Looking like my 2007 Hawk STi Spec D needs HG again. Only rebuilt 30k ago / 5 years. Cosworth head gaskets used.

Bubbling in the overflow tank after boosting, and not all the coolant being sucked back into the header tank overnight.
Better in the cooler weather, but I'm not in any way in denial of the issue / most likely issue.

At 45k miles about 8 or 9 years ago, fully forged rebuild.
At 80k miles about 5 years ago, the block cracked, so rebuild again.

Now HG!
Are Cosworth HG's really any good? Anything better out there, more proven to last?

 

I thought the Cosworth stuff was good and I was recommended them recently. However I was also told about the RCM stuff so decided to go with them. Block hasn't been put together with heads yet though.

 

Cosworth (no longer made) or RCM are the only gaskets we would use (certain std builds we will use OPE SUbaru MLS)
were the heads overhauled / skimmed, what power are you running, were the original head bolts used or converted to studs (both need torqued correctly!!)
who built it??

 

Glad i never touched 2.5l
I went with 14mm studs on my 2.1l to avoid that issue but walls on 2.5l are thinner/more flex ,think cdb conversion is a must on 2.5l

 

Hi Alyn, it's Matt from Aldershot, we've spoken a good few times on the phone, last thing I got from you was an O2 sensor back in Feb this year.
Not sure heads were skimmed on second build, power approx 340bhp. It currently has the ARP head studs from the first built - re used them.

Len at Newbury built it. I may just be unlucky though....

 

11mm and stretched after 1st use
14mm is a way and cdb conversion

 

I didn't think ARP head studs stretched though?

 

They should have checked them for it before putting them back ,even 14mm could stretch.
You do not know if they skimmed a heads before assembly what they should do and bill You for it.
No good You need to do it again

 

they don't in general...........

 

in general but its possible

 

Bit of a plot twist here. Baffled as to why my head gaskets would fail or head lift would occur given I have ARP head studs, I slipped into a bit of denial I will admit.

I hit on the idea that perhaps I had an air leak into the cooling system, and the bubbles I saw in the overflow tank at hot idle were air rather than compression gases.
For the last 6 years I've been running a genuine STI 1.3 bar header tank cap.

I replaced the cap with a genuine new genuine 1.1 bar cap. I reckon my 1.3 bar cap wasn't sealing properly on the vacuum valve and every cool down was sucking in some air and not so much coolant back from the overflow tank.

Hopefully the 5p bubbles will stop after a good few more heat cycles. There's no nasty bubbling in time with RPM pulses, block tester fluid remained blue, no temp spikes. Soft upper radiator hose when cold and on a cold start it doesn't pressurise immediately when cold and so on. True HG would bubble like mad from cold...given the PSI escaping from the cylinder.

Fingers crossed the bubbles are just air escaping and will to a certain degree stop altogether.
Pretty sure my old cap was sucking in air from atmosphere, and the air is still being purged out. There's now so much more coolant being sucked back in overnight, a massive difference.

I was preparing for a £2k HG bill, but I may just be lucky for under £20......

 

Hopefully it will be just that.
I have fresh engine ,5 months driving a car than on trip to EU headers blowed.
Was told burnt valves,prepared for engine out ,took a car to Rally Team in my town and nothing wrong with an engine.
Swapped headers for new rcm twinscroll one and job done.
Was preparing for 3k+ bill.

 

Fingers crossed that's all it is!

 

I really hope so!

I took off my radiator cap today, I cleaned the two seals and the radiator neck, I may just get a new cap from ICP as I think that cap is about 10 years old maybe. Looked ok, but the spring was a bit gritty so washed it out with clean water whilst pressing the spring which has made it a lot smoother. Also cleaned out the overflow tank today - mint now.

I also will most likely do a coolant flush / change. I hear Comma G30 is ok?

 

At work we use uview combustion leak tester. May need to empty out the Header tank a bit in a scoob though.

It can be pretty sensitive: It's confirmed a porous gasket on an 4cylinder ecoboost (common block issue: New short engine required) in its early stages, where the only symptom was an intermittent misfire that would randomly come and go.


There are cheaper kits, but can't say if they are as good.

Comma have narrowed their range, they used to do a hybrid-oat coolant for Japanese vehicles, but now I think the closest is G30, so long as it's flushed beforehand it should be ok.

 

Yes I have one of those, always stayed blue thankfully. The overflow tank bubbles have all but stopped now so it would seem the system has now pretty much purged itself.
My old header tank cap was letting / sucking air in on cooldown, so each drive the bubbles were there in the overflow tank, somewhat creating the impression of combustion gases. Now the system is sealed properly, fingers crossed all is now well.

Combustion leak of any size would've been so much more violent than the gentle bubbles I was getting in the overflow tank. I've had clear as day 100% HG on this car before so I know only too well. They would be constant from stone cold start and so on. The top rad hose would be rock hard almost instantly on cold, would probably stay firm overnight etc.

There was also slimy scum around the seat at the lower part of the header tank neck which I doubt helped. Probably just old coolant from where I've checked the header tank level, but that's all cleaned now with IPA so a nice very tight seal on the new cap.

So far the overflow tank is going up by an inch or so once hot, and going back down the same - to my set level which is about an inch up from the L level on the overflow tank when cold. I think it's now finding it's equilibrium point.

May just get a new radiator cap now too as that's been on there about 10 maybe 12 years. The new header tank cap is a much much tighter / better fit that the STi 1.3bar cap ever was.

In a nutshell I think the coolant system was just airlocked. Slightly blocked vacuum hose and dirty seats where the cap seals should seal. Failing STi cap? Possibly, but back to 1.1bar now for me, fits / seals brilliantly.

Time will of course tell, but I'd like to think with Cosworth HG and ARP I ought to be ok.

 

Following on with a round up of my findings...
My issue was coolant not returning into the header tank from the overflow tank by the radiator - a classic HG sign, but I solved that for mine by replacing the genuine STI 1.3 bar header tank cap with a genuine normal 1.1 bar header tank cap - I think the return valve on the STI one had failed or at least become weak. This cap was 6 years old.

I also had I think a minor blockage in the return line from the overflow tank to the header tank - quite possibly the cause of everything. I took the hose out of the overflow tank, blew through it and I was met with a very minor resistance, possibly a small blockage somewhere, but seems clear now.

Lastly, the actual radiator cap - I took it off and compressed it, the spring loaded mechanism felt quite crunchy and pretty stiff. This cap was 12 years old, I cleaned it with hot water and it freed up enough fine, but for the price I replaced that too. A good clean was perhaps all it needed but...

I now have the same amount of coolant overflowing into the tank by the radiator as is being sucked back into the header tank on cooldown vacuum overnight....about 1.5 inches.

I do still get flash vapour pockets (usually boost related) degassing into the overflow tank at hot idle after a drive (5p sized bubbles) one or two every second but that's normal, especially on a built 2.5 - again can be HG, but...when the cooling fans come in, the vapour bubbles inside the coolant system collapse / dissolve back into a liquid state and stop showing in the overflow tank - true HG would not do this. True HG would turn the overflow tank into a jacuzzi - been there had that. HG compression gases do not change behaviour with cooling fans etc, micro vapour pockets do.

I think most of the flash vapour is from the turbo coolant loop that cools the CHRA - turbo core. The vapour pockets are micro sized (0.3mm) bubbles that when they reach the overflow tank have joined together and become one bigger bubble as they effectively vent to atmosphere by Subaru design - bit like a fizzy drink bottle, you don't see any bubbles, only when you open it. Possibly heads lifting after boost? - not very likely with ARP clamp force. Plus, exhaust gas bubbles don't dissolve like micro vapour pockets do when the fans come in.

Soft top radiator hose overnight - HG would remain hard. No bubbles in the header tank or overflow on cold start up - HG would bubble immediately etc.
At first I was preparing for a HG bill of around £2k but after a ton of research and only 35k miles ago having a fully forged build with Cosworth HG's / ARP head studs, I thought I'd try the caps first....glad I did.

I did not have any overheating conditions throughout any of this by the way. I reckon there's tons of 2.5 STi's out there that bubble in the overflow tank on hot idle after a run, but people don't tend to look for it, they perhaps just look at the level once switched off or check when stone cold.

There's every chance that my old 1.3 bar STi cap began sucking in air on cooldown, the seals on that cap were always a much softer almost silicone type seal, so not so much vacuum was created if they had perished slightly.

I also had heater gurgle for a few seconds when pulling away cold, but I also got rid of that.....
I started the car with the heater on hot, but the fan off. Drove for 10-12 minutes off boost then put the heater fan on about to speed 1. After about 10 seconds I heard the air bubble migrate out of the heater core...not heard it since.
The next morning, I noticed there was a slightly bigger suck back of coolant overnight to cater for the now dislodged air, maybe approx 35ml of coolant, not much on the scale of things.

Also, always keep the necks of where the caps are fitted clean, any slight imperfections or grit can see air being sucked in during cooldown. Both my necks had a very thin layer of scum which no doubt wasn't helping.

Sorry to rattle on folks, but there's some info there that is accurate to my findings, micro hot spot / vapour behaviour versus HG compression gas behaviour and so on.

 

Good news. There's a few cautionary tales about running a higher pressure cap though.

 

Yes. Plus the STi cap was always just an "ok" fit on my header tank, but now going back to 1.1 bar, it's so much more snug. I've no doubt the STi cap was genuine, but perhaps fits better on the metal header tanks rather than my plastic one, don't know.

I think EJ isn't too bad at self purging air out, but you do need decent caps, hoses, clamps / return lines to enable it to do it's thing. Fingers crossed that my overflow tank levels continue to behave like clockwork, time will tell, but with sound caps, the first thing I'll be checking for is a clear overflow / return route if I get further issues, some of those metal pipes and the nipple on the header tank neck are pretty small.
.

 

All of above about cap.
Its 2.5l ,it will go again in short time anyway.