92 Legacy running extr. rough. Failed smog

[salewit]

My girlfriend has a 92 Legacy 2.2 non-turbo. The car has been running progressively worse. Idles extremely rough. Stutters and hesitates upon acceleration. At cruising speed it jerks and has little fits.

 

Mileage is going downhill and is now only 14 MPG. Smog results:

 

Idle: HC 575 (Max is 120)

Idle: CO 1.87 (Max is 1.0)

2500 RPM: HC 49

2500 RPM: CO 1.62

 

Ran compression test today. All four plugs were sooty black. Compression rates all rose fairly quick and were 130, 125 on one side, and 105,105 on the other. When I squirted oil in the 105's they both went up to exactly 120.

 

No overheating, and no visible exhaust smoke. No sign of coolant in oil or vice versa.

 

Vacuum gauge was running around 50 inches but spurting up and down about 10 inches from time to time.

 

Anybody have any ideas?

 

 

P.S. About a year ago, the check engine light would come on sporardically with I think it was a camshaft or crankshaft sensor error. When the light came on, the car would stutter and cough. A little similar to the above. At first we'd shut the car down and restart and the light would go out and all would be fine. Later we discovered that if we just kicked the accelerator to the floor real fast, that would clear it too. The light hasn't come on in awhile (of course we hardly use this thing anymore).

[ericem]

Maybe your 02 sensor, and MAF. First thing that would be good to do is change all the plugs and wires, air filter,fuel filter all the little things. Then you could remove the MAF and try spraying it with MAF cleaner. Also use some fuel injection cleaner, and try some seafoam through the pcv valve if you like. The check engine light might be dead, but try to check the codes, here is a site how. I wonder if a clogged CAT causes this problem as well. All things to look at.

 

http://www.surrealmirage.com/subaru/subaru.html

[Cougar]

Check the temperature sensor for the ECU and see if is working as it should be.

[Legacy777]

Yeah, I'd suggest checking the ECU for codes, instructions can be found on my site with the link posted above.

 

I'd also suggest replacing the coolant temp senser.

 

Does the car run any differently when cold vs. when warmed up?

[salewit]

Thanks all for the help. I've checked the codes numerous times and like I said, it's giving me crankshaft sensor errors. But only for a few seconds, until I kick the gas down. Then it goes away. This is the only error stored in memory. About a year ago we got an oxygen sensor error, and replaced the sensor and never got that error again.

 

This weekend I ran a bunch of tests with info from my Haynes book. I checked the coolant temperature sensor (I think) with an ohmmeter. The chart is showing 176 ohms at 212F and about 4000 when cold. I checked it fairly hot and cool and it was only around 40 ohms! The problem is that I'm not sure I got the right sensor. The diagram in the book shows a sensor with TWO pins. What I checked only has a single tang, so I checked between that and ground. The sensor was almost under the alternator (a little to the left of it a few inches above the crank sensor).

 

I also did the seafoam thing, and I couldn't find MAF cleaner, and when I look at the thing I was afraid to do anything to it. Looks kind of sensitive.

 

I also checked throttle position sensor with my ohmmeter (seemed ok) and the crank sensor. I checked just the static sensor and the reading was right on. I didn't check voltage induction.

[P K]

Based on personal experience and like symptoms, I would speculate that you have a worn slot or key on the timing belt gear that allows play on the crankshaft causing erratic timing. First pull the v-belts off the and see if you can wiggle the crankshaft pulley. Then prolly pull the crank puilley off, and timing cover to see the timing belt gear/cog. Mine was worn and allowed the gear to move a few degrees back and forth - not good. Hopefully the slot is okay and the pulley, gear and/or key can be replaced to fix.

 

Best of luck, anyone else concur?

r/ PK

[Cougar]

The temperature sensor you checked is for the temperature gauge. The sensor for the ECU has two pins like the manual shows. I think it is near the thermostat housing or towards the rear left side of the engine as you face it from the front of the car.

[salewit]

You know I was kind of considering valve timing too. Like maybe the belt jumped a tooth or something. I may just check that next.

 

About the coolant sensor, my book shows a picture of the engine with arrows pointing to the components, and it points to the "engine coolant temperature sensor (on coolant pipe under intake manifold runner)" as being inbetween the alternator and the steering pump reservoir. Ok... I looked up sensor for the temp gauge in another part of the book, and it IS pointing to the ECU sensor in the back of the engine. Damn! That doesn't give me a lot of confidence in this book.

[Cougar]

Another reason why factory manuals are prefered over other sources though, they aren't perfect either. They can have mistakes also.

 

The sensor in the back is the one you really want for the ECU.

[montana105]

Did you replace the black sooty plugs?

[Gloyale]

Ran compression test today. All four plugs were sooty black. Compression rates all rose fairly quick and were 130, 125 on one side, and 105,105 on the other. When I squirted oil in the 105's they both went up to exactly 120.

 

No overheating, and no visible exhaust smoke. No sign of coolant in oil or vice versa.

 

Vacuum gauge was running around 50 inches but spurting up and down about 10 inches from time to time.

 

Anybody have any ideas?

 

 

P.S. About a year ago, the check engine light would come on sporardically with I think it was a camshaft or crankshaft sensor error. When the light came on, the car would stutter and cough. A little similar to the above. At first we'd shut the car down and restart and the light would go out and all would be fine. Later we discovered that if we just kicked the accelerator to the floor real fast, that would clear it too. The light hasn't come on in awhile (of course we hardly use this thing anymore).

 

Right off the bat, That compression is way too low. Specs for this engine are 155-185 psi. could be rings or really dirty valves. Probably rings. That explains alot but not everything

 

Still, there may be another problem. If you got a crank angle sensor code, you should check it's connection and clean it. If you continue to get theat code, replace the the crank angle sensor. The way you're describing *kicking down* the accelerator seems funny. That might affect a Throttle sensor code, but how could it affect the Crank sensor? Are you relying on the same questionable book for the codes?

 

BTW, I think you measured the Oil Light sender, under the alt. Both coolant sensors are located in the back of the manifold. Very obscured by hoses and stuff but it's there

[nipper]

Right off the bat, That compression is way too low. Specs for this engine are 155-185 psi. could be rings or really dirty valves. Probably rings. That explains alot but not everything

 

 

Its not that the numbers are too low, its that they are very uneven. I've seen engines with many many miles on them be around 100 psi and run fun, but they were 100 psi +/- 5 psi.

 

I do agree that the engine has tired or stuck rings. If it was valves or a HG the numbers would not come up on a wet test. This engine is prbbly using some oil.

 

This is part of the puzzle, but not the entire thing. Is it safe to assume the two 105/120 cylinders are on the same side of the engine?

 

nipper

[nipper]

 

Vacuum gauge was running around 50 inches but spurting up and down about 10 inches from time to time.

 

 

Thats impossible. The max reading on a vacume gauge for a internal combustion engine is 30 inches of HG, and thats usually with the throttle closed decelerating. The gauge will read close to zero at full throttle, and 17-21 at idle. The needle shold be steady.

 

http://www.secondchancegarage.com/public/186.cfm

 

nipper

[Gloyale]

Thats impossible. The max reading on a vacume gauge for a internal combustion engine is 30 inches of HG, and thats usually with the throttle closed decelerating. The gauge will read close to zero at full throttle, and 17-21 at idle. The needle shold be steady.

 

http://www.secondchancegarage.com/public/186.cfm

 

nipper

 

 

I was thinking something was weird there too. 50 inches is 2.5 times higher than the FSM spec of 19.5 inHg

[salewit]

Thats impossible. The max reading on a vacume gauge for a internal combustion engine is 30 inches of HG, and thats usually with the throttle closed decelerating. The gauge will read close to zero at full throttle, and 17-21 at idle. The needle shold be steady.

 

http://www.secondchancegarage.com/public/186.cfm

 

nipper

 

Sorry... that was the wrong scale on the gauge. It was around 22 inches.

 

Well I appreciate all the input.

 

1) The scan codes I got were from the website referenced earlier in this thread. And yes, kicking down the accelerator to the floor real fast removes the "check engine light" instantly. I haven't seen the code come back in some time, but I haven't driven it much lately.

 

2) In the past few days I've changed the PCV valve, did the Seafoam thing, checked the throttle body (very clean), listened for clicking injectors, measured resistance of each injector, finally found and checked coolant sensor (was 4000 ohms at 60 degrees), checked throttle position sensor, checked ohms of crankshaft and camshaft sensors (did not check voltage while sensing yet).

 

One weird thing... with engine running, I pulled the injector connectors one by one. On the low compression side, pulling the connectors had almost no effect. Pulling good side did have an effect somewhat.

 

3) I'm wondering about the low compression side. 105 in both cylinders. When I squirted oil in each, they each went up to 120. Does that indicate terrible rings or just mildly worn? The fact that both cylinders reacted exactly the same, could there be a bad head gasket leak between the two?

 

4) I know about the OBD I codes for this vehicle, but is there more info available? If I go to a shop can they connect to the ECM or whatever and just read how everything is performing and maybe cut to the chase about what is not working properly?

 

5) How about a home-made leak down test. Would that help here?

 

Thanks again for all the great help here!

[nipper]

yes those two cyllinders are bad numbers. Like i said, its not the numbers but the differnce. You had minimal change when you pulled those injectors because those two cylinders are not working very hard. They have tired rings.

This can also cause havoc with the emissions.

 

If there was a bad HG they would not have come up on the wet test, unless you did the test wrong. Did you have all the plugs out at the same time and the throttle wide open?

 

JUst out of curiosty, play with a vacum gauge again, and use the link i showed you to diagnose. That will tell you if you have worn rings or a HG problem.

 

nipper

[salewit]

Ok, I tried the vac test again. Good site by the way. My example was the closest to Scenario 5 or 11. A steady 18 inches. Quick acceleration knocks it to 0, and then a quick release jumps it up to about 20, then back to 18. The only abnormality is that about every 2 seconds regularly, the engine would shudder, and the gauge would drop a couple of inches and then return. It was almost like clockwork. A few times it shuddered consecutively, but usually every 2 seconds. This was at idle (700rpm) at operating temperature checked on the power brake booster line.

 

The compression check was tested at operating temperature with all plugs out and gas pedal to the floor for 7 cranks per cylinder. Two bad cylinders I gave 2 complete squirts of an oil can (1/4 or 1/2 a shotglass full?) of motor oil.

 

Obviously I want to make sure pulling the engine apart is the right step.

 

yes those two cyllinders are bad numbers. Like i said, its not the numbers but the differnce. You had minimal change when you pulled those injectors because those two cylinders are not working very hard. They have tired rings.

This can also cause havoc with the emissions.

 

If there was a bad HG they would not have come up on the wet test, unless you did the test wrong. Did you have all the plugs out at the same time and the throttle wide open?

 

JUst out of curiosty, play with a vacum gauge again, and use the link i showed you to diagnose. That will tell you if you have worn rings or a HG problem.

 

nipper

[OB99W]

Ok, I tried the vac test again.[...]Quick acceleration knocks it to 0, and then a quick release jumps it up to about 20, then back to 18.[...]

Take a look at Scenario 4; what's described in your sentence I quoted above fits it (ignore the 15-17 inches part -- a Subaru engine that's healthy typically has a vacuum reading at idle that's above 18 inches, usually closer to 20).

 

 

The compression check was tested at operating temperature with all plugs out and gas pedal to the floor for 7 cranks per cylinder. Two bad cylinders I gave 2 complete squirts of an oil can (1/4 or 1/2 a shotglass full?) of motor oil.

 

Obviously I want to make sure pulling the engine apart is the right step.

Between the low compression, the increase when "wet", and the vacuum readings you've reported, it does seem that the rings aren't sealing well.

[nipper]

I agree your rings are tired. Its time, and can be a major part of your problem.

 

 

nipper

[salewit]

I agree your rings are tired. Its time, and can be a major part of your problem.

 

Alright, well time to yank her then! Thanks everyone for your help. Time to see if my 5 years of auto mechanic training from the 70's will do any good.

[OB99W]

To clarify things -- I think poorly-sealing rings are only a part of the problem, as others have said. In fact, if enough unburned gas is washing down the cylinder walls, especially if the oil is becoming gas-diluted and isn't changed frequently, that in itself can reduce ring sealing and lead to excessive ring/bore wear. Suggestions that have been made in this thread are worth pursuing. Sooty plugs and high CO/HC obviously point to a too-rich mixture (or one that isn't being properly burned, anyway), and the rhythmic shudder and drop in vacuum could indicate valve issues, possibly timing-related; checking that the crank pulley is tight, the crank keyway isn't worn, and the timing belt hasn't jumped (among other things) is worthwhile.

 

Just because there aren't DTCs doesn't mean that a particular part or system is healthy -- pre-OBDII left something to be desired (as does OBDII ). I'd suggest that if you're going to tear the engine down, inspect things carefully before or as they come off; you may find one or more obvious problems in the process.

[salewit]

You know what? Yesterday I ran both the OBD codes again I & II, and I must have done something wrong the first time (I got no codes at all). This time I got a million of them. Some I recognized as part of my testing (injector malfunctions on the ones I pulled the cap off of). So I disconnected the battery last night, re-installed this morning, took a short drive and got a sole code 24. Air Control Valve. Problem is my repair book makes no mention of this device. Any ideas what or where it's located? Or even how to test it?

[OB99W]

[...]So I disconnected the battery last night, re-installed this morning, took a short drive and got a sole code 24. Air Control Valve. [...]

Try looking for info on "idle air control valve" or IACV (or IAC). Have a peek at http://endwrench.com/images/pdfs/Fuel.pdf , page 6. Also, http://endwrench.com/current/spring04pdfs/InsiderInfo.pdf , page 30, last section.

 

By the way, a single short drive after clearing codes may not be enough to get accurate info; some codes aren't set until a problem occurs on two or more consecutive drive cycles.

[salewit]

Hate to resurrect this old thing, but today I started work on pulling the engine out. Just for the heck of it I thought I'd check out the valve timing. I pulled the left and right cover and tried to get the marks to line up. I couldn't, so I pulled the crank pulley and removed the middle cover, lined up the crankshaft mark, and the right camshaft lined up perfectly, but the left was off one or two teeth!

 

I ran down and got a new belt and installed it. The question is, was this my whole problem all along? Should I put everything back or continue on with the engine removal?

 

The IAC code I was getting was because I forgot to reconnect it after doing a resistance check on it.

 

Very blurry pic of marks on left cam pulley. Crank and right cam are lined up and we're at TDC on #1.

[OB99W]

[...]I pulled the left and right cover and tried to get the marks to line up. I couldn't, so I pulled the crank pulley and removed the middle cover, lined up the crankshaft mark, and the right camshaft lined up perfectly, but the left was off one or two teeth!

 

I ran down and got a new belt and installed it. The question is, was this my whole problem all along? Should I put everything back or continue on with the engine removal?

 

[...]

The answer to your question is "maybe". I'll point to two things I said in post #24 of this thread; the first is "... the rhythmic shudder and drop in vacuum could indicate valve issues, possibly timing-related; checking that the crank pulley is tight, the crank keyway isn't worn, and the timing belt hasn't jumped (among other things) is worthwhile.". The other is "I'd suggest that if you're going to tear the engine down, inspect things carefully before or as they come off; you may find one or more obvious problems in the process."

 

In other words, even if the engine is somewhat worn, that doesn't mean finding and resolving a problem such a jumped timing belt wouldn't be enough for you to then be satisfied with its performance.

 

One thing to consider is why the belt jumped. You should think about replacing the tensioner and idlers.

 

Considering the timing misalignment, that cylinder bank would have trouble performing well. If I remember correctly, each tooth corresponds to 15 degrees of cam rotation; even one tooth off would significantly impact performance, and two would make it very poor. In fact, if the other bank was also off as much relative to the crank, I doubt the engine would have run at all.

 

Whether to just put things back together is a judgement call, but it might be worth the gamble. Do consider parts other than the timing belt, however, now that things are apart.