Mystery Misfire (Now w/Video)

[Barnes]

I've got a misfire I cannot figure out. Pretty much desperate at this point.

Symptoms: Car had a code44 off and on for a long time. Misfire is severe at cold idle, sometimes the car stalls. Warm idle the misfire clearly happens, but the engine can barely hang in there. I'm pretty sure it happens at all ranges as my on boost performance doesn't seem to be right.

It appears (and the code 44 would indicate this) that and entire coil is simply not getting the fire signal. This misfire can go for several seconds! Not just a single spark missed, long stretches of seemly no spark.

I'll try to get a video to post.

Config:
Stock coils (*WAS* a COP setup.)
3g MAF
660 denso injectors
Evo3-16g
AFRP set @ 38psi
Stock long block except for a cyclone intake.
Walbro 255 pump
Eprom ecu w/keydiver ecu (CAPS have been replaced)

Diagnostics and adjustments so far:
Boost leak checked
TPS adjusted to 0.5v/10%
Set base idle and timing as best as possible
6-7 ground locations cleaned down to bare metal
FSM check of transpack wiring(except pins 2 and 7, because I'm not sure how to do it.)
FSM check of coil pack wiring
Cleaned the connectors for the coil, CAS, transpack thoroughly.
Listened to the injectors with a stethoscope and they sound good.
Injector resistor pack checked out.

Items swapped with no change
2x stock coil packs
1x trans pack
New Spark plugs (NGK BPR6ES) @ 0.028" (dielectric grease used for sealy)
New Spark plug wires (NGK)(dielectric grease used for sealy)
1x green top CAS
1x socketed ECU with my chip installed (thanks CarRacer)

Per some old posts by BluFalcon (Ed) who had a similar problem, he found a short in his wiring harness. So just to completely isolate that possibility, I used some parts from an old wiring harness, and built some jumpers between the ECU and various parts.

Stock wiring completely bypassed at this point:
ECU to transpack (Pin 54 to 7, and 55 to 2)
Transpack to coil (1 to 1, 8 to 1)
Coil directly to battery positive.
Transpack ground straight to ground.
I've also added several grounding jumpers between the ECU, battery, engine, and chassis.


So after ALL this, I still get the same exact symptoms. The code 44 isn't back, but I haven't done more than idle the car in the garage. There are a few remaining things that have been constant through the whole process, and some other things I can't explain.
1)Keydiver chip (although I swear the car did this on the stock ecu)
2)ECU voltage supply i.e. MFI relay. I haven't checked it.
3)O2 doesn't cycle at idle, runs in the mid to upper 0.9v range. Although I think this is normal if you are rich, indicating spark misfire.
4)Gas is kinda old, BUT I'm pretty sure I put stabilizer in there. Also, that shouldn't affect a code 44 misfire.
5)Right as the misfire starts, I can see the injector pulse-width on the logger jump from ~1.5ms to 3ms. Now, I'm not sure which comes first, the misfire and idle drop, or the IPW jump. My guess is the former because it's dumping in fuel to try and bring the RPMs back up.

One of my last options is to buy/borrow an oscilloscope which would help give a more positive diagnosis. Trying to avoid that however.

I'm going to try and gap the plugs down further. It's an unlikely 'fix', but it's easy and quick. Any other ideas? I'm ready to throw in the towel. /ubbthreads/images/graemlins/frown.gif

 

[G]

So are you running a cop setup or spark plug wires?

 

[Barnes]

Sorry. I *was* running a COP setup, now I'm running a stock coil w/plugs wires setup. Appears to be no change between the two.

 

[FlyingEagle]

Might I suggest a compression test, and leak down test.

Assuming everything in the deliver and management system is up to par, mechanical issues are next on the list.

 

[Barnes]

First start, cold. I didn't press the gas pedal until a restarted the car. So before the stall, i'm not touching anything.

Outside the car after it warmed up enough to keep running.

The other odd thing I noticed, was that if I let the car warm up, and leave it alone for a long time, it seems to idle perfect. I let the car idle for 10-15mins, and eventually it seems to stabilize and have no discernible misfire. /ubbthreads/images/graemlins/confused.gif

 

[ktmrider]

Quoting FlyingEagle:
Might I suggest a compression test, and leak down test.

Assuming everything in the deliver and management system is up to par, mechanical issues are next on the list.

This.

In the Scoobaru world, after doing all you have on the ignition system, it usually means a cracked piston ringland or burned valve(s).

 

[manikbastrd]

This may be a dumb suggestion, and I didn't thoroughly check your posting, but have you checked your Idle air control/Idle speed control? This can give your car idle surge, and if you are running rich, it could throw it off enough to make the engine die. Also, since it runs fine when its warm, that also points me in this direction of thinking. Also, I don't know if this is what you mean when you say trans pack, but the PTU could be taking a dump or not grounded well...Just my $.02...Best of Luck!!! /ubbthreads/images/graemlins/dunno.gif

 

[toybreaker]

Quoting Barnes:

3)O2 doesn't cycle at idle, runs in the mid to upper 0.9v range. Although I think this is normal if you are rich, indicating spark misfire.



^^^ This statement confuses me a little

The o2 sensor can only sense exhaust oxygen content, so I would "expect" the o2 on an engine with an ignition mis fire would read lean as a dog who's missed a week of meals.

No combustion would mean the oxygen content of that particular cycles gas as it passed the o2 sensor would be very high, leading to a low o2 voltage. (the oxygen would NOT be consumed during combustion and would be read by the o2 sensor, driving the o2 voltage down to effectively nothing)

... just something to think about /ubbthreads/images/graemlins/wink.gif



I think a good pic of the business end of the sparkplugs would tell many tales.



It's worth noting that the code 44 parameters are only about the spark side of things.

Jeff O has made several postings in the past that come up in a search (using "code 44" as the search terms) about how the system works

... really miss that guy! /ubbthreads/images/graemlins/frown.gif


In a nutshell, (if'n I understand his postings correctly) the ecu "sees" the inductive kickback on the tach driver circuit.

... no spark = no inductive kickback = a code 44 (under certain conditions)



fwiw, I've had to chase "unsolvable" misfires problem on several cars, and each time it was something different.

Basic maintainance, correctly gapping / replacing sparkplugs, cleaning corroded terminals, replacing bad igntion transistors and addressing flaky wiring "fixed" most of them.


... and swapping in a known good ecu fixed *most* of the rest.

as noted by Jeff, the trace for this on the ecu board runs near one of the caps and that trace may have been damaged / compromised by electrolyte on the ecu board.



It's really cool that James lent you an ecu for the test by substution method, /ubbthreads/images/graemlins/applause.gif

... but I have been burned more than once troubleshooting ecu's by swapping in a not so perfect replacement ecu in similiar scenerios where the ecu is "suspected" to be the problem. I would strongly recommend trying one more known good ecu just to be sure that it's not a damaged ecu coil driver or tach trace.




I did have two misfire problems that kicked my ass pretty hard.




One had too much anti-sieze on the sparkplug threads and so the ground side of things down at the plug was sub-optimal. That car would "burble" a lil at idle (kinda like what you got going on,) and kinda got a lil better when things warmed up, but it really showed up under load / boost.

... the damn thing just felt soft (for lack of a better term) under boost

When NGK (and other manufacturers) changed from cadnium plated plug shells to some kinda nickle type of plating, problems on the ground side of high energy ignitions became more prevelent.


This lack of a good ground phenomenom is most notable when using the silver antisieze products. They are engineered to cut down on diss-similiar metal corrosion / galvanic action on bolted joints, and so by their very make-up and nature make for poor grounding (especially when used in excess!)

You can usually get away with it, but sometimes, it can and will bite you in the ass when the plug shell doesn;t ground well.

Since you have to do something to prevent the steel sparkplug housings from siezing in the aluminum head, anti-sieze is a necessary evil, but it can be taken to extremes.

[pro tip] Try and always use copa-slip or copper based antisieze on your spark plugs (or anything you want to be electrically conductive /grounded)

After it clears up, how does this thing run under higher loads?

If it always runs a little soft, be sure the plug threads in the head are clean and that the plug seating surface in the plug wells are clean and free of dirt / grease / oil / other contaminates.

The entire ignition circuit has to be considered when chasing ignition misfires, and half of that is on the ground side. This means optimising the interface of the plug housings where they ground to the head will pay great dividendes when chasing this type of problem.


Optimising the interface between the plug and the head will pay great dividends, and it should be addressed at each and every plug change.





The other one that completely kicked my ass (as well as the entire dealership full of sharp cats before I saw it ) had a failing / failed ignition noise reducer / condensor / cap (little round guy up on the intake). It was quite awhile ago, so I dont remember all the specifics, but that lil devil can do some strange things, and some of them will result in a misfire (with a code 44 set in the ecu.)

One "tell" for a problem with this component is a whole sh*t-ton of rfi / static on the am radio band. Turn on the head unit, tune it between am stations and rev the motor. If it sounds "sizzly" and increases with rpm, the condensor may be weak / failing


Inspect the wiring to that component, (the wire where it enters the condensor body is often fatigued, and you may have a stray wire strand grounding out things against the intake)

If no problems are found on the can itself, check the the car side of the harness really well, it can chafe and cause grief with little stretches of frayed wires / strands touching the intake manifold / brackets and whatnot

A quick easy test is to just swap it out with a known good one, making sure to get the wiring up and away from the intake.





best advice

Use your logger / other whiz bang device to shut down an injector at a time and see if the problem localises to one injector.

^^^ do this first ^^^


Then, break things down and work thru the system logically

Clean and inspect each and every electrical connector in the ignition system, including the terminals / connectors down at the ecu (clean the grounds while youre there!)

really look at them, make sure there's no corrosion, and that none of the female connector pins are spread open / distorted. You can use "De-Ox-It" (available at radio shak) on any that look corroded / wanky to get them cleaned up and serviceable /enhance conductivity.

Inspect the wiring throroughly, especially at the back of each connector. Stress concentration can break the individual strands where the wire enters the connector body, resulting in reduced continuity



Make sure there are NO vaccum leaks!

Check each and every hose / connection and solenoid!

Some solenoids have cracked nipples, and a cracked fpr solenoid will flat ruin the fuel pressure compensation, making things run fat enough under vaccum to give a misfire like you are describing

... although it would probably get worse as the car warms up, (and your car does not appear to be doing that), it's still worth checking the fpr solenoid and the diaphragm in the fpr itself, just to rule out any problems there.



Dump the fuel that's in the tank if it's at all old, questionable (or possibly contaminated)

Fill it with some fuel from a busy gas station.

Get the engine good and warm, boost it a few times to get the firing ends cleaned up.

Shut the thing off, and then the next day, cold start it, let it run for a minute or two (while it misfires) then pull them out and take good pics of each firing end

Number them so we know which plug came from which hole, and that will help us determine what's going on in the combustion chamber and where to go next.

[edit] post up what the fuel trims are doing

 

[Barnes]

Haven't read your entire post Toybreaker, but your comments about anti-seize jumped out at me. I've *always* used nickel anti-seize on my threads, and done a VERY thorough job of covering the threads. The fact you had an actual misfire issue related to that means I'm going to try that next!

 

[Barnes]

Tried cleaning off the anti-seize. No change. :/

Next step I guess is gap the plugs way down. If no change, then I'm putting in a stock ECU, injectors, MAS. I'm also going to try and attach the intake pipe straight to the TB elbow per Mitsuturbo's suggestion. Just for good measure eliminate the turbo/ic piping/ic track.