I'm losing coolant from my STI-5 and have the annoying task of figuring out why and from where! There doesn't appear to be pools of water forming under the car when you leave it in the same place for a few days and when sitting at idle water doesn't appear to be leaking from anywhere.

My question is does the cooling system opperate at different pressures according to revs and/or engine load? Could it be the case that at higher pressures its forcing its way out somewhere? The useage does seem related to how hard I drive the car (hard=more leakage). If the main reservoir drops more than 2-3" from the top she can no longer circulate at idle making traffic a real concern.

I'd therefore appreciate some background on how the Impreza's cooling system works and where any known weak spots are.

 

The usual suspect for the loss of coolant is via header tank, pressure builds up and forces it to overflow from there. Coolant system is simple really, higher temps = more pressure, it may be possible you have a leaking headgasket which is pressurising the coolant system.

Andy

 

Can anyone recommend a sequence of checks/tests to methodically identify the likely culprit? I'd obviously prefer to start with the simple, easy, cheap checks before stripping the heads off to check the gaskets, etc

 

radiators are a weak spot. when car is stationary they tend not to leak, however under presure you get a jet coming out from the end caps which can be difficult to see. Best bet is in daylight feel the side of the radiator and look for any apparent cracks =(this can be difficult to find), another way (can be dangerous) is to get the car upto operating temperature, open the bonnet and open the throttle by hand to see if you can see any jets squirting from anywhere. The top rad hose and bottom rad hose can also perish with age as well.

 

I had a mysterious leak which I traced to the coolant hose to the turbo being split. The coolant was simply evaporationg on the turbo (could smell the coolant though).

Why not have a garage pressure test it for you ?

Midlife.....

 

This would suggest that the pressure is related to revs rather than engine load. Yes/no?

Midlife, there is some residue on the turbo heatshield which might be burnt off water from that pipe. I'll give the shield a good scrub with a toothbrush and drive it on a dry day to see if it comes back. Pressure test may have to follow if I can't find the bugger.

 

No, its just easier to put your head under the bonnet & look whilst stationary than to try it whilst you are driving

 

Agreed!

Seriously though, as revs increase so does the pressure of water being circulated around the engine etc

 

Pressure test would be the most straight forward option.