A long sad tale and a cry for HELP!

Headlights to tailights and everything in between.

Moderators: Helpinators, Moderators

Post Reply
Coyote
In Neutral
Posts: 12
Joined: Mon Jul 23, 2018 5:52 am
Location: Texas

A long sad tale and a cry for HELP!

Post by Coyote »

If I were an Aussie, I'd say ”G'day, mates”, but since I'm just a Texan, I'll try, ”Howdy, y'all”. . .

This is my first post. I've lurked and read for several months and finally registered, but up 'til now I haven't felt that I had anything to say that was worth anyone's time to read.

I had also hoped that my first post would be contributory in some way. That would be tough since most every post I read here only shows me more of my own ignorance about Subaru, Legacys in particular and makes me feel like a midget trying not to be stepped on in a crowd of giants. Unfortunately, I come with hat in hand asking for advice and technical help.

First, the vehicle: 1994 USDM Legacy L or LS (not quite sure which) Wagon, 2.2L N/A 4EAT 200K+ miles.

Problem: Limp mode

The Sad Story:

One day my wife comes to me (her car) and tells me she thinks I should check the transmission fluid because she thinks the transmission is “slipping or something”.

I check the dipstick and find nice clear, pink fluid at the proper levels both hot and cold, so I go for a test drive.

The transmission id definitely NOT slipping. Neither is it shifting. I have only one speed forward and reverse. A little research tells me that the car is in 'limp mode'. A bit more research tells me what limp mode is, why it happens and how to determine specific reasons.

I built vrg3's B10 Scan tool and used it to determine that I have no stored CEL codes (no active ones either).

I used the procedure from Legacy 777's website to check for transmission error codes and found Code 14, faulty Shift Solenoid #2. The procedure worked flawlessly (providing I made no 'finger checks' while going through it), was consistent and repetitive. Code 14 every time.

Since this made perfect sense, I found the solenoid pack and a place with a lift.

Before I got to the shop with a lift, I put plugs and plug wires in because I was doing them on one of my other Legacys (same everything except it's a Sun Sport wagon). I went through Legacy 777's 'write down sheet' for the select monitor with the B10 scan tool before I started and again after the plug and wire install (some expected improvements but still in limp mode and I didn't expect plugs or wires to fix that) and no CELs.

I decided to leave the scan tool connected for my drive home so that I could monitor coolant temperature and compare it with what the gauge on the dash was telling me. I cleaned up and closed up the shop (not the shop with a lift) and headed for home.

Everything was fine (except for limp mode) as I turned onto the highway (60 mph road) and I glanced at the laptop several times as I drove and everything was still OK. But here, the plot thickens: by the time I got to the first stop light, the scan tool wasn't showing me any temperature reading. B10 was running because it showed a page titled Coolant Temperature but there was no temperature. There was no value at all—just a blank. I was stopped at the light long enough to reboot the laptop, but I couldn't even read the ROM ID so I decided to wait until I got home to mess with it and shut the laptop down.

When I got home, I tried again. I still couldn't read the ROM ID. Since there were only four possibilities (bad B10 software, bad computer, bad cable or bad car), I decided to find out which.

The B10 software was on a thumb drive, so I went to 'Plan B' and booted off of the B10 CD but there was no change—still stalled at trying to read ROM ID.

I then unplugged the whole setup and carried it about 10 feet to my other wagon and tried that. Everything worked as advertised so I knew it was something in the first wagon doing the dirty work (or not doing it as the case may be).

I decided to quit for the day and deal with the problem after the transmission issue was taken care of.

When I got to the shop with the lift, we raised the car and drained the transmission and dropped the pan. There was a small bit of sediment in the bottom but nothing to worry about. Before dropping the valve body, we checked the solenoids with an ohm meter. They all read almost exactly the same as the new ones (26.x ohms, spec is 20-30 ohms). No other continuity to ground from the high sides, zero ohms to ground from the low sides. We decided that dropping the valve body was pointless at this stage, replaced the pan and filled up with Dexron. When we started the car, we were almost back where we started—limp mode and no CELs but now, no power light at ignition on and no 16 flashes and we were unable to do the TCU check procedure (never got a power light).

Next, we 'Easter Egged' the wiring harnesses. We found nothing obvious except that in general the wiring was in excellent shape. We found no brittle wires, all were brightly colored in the proper colors and in general, remarkably preserved in a car this old.

At this point, the TCU was looking like a good suspect, so we pulled it and popped the cover. AHA!!
There were two bulging electrolytic capacitors with black 'pookie' under them (IIRC, C602 & C604).

We called it a night.

The next day, I 'visited' my parts car (a 3rd '94 Legacy wagon, also L or LS, also 2.2l N/A USDM) and pulled the TCU. It had the same P/N which was no surprise as the parts book (again thanks to Legacy 777) said that '94 wagons all had the same P/N TCU unless they were turbo.

We may have screwed up here as we did not at that point understand that there was a difference between 'clearing the codes' and 'resetting the ECU/TCU' but the car had sat overnight with the battery disconnected so we thought we were good to go.

We inspected the TCU circuit board from the donor car and it looked brand new. There was nothing that looked even slightly suspicious, so we plugged it into the car (with battery disconnected) and started the car.

To our surprise and disgust, nothing changed. We were still in limp mode and couldn't get a power light to do the TCU check.

For the rest of the day, we checked continuity of the harness (according to what we believe is the correct wiring diagram and found nothing wrong. We checked the bias resistor on the front of the passenger strut tower—12 ohms and at various times system voltage and intermittently 4.7 volts. We swapped the two TCUs back and forth to get comparative readings (ALWAYS disconnecting the battery BEFORE pulling and plugging).

We ran out of daylight and most of the night with only confusing and tentative conclusions so we decided to rethink things and start fresh the next day.

When we started again, on a whim, I plugged in the B10 scan tool and to my utter surprise and total confusion, it worked perfectly! We thought then that maybe there was something going on in the ECU that might be doing us wrong but everything looked good. Same readings as immediately after the plug/wire change. I understand that the ECU has dedicated circuitry that does nothing but communicate with and command the TCU and that B10 tells us nothing of that but without being able to run the TCU check, we knew of no way to find any of that information so we continued checking continuity and scratching our heads and trying new ideas about what 'might' be going on.

Our continuity checks were constant and never indicated bad wiring. The voltage checks however were intermittent. We'd measure a pin, record the reading (for example the voltage on the shift solenoids) and without doing anything except moving the meter probe to another pin and back, we'd get a different or opposite reading.

At one point, Charley (a 30+ year transmission specialist) said, I've got to know something”. We'd just measured a voltage to the #2 shift solenoid but nothing to the #1 shift solenoid. We disconnected the battery and cut the wire to the #1 shift solenoid. Charley then applied battery voltage to the harness end of the cut wire (the wire going to the solenoid) and the transmission went into 1st gear! We shut down, disconnected the battery and jumpered the cut wire and powered up again. NOW the car went into 1st gear again!

Without shutting off the car, we drove over 5 miles, reaching up to 50 mph, stopped and started several times and everything worked perfectly! All gears shifted, at the appropriate rpm, down shifted and up shifted perfectly. This was with the TCU from the donor installed.

Man, we thought we were in tall cotton!

When we got back to the shop, we pulled in and shut down and thought we were going to solder and heat shrink a wire but we were back where we started—limp mode, no shifts, no power light and the same intermittent voltage checks. B10 still works but no change there—no CELs, active or stored.

From that point to now, we've not been able to duplicate the functioning or get anything to 'stay put'. If either of us had any hair, we'd have long pulled it out.

Tentative Conclusions:

We believe the wiring is good. We have seen nothing to prove otherwise.

We're 100% sure that the original TCU is bad (bulging, leaking capacitors)


We're 99.9% sure the TCU from the donor car is good. The.1% uncertainty is because I've never seen the donor car driven. I bought it just after the PO's girlfriend hit a deer with it and the front end was all smashed. I did, however pry stuff out of the way, connect a battery and start the engine for a minute or two only (no radiator). I subsequently pulled that engine and put in the Sun Sport wagon. That was over 100K miles ago with no issues. I have no reason to think that the TCU from that car might be bad but there is always that .1% that can bite you in the ass. . .

It seems like the only likely candidate left is the part of the ECU that talks to the TCU. Unless someone here yells STOP, DON'T DO THAT!, that's going to be the next move.

We're surely hoping that someone on the board can and will provide some help or insight that we've missed. . .

If you've read all of this,THANK YOU, whether or not you can help and I promise to try and not be so long winded in the future.
Any 20 minute job is only one broken bolt away from a three day disaster. . .
Legacy777
Site Admin
Site Admin
Posts: 27884
Joined: Tue Oct 15, 2002 11:37 am
Location: Houston, Tx
Contact:

Re: A long sad tale and a cry for HELP!

Post by Legacy777 »

Welcome to the BBS!

What part of Texas do you live in?

What you've described is interesting and the fact the readings are not consistent. Out of curiosity, have you checked just the wiring for continuity with the end connectors disconnected from the TCU/ECU and end device/solenoid? What about checking for shorts in the wiring? Also, what about grounds; have you checked your ground potential differences at the battery, ECU/TCU, engine, transmission, etc?

I once saw an issue with a solenoid throwing a code because there was a ground wire between the chassis and transmission that was disconnected. This caused a ground potential variance. If you find a ground variance, running some wire from the chassis, trans, engine, etc to the battery is a good way to resolve that.
Josh

surrealmirage.com/subaru
1990 Legacy (AWD, 6MT, & EJ22T Swap)
2020 Outback Limted XT

If you need to get a hold of me please email me rather then pm
Coyote
In Neutral
Posts: 12
Joined: Mon Jul 23, 2018 5:52 am
Location: Texas

(SOLVED) A long sad tale and a cry for HELP!

Post by Coyote »

Thanks for the welcome Josh! I saw that a fair number of folks had read my post but I figured they were all as stumped as I was.

I live in College Station. It's not too big but big enough to have most of the city amenities. It seems as if over half of the population is A&M students and it sure turns into a ghost town when school's out!

Yes, we isolated the wiring harness by disconnecting every connector in the circuits we were looking at and checked for shorts to ground and between the different wires of the harness. As I noted, we found nothing to indicate bad wiring.

We also worked out on the grounds. Despite having good and consistent ground readings everywhere that we were supposed to, we took off every ground we could find, wire brushed them, re-tightened and checked again, all without result. I am familiar with ground potential variance (we called them 'ground loops') on the computers I used to diagnose and repair for a living. They had 3 phase, 400 cycle (I still can't get into calling it Hertz), 60 amp service and ONE AND ONLY ONE electrical ground to chassis ground connection. If you somehow got another, current would flow from one ground potential to another (thus the 'loop') and random circuitry would start telling you that 2+2=3, or maybe 7 or whatever!

The good news is that the problem is resolved! I know you're familiar with Murphy (of Murphy's Law renown) and since he's made my life miserable on numerous occasions, I decided to see if I could thwart him.

We found a '90 Legacy wagon, FWD, N/A at a local junkyard and it had its TCU still aboard. When we found the mounting nuts loose, I heard Murphy laugh, but knowing him, I took it anyway. .1% is a long shot, but still a shot and those are some of Murphy's favorites. The junkyard TCU fixed all problems!!! Swapping my parts car TCU back in reverted to the same symptoms and going back to the junkyard TCU fixed everything.

Lessons Learned (AGAIN):

No matter how experienced you are at troubleshooting, a single bad premise (in this case the 99.9% belief in the donor car TCU) takes you to never-never land and if you happen to get the right answer, it is an accident.

Circuit diagrams of the TCU would certainly have helped but the rule is this: When everything is working correctly, you know what the circuitry is doing. When it's broken,ANYTHING can happen and it often makes NO sense.

Murphy gave me (and Charley) some bad days for sure, but it won't be the last time and every time he does, that's one less trick he has in his bag.

P.S.

Murphy got the last laugh.

As Charley and I were looking at the car after the successful test drive, we noticed coolant dripping out of the front center, under the lower radiator hose. it was dripping out of the weep hole on the water pump. . .

Murphy gave us a sloppy salute accompanied by maniacal laughter and vanished. . .
Any 20 minute job is only one broken bolt away from a three day disaster. . .
Legacy777
Site Admin
Site Admin
Posts: 27884
Joined: Tue Oct 15, 2002 11:37 am
Location: Houston, Tx
Contact:

Re: A long sad tale and a cry for HELP!

Post by Legacy777 »

Well I'm glad to hear that the different TCU fixed the problem. Sometimes it's those random things that cause you nightmares. I had an issue with a high idle after cleaning the TB and searched and searched for the problem. What I ended up finding is a little needle valve type screw on the top of the TB that was obscured by the engine cover bracket. Adjust that screw and all is well.

Anyway, good luck with the water pump it should be an easier task. I would suggest doing the timing belt and idler pulleys if they haven't been done in a while. Also may want to look at the front crank & cam seals, rear cam o-ring, and maybe resealing the oil pump if there's leaks.
Josh

surrealmirage.com/subaru
1990 Legacy (AWD, 6MT, & EJ22T Swap)
2020 Outback Limted XT

If you need to get a hold of me please email me rather then pm
Post Reply