In need of suby doctors to revive my baby - Very bad misfires

Here's my story, I have been a silent reader of this forum for a few years now, it has always been very useful and filled with very knowledgeable people so thanks for all the help from the active members of this community. I would describe myself as a somewhat skilled parts changer. I've replaced engines, head gaskets, timing belts, done suspension work, brakes, fuel pumps etc. I feel confident in my wrenching skills, where I lack is with diagnostics.

 

I've owned a 2006 LGT for two years, a year after I bought it I decided to do a leak down test on it since I felt it had a pretty bad piston slap. The numbers were horrible so I decided to look for a new (used) engine. I did check if I could simply rebuild it but the bores were already oversized and the machine shop told me they would not be able to get rid of the damage. Got a used engine from a reputable performance shop, leak down numbers on it were good so I went with it. Fast forward a few months the engine is now installed, I nervously turned the key and to my highest excitement and relief it fired right up and sounded waaayyy smoother than my previous engine. I let it run for 5 minutes or so to get it to temperature and burp the coolant system. Everything is still fine so I decide to take it for a short test drive, while backing out of the garage it starts to sputter a bit, than a bit more, until it basically sounds like it's running on only one cylinder and stalls. Now I can't get it to idle more than 20 or 30 seconds, same thing happens everytime.

 

The car is in my father's garage, which is a two hour drive from my place so I try to make my trips worthwhile. Too many times I got there only to find out I'm missing a 2$ part and all the shops are closed because I can only get there on the weekends. I'm trying to make a plan of all the things I should be looking at to try and identify the issue. Here's what I'm thinking :

 

1. Check the spark plugs / Coils / Wires

2. Replace the 02 sensor with the spare one from the previous engine

3. Compression test all cylinders to make sure the engine is healthy

3. Clean Maf sensor? From here I'm kind of clueless...

 

To be honest I don't really know where to start looking, I don't want to blindly throw parts at the problem I'm looking to sell the car soon I just want to make sure it is in good working condition. I would love to hear your theories on this and what you guys would recommend I look for. My guess is it's probably not mechanical since it ran just fine for a while. Maybe the plugs or coils but why would all of them start to fail at the very same time, seems pretty unlikely. At first I thought it might be related to the 02 sensor since it started to misfire only when the car got up to temp but I've tried starting the car up again when cold a week later and it ran like crap instantly and stalled after a few seconds.

 

Anyway any input would be greatly appreciated, sorry for the long post, cheers from Canada!

I'd get a means to datalog the car. You'll be able to see which cylinders are misfiring, as well as other trouble signs in the ECU's response. An OBDII Mode $06 capable scanner will give you some capability, but something like the BtSSM Android app and bluetooth dongle, or a Tactrix cable, is best.

 

If you're selling the car soon, might try and find someone local on here or on FB groups or something with the capability to lend you a hand in exchange for a 6 pack.

Thanks for your reply, I have a cobb AP to read codes and so far I got P0302 P0303 P0304 as well as PFFFF. I took the plugs out of cylinder 1 and 3 before I left last week and they were both wet with fuel so my guess is they all seem to misfire. I'll put new plugs in there anyway since the ones I pulled are covered in carbon residue. The car started to heavily smell like fuel when it started to misfire also...

OK, so you're misfiring on 3 out of 4 cylinders badly enough to set the DTC, and I'd bet also missing on the other cylinder as well just not quite as bad when the codes were set. It takes something like 80 misfires in a 10 second period to set the code for a given cylinder.

 

Your Cobb AP has the capability to monitor and record individual cylinders for roughness, as well as a plethora of other useful diagnostics data. I'd learn how to use those features, way better to make data driven decisions than to try and rationalize what could be wrong based on physical symptoms.

 

I'd take a look at individual cylinder roughness counts, as well as AFR to make sure that you're not running super lean or super rich at idle. Finally, take a peek at long and short term fuel trims/corrections and see if they're swung very far positive or negative. The long term trim is a 'learned' value, that the ECU will store if the short term trims are way off-zero for a long enough period of time. This gets reset every time the battery is disconnected or a trouble code is cleared. However, given that the car won't stay long enough to 'learn', you might have to work with the short term trims.

 

The idea is as follows:

 

If you see the AFR swung way rich or lean (normal at idle should be 14.7, rich is lower than that, and lean is higher than that), take a look at the fuel trims. This will tell you if the car is at the limits of its ability to add or remove fuel to compensate for the off-target AFR, and go from there. Maybe post results here and we can advise.

awfulwaffle thank you so much, I'll look into how to use the AP to get this data and post the results here. This is the kind of stuff I know I should learn, I guess this is a good opportunity. So if I understand correctly, I'll look at the AFR to make sure it's not way out of wack. If it is, then I look at the fuel trims to see if the ECU is maxing out it's ability to adjust the qty of fuel injected.

 

Now say the short term trims are far from 0 I would guess it's because either it's getting a wrong read on the measured AFR thus trying to compensate for false data or it's physically not able to add or pull fuel in the cylinders which would indicate a problem with the fuel system? Am I understanding this correctly?

awfulwaffle thank you so much, I'll look into how to use the AP to get this data and post the results here. This is the kind of stuff I know I should learn, I guess this is a good opportunity. So if I understand correctly, I'll look at the AFR to make sure it's not way out of wack. If it is, then I look at the fuel trims to see if the ECU is maxing out it's ability to adjust the qty of fuel injected.

 

Now say the short term trims are far from 0 I would guess it's because either it's getting a wrong read on the measured AFR thus trying to compensate for false data or it's physically not able to add or pull fuel in the cylinders which would indicate a problem with the fuel system? Am I understanding this correctly?

 

More or less, but the possible causes are pretty numerous. The car will try to compensate for any number of things including bad AFR or airflow sensor data, but also for mechanical issues like a vacuum leak, bad fuel injector, low fuel pressure, etc. The particular scenario in terms of AFR and fuel trims will indicate where in that particular set of faults to look closer. If the trims and AFR are fine, then you rule out those issues and look at other possibilities.

If it starts and runs smooth until it gets fully warm, then I would guess it's a sensor going bad somewhere. Monitor the coolant temp numbers. They should slowly rise and get to 180-190 and stay there without any erratic movement.

 

IF it's misfiring all of the cylinders at the same time, then it's something that is telling the car to alter fueling, or cutting power to all 4 coils at the same time. But what you describe sounds like more of a flooding out situation. Monitor things like, cam and crank sensors, TGV motors (if they are staying closed), AVCS operation, and clean the MAF.

Thanks for your input Redwagon. Actually it did run great until warm the first time I started it. Since then I can't get it to run for more than about 30 seconds, it'll run rough for a few seconds (maybe 10-15 seconds) then it gets worse and worse until it basically is only running on 1 cylinder and dies. I'll make sure de take a look at all of these. TGV's are deleted on my car if that is even relevant.

Ok, and I'm assuming vac leaks have been ruled out? Different plan because if it does it from a cold start, then it points toward being mechanical. One way to check that is to reset the ECU and give it a cold start. But keep in mind that if you do this, you lose the stored codes from the previous diagnosis.

 

My thought here is: If it's mechanical, then it will do the same thing with a fresh reset of the ECU, because the ECU wouldn't be running skewed values from trying to correct the condition previously. If it's electrical, it will either start clean and deteriorate, or throw a code for a sensor straight away. I also realize that you won't be able to monitor the operation of the sensors if the car doesn't stay running.

 

Another test is to unplug the MAF and try to start it. That will force it to run in open loop and take the MAF/O2 sensors out of the equation.

I have not done any tests for vac leaks yet. I'm waiting on a vacuum gauge I ordered last week. As for ECU resets, I remove the battery between my visits to put it on a trickle charger since I get to go work on the car only once every other week. So what you're saying is I should worry about a mechanical issue since I get the problem even when starting cold and with the ECU freshly resetted?

 

What kind of mechanical issue could cause this bad of a misfire? Timing? Although if timing was off it would'nt have run correctly the time I started it.

 

I'm feeling kind of silly not to have tought about this earlier but what about bad fuel? With this covid thing I could not work on the car for a while and the fuel in there must be around 3-4 months old...

 

I'll surely give the unplugged maf a try and clean it aswell.

Mechanical as in, not a sensor reading. Unless of course the car throws a CEL for a sensor straight away. Things I'd check on the mechanical front: Vac leaks, Fuel pump, something that would block air from the intake manifold (vac leak or obstruction), or cut power to all coils at the same time (seems unlikely, unless coolant temp sensor is reading something like -40*).

 

So before starting, I would clean out the MAF, and check with the AP, that the coolant temp and ambient intake temp is reading something in the ballpark of ambient. Also check that the stock front O2 is reading 14.7 or leaner (numerically higher). Then smoke test it. if you don't have a smoke tester, trace ALL the air lines including emissions paths (secondary air pump, PCV, etc.). Double check all hose clamps, and make sure a hose isn't kinked shut. Next make sure all coil packs are fully plugged in (sometimes they're on, but don't click).

 

It could also be the fuel pressure regulator. If you have the one from your last engine, try swapping them.

You guys are awesome, I'm sort of getting an idea on how to start attacking the problem now. As Awfulwaffle suggested I started doing research on fuel trims and different sensor data and how to interpret them. I looked at a few hours worth of content on youtube by a channel called GoTech, pretty awesome stuff. Their videos are in depth, concrete examples and case studies on different cars and problems worth a good watch if someone reading this is in a similar situation.

 

I'm pretty sure I have a small party smoke machine from when I was in high school, it's probably been 15 years since I've seen the thing but it must be laying around somewhere I guess I found a new purpose for it.

 

So based on everything suggested here, here's my new plan :

 

- Replace spark plugs (I got new ones, old ones were full of carbon and gap was at 0.060") might not be the problem but definitely won't hurt either.

- Recheck coil pack connectors (I checked them last time, the clip on #3 is broken, I'll replace the tie rap holding it with a working connector from the spare harness)

- Clean maf

- Check for sensor values on coolant temp, intake ambient temp, as well as front 02 sensor I'm assuming I can check these values with the key "on" without the engine running? (02 I'm not sure about since it is meant to work when hot?)

- Smoke test it to check for vaccuum leaks

- Give it a start and data log the fuel trims, AFR, Target AFR, temps etc. Long term will be at 0 since I removed the battery but if the short term is still off more then 10% either positive or negative then something is still wrong. Depending on the results I'll probably have a better understanding of where to at.

 

Here's where I'd love for someone with a bit more knowledge than me to shime in. The examples I've looked at on youtube are all for N/A cars. I understand a n/a car will always have a manifold pressure below atmosphere which means a vac or intake leak will make it run lean. I am right to assume at idle my car will behave the same since it is obviously not making boost? For a turbo car if I was to have a leak in the intake system (after turbo) my understanding is it would be running lean at idle and rich under boost since the system would be under positive relative pressure? And if it is a vac leak or intake leak before the turbo then it would be running lean regardless as it would on any n/a car?

 

Also, how much time does it usually take for a car to start running clsoed loop? During open loop operation, it is my understanding that the 02 sensors are not taken in consideration. If so what data is used to adjust the trims during open loop, is it only running based off of maf, temps, rpms and fueling tables? If so then it basically doesn't have a clue of what is actually happening in the engine without the feedback from the 02 sensors so it can't be compensating for a problem occuring after the maf? Could it be I can't even reach closed loop operation, and if so are fuel trims in open loop even relevant?

 

I know this is a crap ton of questions, I'm soaking up as much info as I can and trying to make sense of all of this. Once again thanks for shiming in, I'm enjoying the learning process.

Sounds like you're learning quickly!

 

Yes, you can just have the key in the on position to read those values. The O2 will read whatever the leanest it reads (~20.0). If there is still gas fumes in the exhaust it will read closer to 14.7, but it shouldn't ready any richer than 14.7. But the O2 sensors need to heat up first before they give a reading, so you'll have to keep the key in the on position for more than 10 seconds or so to get it to read.

 

Yes, a turbo car is in vacuum unless you're up in the rpm range. At idle, your boost gauge should read in the neighborhood of 17-19Hg. Any less than 15/16, and you have a boost leak somewhere. If you're monitoring it on something like Torque or Dash Command, it will read negative boost and peg itself at -7psi.

This won't be helpful, read it off your AP. And yes, those would be the characteristics of where the boost leak would be if you had the car running, but it doesn't sound like we're there yet. It sounds like if your car has a vac leak, iit would be a BIG vac leak, like the throttle body coupler coming off or something. Also check the turbo inlet, make sure there isn't a huge rip in it. The smoke machine will show you this if you can get that going.

 

Open loop just reads load and RPM. It derives load from the MAF reading. It fuels based on the programmed tables only, no real time adjustments from the O2 sensor, intake air temp, coolant temp, etc. O2 sensor is the main thing that it uses to make it's adjustments. And as mentioned, the O2 sensor doesn't come online until after 10 seconds of heating up, so for the first 10 seconds it's in open loop. Once it comes online, it starts the adjusting process, and adjusts the car into the idle/warmup mode it "settles" into. For those 10 seconds, you're in open loop. And yes, the MAF reading does come into play because the MAF scaling (airflow reading) determines load, which is used for open loop fueling.

 

If the fuel trims have been zeroed, then you won't read any fuel trims until the O2 sensor comes online and the car starts closed loop.

Edited June 21, 2020 by lil'redwagon

I would do a vacuum leak test first, it's free if you have a air compressor.

 

I use a pint size paint can to plug the hose coming off the air filter box, remove the small hose from the blow off valve on the tmic, give that hose a short blast of air and put your thumb over the hose, listen for leaks.

Finally got to work on the car yesterday. I tried to smoke test it but it didn't actually work like I thought it would. I got some smoke in there, but the machine isn't meant to put out any pressure so it filled the hose it was connected to but then started leaking out of the actual body of the machine. Anyway decided to skip this and see what the vaccum gauge would read when running.

 

So I replaced spark plugs, replaced the broken coil wire connector on #3, installed a vaccuum gauge, cleaned the maf and setup the AP to log some data. The car started fine and Idled fine for several minutes, I did not get a CEL this time, yay no misfires! (or so I thought). I remembered the car started to run poorly the first time when I was backing out of the garage so I gave it a couple quick taps on the throttle to see if it would start misfiring. It sounded it bit rough at times and had a bit of a hesitation but overall much better than the previous times. The AFR started to progressively lean out until it reached 18 at which point I shut the car off as it was obviously not able to correct the issue.

 

I was a bit confused with the sensor names on the AP, I was expecting to see fuel trims but instead found A/F correction 1 / 3 / learning 1 etc. I assumed 1 and 3 refered to the front and rear 02 sensors and "correction" would probably be short term trims and "learning" long term trims? I'm still not completly sure about this but the data seemed to make sense.

 

I attached the log to this post, I'm hoping I logged enough valuable info and the correct sensors. If anyone with a bit a knowledge could take a look at it and interpret what is going on I would be very grateful. From the very little I understand, the mixture started to lean out and the short trims start to rise to try and enrich the AFR back to stocheometric but the AFR didn't seem to respond. The values just diverge until the fuel trim is saturated at +25%. The fact that the fuel correction seems to not have any effect on the AFR is wierd to me. Could it be that the 02 sensor reading is wrong and the ECU is trying to correct for a bad reading? Or it is physically not able to send more fuel in the engine? Altought this seems less likely since I can get the engine to rev which would require more fuel than correcting mixture at idle. I could also be very wrong

2020-06-27 - Datalog.csv

Ok, so it's fuel. I'd guess that it's either the fuel pump, or one of the injectors being clogged and not firing. Did it smell like fuel when it started sputtering again? Do you know the history of the injectors? Have those injectors been run on E85? Apparently injectors get clogged with only 10K miles of E85 use. Depending on the answers to these questions, I would guess that the next steps would be:

 

To test the fuel pump, get/rent a fuel pressure gauge and put on the fuel feed line going into the engine and see if it holds steady pressure from then it runs smoothly to when it leans out. If it does, then your pump is ok, and you could look at your injectors.

 

Some indicators of an injector going out: If it's only one injector, your car will lean out, but you will smell fuel. What happens here is when one injector doesn't get fuel, the O2 sensor starts to read lean, and tells the ECU to add fuel. The other three injectors start going rich to compensate for the one cylinder not having any fuel in it. The car will read lean, because it's a single reading at the collector. So the car will read lean, but will smell rich.

 

A way to test which injector is going out:

1) while it is running rough and stumbling, go around to each injector and unplug the wire connectors to each injector one at a time. Whichever injector isn't firing won't make a difference if it's plugged in or not.

2) you could buy one injector and swap one out at a time. Move the one known good injector around to each of the four cylinders until you replace the bad one.

3) send out your injectors to be tested and cleaned.

Edited June 29, 2020 by lil'redwagon

Maybe a dumb question.. were the butterfly valves removed when the tgvs were deleted?

How's fuel pressure?

The TGV's are gutted, welded shut and ground where the shaft was for the valve plates. The injectors are the same I had on the car before replacing the engine (I kept my original intake) I don't believe they were ever used with E85. It did smell like fuel the first time I ran the engine (before my orignal post) when it was misfiring badly but this time it didn't seem to smell. I'll find a way to look at fuel pressure next time I get to work the car. In the log we can see some misfiring on cylinders 1,3 and 4 at about 395 seconds can this be caused by low fuel pressure? Also is there a reason (other than randomness / luck) why cylinder 2 would be the only cylinder not misfiring? I also have the injectors from the used engine I bought I might try to swap them and see if that makes any difference. Thanks for your inputs guys! Edited July 2, 2020 by divadgagnon

So does the car have a tune to delete the CELs set by gutted TGVs?

 

 

The TGVs apparently aid with cold start emissions. Not that it's cold this time of year, but when the car is running in open loop perhaps this is part of the issue.

 

 

However, TGVs are gutted/deleted on my car too and it carries the CELs, and it runs fine.

 

 

Fuel first, if spark has been verified as good which it sounds like probably Yes.

 

 

O2 sensors, when running open loop after it first starts, shouldn't be able to contribute to that bad of a problem.

 

 

Swapping in the other set of injectors is a good idea since they're on hand. Just mark them and associate where the misfires come and go from.

Looking at the log, it looks like the car is just getting leaner and leaner. The fact that there are misfires on multiple cylinders at the same time, indicates to me that it's a fuel delivery problem, not a particular injector. And all the different cylinder misfires seem to correspond to when it goes lean. It's not likely that all injectors start to crap out at the same time.

 

Another symptom that your log shows is that you're *gradually* going lean. Not all of a sudden. Which tells me that there is something slowly dying (like a fuel pump), or overall voltage in the car slowly tapering down to below ~8v, or something getting clogged under pressure (like a fuel filter).

 

If you suddenly went super lean, that would point to an electrical connection with a weak contact. Or an injector that either had power to it or it didn't, so it would spray fuel or it didn't. But in this case, it looks like an electrical motor getting weak (fuel pump), or if there's already a ton of junk in the fuel filter, it all gets stacked up at the output port when fuel is flowing through it. So at first it's fine, but the collected debris gradually impedes flow. It's also possible that your fuel pressure regulator is clogged and impeding the fuel rails. I'd say to rent a fuel pressure gauge and T it in to the fuel line that goes from the firewall to the fuel pressure regulator. Start the car and watch the gauge through the windshield. When it starts to stumble, pay attention to see if pressure drops. If so, it's the fuel pump or filter. You can get a whole fuel pump sending unit as a low mileage STi takeoff for ~$70 on ebay. You'll have to swap out the fuel level sender, but it bolts right up to the outside of the STi cage. Then drop the whole thing in, and you're good to go! But confirm it with the fuel pressure gauge first.

That would make a lot of sense. Does anyone happen to know what's the inside diameter of the fuel hose connected to the regulator? I'll try to swing by the hardware store and pickup a few fittings and a pressure gauge before the weekend I think it would be cheaper than buying a kit off of amazon.

 

As for the fuel pump, I did remove the sending unit in february to siphon out the old gas that was in there for over a year and put fresh gas in. Maybe I did something wrong when installing it back in. Or maybe just the fact that I took it out unlodged some crap and clogged the filter. Might try to take it out again to double check everything. It would suck that my fuel pump is going out... I had a AEM high flow pump that I never installed and just sold it a few weeks ago...

Edited July 2, 2020 by divadgagnon

Unplug the MAF and run the car. It will run off the default values instead of sensor input. If it runs fine, start looking at cleaning/replacing the MAF.

 

From the end of your log the MAF is reading at idle about 40% higher than it should be at idle. roughly 2.5 g/s is normal on a 2.5L.

 

You shouldn't be reading ~4 g/s at ~850 RPM

Edited July 3, 2020 by poconoracing

^ Good catch! But that still wouldn't explain why it's drifting lean - if it was reading more air than it should, then it would be sending more fuel than normal, making it drift rich. It's possible that there could be TWO problems going on, it wouldn't be the first time I've seen coincidental failures that make diagnostics really complicated.

 

It would still be a good idea to test fuel pressure. And if you get it to stumble, try unplugging it while it's stumbling and see if that smoothes it out. And keep logging, it's helpful.