good compression numbers. I've had so many hold ups pulling the head like a stubborn inside manifold to turbo bolt that I broke/cut/heated a few tools to get out and then of all things a stripped lower (yeah the only you can hardly get at) water pipe hex bolt that I had to weld a cut off bit to remove. I swear these cars can test your patience when things don't go correct, but if you have a lot of tools almost any problem is fixable. I'll post an update after getting the valve job/valves cleaned up.

[slow ride]

I had a slight miss at idle that was caused by a leaky valve stem, so if everything else has been ruled out that's certainly a potential culprit.

How did you determine the valve stem seal leak? I assume you pressurized the intake and noticed it with the valve cover off? Did the car smoke at all also or misfire off idle?

[slow ride]

So after a bunch of messing around and cleaning up the valve job on the drivers side head the pos still misfires on #4. I pulled vacuum on all the intake ports and chambers after doing the valve job and they would actually hold vacuum and I had leakdown under 1% using a blockoff plate so basically nothing leaking in the head. I'm going to re-test the leak down on a few cylinders, but 4, 5, and 6 were 4%, 2% and 6% after cleaning up the valve job and installing the head so most of that would be rings. Engine basically runs exactly the same as before and just misses at idle. Probably 50-80 per reset, but it varies some. I did do a running compression test this time and I was seeing about 60 psi on 4, 65-70 psi on 5 and 60-65 psi on 6. It's really been a kick in the nuts as I was so sure the problem was the leaky exhaust valves on 4.

With the running compression test so close I can't see how the misfire isn't ignition or fuel related (I also swapped injector 3 and 4).

[wbrunner23]

With the running compression test so close I can't see how the misfire isn't ignition or fuel related (I also swapped injector 3 and 4).

What a pain in the ass after all that work.

So, lets see what you've got so far:
1. Injector & coils swapped, no change
2. Valves sealing well - leakdown improved from ~8-9% to 4% on that cylinder
3. Ring sealing well - compression, leakdown, & idle dynamic compression results

Since the misfire isn't moving when you swap coils/injectors, the cause has to be something fixed to that cylinder.

Did you gather snap throttle running compression results? In the link Zba provided above, snap results when compared to static & running compression results can indicate other issues, such as cam lobes, valve guides, etc.

Not sure if any of us will be able to figure this out for you from over the interwebs, but I'm leaning towards an issue with wiring (wires shorting under sheathing somewhere further up the loom), potentially cam lobes on cylinder 4 are worn excessively (snap compression results would help here), and then lastly, the easiest potential solution could be a ground issue.

I was having misfires on cyl #4 as well (both idle and above). I did everything I could think of short of pulling heads: compression test, boost leak test, swap injectors/coils, new injectors/coils, blah blah, I spent a few consecutive days looking trying to figure it out. Turns out my misfire was caused by bad ground at the post under the the coolant reservoir that popped up randomly (I hadn't touched that ground for a while at the time), believe it or not. Pulled the ground, cleaned the terminals, and reinstalled - misfire gone.

Hope you are able to sort it out soon. These cars can be frustrating as hell sometimes.

[spacecadet]

lol you ground all the valves and its still missing? that sucks man. Good luck I hope you get it. hows the harness look?

[Zba]

The numbers are only part of the story. What was the gauge doing during the testing?

[slow ride]

I only did the snap test a few times and they shot up to around 100 or so on a few cylinders including 4. I plan to redo the test tonight or this weekend and write down everything as I was just busting ass trying to figure shit out. Nothing is a smoking gun on this thing other than being on 4. I know when I had the engine out for the turbos I did look at the grounds under the coolant bottle and last night I looked at the head/harness ground on the passenger head. It's getting pretty frustrating as you can tell, but I will figure it out.

One thing of note is a few years back when the car was stock turbos/Daz tune I popped a ICM on a trip to CO and it was the one for cylinders 4/5/6 and the issue was 4 went 100% dead which I find strange since that is my problem cylinder. The car did have a a light misfire at idle back then on 4 as did my old GAIC ECU did, but not this bad so I didn't look into it.

On this drivers side head I only cleaned up the valve job on the exhaust side on 4/5/6. The seats were wide so I lightly recut them and lapped the cleaned up valves in. I tested all 3 chambers and they would hold vacuum very tight and leakdown was basically nothing on all 3. The only thing I can even suspect is the lifters (god I wish I at least moved them around) or ring seal.

[slow ride]

The numbers are only part of the story. What was the gauge doing during the testing?

As you reset the gauge it would pop up to 50 quick and jump a few times to 55 and settle to 60 with a few more pulses. It was very close to the others, but was down a few % compared to 5 and 6. The pressure pulses didn't seem very erratic or anything on 4 vs the others. I will warm it up and do a full evaluation tonight.

[Zba]

Have you tried swapping icms again?

[slow ride]

Well not this last time, but I will be. I just find it weird that the only ICM failure I have had in the past was also on 4 (dead cylinder in the mountains how fun is that lucky for spares :). I suppose it could have been a fluke, but I really need to go through the wiring again with a fine tooth comb. Whats your thoughts on the running compression? After doing the t belt this time with new chain guide pads I'm at 4 and -4 on intake cams (was better before). I know offset on the camshafts can change compression some, but it's still in spec.

[getslideways]

Does the misfire rate change if you push in the clutch? Such as when watching the misfire counter in vagcom?

[chaos2984]

Did you flip the head over and put liquid/ solvent in the combustion chamber of the head and see if you were loosing any fluid through the intake or exhaust valves when you were doing the head?

I would make an overlay harness for testing that goes from the coil to the ICM and see how that does.

[slow ride]

I plan to power the coil from the battery at this point and run the signal from the ICM direct. I did a leakdown and vacuum test on the head ports/chamber before and after the valve job, they sealed just about perfect after. In this condition basically no fluid would ever even leak out after a long time. I did cold compression last night and got 158/162/160/149/152/155 so I did end up with better compression from fixing up the exhaust valves. I borrowed a o scope that I plan to view the coil and injector pulse with so I hope that might shed some light also.

I'm sort of kicking myself for not swapping or getting new lifters for that cylinder, but they appeared to be fine and I did take the exhaust side apart to bleed them down/inspect them. I let them pump back up when warming up the car to be sure they were adjusted after the exhaust valve job. Had I swapped them from say 4 to 5 at least I could rule that out (stupid me). Cams looked really good as did lifter surfaces so that is a good since I have used Rotella over 10 years.

[julex]

Are you on stock manifold, how about throttle? I am on enlarged plenum RS4 manifold and HEMI throttle so the problems I describe might be exacerbated by my set up. I am also running bosch 0280158040 injectors which are 1000cc @ 5bar.

I'd been chasing similar issues with #1 and #4 (front cylinders) and came to the conclusion after a lot of troubleshooting that this is simply physics at work here. At idle, when air velocity is very slow, the front cylinders simply get more air than back of manifold and a such, they simply run leaner than rear. In practice, it is only the middle cylinder that get the prescribed 14.7:1 at idle, front gets a little higher mixture ratio, rear lower but it is all averaging to one number at the O2 sensor of course.

I "solved" this by tweaking tune and giving the front two a little bit more fuel at around idle and subtracting a little from rears. My car idles butter smooth now.

[slow ride]

Glad you showed up here as QW is impossible to dig up my old post for shit. I was going to go over everything you spoke about on QW with the cylinder trimming and possibly talk to Blake or mess around more myself. Does the mixture tweaking work across the board or just idle/low load? With these ECU's can you move the target AFR at idle while keeping off idle 14.7ish? My TB is stock, manifold is a enlarged plenum and ported bells/tb exit early gen (2000 flat top) manifold. I'm running the smaller 298 gt500 style injectors that are clocked correctly (plugs facing valve covers). Thing is my #1 cylinder is 100% fine, it's just 4 that goes crazy high misfire and I had a few extra misfires even with the stock injectors on 4. I suppose I would be willing to test this out on my current file which runs good on pump gas.

This has been such a frustrating problem to deal with as nothing seemed to help. I used a o scope last night and the ignition signals before and after the ICM seemed to be identical between the 4&6 coils, they also match the #4 at the coil, but it's hard to get a trigger on this old ass scope with the rpm fluctuation. I did another running compression test on 4 and 5 last night which was 4=61 and 5=69 psi after 10 seconds or so of running. I tried to open up the plug gap to .032" just to see what would happen, but no change. I also pulled the FPR vac line which should also enrich all the cylinders and it didn't change much is anything. I was going to view IPW also, but it was getting late and I had enough of the fumes in the garage haha.

[getslideways]

When watching the misfire counter in vagcom does the misfire rate change if you depress the clutch pedal vs just being in neutral?

[julex]

Glad you showed up here as QW is impossible to dig up my old post for shit. I was going to go over everything you spoke about on QW with the cylinder trimming and possibly talk to Blake or mess around more myself. Does the mixture tweaking work across the board or just idle/low load?

There is an adjustment table per cylinder but the only available axis is RPM... But it is good enough imho.

With these ECU's can you move the target AFR at idle while keeping off idle 14.7ish?

In theory yes but in practice no due to narrowband O2 sensor. You can move the target ECU's "stoich" point to anything you want but you're at a mercy of O2 accurace and with narrowband O2 sensor you would have little chance to have a stable mixure beyond couple of decimal point around 14.7

When watching the misfire counter in vagcom does the misfire rate change if you depress the clutch pedal vs just being in neutral?

I probably proposed this in the other thread too. Bumping torque reserve up (which is what pressing the clutch pedal does at speeds below 5 or 4km/h) might also mitigate the misfires.

[slow ride]

I have messed with depressing the clutch and Blake did turn up the torque reserve on a few last revisions(visible in a lower vacuum signal) so I follow you guys on this. In the past with stock fueling/turbo type setup it might have a few too many misfires on 4, but it was nothing crazy 15-20 tops and they would clean up with the AC on or after a long run time, etc. now its 80+ misfires on 4 between resets with only 0-10 on any other cylinders. I think bumping the torque reserve helps out slightly now, but the counter is still racking up a lot of misfires.

Julex, how many misfires were you getting on 1 & 2 before you tweaked the mixture distribution? I guess when I say idle I mean no AC and other stuff running so the lightest load on the engine possible. I might induce a vacuum leak in the back of the manifold at the booster hose just to see how it acts. I know it will lean out some, but just for a quick test.

[slow ride]

I figured these were some good quotes from Julex from the other forum thread here http://forums.quattroworld.com/s4/msgs/225731.phtml

You could correct it by tuner either upping the torque reserve for idle (tables KFMRES, KFMRESK)(more air and fuel coming in during idle will smooth the difference) or diving deep in and adjust fuel correction for this specific cylinder for RPMs you're interested in (FZN0 to FZN5 and FZNW0 to FZNW5)

If you're NOT getting misfires when you're with warmed engine at a stop and with clutch pedal pressed in (thus in clutch reserve mode dictated by KFMRESK table), then simply using bludgeon approach of replicating that table into KFRES table and calling it a day will do the trick for you. You will consume a little more fuel in idle but your problem will be solved...

My specific car with RS4 modified intake manifold (ported, enlarged plenum) and 2.8 heads was routinely misfiring on #1 and #4 (first two cylinders from throttle) and very lightly on #2 and #5 (middle) with rear being rock solid. I modified FZNW0 (cylinder #1) and FZNW1 (Cylinder #4, internal ECU variables are following firing order not front-to-back order as we refer the cylinder by) for 1.04 multiplication at 520 and 760 RPMs, middle ones to 1.02 and left rears at 1.00 and my car never misfires now.

Just keep in mind that when you say 1.04 for a given cylinder in a bank, it multiplies the amount of fuel injection time by 1.04 for that cylinder but when you do that, the bank as a whole will get less fuel positive fuel trims now so it effectively takes a little bit of fuel from other cylinders.

This might be something that you may benefit from.

And they all comprise from values calculated for different miscalibration, leaks, etc, situation.

What we think a idle trims are really applied even to part throttle trims but then there are two sets of part throttle trims one for low load and one for high load with a continuous interpolated value generated for all load situation, which is then used in fueling calculations.

You can see this when you run VCDS block 032 and observe how part throttle trims change the more gas/load you throw at the engine.

Having 5% disparity between banks sound high but maybe you should first look at your blocks 090-093 to make sure your cam phasing is ok. If that's way off between banks then you will have problems.

I've about exhausted everything else in this pile that could cause the problem. Also curious what the difference between (FZN0 to FZN5) and (FZNW0 to FZNW5) also. Julex I sent you a PM.

[julex]

FZN0 to FZN5 - fueling multiplicator for normal operation (this one is pertinent to idle area)
FZNW0 to FZNW5 - same but during intake cam advance (med-high load from about 2k-4k rpms, there is a table to dictates "NW" state of engine per load/rpm axes)

0-5 is firing order NOT visual order as we are used to when talking about cylinders.

[slow ride]

FZN0 to FZN5 - fueling multiplicator for normal operation (this one is pertinent to idle area)
FZNW0 to FZNW5 - same but during intake cam advance (med-high load from about 2k-4k rpms, there is a table to dictates "NW" state of engine per load/rpm axes)

0-5 is firing order NOT visual order as we are used to when talking about cylinders.

Thank you, I think we were trying to adjust FZNW and not FZN I believe and not seeing any results (for good reason it appears).

[slow ride]

So far we have tried in order of firing (1, 3, 5 ecu lingo which are cylinders 4, 6 and 5) 1.06, 1.00, 1.00.
1.15, 1.00, 1.00.
0.90, 1.00, 1.00.
1.10, 0.95, 0.95.
And I can't tell a bit of difference with how the engine runs or see any real change in misfire count. Surly I would see something, but no smoking gun. I'm trying to think of a vag channel to view and verify our changes are doing something, but don't really think anything would show this.

What should timing be at this light load as I'm sitting around 0-5 degrees at idle bouncing. Airflow from my b6 s4 maf is around 5g and engine load is around 20-21% and measures 20-21" if we don't alter torque request. For a quick test I can feed air into the brake booster line to change distribution and the misfires go way down to just a few, but airflow is up to 7-8g, timing 10-15 and load is around 27-29%. I do this by teeing into the y pipe pcv line so no unmetered air enters. Basically the engine pulled in air through the brake booster line and closes the throttle body more. 0.0-0.4% is all the tb is open.

I have verified the coil is firing as I removed the coil boot and hooked up a spark plug cable so I could attach my timing light. The engine runs normal this way and I can hear the mis just as before, but I don't get any missing pluses from the light even though the misfire counter keeps going up.

At this point I'm getting ready to just turn off misfire detection and run it. One would think altering the tables Julex mentioned could fix this, but I'm just not seeing any change at all. Overall pretty damn frustrated with the pile

[slow ride]

FZN0 to FZN5 - fueling multiplicator for normal operation (this one is pertinent to idle area)
FZNW0 to FZNW5 - same but during intake cam advance (med-high load from about 2k-4k rpms, there is a table to dictates "NW" state of engine per load/rpm axes)

0-5 is firing order NOT visual order as we are used to when talking about cylinders.

Julex, do you know is FZN1 and FZNW1 the correct cylinder to change for the actual #4 cylinder? I keep going back to this internet image of firing order for the 2.7T which I thought was wrong, but if not it shows that #4 cylinder is 4th in order which would be FZN3 and FZNW3. Can you clear this up?

[julex]

I'll look at it but I am pretty sure you follow firing order and prepositionally match to N0 - N5. So cylinder 1 is FN0, 4 is FN1..etc. Everything in tune that related to specific cylinder is indexed to firing order, not to visual 1-2-3 4-5-6 as banks go.

[slow ride]

I see, Blake also tried retarding timing more just at idle and it cleared up the misfire a lot. I just figured if we were not altering the correct table that could have been an issue on our part for not getting the trimming to fix the issue. I still have some goofy bank to bank offset in part throttle LTFT's (about 4-6% depending on rpm rang), but I'm throwing in the towel on that and tuning away.

[kingkhalilz]

Someone can calibrate SPDR and DTRR :) these calibrators be lurking

[getslideways]

Randomly thinking about this thread the other day, so just bumping this to see if there was ever a solution or progress.