The timing was off one tooth on each cam in different directions. I was about ready to pull the engine and thought the timing was the only thing that hadn't been triple checked. I still don't know why it was running smooth when cold. Strange.

 

Happy Ending!

 

 

 

Background: I swapped a 92 2.2 and 5sp into a 93 that had a bad auto trans. Note: the 92 engine and 5sp did come out of a car with a front end wreck and I could not check it prior to installing it in the 93. I replaced the cam and crank sensors when I did the TBelt.)

 

The 92 engine runs fine prior to getting all the way up to temp (below 1/2 way on the gauge). Once its to temp, it will not idle and runs rough, like its missing. The temp stays below 1/2 always.

 

I've been through things may times, checking and rechecking. What am I missing??????

 

The only indication I have is #4 is running clean, 1 2 & 3 are rich, as indicated by the plugs.

 

Things I've done and checked:

 

Compression cold and hot 140 all four (good for 6500' CO Alt.)

 

New timing belt (runs fine when cold so timing is good, I believe)

 

Swapped coil pack, wires, and plugs with a good 2.2 no change on either car.

 

Changed Map Sensor, Mass Air Flow Sensor, MAP Sensor. no change.

 

TPS checks out for resistance and power, with in limits.

 

O2 sensor has good cross over once hot 0.2 to 0.8, and when cold its a steady .33 ohms.

 

Swapped out fuel rails, complete with injectors, no change.

 

The engine runs great when cold. With the good compression, I believe its a good 2.2. What am I missing???

 

I'm about ready to stick the high mileage (250K) 93 motor back in, but it would really bite to have this be an electrical problem.

 

Thanks for the help in Advance

i don't know this engine, but how's the coolant temp sensor? that could make it run real rich, since it never thinks the car has warmed up...

The problem continues!!!

 

I've checked the intake for blockages as one person suggested; its clean as can be. I miss stated the clean running cyc the other day. Plugs from Cyc 1, 2, & 4 have heavy dark fuffly deposits; Cyc 3's plug is as clean as can be.

 

I will swap out the temp sensor, but with #3 running right, I feel it is not a problem with a single sensor.

 

Ideas???

Located on its own bracket directly behind the engine on the firewall.

Same simptoms, got to the point it wouldn't move, oh, it would idle rough, but when you put it in gear it would cough and rack violently for a sec, then die.

Got progressively worse til it wouldn't idle after it warmed to op. temp.

 

 

 

Jay

Problem continues :>(

 

Changed the Temp Sensor - no change

Changed the ignitor unit - no change

Changed the throttle body - no change

Opened the air screw on the (90 Throttle body) it helped smooth out the idle, which is now up to 2000 rpm. Just confirms the rich condition.

 

Checked the intake for blockage - none

No vacuume leaks found

Checked the fuel pressure - 35psi

 

O2 sensor is reading in the high .8 volt range. Used a fluke 87 to record it as i drove at 40 -50 mph the average was 0.856v, very rich.

 

Number 3 is still very clean while 1, 2, & 4 are carbon black from the rich condition.

 

This is what has me stumpped, if it was a sensor, they would all be rich or lean (not 1 of 4). The way I see the wiring, all injectors have a common ground and are receiving the input to open from the ECU. All Injectors have the same voltage readings 13+ volts.

 

How can one cyclinder be lean and the others way to rich?

 

I'm taking a break on this one, but if you have an idea please post it.

Did you check vacuum? If you have a leak your O2 sensor would over compensate for running too lean and richen the mix. it's an easy check with a gauge and something else to rule out. Other than that I feel your pain about unexplicable problems:banghead:

D.

Try swapping ECU? May have a different fuel curve for AT vs MT after at temp?

I believe from another post (from memory) someone that did the AT-MT swap had to ground out one wire on the ECU that was creating the idle problem at temp, look around - or maybe someone will chime in.

Vacuum is steady at 23psi with normal throttle response.

 

I've tried both the AT MT ECU's they do run a little different, but both have the same problem--#3 correct other cyc rich.

 

The only thing I haven't changed is the engine wire harness, but both engine were running fine before the change as indicated by the spark plug condition (new non-running engine, due to the front end accident) and running the 250K motor.

Hi,

Have you thought of getting it onto a 4 gas emission analyser - if possible get two readings, one when cold and one when it's warmed up, I think you may be concentrating too much on the "too rich" idea, just because cylinders 1,2 & 4 plugs are black doesn't always mean that it's too rich - it could be that they are not firing efficiently enough so that they soot up - after all number 3 is ok !.

If possible please post 4 gas readings here !.

Regards

Dave H.

Also thermostat may be worth checking - check top and bottom radiator hoses to see if they both feel the same temp when guage says warmed up.

Dave H.

Both the upper and lower hoses are warm. No problems here.

 

The O2 sensor is reading .85v average when driving and it should be closer to .4

 

Thanks for the ideas, Larry

here's a hack, idea-- try it at your own risk--- keeps you going until you can sort it out.

 

Record resistance of the temp sensor driving the ecu. Find the highest temp where it still runs ok, note the resistance R1. Then check resistance at normal running/higher temp, if practical R2. Take the difference (R2-R1), add 50% to that value and buy that resistor from ratshack. Add it inline with the sensor wire. Will spoof the ECU to think it's still a bit cool.

 

Crappy mpg, not-great emissions, but it'll buy you some time.... note: your cat might not like it...

 

Mike

I said in another post you need to snip the mt/at identifier wire, did you do this?

 

Connector B48, pin 20

 

http://main.experiencetherave.com:8080/subaru_manual_scans/FSM_Scans/ECU_I-O_page1.jpg

http://main.experiencetherave.com:8080/subaru_manual_scans/FSM_Scans/ECU_I-O_page2.jpg

 

On the first page you can see the connectors, on page 2 near the top is shows the MT/AT identifier pin.

I did check the ECU inputs and where Pin 20 is in B48 there no wire. Nothing to cut.

hmm....maybe it needs a wire.

 

Anyone have a 90-94 legacy with MT that can look on that connector to see if they have a wire in that spot?