Questions About 1993 Legacy 2.2 Engine for an Expert

[lavabit]

If you are an expert on the 1993 Subaru Legacy 2.2 engine, please help me. I have a bunch of questions.

There is a module located under the dash panel close to the steering column and it has 3 connectors on it: the connector on one end has 16 pins, the connector on the
other end has 20 pins, and the connector in between those connectors has 12 pins. Is this the engine ecm module?

Is the sensor at the left front of the engine close to the left end of the timing belt cover the cam sensor?

Is the sensor located at the middle front of the engine close to the timing indicator the crank sensor?

There is a sensor at the top rear center of the engine, close to being underneath the throttle body air intake butterfly valve (connected to gas pedal).
Is that the knock sensor?

More or less underneath and close to the air bypass valve (the valve used to control idle air intake) are what appear to be 3 sensors/devices: is the one with only
a single electrical contact the temp sensor that goes to the gauge on the instrument panel? Is the connector that seems to have 2 pins on it the temp sensor that
feeds the engine control modules? There is a third device in that part of the engine. It has an electrical connector on it and 2 vacuum hoses also connect to it.
What is that device called and what does it do?

I removed the module that was near the steering column under the dash and the engine still runs. Assuming it is the ecm module, is it normal for the engine to still
run with the ecm module removed? If it can run with the ecm module removed, then what tells the fuel injector how much fuel to put out? I thought only the ecm
module could do that.

It would be really great if this web site could put pdf files of all the Subaru service manuals, parts books, and wiring diagrams online along with close up photos
of all the major engine control sensors/actuators/modules/etc showing where they are mounted on the engine.

Here is the problem I'm having with the engine.

When the engine starts cold, it runs like a brand new engine for about 20 to 30 minutes and then suddenly it starts intermittently missing and continues to
intermittently miss forever unless it is shut off again for a while and then restarted at which time this process repeats. When I say that it is missing, what I
mean is that the engine runs roughly for about 3/4 of a second and the rpm's drop. My guess is that the spark plug firings are missing or maybe the timing is being
radically changed. However, I am not positive that this is an electrical problem and not some sort of fuel problem. All of this happens if I simply leave the
engine idling in park. However, even on the highway, I can notice some wavering in the engine power once this intermittent problem kicks in.

If the module I removed really is the ecm module, then the problem can't be anything to do with the ecm module or any sensors or actuators that are connected only
to the ecm module because even with it removed the engine still runs great for the first 20 to 30 minutes and then starts the intermittent missing.

Next is a list of all sensors/actuators/electronic modules that I think control the engine. Please tell me if my list is missing anything, has anything on it that
should not be there, or otherwise has any errors.

ecm module
igniter module
ignition coil pack
mass airflow sensor on the air cleaner
crank position sensor
cam position sensor
knock sensor
temperature sensor that feeds the engine control system (as opposed to feeding the instrument panel temperature gauge)
throttle position sensor
4 fuel injectors controlled by the ecm module and/or the igniter module
air bypass valve which allows more or less air into intake in order to control the idle speed
oxygen sensor mounted near the catalytic converter
the device on the engine near the air bypass valve which has an electrical connector and 2 vacuum hoses attached to it


Please list the sensors/actuators that connect only to the ecm module, only to the ingiter module, or which go to both modules (if any).

Thanks very, very much!

Bob

[Aerotech]

Automatic or Manual trans?

Turbocharged?

The ECU has 4 bright yellow connectors all in a row and is the largest module under the dash, so whatever you unplugged, it wasn't the ECU.

You have the sensors identified correctly, except I'm not sure what the device w/ vacuum hoses is...

[lavabit]

It is an automatic transmission and is not turbocharged. Has the 2.2 engine.

Also, does my car have an atmospheric pressure sensor and if it does, where is it located?

Thanks again very much,
Audioresearch

[lavabit]

Is the ecm module located near the steering column under the dash panel?

[93forestpearl]

You probably unplugged the TCU.

The thing with the hoses to and from it underneath the idle air control valve is the purge control valve.




Josh (Legacy777) has the engine and wiring parts of the factory service manual on his website.

[93forestpearl]

You probably unplugged the TCU. It is either next to or in front of the ECU. I haven't had an auto car in a while.

The thing with the hoses to and from it underneath the idle air control valve is the purge control valve.




Josh (Legacy777) has the engine and wiring parts of the factory service manual on his website.

[Arctic Assassian]

you unplugged the tcu, the ECM is behind it, it's a freakin pain to get to. if instead of letting the car sit and warm up, if you just get in and go does the problem develop faster?

If so, you probably just have a cracked coil or something, that once it warms up, you develop a break.

Also, if you unplug the ecm, your car won't start.

[lavabit]

I did finally find the ecm up where you said it would be. I can see there is one nut on the bottom that holds it in. Where are any other bolts/nuts that hold it in or is the one on the bottom, which is the only one I can see, the only one? Should I get the item out just be reaching it from under the extremely crowded dash panel or
should I do something else like removing part of the dash panel? I haven't yet found a removal procedure for this in any of the manuals I got off the Net.

Do you know exactly what the igniter module that is on the engine compartment firewall about at the top middle does? Is it just an amplifier to boost the voltage of the spark pulses or does it also generate the spark pulses? If I put a scope on the outputs of the igniter that go to the ignition coils, what should be the values of the most positive peaks and the most negative peaks I should see?

I downloaded a service manual for a 1992 Legacy which I figure is close enough to mine and it mentions something called "selective monitor". Is that some sort of external test equipment? Are there any scan tools or other monitors (maybe including the "selective monitor" if that's one of them) I can buy that I can connect to
the car? Where do you connect them?

The aim of all this knowledge is to figure out why my 2.2 engine runs great all the time for the first 20 to 30 minutes and then starts intermittently running very rough for burst of about 3/4 of a second during which the rpms drop (often enough to stall car if it is at idle). When it does this, the check engine light does
not come on.

Do you think if I put the ecm into its "D" (dealer) mode it would give me any trouble codes for this problem even though none show up in the "U" (User) mode?

Thanks very, very much!

Bob

[GeneralDisorder]

Have you changed the plugs and wires? These are sensitive to bad plugs and wires and most folks find they need to be replaced annually.

If it's not throwing a code it's likely that the problem is ignition related and when the engine gets warm and the mixture starts to lean out you have a misfire condition. Pull the plugs and replace with NGK's. These engines don't like other brands of plugs. Get wires from the dealer.

The ECU is self-diagnosing for a reason - if it's not throwing a code then all the sensors are reporting correctly. Chasing problems with the ECU or it's sensors when there are no codes present is silly unless you have exhausted all normal lines of troubleshooting. The ECU can't tell you if the plugs or wires (or coil) are bad. Start there.

GD

[Aerotech]

+1 for the last advice, sounds like something's heating up, and plugs are misfiring. Change plugs & wires before you go digging into the wire harness! Check the coil connectors too, I had one that turned green inside, caused misfires... caused by a degraded plug wire.

The ECU has two fasteners on it, by the way, you probably need to pop out the mirror switch and look in there to see the other one.

The ignitor is basically a signal amp, ECU generates fire signal, goes thru ignitor to the coil.

[Buffman]

Change the fuel filter also if it hasn't been changed. Mine looked fairly new, but what came out of it said otherwise....

if you have no CEL I'd suspect general maintenance items as being cause.

[Legacy777]

I didn't read all the posts, however I would suggest you check out the stuff you need to know forum. There are a couple posts/threads in there with links to service manuals.

This info on my site is pretty handy to identify where all the sensors/ECU's/etc are located

http://www.surrealmirage.com/subaru/swa ... l#location

I know it was already discovered that the control module previously noted is the TCM, however for anyone searching in the future, this picture may be of help.

http://main.experiencetherave.com/subar ... cuhide.jpg

[RJ93SS]

buy a haynes manual and it will tell you what and where those sensors are and will do

[lavabit]

I've done more diagnosing on my 1993 Legacy Wagon with AT and no turbo.

The problem I've been having is that the idle speed takes "hits" where the rpms drop intermittently, sometimes by enough to stall out the car.

The problem seems restricted to idling and doesn't show up when driving on the road as long as I don't come to a stop.

I removed the throttle position indicator from the throttle shaft, but kept it attached to its electrical connector. Using a screwdriver, I rotated the rotating portion over all the way to the place it should be in at fully closed throttle (idle position).

Then, with the tpi hanging in air from its wiring harness, I started the engine.

The problem of the fluctuating idle speed went away!!

However, if I try to re-mount the tpi where it belongs and position the tpi in the correct orientation so that it's idle switch is closed when the throttle is closedm then the mounting screws don't line up with the slots of the tpi they are supposed to go in. Plus, the orientation of the tpi is such that the tpi would hit the rear bracket that holds on the plastic cover piece that covers the ignition coil area.


Also, the resistance of the tpi when it was rotated between full closed throttle to full open throttle was way out of spec per the resistance specs in the 1992 service manual.


So, I have to wonder if the tpi in the car is not the correct one for the car.

Question: Does the 1992 service manual cover my 1993 Legacy?? It appears that there is a 1993 supplement to the 1992 service manual. The 1992 manual is on the Net, but I could not find the 1993 supplement. Do I really need it?

In particular, are the resistance values that are shown in the 1992 service manual the correct ones for the tpi of the 1993 Legacy???

The resistance I was supposed to measure across terminals 2 and 3 of the tpi was 12K ohm, but the actual resistance was about 1 Megohm!

The resistances I was supposed to measure across pins 2 and 4 was supposed to vary between about 4K to 12K, but mine varied between 9K and about 500K (I hope my resistance meter isn't no good. If the values I measured are correct, I wonder how the engine even ran).

THe numbers stamped on my tpi are

A22-000R25

What number or numbers should be on the tpi?

I also measured that pins 1 and 2 were shorted only when the rotating portion of the tpi was at the very far end (the closed throttle end).

Is it normal for pins 1 and 2 to short out only at the very end of travel and not, say, within a range of so many degrees of being at the end?

I'm going to go to a local salvage yard tomorrow that has a 92 and a 94 Legacy and try swapping out the tpi.

[93forestpearl]

Sounds like you may have found your problem. The first gen Legacy is weird from other Subarus in that it has a reversed signal.


And they do go bad from time to time. Let us know how swapping in a different one pans out.

[Legacy777]

The TPS info in the 1992 manual should be the same for the 1993 model years. From what you've described you may want to swap in another TPS.

Also to note, if you remove the TPS, it needs to be calibrated, which is very easy, and needs to be done following these procedures

http://main.experiencetherave.com/subar ... sting3.jpg

[lavabit]

I swapped salvage yard tps, but it seemed to be same as my tps, so I put my tps back. I simply rotated my tps all the way over so that its idle switch is activated the easiest (looking at tps from pax side of car, I rotated it as far counterclockwise as it would go).

I thought my idle problems were gone, but I was wrong. Now instead of showing up about 15 mins after engine is started, they show up more like 30 or 40 minutes afterwards and seem to be less severe, although still objectionable.

Both Legacy's in the salvage yard had their tps set to what seemed to be the same place as mine had been set before I rotated mine all the way over.

So, apparently, fact that the tps can't rotate all the way to the end of its travel and still have its mounting screws line up with the mounting slots is normal.

Since my tps rotation helped, but did not 100% fix problem, now I suspect most that the problem is the oxygen sensor since ecm does not pay attention to it until engine is fairly hot.

I opened another topic about isolating/testing the oxygen sensor in the engine forum.

If I finally fix this problem, then I'll go back and adjust the tps the "right" way. But right now, I'll take anything that makes this engine stop running so poorly.

[lavabit]

Update:

I did disconnect the electrical connector from the temperature sensor that feeds the ecm (not the one that feeds the instrument panel temp gauge).

Unfortunately, the erratic idle rpm problem has remained.

Because this problem only starts up after the engine has run quite a while, I'm really suspicious of the oxygen sensor and I may just get a new one and try it.

I'd sure like to know whether or not the ecm stays in open loop mode indefinitely when the wires are removed from the temp sensor.

[lavabit]

Josh,

I just edited this post to bring it up to date.

My car is a 1993 Legacy "L" wagon with auto transmission, no turbo, no AWD, and with the 2.2 engine.

The problem my engine has is that if I run the it for somewhere between 20 to 40 minutes, it starts intermittently missing. The missing is sometimes bad enough to stall out the engine and it happens both at idle and at highway speeds.

Before the missing begins, the engine runs like it it is brand new!

Also, the exhaust is a little sooty and the spark plugs have a thin film of carbon on them.

Here's a summary of all I've done to try to diagnose this problem so far.

I've swapped out these parts with others I got from a local salvage yard:

cam position sensor

tps

air flow sensor

igniter module

ignition coil pack

ecm

I not only swapped the salvage yard ignition coil pack and igniter module and observed the identical behavior, but I also swapped them out after the engine was already missing and even with those 2 parts put in cold, not hot like the engine, the engine immediately started missing. Therefore, the salvage yard coil pack and igniter module were not failing because they were being heated. I think there is almost no chance whatever that the coil pack or igniter module could be the cause of the missing problem.

The ecm I swapped from the salvage yard was not the right one for my 1993. The salvage yard ecm was from a 1994. Nevertheless, the engine still ran great with it for 20 to 40 minutes and then started missing, all at idle.

I eliminated the oxygen sensor and the idle air valve as suspects by removing the electrical connectors from both and also by blocking the intake hose that goes to the idle air valve. I adjusted the throttle plate stop screw so the car would idle at reasonable rpms even though the idle air valve was disabled. The problem remained even after doing that.

The spark plugs are new and were gapped to about .45". The plug wires are new.

I disconnected the electric connector from the temp sensor. Problem remained. I put connector back on.

I disconnected the electric connector from the knock sensor. Problem remained. I put connector back on.

I disconnected the vacuum hose going to the purge valve. That made no perceptible difference in how the engine ran and the problem remained, so I put it back. I did the same for the vacuum hose going to the power brake booster with the same results.

I removed the electrical connector going to the airflow sensor. Problem remained, so I put it back. Also, because this particular sensor is so critical, I swapped it out with 2 different salvage yard air flow sensors also.

One at a time, I removed the spark plug wires from each of the 4 cylinders. As expected, the engine ran rougher but the intermittent missing problem remained regardless of which cylinder had its plug wire removed. All plug wires are back to being connected again now.

I put a fuel pressure gauge in the system at the output of the fuel filter. It held rock solid at about 30 psi, both when the engine was missing and when it was not missing, so there is no fuel pressure problem. I had put in a new fuel filter. I've also swapped out both fuel rails, including fuel injectors and the fuel pressure regulator. All the gas was drained from the fuel tank and replaced with new fuel. The fuel tank looked clean inside when its access cover was removed.

I cleaned 2 ground connections on each of the two strut towers inside the engine compartment and another ground connection right near the battery. I installed brand new battery cables and terminals (1 cable goes to engine ground, the other cable goes to the starter). The problem remained unchanged.

I haven't disconnected all vacuum hoses from the engine (but I did try disconnected and capped off the purge valve hose and its manifold nippleas mentioned above), however I do not hear any vacuum leak sounds. Did same for brake booster vacuum hose & its manifold nipple.

I plan to buy a spray bottle of the stuff you can spray onto an engine to detect vacuum leaks (the idle is supposed to go up briefly when you hit a vacuum leak with it) and test with that.

Drained almost all the gas out of the fuel tank and filled it about half full with new gas.

Put on a new fuel filter.

Pulled off the pcv valve hose. That made no perceptible difference in how the engine ran nor did it make the problem go away.

Compression test showed all 4 cylinders reading between 175 and 177 psi.

I backed off the throttle stop screw so the idle rpms dropped to the correct 700 rpm range and then I did a vacuum test. The vacuum gauge showed vacuum at about 20.5" Hg. The needle rapidly vibrated between about 20.2 and about 20.9 to the point where it looked like a blur. I was told that such a small range, 0.7 of 1", was normal and nothing to worry about. I don't remember if I kept looking at the gauge after the missing began however I did the vacuum test a previous time when I still had the idle speed set high (about 1400 rpm) and I definitely did keep looking at the vacuum gauge for not only the initial 20 to 40 minute period when the engine ran great, but also during the missing that occurred later and the gauge steady the entire time at about 13.5" Hg. The gauge absolutely did not budge during the missing.

The only item external to the engine that I haven't swapped out is the crankshaft position sensor. I can't get the one out of my engine without putting enough force on it where I might break it.

I also plan to test all resistances and voltages shown in the service manual for all the sensors/actuators/electronics modules. I think I've either swapped out or tested everything else there is.

Out of desperation, I'm going to buy some air flow sensor cleaner and use it even though, because that sensor is so critical, I've already swapped out not just one, but two air sensors from the salvage yard.

I took car to a Subaru dealer. They gave up after 5 minutes. They thought the problem might be moving valve guides and were about to drop exhaust pipes so they could look into the engine and watch the valve guides, but they said exhaust bolts were too rusted and they might break exhaust studs and so they did not want to go ahead.

Do you think moving valve guides could really be the cause of my engine's problems?

I badly need to diagnose this problem and especially figure out a way to determine if the problem is an internal engine problem that will require me to remove the cylinder heads to diagnose/fix.

Do you think that due to the results of the compression and vacuum tests above, the engine could not possibly have an internal problem??

Do you think I missed anything?

Do you know of any diagnostic tests I could run to at least isolate this problem to a bad spark, bad fuel delivery, valve/camshaft train problem, crack or hole in cylinder head or block or blown cylinder head gasket?

[Legacy777]

Goin to bump this thread.....I haven't been around at all lately, and need some more time to look at what you wrote.

[glennda5id]

The fact that you cannot remove your crank sensor suggests a problem. The crank sensor failing after heating up is defiantly possible. This may sound like a lot of work, but if you remove the crank pulley and the timing belt service covers you will be able to tap the crank sensor from the button and push it out.

I also would consider having a compression check done, but your problem sounds electrical.

While the service cover is off, check your timing belt alignment.

[jp233]

Yeah, there is a chance that your crank sensor goes bad when it gets warmed up, however I'd think it would happen a lot sooner than 20 minutes. That sucker would get hot within 5 minutes of running.

Anyways yeah, might as well take some PB BLASTER and soak anything that is stuck, really good. It helps to do this when things are warm. I've tried every damn rust buster spray, lubricant, whatever, and PB Blaster is the damn best.

Might as well try squirting some on your crank sensor hole and of course on the exhaust studs. Do this several times if its still tough. Actually, you can get some naval jelly (rust remover gel, sold in auto parts stores) and wipe it on the exhaust studs/nuts first for a while to try to break up the rusty surface as much as possible, then clean all that off and soak it in PB. Best to soak really impossible things (like the New York rusty underside of my friggin truck, I'll soak suspension/exhaust/fuel tank components for 3 days before trying a wrench on anything).

on the crank sensor yeah, it would be worth the time to tap it from the bottom.

well, on the exhaust studs at least you can just get those out with a gun. Mine zipped right out on a rusty, Masshole 93 model, and one of the studs came out as the nut was rusted solid to the stud.

[lavabit]

I found somebody who towed the car away and crushed it and gave me $100.

All problems with that car are now cured.

Worst piece of junk I've ever seen. This brand should be banned from the USA.

[tahiti350]

Lavabit,
Sorry you had so many problems with your Subie, they are good cars, and even though I've had a few problems I've been able to sort them out and fix them. If you had posted the car for sale here you would likely have gotten considerably more for it, especially with everything you had done to troubleshoot it, and all the known good parts.

Better luck with whatever you get to replace it.

[tturnpaw]

Sadly this was not resolved properly. I had similar issues with my subie. Simply enough, it was just the coil pack plugs and wires.