Pulled the CEL codes... Now the fun starts

[86BRATMAN]

CTS=Coolant Temp Sensor

 

Have faith, the car is 20 years old and problems will arise. My most recent purchase was the cause of grey hair for an older fellow for 5 years that he couldn't get it to run right. Turns out it had a corroded junction for the wires to injectors 1&2. All it needed was a fresh set of eyes, solder and a two inch section of wire to be a reliable daily driver. Keep your cool and go at this one logicly. It will work out.

[psychsurf]

Thank you, I appreciate the encouragement. I'm certain it is something simple, it is just the diagnostic process that intimidates me. I'm more of a concrete mechanic, "head gasket blown? Let me get a torque wrench!". Electrical issues just make my butt pucker.

 

Now, to get it home... Tow rope, tow dolly, or tow truck?

[Gloyale]

My most recent purchase was the cause of grey hair for an older fellow for 5 years that he couldn't get it to run right. Turns out it had a corroded junction for the wires to injectors 1&2.

 

 

This is excactly the type of issue I was thinking he may be having. Those crimped jujnctions are a real weak link in the electrical system. I've seen them cause problems with power windows, TCU function, and fuel injectors.

 

I can't think of where there are junctions that would affect KNock, CTS, and Air Flow meter all together. Hmmmm.......

 

Ah HA! Of course! Intake manifold bolt ground wire!..........maybe

 

It junctions with about 8 little black wires all crimped toghether somewhere in the harness

[psychsurf]

Ok, called a tow truck and went to go pick it up. Got there and the shop owner is like, "well you could just take a chance on driving it home".

 

I said, but I thought wasn't running?

 

"well, we had the computer hooked up the oscilloscope and it was outputting nothing. Then we tapped on it, and it started outputting. So we just turned the key and tapped the ECU until it started. Need a new ECU, probably got a loose transistor or something"

 

I say, are you sure it wasn't a short somewhere under the dash?

 

Nope, definitely the computer.

 

Now it's starting just fine, so I'm back to get my wife out of bed to give me a ride to pick it up.

 

The big question is, does what the mechanic says about tapping on the "bad" ECU make sense? Or is it as simple as I just got worked by an unscrupulous mechanic for an hour of "diagnostic" time?

[psychsurf]

Ok, got it home no problem... Not a hiccup along the way.

 

I'm scratching my head.

 

Anyone know if that ECU stuff was bogus?

 

Anyone want to buy a subaru?

[Cougar]

I tend to think the ECU is ok myself. It would be nice to know if there was any spark to the plugs when this problem occured. I would suspect the ignition relay may be the real problem but that is a guess. If the ECU really is the problem then you should be able to find a replacement for around 30 dollars at a salvage yard.

[Skip]

I tend to think the ECU is ok myself. I would suspect the ignition relay may be the real problem .

 

I'm on Cougar's team here.

 

At least find this relay (way up under dash above the ECU)

Maybe "Fonzi-ing" the ECU closed the relay.

[psychsurf]

What about what the mechanic said about ECU output to the oscilloscope? Does that sound like a legit thing, or blowing smoke?

 

Its funny, it is starting and running as good as ever now, like it never had a problem at all. Gonna make it hard to figure out what went wrong before.

 

I actually figured out one thing today, though. The connection to the knock sensor is really loose. I went to check it, and it came off the sensor with almost zero resistance. I can't imagine it's making a good contact to that pin on the sensor. Figure I can just crush that barrel connector a little to make it grab the pin better?

 

 

On the downside, I'm getting major stress from my wife about this car still. Not a good decision from a marital bliss perspective. Seriously, anyone want to buy it?

[daeron]

At least find this relay (way up under dash above the ECU)

Maybe "Fonzi-ing" the ECU closed the relay.

 

Thanks, Skip!! I needed that tonight, it had me on the FLOOR...

 

What about what the mechanic said about ECU output to the oscilloscope? Does that sound like a legit thing, or blowing smoke?

 

Its funny, it is starting and running as good as ever now, like it never had a problem at all. Gonna make it hard to figure out what went wrong before.

 

I actually figured out one thing today, though. The connection to the knock sensor is really loose. I went to check it, and it came off the sensor with almost zero resistance. I can't imagine it's making a good contact to that pin on the sensor. Figure I can just crush that barrel connector a little to make it grab the pin better?

 

 

On the downside, I'm getting major stress from my wife about this car still. Not a good decision from a marital bliss perspective. Seriously, anyone want to buy it?

 

It sounds to me, that his comment about the oscillioscope indicates that there was no coil activity (ie no spark) while this problem was occurring... and it sounds (although I may be wrong) like he DOES know an 80's subaru from a hole in the ground.. That bit of knowledge about the 87 is a fairly subtle bit of trivia to know (and not so trivial, at that.)

 

Patiently and carefully un plug every single plug you can find, and look for corrosion and green. If you comb through the car and under the dash at the ECU and find nothing, then start the car. Wiggle every inch of every piece of wire you can lay your hands on, and listen to the engine for bucking and hiccupping. Pay especial attention to the circuits which have thrown you codes. Replace your battery cables if they look even remotely questionable; the factory terminals were made out of stamped copper and bolted onto a large cable lug.. so both old cables could be used as alternate grounding straps, (battery to chassis/rad mount, engine to chassis, or alternator to chassis) which can ONLY be a good thing.

 

http://www.ch601.org/engines.htm Not sure how much relevance the EA82 partial FSM there has to your turbo car, but it will be of SOME assistance in the meantime. The FSM that the PDFs were excerpted from covered the turbo models, but not all the turbo troubleshooting info is there.

 

Don't kick yourself too hard for getting the turbo.... its hard to say no to some words. "Turbo" can be one of them; I am sure you are familiar with others

 

Interesting revenge idea: find a valve core wrench, fly back to dudes house in the middle of the night, and pull all four valve cores from his tires. Dude would TOTALLY never guess it was you, heh....

[Cougar]

The ignition relay supplies power to the ECU. If the connection in the relay is bad the scope would have showed there wasn't any output from the ECU but that doesn't mean the ECU was really the problem. It depends how well they tested things. I'm am going with the percentages by suggesting the relay as the trouble.

[psychsurf]

Ok, sounds like it is time to dig up into the dash and play with that relay a bit... Mabye it just had a bad ground or something? Add relays to the list of things I know nothing about.

 

I've been working every day, so I haven't had much time to play with it. But interestingly enough, I haven't had a single problem with it since it came back from the mechanics. Not one. Starts, runs, not CEL, no lurching, no nothing.

 

On the progress side, I'm 99% I've got the knock sensor issue licked. It was so forehead-slappingly simple I almost hesitate to mention it. But I will, just the same.

 

I went to check the connection on the KS, and it was loose. Like really loose, like it came right off the sensor without any resistance at all. So I fiddled with it and fiddled with it and fiddled with it... Finally I noticed the o-ring that goes around the top of the KS. And I noticed that when I was trying to seat the connector, it was stopping right at the O-ring... A little grease and a LOT (like a LOT!) of downward pressure, and it pushed down that last 1/4" to fully seat. I disconnected the battery for a half-hour to clear the codes (this IS one of the models that works on, right?), and tomorrow I'm going to go run all my errands in it and see if any of the codes come up again.

 

I'll keep you all updated, and thanks again for the help!

[psychsurf]

Ok, I did a little more searching and reading on the ignition relay, hopefully this will help me isolate the problem (which doesn't actually exist right now!).

 

If it was the ignition relay, this would explain the ECU not getting power (not showing any output, as the mechanic said), and it would explain why the check engine light was clicking on and off when the key was in the "on" position. If the relay is bad, the CEL will be off with the key on, right? It would also explain why a knock to the underside of the ECU might have caused the relay to switch on again.

 

Ok, I think I've got that. I'll definitely check into that.

 

Question on that: Can the relay "kind of" fail? Like where it will go on and off, and refuse to push power through, and then start working flawlessly again the next day? Is there anything else that will cause the CEL to go on and off with the key in the on position like that? When I read that part about the ECU getting power, it did seem like a strong indication that it might be the relay or the ECU.

 

I almost wish it was still not running, so I could more easily figure out what I'm supposed to be looking for! I almost wish that. :-\

 

Anyway, that's a good start for the issue it was having at the mechanic's.

 

 

Speaking of problems its not having anymore, what about the bucking/power loss it did the day before I took it to the mechanic? Could that potentially be caused by the same relay, or do I need to look elsewhere for that problem?

 

Again, the bucking took place right when I let off the clutch and pushed in the gas on upshifts, and the CEL would flash on as it bucked, then go off when it began running normally again. It did this a couple of times, then died at a stoplight with 1/8 tank of gas, again as I was letting off the clutch and taking off. After starting it (it started after maybe 20-30 seconds of cranking) I drove it to the next gas station, keeping the rpms high to keep it from stumbling. After filling the tank back up, no more problems. It seems like a fuel problem, but before I go spending the money on a fuel pump, I want to triple-check. I have a can of seafoam, should I run that through it and get the tank back down to 1/8 again and see if I get a repeat performance?

[94Loyale]

Well, if it helps at all, here's the symptoms I had when my Ignition Relay went bad. On quite a few occasions, when I would try to start the car, it wouldn't fire up. It would take a few on,off,on,off with the key until it would start. Then while driving, the engine would just die completely, out of no where, and then not start. Then I would have to do the same thing to get it running. Then, eventually, it would not start at all. Keep in mind, when the car was running, it was perfect, but sometimes it would die, and when you pushed the clutch in, it would idle, and then drive fine again. It took me a while to figure out, but with Cougar's expert advice on electronics, I found the problem to be the ignition relay. I was not able to find it at any parts stores around here, but it is on http://www.rockauto.com for like $30. Its brown, and has 6 pins, located above the ecu. Once you pull the ecu down, you can look up and see it, its up there pretty good,haha. But your problems seem alot like mine. Maybe just go to a junkyard and grab one or two. They don't go bad very often.

[psychsurf]

Actually, 94loyale, I just read your diagnostic thread. Very informative, and good on you for keeping after it like that. Have I mentioned I HATE electrial problems?

 

As for getting new used relays, unfortunately, I'm in the south where we're not blessed with good subaru boneyards. That, along with my wife giving me hell for each additional dollar I put into the car, makes me want to do as much diagnostics as possible before buying parts.

 

Oh, I thought of one more thing I remembered from when it wouldn't start at the mechanic's, just want to make sure I mention all the symptoms I can in case something jogs someone's memory. For whatever its worth, when the car wouldn't start and the CEL was clicking randomly, the fuel pump was coming on and off with it.

[Gloyale]

For whatever its worth, when the car wouldn't start and the CEL was clicking randomly, the fuel pump was coming on and off with it.

 

That's what it does when you have the green connectors hooked up. It is a test of both the CEL light and the Fuel Pump. It is part of the D-check mode. Didn't I say that already?

 

And there is no need to Unhook the battery to clear codes. Simply plug in both green and black connectors, tart the car, run engine over 2000 rpm for 40 seconds or more. the CEL will then blink signaling cleared codes(unless an active code is still setting)

[psychsurf]

Ah, yes, I do remember you saying it sounded like the test connectors were hooked up. I thought the light and pump were supposed to go in one-second pulses, though.

 

Man, I wish I'd opened up the dash when it was doing that and checked to see if the connectors were connected. That possibility really throws a wrench into my figuring.

 

I didn't connect the green connectors, so if they were hooked up at that point it would be because the mechanic had taken the initiative to presumably check for codes, then forgotten to unplug them before buttoning the dash back up. All this before even telling me the car wasn't starting well or asking me if I wanted him to check it out. Or maybe he was wanting to work me for some more cash, and did that to screw with me. It is "interesting" that the only time the car hasn't started RIGHT up was when it was at the mechanic.

 

I almost hope that was the case. It would certainly be the neatest solution, but there being no way to verify it is pretty worrisome. That, and when I did open the dash to get the codes again when I got home, it had obviously been dug around in quite a bit.

 

Ugh, this is making my head hurt.

 

So, just to rule it out, is there anything other than D-check mode that could have caused what I was seeing there?

[Cougar]

A word about simple relays.

 

They are used mostly to switch power on to circuits that require fairly high current. They are switched on using a low current control circuit that energizes a magnetic coil. When the coil is energized it usually closes (or in some cases, opens) some contacts that complete the circuit for whatever is being controlled. These contacts can get worn over time due to arcing and can fail intermittently or completely. Tapping on a bad relay can sometimes reveal an intermittent problem. They are a very common source of trouble in automotive electrical circuits.

[Gloyale]

You can make you're life easier in the future by doing 2 things.

 

First, fish out the test and read connectors(green and black) and pull them through the fuse box opening. Then they are right there and easy to access without pulling the dash panel

 

second, knock out the little plastic recess in the very center of the lower panel. This will give you a window to view the ECU monitor light.

 

Now, the green connectors make should make the CEL and fuel pump cycle on and off, I believe, in 2 second intervals. If it truely is more random that that, it could be a sign of a short or a relay problem.

 

Think about the way the ECU functions at start up. When you first turn the Key "on"/eng. off. The CEL comes on and the fuel pump pressurizes the system with a short pusle. Once the car start the CEL goes off. But if the car were to stall, the CEL comes back on.

 

SO... if the ignition relay truely is bad, then the CEL could flash for a moment, car stumbles, as the ignition relay cut's power to the ECU. Happen when you let the clutch out? Perhaps the clutch pedal linkage is rubbing against omething......like the ignition relay.

[psychsurf]

Well, I don't know...

 

I cleared the codes, then started it up and went through and wiggled every connection I could lay my hands on. Not a hesitation, not a stumble, not a problem.

 

Took it out for a drive and ran the snot out of it, trying to elicit the same response on shifting. Not a hesitation for a second. Got back home and checked for codes, and they're all clear.

 

I'm stumped. I'm going to have to wait for it to die again before I can fix it, aren't I?

 

Guess I'll just put as many miles as possible (I commute 80 per day) on it, and see what happens.

 

Here's an alternate theory... The guy I got it from suggested that the problem sounded like a loosening disty set screw, then when it wouldn't start it was totally loosened, and the mechanic just didn't want to let on it was something so easy (=cheap) to fix.

 

I dunno. I guess it is running good now, I should just be happy about that.

[94Loyale]

Well, hopefully this week I'll head up to the junkyard and pull an Ignition relay for my car. While I'm up there I'll grab a spare or two. If you want to try one, I'd be more than happy to just mail one down to you, after I make sure it works in my car.

[psychsurf]

Hey, I'd love that, 94Loyale! Just let me know what you need and I'll paypal you asap. If nothing else it would be good to have on hand as a spare.

 

Oh, and the report from my commute to work tonight... It's actually running better even than when I picked it up! No codes, the throttle response is WAY better, and it doesn't feel like it's lugging at 55 in 5th gear. It actually responds to wide-open throttle now! My butt-dyno says I've gained about %20 in performance.

 

Since the only thing I KNOW I actually fixed was the knock sensor connection, could it be that with the ECU detecting a fault in the circuit it retarded the timing as if the knock sensor was actually indicating a knock? That's the only thing I can think of, because it's running like a whole new car and there's really not much reason for it!

 

I've got my fingers crossed it will stay that way.

 

I just wish the other issues I had could be attributed to the same problem.

[94Loyale]

I don't mind at all Our shop is closed between Xmas and New Years, so I've got all week to do nothing. When I grab them, I'll let you know!

[daeron]

if you feel like doing some reading, I did a fairly exhaustive write-up on relays, how they work, and how to install a simple relay for use in a fan or lighting circuit, and also for starter control (read: the modification I did in my write-up involved my STARTER, not my ignition relay)

 

http://www.ultimatesubaru.org/forum/showthread.php?t=74632

 

Not to sound immodest, but IMO its a good explanation of what a relay is, and how it works, and its in the context of installing it for use on a new circuit. (even if it IS a bit wordy, heh) The knowledge is there to help you understand how simple the things really are. It might help things "click" into place in your head (no pun intended )

[psychsurf]

You shouldn't be modest, Daeron. I actually read it last night when I was searching for information on the possibility of a bad ignition relay. I've been working on my own cars for 15 years, and that was the first explanation I've ever read of a relay that made me say, "Oh, now I get it."

 

Amazing what it simple concept it is, and how voodoo it all seemed before.

 

So yeah, toot your own horn by all means, it was a very good nuts-and-bolts article.

[Gloyale]

Oh, and the report from my commute to work tonight... It's actually running better even than when I picked it up! No codes, the throttle response is WAY better, and it doesn't feel like it's lugging at 55 in 5th gear. It actually responds to wide-open throttle now! My butt-dyno says I've gained about %20 in performance.

 

Since the only thing I KNOW I actually fixed was the knock sensor connection, could it be that with the ECU detecting a fault in the circuit it retarded the timing as if the knock sensor was actually indicating a knock? That's the only thing I can think of, because it's running like a whole new car and there's really not much reason for it!

 

Maybe it just needed to be shaken down a bit. And a poor Knock sensor connection could casue alot of trouble, epecially if it was making contact, then not, then contacting, then not...... you get it.

 

Keep us informed. And don't take TOO much crap about it from the wife, it's running now ain't it?