Random Misfire

[Jarsky]

Hey all, figured i post here while I'm looking around the internet. I have 93 Subaru legacy that's developed a weird random misfire issue. I'll be driving sometimes and shell be fine, other times it vibrates like hell and sounds like trash, and then it suddenly goes away. I just took the engine out of the car to clean off all the 300K dirt and oil that was on there, replaced the intake gasket and put on an aisin timing kit, and threw the engine back in. At first i thought the issue was a bad alternator, (my alternator was bad), but it turns out this isn't the issue. Crank and cam position sensors were replaced last year as well as a fresh set of NGK plugs and wires. I've read that sometimes the oil pressure sensor can leak onto the crank position sensor so I'm going to check that, and make sure everything is tight but if anyone has any ideas they would be greatly appreciated, thanks for any help.

[jonathan909]

Oh, my.  "Random" - the word that strikes fear in the heart of anyone who's ever had to debug anything.

I should know this, but is '93 pre-ODBII, meaning there are no diagnostics to be had?

Fwiw, I don't think an oil drip is going to bother the crank sensor.

[Jarsky]

40 minutes ago, jonathan909 said:

Oh, my.  "Random" - the word that strikes fear in the heart of anyone who's ever had to debug anything.

I should know this, but is '93 pre-ODBII, meaning there are no diagnostics to be had?

Fwiw, I don't think an oil drip is going to bother the crank sensor.

yes the random is the worst part, and yes 93 is pre-OBDII, not only that but my car has been manual swapped so i have a check engine light 24/7. I tried to see if anything was lose and the only thing i could come up with was the coil pack, but I don't think it was lose enough to cause any wires to come undone. Regardless, tightened that up and went on a drive around the block and it was running right as rain but I have a sneaking suspicion its not the culprit. Hopefully I'm proven wrong. As I said though it is intermittent. I have no idea what causes it to start or to stop, sometimes a good rev to almost redline will make it fine, other times ill get home and it'll just not stop, yesterday I drove down the highway for 5 minutes and it didn't stop till I stopped a stop light a minute from my house.   

Edited May 7, 2021 by Jarsky

[jonathan909]

Man, something that random is going to be a b@st@rd to debug.  I know this is a mantra, but there's a lot of truth to it:  Bad grounds.  Hunt them down and make sure they're all sound.  Bad connectors - but you had the engine out and back in, so you've cycled a lot of them and they should be making good contact.  You have a new alternator, so your power should be solid.  Maybe a flaky ECU?  Is the misbehaviour isolated to the engine, or are there any other peripheral symptoms?

[1 Lucky Texan]

yeah, came here to suggest refreshing grounds.

 

if you could rig a vacuum gauge to observe when driving, might spot something when it misbehaves....?

[Jarsky]

1 hour ago, jonathan909 said:

Man, something that random is going to be a b@st@rd to debug.  I know this is a mantra, but there's a lot of truth to it:  Bad grounds.  Hunt them down and make sure they're all sound.  Bad connectors - but you had the engine out and back in, so you've cycled a lot of them and they should be making good contact.  You have a new alternator, so your power should be solid.  Maybe a flaky ECU?  Is the misbehaviour isolated to the engine, or are there any other peripheral symptoms?

I refreshed a decent amount of the grounds a year ago when I did the awd manual swap but ill look again. the only symptom that ive seen, yet again a random one, is that sometimes ill be driving and the check engine light will flash, and for the faintest second the rpms drop to 0 the rocket back up to where they were. besides that radio works fine, I have a aftermarket rad that I had to wire myself but that's also fine. one thing ground wise, the grounds toward the front of the car on the lower flat bar of the engine bay, they would be right in front of where this picture ends if that makes sense(that was a crap description but idk how else to explain), i believe they are for the oem rad, I do have those unhooked because my rad has its own grounds, could that possibly be the problem? I figured because there's no rad in the car it would be a big deal but now I'm not too sure.

[jonathan909]

First of all, what the hell does a rad need grounds for?  I get where you mean, but they're scattered elsewhere as well, e.g. on the shock towers (aren't they the two bolt heads I see peeking up there over the rim of the stamping?  My '99 OBW has a couple there, on both sides.).

But that CEL flash and the brief drop in revs?  I'd be thinking bad ECU, maybe bad power.  Investigate your relay box.

Edited May 8, 2021 by jonathan909

[Jarsky]

35 minutes ago, jonathan909 said:

First of all, what the he11 does a rad need grounds for?  I get where you mean, but they're scattered elsewhere as well, e.g. on the shock towers (aren't they the two bolt heads I see peeking up there over the rim of the stamping?  My '99 OBW has a couple there, on both sides.).

But that CEL flash and the brief drop in revs?  I'd be thinking bad ECU, maybe bad power.  Investigate your relay box.

Well I'm guessing so it doesn't get fried. Yes those are the shock tower ones, i looked those ones over but ill have another glance. Ill take a look at the relays tomorrow. One other thing i just remembered right behind the battery is where the harness connects from engine to chassis, the bolt was a bit loose so there was slight play, as i was tightening it up the rpm went up and down a bit. I don't think it was that as I was driving with another engine in the car and it had no bracing to the chassis but i figured it was worth mentioning. 

[jonathan909]

Rad:  Fried by what?  Rads aren't electrical any more than the hose running to your lawn sprinkler is.

Yeah, you'll want to look seriously at wiring e.g. the ground you describe.

[edit]

And don't ignore the possibility that you could have a bad keyswitch.  It's a single point of failure for the entire electrical system.

Edited May 8, 2021 by jonathan909

[john in KY]

Fuel pump?  If still original. it's what, 30 years old?

[jonathan909]

Anything's possible, but the blink-of-an-eye outage smells electrical to me. 

Still, the "vibrates like hell and sounds like trash" could be momentarily running too lean, no?  Any backfiring?

Edited May 8, 2021 by jonathan909

[Jarsky]

2 hours ago, john in KY said:

Fuel pump?  If still original. it's what, 30 years old?

its got a sti tank and a basically brand new pump. 

[Jarsky]

2 hours ago, jonathan909 said:

Anything's possible, but the blink-of-an-eye outage smells electrical to me. 

Still, the "vibrates like hell and sounds like trash" could be momentarily running too lean, no?  Any backfiring?

backfires basically all the time, stock headers and cat and some aftermarket mid pipe but i forget the brand, i need to put a muffler on it but it does make quite a lot of pops for having basically a stock exhaust. Might have something to do with having a sti/wrx fuel pump but im not sure

[Jarsky]

3 hours ago, jonathan909 said:

Rad:  Fried by what?  Rads aren't electrical any more than the hose running to your lawn sprinkler is.

Yeah, you'll want to look seriously at wiring e.g. the ground you describe.

[edit]

And don't ignore the possibility that you could have a bad keyswitch.  It's a single point of failure for the entire electrical system.

just so i don't feel dumb, Like where I put the key in right? i never thought about that if it is that. 

[jonathan909]

Yes.  Not saying that's the problem, just that it's a mechanical switch like any other, the contacts wear over time, and yours is really old.

[forester2002s]

+1 on the ignition switch being a possibility. It's easy enough to wiggle the key when the misfire occurs, to see if it makes a difference.

[1 Lucky Texan]

cooling system electrolysis can be problematic I think.(haven't read-up on it, just seen it mentioned in the past.0

I could see where grounding the rad could be decent insurance to help maintain the coolant at ground potential.(relative to the engine)

 

 

Edited May 8, 2021 by 1 Lucky Texan

[jonathan909]

I don't want to get into the weeds on this, but I'm not buying it.  Lots of things get "mentioned" that are complete nonsense.

The coolant is constantly circulating through the engine, so I think it's succeeding in keeping itself pretty damn grounded wrt the engine, regardless of whether the rad core is insulated from the chassis.  I'd need to hear a pretty convincing argument to think otherwise.

[el_freddo]

2 hours ago, jonathan909 said:

I don't want to get into the weeds on this, but I'm not buying it.  Lots of things get "mentioned" that are complete nonsense.

The coolant is constantly circulating through the engine, so I think it's succeeding in keeping itself pretty damn grounded wrt the engine, regardless of whether the rad core is insulated from the chassis.  I'd need to hear a pretty convincing argument to think otherwise.

The issue is when there’s a stray electrical current in the cooling system that causes electrolysis. It legit. Loads of science behind it. 

Other reason to ground your radiator (not warranted here) is when your radiator temp switch is earthed to the brass radiator as they are in the EA81 MY Subaru’s. 

As for the issue at hand: almost sounds like some broken wires doing dodgy stuff. That’ll be fun to find. Has the vehicle been in a crash at some point? 

As for the manual conversion. Read up on the wiring diagram and delete the unnecessary pins at the ECU that relate to the auto (or is it ground a certain pin to tell the ECU it’s in manual mode, I can’t remember but a gregory’s manual has a good wiring diagram I follow for loom cut downs) and remove auto TCU if you haven’t already. 

What gear does the ECU think the auto is in (if that’s even a thing)? 

Cheers 

Bennie

[jonathan909]

6 hours ago, el_freddo said:

The issue is when there’s a stray electrical current in the cooling system that causes electrolysis. It legit. Loads of science behind it.

I have plenty of experience with trouble from ground loops and very small circulating (as in electrical, not liquid) currents, so I get what you're talking about.  What's important is that these problems are almost always solved by selectively removing excess grounds, and never by tossing in a bunch of extra new ones without very carefully analyzing the problem, or you can often make it worse.  More grounds means more possible paths for those circulating and stray currents, and in this case adding a location (i.e. the rad) for possible electrolysis and/or corrosion that didn't previously exist.  In other words, let the rad float (electrically speaking) and it won't give you trouble, even if it is a few millivolts away from the engine.

(If there are any vintage microcomputer geeks listening, a formative experience for me in solving this kind of thing was about 40 years ago, working with a client who was building an early GIS system - it had to do with potential oil spills.  Their system had outgrown the original (10-slot S-100) North Star Horizon chassis thanks to a whack of Matrox graphic cards.  So I migrated them into a big rackmount (22-slot) Cromemco Z2-D box, which also meant duplicating in wire-wrap the serial and parallel I/O that were on the Horizon's motherboard (unusual at the time, as most S-100 backplanes had no active circuitry other than (perhaps) termination).  They needed a hard disk, and this being a few years prior to the introduction of the very first (Shugart) 5meg 5" Winchester, we were using a 15meg 8" drive - a Priam or something.  That's the part that was giving them trouble - the disk just wouldn't settle down and behave reliably.  Disk read amplifiers are quite sensitive to noise, and I traced the cause to a couple too many grounds in the disk system that had been added and/or neglected by the North Star designers.  Once I'd had some time to stare at the whole picture, I took a reasonable guess at which grounds had to go, and when I removed them the noise disappeared and the system sprang to life - and gave them years of pretty impressive service.)

So I'm still skeptical, but might be convinced if someone can direct me to any solid references on the subject describing this application.

Edited May 8, 2021 by jonathan909

[1 Lucky Texan]

well, Not sure my radiator has any specific ground wire. And yes, theory is not the same as practicality. But moving fluids do create friction and that can lead to static charge.

 

Maybe measure from the battery or chassis ground, and a probe touching only the coolant. Maybe try with engine off, vs running at operating temps? Still, there are multiple formulations nowadays for coolants as well so, yeah - could be 'lost in the weeds' time.

Edited May 8, 2021 by 1 Lucky Texan

[idosubaru]

OBDII is meaningless here - pre OBDII stuff still gives codes. When you read the check engine lights what does it show?

If it stumbles take notice if its:

1. recently wet/humid

2. happening when you let off the gas entirely (throttle closed)

3. is it ever hard to start or restart?

Do you have access to spare parts?

Hate to suggest guessing but MAF sensors are easy to swap on EJs and I’ve seen them cause starting and drivability issues without throwing a code  

The grounds are probably to the frame rail, I do t recall Subaru grounds on any radiator I’ve worked on. But maybe that’s indicative of my memory?!? Lol

It’s unlikely the grounds anyway. Almost every Subaru I see and drive has corroded and detached grounds and runs fine. Yes they can be problematic but that pic looks like a clean car, engine came out and you had a good look at it and it starts.

[1 Lucky Texan]

would a '93 have a blink code?

[el_freddo]

22 minutes ago, 1 Lucky Texan said:

would a '93 have a blink code?

For the want of a better name, yes. Plugs under the dash need to be connected and the long and short flashes counted to make the code. This is how I’ve always done it. 

Works on OBDII too from memory. I’m sure I’ve used this on my sister’s Gen3 liberty and my MIL’s SG foz.

Cheers 

Bennie

[Jarsky]

19 hours ago, el_freddo said:

The issue is when there’s a stray electrical current in the cooling system that causes electrolysis. It legit. Loads of science behind it. 

Other reason to ground your radiator (not warranted here) is when your radiator temp switch is earthed to the brass radiator as they are in the EA81 MY Subaru’s. 

As for the issue at hand: almost sounds like some broken wires doing dodgy stuff. That’ll be fun to find. Has the vehicle been in a crash at some point? 

As for the manual conversion. Read up on the wiring diagram and delete the unnecessary pins at the ECU that relate to the auto (or is it ground a certain pin to tell the ECU it’s in manual mode, I can’t remember but a gregory’s manual has a good wiring diagram I follow for loom cut downs) and remove auto TCU if you haven’t already. 

What gear does the ECU think the auto is in (if that’s even a thing)? 

Cheers 

Bennie

the car hasnt been in a crash, and basically all the ecu side of things is good, the ecu thinks the car is in neutral 24/7 but ive been driving the car with the swap for months.