[chrism]

I've exhausted the brain trust that is my coworkers (we're mostly mechanical engineers in the aerospace field, but we all do our own car maintenance) so now I'm turning to all of you...

It's about 20miles from my home to my office. I drive a 99 Outback 2.5 DOHC.

A while back the engine overheated about 2/3 of the way home from work, no warning, just the gauge spiked, and stayed high. The reservoir tank filled up and I pulled over. Letting it rest for about 20 min, I drove another few miles and it overheated again. Let it rest and made it home with the needle climbing again. So, I replaced the thermostat with an OE part. Didn't fix the issue. I would again get 5-8 miles (with the gauge reading normal for a while, and then climbing) and it would overheat and the reservoir tank would overflow. The upper radiator hose would be blazing hot and the lower quite cool. So, replaced the thermostat with a generic one from carquest. Same thing. I then went deep and replaced the timing belt, tensioner, idlers, and waterpump and put in the OE thermostat. After putting it all together I got one full day (to-and-from work) plus the test drive with no problems. Engine warmed up and the gauge read normal all the day long. So that's about 60 miles of normalness, and a friday. The car didn't move all weekend. On monday, I make it about halfway to work and the stupid thing overheats, reservoir tank overfills and all. Let it rest for about 30min, then get another 6 miles before it overheats again. So it rests until I go to drive home, get about 6 miles in, bam overheats! Rest about 15min, then get a few miles and overheats. I limp home. I get home and pull the thermostat. I've been driving for a few days now with no thermostat, the engine will warm up but the needle never rises above "normal". I tested the thermostat in boiling water - inserting and retrieving for about 20 minutes. It operates without fail.

Sorry for the long story, but does anyone have any idea what's going on? And no, the thermostats were not installed backwards.

 

[dralb]

I'd guess head gasket. You've pretty much eliminated everything else. (assuming you've bled the system thoroughly, they can be tough at times.)

 

[chrism]

Thanks for bringing that up, forgot to mention...had the coolant tested for exhaust, results came back negative. I also don't have any of the other symptoms, no thick smoke exhaust, no water/oil or oil/water mixing (drained the oil just to make sure). Head gaskets are leaking a little oil but no coolant. Runs fine other than the overheating issue.

I've also replaced the radiator cap. And attempted to bleed best I can. Since I yanked the t-stat, no loss of coolant over the last two days, I haven't had to add more. I've also back flushed the radiator, block, and heater core (all separately) and the water just flowed right through with virtually no debri.

 

[cardoc]

Clogging in the radiator is one option. You would have to be able to control pressure to force the water out the lower hose port only to get a good backflush.

You can't install the thermostat backward if you are using a Subaru stat, so unless the thermostat is sporadically failing, there shouldn't be a restriction there.

Since the lower hose is cold and the upper hose hot, it does lead to an indication that the stat is bad. Did you try a different Subaru stat? Getting bad ones out of the bag is possible.

 

[rasterman]

Generic stats are bad for Subarus. Get an OEM one.

 

[1 Lucky Texan]

Are there bubbles in the radiator or o'flow tank at idle when it overheats?

what test was used for the HG test? the color change test - or did someone use an exhaust gas sniffer?



there really needs to be a stream of bubbles to get a response on those - and if there's a stream of bubbles...welll...wouldn't that mean the HG was bad?

point is, people often have odd situations that lead to the overheat as you described, some odd load like an incline, or sometimes when exiting on a deceleration ramp - but may idle OK.