Help Diagnosing coolant loss - Turbo Subaru EJ25

Hey all,

I have a MY09 Subaru Forester XT, and I am about at my wits end.

Long story short;

The car over heated due to a air lock from replacing turbo's.

Engine was pulled out, heads were sent off to be reco'd

Purchased a brand new shortblock and put on the reco heads

same thing - still loosing coolant and now smoking

Took the engine back out - got the heads checked over again, machine shop found a crack that said was near a oil gallery and exhaust valve (they did not pressure test the heads during the reco), so could be my issue for smoking. but were not confident it would cause coolant loss.

Sourced another set of heads, they were pressure tested, tested ok, then everything was reco'd again.

Fresh gasket kit, used the same shortblock.

Same thing - Still loosing coolant.

Current setup;

2nd set of Reconditioned heads

Brand new STI Shortblock EJ257, would be 2.5 tanks of fuel old now.

Silicone Turbo Inlet

PCV valve Deleted, crankcase and valve covers venting to atmo till can sort this issue.

VF52 Turbo (have put the stock TD04 on as well, same stuff)

ARP Head Studs, toruqed down in three steps to 90 ftlbs as per instructions, in sequence shown.

Genuine head gaskets

Air Pump/TGV Deletes

Throttle body coolant bypass (used little rubber caps on the two lines)

Where do I go from here?

The bottom end is brand new from subaru, the heads have been reconditioned and I have been assured these ones have passed the pressure test.

Is there any other way I could be loosing coolant? I have vacuum filled the system and pressure tested the cooling system. I can get it up to about 17psi (just before the rad cap opens) and it will drop a psi every two minutes or so (didn't time it exactly).

I can see some little bubbles in the expansion tank through my spill free funnel when I run the engine.

Car seems to drive fine, just needs about 100ml in the overflow bottle after almost every drive....not right for basically a brand new motor.

When you pressurized the system do you see anything leaking?

Thanks for the reply,

Unfortunately I do not see anything leaking while having the system under pressure.

Have you checked the oil to see if it has gone milky like coffee?

I have the same dumb car, albeit with a stock turbo and now a missing connecting rod. (well, not missing, just not connected to the crankshaft anymore.)

Here's an easy test for bad head gaskets or a crack in the engine

Fill your coolant system to the brim with water (we're expecting it to leak so no need to waste coolant)

remove all 4 spark plugs (easier from the bottom)

disable your fuel and spark (or, don't worry if you can easily access the starter

pressurize the coolant system to 15psi - give or take, we just need some pressure in it.

wait 5 to 10 minutes or until you lose 10psi

A.) Unable to access the starter?

A.1) Pull the injection and ignition fuses

A.2) crank the engine and listen for water being forced out of the plug holes, then go take a look. Are any the the spark plug holes wet?

B.) able to access the starter?

B.1) leave the key out, don't worry about fuses

B.2) Remove the "s" terminal from the starter (the little wire)

B.3) Put on safety goggles if you've forgotten so far

B.4) Keep your head away from the spark plug hole and bridge the starter "s" terminal to B+ to crank the engine

B.5) watch / listen for any coolant escaping from the spark plug holes

Hopefully that tells you which cylinder to chase.

Anyway, I replaced the rad before the engine went, yet can't remember if there were coolant lines into the rad from the automatic trans. I had cars pop the internal cooler before and leak coolant into the trans.

Also check the oil filter pedastal as I believe that is water cooled and could be cross contaminating. Check your oil as Michael_Au recommended to help narrow this down.

Last thing I can think of is to check the caps on your coolant surge tank and radiator. The radiator should have the lower psi(kpa) rated cap.

Maybe even the possibility of a cracked turbo center housing, leaking coolant into the exhaust?

If none of those prove fruitful, remove the exhaust manifold and pressure test the coolant system again and you may see coolant leaking from an exhaust port.

I hope one of those helps!

-Pat

Thanks for the reply's fellas,

I have checked the oil, it looks alright but I have made some progress.

I ended up doing a leak down test on each cylinder, here are the results and some pictures down below.

Drivers side rear, 2% leakage. 90 psi in 88 psi out

Drivers Front, 90 psi in 87.5psi out

Passenger rear - 90 psi in, 90 psi out

Passenger Front - 90psi in, 90psi out

So looks pretty good to me - No bubbles could be seen in expansion tank during the leak down, fingers crossed I am in the clear.

While I was under the car, i noticed a small dab of coolant on my sub frame...

Upfron further inspection I found a leak! (Actually found three) - Just to confirm they were indeed leaks and not just spilt coolant, I wiped everything up and pressurised the system. Sure enough they started to drip.

So I have tightened up the clamps on the leak points and so far so good, pumped it up to about 17 psi and left it for about 25 minutes. It dropped half a psi of pressure, I assume this is pretty normal amount of pressure loss?

Now I believe I have another issue - the car seems to be running a tad on the warm side. I have a gauge reading the obd2 coolant temp, from my general reading (especially in our current weather - winter) mid to low 80c is normal during high way drive (100km/h) but my car is reading about 94c.

The temp drops radically during higher rpm driving (hard acceleration) but during normal cruise rpm the temp creeps up to about 94c.

Keep in mind the fans turn on at 97c, having a car driving at 100km/h with the fans on is not right.

My bottom radiator hose is cold and not warm at all, as is my thermostat area.

Any ideas on what could be causing this?

Some pictures

https://imgur.com/a/j0wbRUm

Did you use an OEM thermostat and it's clocked correctly?

I think you said you're using a spill-free funnel so the cooling system should bleed decently. After it's warm and the thermostat has opened, try holding the RPMs around 2,000 for 30 seconds to see if you can find the last bit of air.

Have a OEM thermostat and it is installed correctly. They can only go in one way, spring towards the engine, jiggle pin up top.

Will try another bleed, hoping its just got some air stuck in it.

What's the ambient temperature? I don't see anything wrong with that temperature if that's the max you're seeing, unless it's really cold out. It's not even 100C water temperature. What is the thermostat opening range? Check the service manual. For example, some open at 82C but not be fully open until say 95C. Fans running at 50 km/h or whatever isn't a big deal, that's normal on lots of cars, unless you are splitting hairs on fuel economy. Hell old cars had crank driven mechanical fans without clutches that just ran all the time.

I'm not saying there are definitely no cooling issues, but running 97C/207F is not cause for alarm. Lots of cars are designed to run that hot or hotter, especially things like new BMW's that will run at 105C.

The thermostat that is in the car is rated to start opening at 78c. When I tested it in the saucepan with a thermometer it began opening at 78c and was fully open between 85c-90c.

That might be true about most cars, especially modern ones - they would run hotter for emissions/fuel purposes. Those are what I am concerned about for my car.

I have seen reports of low to mid 80c~ during highway/freeway driving, mine is at least 10-15c higher then the others, so I feel something is up.

that said I ended up putting in a replacement used radiator from the parts car I have (same car but auto). This radiator was a genuine style one.

Temps are down a little bit, it seems to be cooling better then the eBay spec one - though this seems counter intuitive as the thermostat should be regulating the temperature, not the radiator?

I am getting about 90c just normal driving, but the odd thing was, that when I was going down hill the temps would start to drop quite a bit, down to 80-82c.

Now any normal car will drop temp during a down hill, no/minimal load and the rpms are up for engine breaking. I thought i would try shifting in neutral to take the rpm's otu of the equation and the cars temp still went down.

I tried it while coasting on a 100km/h stretch of road, sadly it did not do the same thing, it seems to be related to the angle of the car. Up hill the cars temps would go up, with 97c being the most I saw going up hill. Mind you these are still light throttle/load applications.

The temps did not spike down 10c after hard acceleration with this radiator installed and the bottom radiator hose is now hot.

Ideally in the 80c range would be where I want the temps to be during normal freeway type driving.

I am a little confused bout the down hill driving and how it would cause the temperatures to come down quite quickly - perhaps there is a flow issue somewhere?