nmos

When I warmed it up before the block test (got the temp up to around 170) I did get a few bubbles but not many.

This is a 97 Subaru Impreza OBS with a 2.2l engine. Long story short, I've been chasing a case of excess coolant going into the overflow container and ultimately being lost followed by mild overheating if I don't top off the radiator each morning. I first tried replacing the radiator cap with a new genuine Subaru cap but no change. I tried flushing the cooling system but when I removed the lower hose to drain the old coolant/flush I found a large "dent" in the lower outlet of the radiator which was restricting flow. I replaced the radiator but used a Autozone/Stant thermostat (shame on me) and the problem remained. I tried running with the thermostat removed while waiting for a genuine Subi thermostat to arrive and that seemed good for a couple of days but now the problem is back (and it has been warmer lately). At this point I started thinking head gasket. I know the 2.2 isn't known for this but mine does have 230k miles on it. Today I borrowed a block test kit (the "turkey baster" with the blue fluid) from Autozone and followed the directions. I tested for 3+ minutes with the engine getting over 190 degrees and saw no change in the fluid suggesting that it's not a head gasket leak. One other piece of evidence is that I've been noticing a buildup of "black crud" in the overflow tank but no obvious signs of oil in the coolant or coolant in the oil. So, my question is what is going on? Is it possible that the head gasket could only be leaking under actual driving conditions but staying sealed at idle? Could I really have gotten THAT much air into the system otherwise? Any ideas?

Ok, the final outcome was that after replacing the relay the new fuse popped again so I started looking around that circuit for shorts and it turns out that the wire (yellow/red) to one of the O2 sensors was rubbing/shorting against a metal bracket in the engine compartment. After repairing that it's running great. FWIW I gave up getting the old relay out and just disconnected the plug and attached it to the new relay. It's also worth noting that even the Subaru dealer didn't have the fuses so I ended getting a bunch as well as the relay from the local U-Pick-It for a total of about $20. One other lesson learned. I made things worse early on. It was a holiday weekend and I couldn't get a hold of proper fuses so I used a larger 60A fuse on SBF2 for testing. That's probably what really did in my relay. I'm going to grab an extra fuse holder for the normal blade fuses and put it in my tool box for emergency use.

Maybe the AC Compressor is worn and putting more than normal strain on the engine when it kicks in? That jolt from the AC clutch kicking in could cause the false misfire reading. BTW, a transient error like that misfire or the knock sensor will cause the check engine light to stay lit for about 24 hrs after the last occurrence.

Imdew: Thanks, hopefully I'll be able to fit a nut driver in there and get the bracket down into a more accessible position. You wouldn't happen to know how that main relay is attached to the bracket do you?

I need to remove the main relay on my 97 Impreza (see my thread titled "Won't start. No Fuel and No Spark") but it's tight in there and I can't see how it's attached to that metal bracket. Any tips on getting this thing out of there?

johnceggleston : Yep, it turns out all of those codes relate to things that are powered by the main relay via the yellow/red wire and I confirmed that there is no power to that wire with the ignition on "run" however things that are powered off a different set of contacts on that same relay are working. This whole thing started when I was chasing, what turned out to be, a blown fuse (sbf-2). It could be that when these contacts failed they caused the fuse to trip but it could also be that I shorted or pulled out a wire while trying to cram my large size hands up into that small size gap where those relays live. At this point I just want to get that relay & connector out so I can test but there isn't much room and I can't tell how it's attached. I can get one hand on it pretty well and I'm half inclined to just snap it off but I really don't want to break the connector. I'll probably start a separate thread on that.

Ok, so I checked the voltage on the middle pin of the IAC cable and there was no battery voltage as Haynes manual says there should be and it looks like that is powered via the main relay under the dash right below the fuel pump relay. Either that relay has gone bad (unlikely since the fuel injectors are clearly working) or I pulled the wire lose while I had my hand up there trying to get at the fuel pump relay earlier.

Not so fast! I replaced the fuse for sbf2 and the car starts but I'm getting several error codes. I've cleared them but they keep returning. The codes are: P0505 Idle Air Control P0141 O2 Sensor Heater Circuit P0443 Evaporative Emission System Purge Control Valve Circuit P0135 O2 Sensor Heater Circuit Any thoughts?

It looks like I found the problem. The fuse for SBF2 was broken. It wasn't blown in the usual way but broken near one of the blades. I would never have been able to tell if I hadn't actually pulled and tested it. These appear to be an oddball fuse (fl30a ?) so I may have to wait until tomorrow (I'm posting this on Labor Day) to get a replacement.

I have a 97 Impreza OBS. Yesterday I had pulled off the road and as I was I was getting back on a chunk of loose asphalt hit the underside of the car. I think it hit the suspension just inside of the drivers side wheel but I could be wrong. Everything seemed fine at first, engine ran fine, temp stayed normal and no trail of fluid behind me but about a quarter mile down the road the engine just quit and wouldn't re-start. The starter turns the engine over but not even a cough from the engine. Here is what I know so far: * The fuel pump is not priming when I turn the key to run and there appears to be no voltage at the pump either then or when cranking the engine. * The fuel pump does run if I power it from the cigarette lighter and will happily spit fuel from the fuel filter in this configuration but the engine still won't start. * I don't believe I have spark but it was hard to tell under full sunlight. Can I use a regular 12v test lamp between the coil and ground or will I fry something? * There are several relays that ARE working, the one mounted on the kick panel under the dash, the "Main" relay (brown plug) and whatever that one just below the Main relay (white) that I can feel clicking when I turn the key to run but I don't think the fuel pump relay is activating but it's really hard to get more than just the tip of a finger on it. The wiring for this relay is equally hard to reach but maybe there is somewhere else I can test from? * There is NO check engine on with the key in the "run" position (there should be right?) * The OBD port appears to be dead. It's been a couple years since I've used it but I could have sworn it would light up elm based scanner without the engine being on. * I checked all the fuses except the one marked SRS with the "special" cover over it and they all looked good. BTW should I be pulling that one while I'm fiddling with the wiring under there? Is it powering anything else important? My plan for tomorrow is to have a look at the ecm. I'll attempt to check power and ground for sure and try to find and test the wire for the fuel pump relay. Any thoughts on where I should begin with this thing?

You are pretty much on the mark. It's got around 230k miles. Water pump was replaced at the same time as the timing belt but nothing else cooling related has been done since except for radiator hoses so the cap may be original. I haven't caught it overflowing again but it's my wife's daily driver and her last mile on the way home is slow but with no stops so it would probably be back down by the time she gets here. I'll have her leave it idling when she gets home and see what it does. In the mean time I'll order a radiator cap. I don't know if the t-stat is origional or not but I've never replaced it so it's at least 100k+ old however the temp mostly seems really stable. In other vehicles where I've had a bad t-stat the temp take forever to warm up (stuck open) and then would wander quite a bit but this one isn't doing that. If it were stuck closed the trip home that night would have overheated it for sure since it was 75+ mph driving up long hills etc (4k+ elevation increase). I've been putting off the timing belt because back 100k+ ago I made the mistake of having someone else do it and he didn't torque down the nut holding the harmonic balancer enough and it started to wobble loose damaging the key way and the belt did skip a tooth or two so I ended up re-doing the job and was forced to use loctite on the harmonic balancer due to the keyway issue (it kept wobbling loose otherwise) so I know I'm in for a battle getting it off.

Yes, both fans work.

I have a 97 Impreza OBS with the 2.2L engine. About 11 days ago I was half way through a 2 hour drive back home on the freeway and I decided to pull off and get something out of the trunk and I left the car idling. While back there I heard the cooling fan kick on and when I got back in I saw the temp had rising to maybe 3/4 of the way up the gauge. I drove maybe 50 feet further to get under some better lighting and the temp immediately dropped back. When I popped the hood I could see that the coolant overflow bottle was full and had probably already spit out some coolant. I kept a close eye on the temperature the rest of the way home and it stayed stable right where it normally is but the coolant level was definately low after I got home and let the engine cool off. I topped it off and have been keeping an eye on it but today my wife mentioned that she heard a gurgling sound when she started it in the morning and sure enough it was low again. I'm NOT seeing any bubbles coming up through the overflow bottle so I don't think it's a head gasket. I should mention that maybe a week before this incident I had noticed that the engine seemed to be idling just a little less smooth than normal. It is due for another timing belt and I thought maybe the old one had stretched a bit throwing off the timing. My two theories at this point are A) loose timing belt slipping a bit on the water pump or B ) radiator cap. Any thoughts?

I know this thread is ancient but I just wanted to say thanks to Numbchux for actually posting the correct size socket once he found it. I'm in the same boat as he was when he first posted and he probably saved me at least 1-2 trips to Sears.