Car misfiring, random misfire code, constant knock counts, oh my.

Alright so my car has developed some misfiring problems.

[Update 1] Halfway fixed, misfiring ended up being a bad PTU but the knock/dogging not fixed yet[/Update]
[Update 2] knock issues seem to be fixed with a new sensor.[/Update]


Cliffs:

Started misfiring the morning then went away never to return all day.
Could drive for 2hrs no issues once it went away
Throwing random misfire code
Multi-bank issue, front/rear afrs peg at 20-22
Also sometimes 7-9 Knock counts as if knock sensor was disconnected but went away never to return
Seems to be getting worse
New plugs did not fix
Swapped coills didn't fix.


First instance:

Started the car in the morning, went a mile and stopped at a light. Check engine came on and the car ran rough. It completely went away and the car was fine the rest of the day.

Second instance:

Car was fine in the morning to work, at lunch it did the same thing but this time, it wouldn't go away quickly. I pull over and check ignition related connections. I was only a mile to my food place so I started limping it. Then all of a sudden it came alive. It had a "random misfire" code stored.

Car ran fine the rest of the day and took it on a 2hr trip.

Third instance:

Did it again on the way home from 2 hr trip a couple days later of the car sitting. It took a few miles for it to clear up.

Decided to change the plugs. This didn't fix anything.

Fourth:

Happened one more time on the way to work and went away but I turned around and parked it. This time the check engine light was kind of flickering. I figured it was just the misfire code.

Fifth:

This morning the car started up just fine, went maybe a mile and the check engine light popped up. I pull over and checked the code, no code? Put up the logger and it's 7-9 counts of knock and my knock light is triggering. I check the knock sensor (it's <20kmi old) connection and all seems good visually. Start the car back up and a couple hundred feet and it goes away and the car come alive again. Odd...

A mile down the road the misfires start to happen. Both banks run lean 20-22 so it's not just one cylinder. No code right away. This time I drive for 4-5 miles and it's not going away. Pull over and check connections etc so I drive it home.

I have a spare coil pack, luckily I can reach all the connections so I plug it in. No changes. I also unplugged my DLI just in case that was messing things up. No change.

Sounds to me an intermittent injector... Or a bad signal wire from the O2s to the ECU...

Sounds like fuel issue. Check fuel pump and hotwire?

Fuel pressure was good on the gauge.

Was talking to Hans and he linked my ECS issue with my misfire issue. Could be a possible bad ECS unit. I replaced the bad caps like 7 years ago but maybe it finally gave up the magic smoke.

damn, I hate intermittent problems like this. Did you make sure the MAF sensor is plugged in well? I had a gasket that was pulled out in the plug a little bit and was keeping it from plugging in all the way.

I am hoping since this last time I couldn't get it to go away, it's here to stay and might be easier to positively fix it.

What knock sensor is in there? OEM, or alternative? I had similar issue with an alternative. Came and went with no logic.

Stock one. Nothing crazy here.

Of course, get home and it runs fine. Gonna try to mist the wires tonight and see if anything is jumping. Could be just that simple.

Quote:

---

Originally Posted by Toni

Stock one. Nothing crazy here.

---

OK, just wondering if it was an oem "like" part.

Flakey MAF sensor?

It'll cause the lean issues and the lean issue will cause misfires. The knock may or may not be a byproduct of this (lean AFR under power). Log the conditions to verify.

Jeff

Well made it to work today no problem. Seems if it doesn't happen within the first 5 mi/5 min, it won't happen at all. Getting on the boost is no problem either.

This doesn't feel like a fueling issue to me. I've had problems with the maft gen II in the past because it used the same shitty connectors as the maft pro. Once you play with it, it comes alive. They sent me gold plated pins to fix it. So that was the first thing I checked. It was essentially a mas problem. However, when that happened, it felt much much different. This feels like an obvious imbalance, like I turned off an injector. Thing is it's happening on both banks.

So now my plan is to log the early starts every time to catch the situation. I have spare coils, PTU, and wires on hand. I think I need to catch the problem quickly because it's probably flooding some cylinders. I'll first try pulling the ECS cause that's easy. Then try swapping things without running the car much. An engine restart hasn't fixed the problem yet so I can be somewhat sure if it goes away, it's something I did.

If nothing fixes it, then I'll let it run for a while and pull the plugs to see which or if all are wet to get an idea if it's a global issue or all.

I could get a scope on the lines to the PTU to see if the ecu is spitting out junk but that will take some setup.

Is your Gen 2 mounted solidly? If it's moving around, that could very well be the problem. I know if I pick mine up while running it'll get stupid from time to time.

Jeff

It's your ECS computer. Disconnect it and the problem will go away. I believe they can be rebuilt, but you can usually find used ones for super cheap. In my experience, the ECS computer going bad is likely to take the ECU with it shortly thereafter.

Quote:

---

Originally Posted by HellBringer

It's your ECS computer. Disconnect it and the problem will go away. I believe they can be rebuilt, but you can usually find used ones for super cheap. In my experience, the ECS computer going bad is likely to take the ECU with it shortly thereafter.

---

Maybe not, don't be so sure of yourself :) Although you do make a good point so I debugged the ECS light last night and it's definitively the computer so I left it unplugged til I get a new one to prevent other issues from it. I've already rebuilt mine once.

Today it happened again even with the ECS unplugged. Drove the car all morning, came back from a dyno day and it happened. Stopped and started it and it didn't go away so I swapped the PTU. Started up fine, no misses, night and day.

Hooked up the old one again and I'll see if it happens again to swap it again to confirm, it may have just be coincidence.

Well it happened again the next time I started the car. Swapped the PTU again and it went away.

PTU provides the rpm signal so that may be what you're fixing.

If I recall Greg has told me the RPM the ecu uses is from the CAS. Actually, a few months back I had a crazy RPM signal and car ran funny but a bit different than this. That's when he told me about it. I'm betting that was the first sign of the PTU going whack on me.

excellent thread. Never knew the POS ECS computer could cause issues...I'll be disconnecting mine as soon as I get a chance.

Quote:

---

Originally Posted by Toni

If I recall Greg has told me the RPM the ecu uses is from the CAS. Actually, a few months back I had a crazy RPM signal and car ran funny but a bit different than this. That's when he told me about it. I'm betting that was the first sign of the PTU going whack on me.

---

Nope if he said that it was wrong. The CAS on the 91/92 provides the crank angle and TDC not rpm. RPM has always been from the PTU.

Well one of you is right that's all I know lol :)

Well, looks like there were two simultaneous issues :( The misfiring and the knock counts.

(For those that may not know, Chrome ecu can light the check engine light on knock counts. Mine is configured for 6. So my 7-9 counts I am getting make the check engine light go on)

Started up the car in the morning and went out about a mile. Knock light came on and the car felt like it had no power. Then it just goes away and the car comes alive.

The misfiring was happening with more frequency and I only saw this knock light issue twice. Three times now.

Quote:

---

Originally Posted by HellBringer

...In my experience, the ECS computer going bad is likely to take the ECU with it shortly thereafter.

---

I fear you may be right and this may have already done damage even though it may not have been connected to my misfiring :(

7-9 Knock count is the sensor failing...

Yeah, or the wiring is shitty, or....the ecu has some gremlins.

Car lasted for like 17 years with it's first sensor and the second goes bad within a couple? lol. Possible though.

Mine was bad out of the box.

Wahhhhat? Really? Would prefer not to spend 100 bones again haha.

Yeah. It was the aftermarket brand one that Ninja sold. He took them off the site after 2 failed new.

Oh, well, tough to compare that to a new oem that has been working. Had me worried that I'd get a bricked oem one.

Hey Toni, sorry it took me so long to post here but the topic of the RPM signal coming into the ECU from the PTU got me second guessing somethings. It's just been so long since I did this that my memory isn't as sharp as it was on the various functions.

At any rate, the RPM signal does come from the crank sensor. Maybe the earlier cars did things differently but with the flash ECU, the crank sensor is what is used to calculate RPM. There is a math function which simply takes the crank pulses and divides them by 3 to get RPM. Its actually a bit more complicated than that but don't feel like writing a novel on my phone about it.

As for the PTU RPM signal input into the flash ecu, there isn't one. The output goes to the tach.

The ECU uses the crank signal pulses to time when to ground the PTU so it knows when to switch coils.

BTW, yes I'm aware of the 91/92 sensor being mounted on the can but that photo style sensor has two outputs. One is the cam signal, the other is the crank signal. On the 93 cars they split the two into individual sensors. The method for RPM calculation is still the same.

Great, thanks for the clarifications. May help someone in the future.

I have a new or different symptom now and I found a log.

New symptom. I normally don't drive with my fuel pressure gauge in, I usually have my oil temp(I can hot swap the gauges in the same location). On the highway from an hr drive I notice the low light comes on and I see the needle freaking out. It was dropping below 2bar differential pressure across boost. Car was running fine but I was just cruising so not sure if I'd notice a quick blip. Could be the gauge, boost/vac signal, or actually dropping pressure. If it's dropping pressure I guess it could be the mfi/ecu, pump, or wiring. Not sure if this is connected or a separate issue.

I found a log of the knock sensor going whack. What's interesting is I get a voltage when it's showing 7-9 counts. It also shows no voltage in some sections. I need to do some logs of it working fine and also unplugged to see how the logs compare.

It does seem to follow the base, so I dunno. Might need your two cents (or a quarter) if you have it Greg.

https://dl.dropboxusercontent.com/u/...cksensor01.png

That sensor is losing power, or the connection is dropping or the sensor itself is bad. No way will you see 0v with the motor running.

I'm surprised that the knock sum didn't go away when it did have voltage.

The zero volt counts are bringing down the average noise (knock base) so when it comes back to life, all of a sudden its out side the norm.

Alrighty. The new fuel pressure symptom was a false positive. Apparently my pressure sender is working fine but the gauge unit is not. Phew!

Misfire problem is gone. 100% the PTU.

The knock sensor issue did not resurface for a long time. I've driven some 3000 miles and two track days at Watkins Glen. However yesterday morning it came back. It was a cold morning and it came on about 2 miles going to work. Man the car dogs without timing lol. I let it cruise on the highway to see if it would go away. Ten minutes, still there. Ok so I have a new sensor with me and I figure it's not fixing itself this time and I'll plug it in when I park. Of course the damn thing goes away as I pull into work. Sure I can replace it and see if it goes away but I generally hate throwing parts at a problem.

There appears to be no pattern with this other than it only happens when I am on my way to work in the morning. Possible temperature thing.

It happened again today and I turned around got home and plugged in a brand new sensor. New sensor didn't change anything. I expected it to fix the problem within seconds or after a restart. Neither occurred. So I took off with the old sensor and it fixed itself within a couple miles. I'll unplug the old sensor, drive a while, and plug it back in to see how long the ecu sees the sensor again just to make sure it's as instant as I think it should be. I fear that it's the ECU being a bitch hole. Maybe I'll setup some probes at the ecu to test it with my meter when it happens to see if I'm getting anything at the harness or if it's ECU internals.

so wait, you are just plugging in a new knock sensor and not installing it? did you reset the ECU to clear the learned values?

also what flash ECU are you running, i had very strange issues with my knock sensor on my first flash ECU, but the sensor was fine.

Alright, I gave in and swapped in the new knock sensor I've had waiting a couple weeks ago. I've tried driving the car every day to work and so far in 2 weeks it hasn't given me troubles. A good mix of dry/humid cold/warm mornings. So far so good. I think it was either the sensor or the connector. I'm going to replace the harness end of the connector to get some new pins to ensure a good connection. So far I think I can say all my issues in this thread are fixed.