92 Legacy Stalling on Hot Days

[79er]

I've got a stalling problem that has got me baffled--I see a number of other threads that detail stalling issues, none quite fit what I'm experiencing. The common conditions seem to be hot afternoons (above 90) and a heat-soaked engine--the problem doesn't surface until at least 30-45 minutes of stop and go city traffic. Today for example, I was out and about with the AC on for about an hour when I first detected flat acceleration and stuttering or shuddering til about 3500 rpm (manual transmission) starting from a full stop at a traffic light. Turned off the AC but not much help--this insecure "will it go or not" through a few more stops, then total stalling. I was on a slight uphill incline and letting out the clutch--zero power--floored it--nothing. Missed the green cycle and pissed off all the folks behind me. Started revving the engine as the cross traffic got its orange and held it at 5,000--same thing--slight forward movement and stall--kept working the accelerator and managed to hick-up across the 4-lane intersection. Response improved after that a bit and I manged to get to my driveway a half mile away. Puzzle in all this is the engine never dies--idle is fine. Its just when any load is added it just goes flat with no response at all once the clutch is out. With the clutch in, revving from idle sometimes is a struggle but usually not. Short trips, or cooler mornings and the stalling does not occur.

 

Now a bit of history--I experienced the same problem for the first time (in spades) about a year ago driving I-80/I-76 from Nebraska to Denver, Colorado. Problem started at the state line after I filled the gas tank about 2pm on a hot July afternoon. The response was a little different from what I described above--I could get up to interstate speeds but slowly power just seemed to ooze away, I would downshift and keep the rpm at 4,000 to 5,000 and lose power, downshift again, and then when I could get no more than 25-30 mph out of it, stop for about 10-15 minutes for a bit of a cool down, and repeat the whole cycle all over--contemplating the cost of a 200-mile tow I figured I might as well just keep up the cycle and buy a new engine if that is what it came to. So 160 miles of that (about 7-10 miles at a time--7 hours) and it was 9 pm and as suddenly as the problem started it disappeared. Never experienced the same symptoms again until now just over a full year later. Initially a year ago I thought the problem was bad gas, but I rather doubt that now. Thought the O2 sensor might be at least part of the problem though I'd replaced it a year earlier. But in the semi-annual emissions check I had this last May, the numbers were pretty close to those on the test from two years ago when the sensor was new, and as I already mentioned response and driveability has been normal after that trip last July until the last week.

 

As far as engine diagnostics, the Check Engine Light has never come on. Last summer there were no CEL codes; a week ago I got an 11 which translates to defective crank sensor, but the service manual suggests I shouldn't be able to start the engine if either the crank or cam sensors are defective, and that has not been an issue. I've wondered if the knock sensor might be the problem, but is the retard table in the computer "wide" enough to actually retard the engine to a stall? I also thought fuel pump or filter might be the problem, but if I can get fuel to support 5000 rpm, 2-3000 should not be a problem. So I'm at a total loss at this point--I'm guessing it has to be a sensor issue that is going bananas at high heat and the computer responds by trying to shut the engine down. Or might it be some problem in the fuel vapor purge circuit? Any one with a solution to propose?

 

Woody

[Manarius]

This sounds to me like a dying Mass Air Flow Sensor (MAF). MAF's cause stalling and all sorts of weird behavior.

 

You can try cleaning yours with some MAF cleaner. I used 90%+ rubbing alcohol and a q-tip and it seemed to work okay. Unfortunately, if that doesn't work, you'll be forced to replace it. The sensors are not overly expensive as long as you pull them from junkyard cars.

[Fairtax4me]

That's a fun one.

Knock sensor sounds possible. So does a dead TPS.

[wtdash]

I'm sorry that I don't have a "Been there, done that- Fixed", but something I've read about to try, is pouring water on the sensor (not the wires/plug) to see if it helps eliminate possibilities 1 by 1...I'M NOT sure this is the smartest idea, but a local Subaru mech I trust DID tell me 'Subarus are very resilient to water'....his reference was to spraying down the engine bay to clean it, but should apply here, too.

 

Also, check the knock sensor for cracks. Common issue w/the <'95 EJ22's. A bad KS will put the car in 'limp mode' but should still drive.

 

GL,

TD

[Olnick]

I wonder if a bad coolant temperature sensor could be part of the problem.

[Legacy777]

Well this sounds like an interesting problem for sure.

 

What sort of maintenance have you performed recently, if any?

 

Just because you can rev the engine to 5,000 rpm, may not mean you're getting enough fuel when you start to load the engine up. I would probably start by replacing the fuel filter if you haven't done it recently.

 

Some additional tests/info, like fuel pressure would probably be helpful. Short of that, my stab in the dark would probably be MAF sensor.

[Legacy777]

Actually....the more I think about it.....the MAF sensor is probably a more likely culprit than I first gave it.

 

I experienced some stalling/hesitation several years ago. When I was crusing at freeway speeds the car would hesistate and lose rpms. Most of the time I was able to put the car in neutral and bring it back. On rare occasions I would completely lose all power, and the car wouldn't run. If I let it sit for a few minutes and start it up, everything would be fine.

 

I experienced this problem for over a year. It was intermittant and was pretty much non-existent during the winter months. It was the hotter summer months that I experienced the problem. It finally left a code to confirm my suspicions. I replaced the MAF and never had the problem again.