One or two bubbles in overflow tank - normal?
 

I had an issue previously on a newly built 2.2 stroker engine where it would lose coolant when driven hard. Coolant was not being forced out the overflow, but the overflow would usually bubble constantly:



Initial symptoms post engine build:
- Coolant loss on boost
- No overheat
- Overflow bubbling


Eventually, had the engine decked and gaskets replaced (workshop confirmed the HG were starting to go and there were signs of coolant entering the cylinders). However, post rebuild, the engine continued to consume coolant but without any bubbling observed.


Symptoms post rebuild:
- Coolant loss on boost
- No overheat
- No bubbling observed


We tried using worm clamps on the turbo coolant hoses, and changed the combination of rubber tubes and metal pipes leading from the upper reservoir to the overflow tank to a single silicone tube. Miraculously, this seemed to fix the problem and the engine was no longer consuming coolant. However, although I still generally don't see any bubbles, I have twice so far seen one or at most two pop up sometimes in the overflow (approx. 20% of the time). Most of the time I don't see any bubbles. And when I do, it is just a single bubble on its own, unlike the previous rhythmic bubbling as per the linked video (which turned out to be a bad HG)


Symptoms post changing hoses and clamps:
- No coolant loss on boost
- No overheat
- Occasional one or two isolated bubbles observed


Any ideas on what could be happening? HG starting to slowly go? Cracked block? Or just normal operation? Am almost at my wits' end. Maybe will try a leakdown test next. Any advice will be much appreciated!

Spec?

Turbo, boost, engine internals? Head studs? Ecu and map?

Could the heads be lifting on boost?

Thanks for the reply. Base car is a 2007 STI. Parts used:


IHI VF42 turbo, boosting at 1.6 bar (Defi shows peak at 1.7 bar but they are known to over-read).


Mahle 4032 pistons, Carrillo Pro-H rods, 79mm factory crankshaft, JDM GRB EJ207 block (bored to 92.5mm) and GRB EJ207 heads /w exhaust cams sealed for single AVCS compatibility with stock ECU. Stock valves. ARP head studs, ACL bearings. Mapped using Ecutek (will try and get a recent chart but it is making about 370hp at the crank estimated).


Things I'm considering are heads lifting, or head gaskets (again?!) or a crack somewhere in the block.


Other than the one or two bubbles on maybe 1 out of 4 times I check, the engine seems perfectly fine. No coolant loss, and water temps even drop when I am running hard (high boost, high rpm). No overheating whatsoever (not yet tested on track though). I recently swapped fan motors and saw the mechanic taking the hose out of the overflow briefly to place it somewhere while the car was cooling down. Could that have introduced a bit of air into the system? That was about 50km ago however. How long (or how many km) will it usually take for air to get purged out of the system?

Maybe small leak somewhere letting some air in? I get bubbles in mine sometimes like yours after driving it really hard, but it's my turbo boiling the coolant. I can feel the bubbles in the coolant line from the turbo to the turbo tank.

I'm beginning to think it's the head gaskets, unfortunately. The bubbles are getting more frequent, and after a recent hard drive, I gave the engine a few revs and checked and could see a stream of bubbles pop up in the overflow. It doesn't bubble all the time, but usually after hard driving and some revs.


When it bubbles, coolant level will drop slightly. Also I notice that the boost builds slightly more slowly, and peak boost pressure is about 0.1 to 0.2 bars lower than before. This seemed to coincide with when the bubbling started to get more frequent.


The engine still doesn't overheat, and in fact the coolant temp drops as I rev it higher. Will look into doing a leakdown test and the sniff test, but based on my experiences on the previous HG failure, this looks like early stage but heading in that direction as well.


Anyone knows what else could cause these bubbles besides a failing HG?

Headgasket will normally blow the coolant out of the overflow on boost.

That is what I thought as well, and from what I've read almost all cases of coolant being pushed out the overflow have been head gasket failures (a few being misrouted coolant hoses).


However, I'm not sure what else could account for the bubbling in the overflow that seems to coincide with driving under boost and engine revving?

I have a small amount of bubbling in the over flow tank, my mates is the same. Think i noticed that after changing the radiator on mine and same on my mates. had that for years with no loss of water. As long as its not over flowing as that would be a sign of head gasket fail.

It's not overheating and not pushing coolant out, but the size and quantity of the bubbles seem to be directly related to boost.

If I drive it hard, then stop the car and give it a few decent revs before checking, the bubbles are numerous and about 1cm in size.

If I take an easy drive and give it a few revs and check, the bubbles are few and about 2mm to 3mm in size.

I'm hoping otherwise but it does seem like HG, doesn't it? Or could it be cracked heads or block?

If its loosing water maybe, if not then should be fine. Pipes tend to go, my mates just went on his few weeks ago, top hose.

Under hard boost in some cases the heads can lift using standard size head studs regardless of being a top brand such as arp ,RCM etc which is usually the reason for being fine day to day normal driving then losing coolant under boost , hence why on big builds with big turbos 14mm head studs are used to keep heads clamped down

The gasses are only coming from one thing.!

Had this twice now and it will only get worse.
It might last a year or a week. Eventually it will end up pushing coolant and the expansion bottle will be full all the time. Never had a overheating problem tho.

Both times Hg had gone. You may also find you hear bubbles in the heater matrix when starting the car from cold.

Just get a sniff test done, this will tell you what's in the bubbles, if it is hydrocarbons then your HG is failing,simples.

Sniff test wont show if slight hg fail.

Why not?
The bubbles will contain gases if the HG is failing, this is what the test is for, testing for old combustion chamber gases in the cooling system, if it is just air then the test won't change colour.

The test didn't work for me when I had my coolant spilling out of overflow with loads of bubbles.
Tends to let gases in on boost so when doing a test the car is at idle so won't let gases in then.
If hg fail is quite bad then I'm sure the sniff will change colour, but not if slight failure.

If the bubbles are caused by HG failure the bubbles should be exhaust gases and hence fail the sniff test.

Should but not always will. By the time you have bubbles, they have already reached the overflow, so when testing on idle there won't be no bubbles in header tank at that point, until you put car under boost then it will return.

I've tested the bubbles on mine whilst its been bubbling away and it will not fail the sniff test.

Well mine didn't change colour, maybe I should of gone for a good blast before doing test, but then I wouldn't be able to open cap without it killing me :)
Maybe yours was entering gases while idling.

Mine isn't blowing coolant out.

Mine passed a few sniff tests at the garage.

I still wasn't convinced so I hired a hydrocarbon test kit and it failed,
I took the car out and give it some prolonged boost in 4th then pulled over at the side of the road,

There was lots of bubbles in the expansion tank and at first the test was still negative.

I then left the tester sealed in the top of the expansion bottle for a few minutes and let the gasses to continue to fill the expansion tank.

I then squeezed the tester again and the fluid instantly turned green confirming the hg had gone.

Im guessing that removing the lid off the expansion bottle let out the gasses and caused it to initially pass the test.

^^^^thanks for clarification to what I was saying. I had to just pull it apart and see was my option and was then confirmed.
But tbh I only block tested it the once so.....

I'm 99% sure mine is actually a leak.

I just need to find it.

Ok so as an update, it got to the point that the overflow tank would bubble even on a cold start, before even driving an inch.


I brought the engine into the shop to take apart, and apparently, what I'm told is the HG shows no signs of coolant leaking into the combustion cylinders, and the heads also appear flat based on a quick ruler test.


I'm pretty stumped now as to what could be causing the bubbling, the loss of coolant, and the loss of power, if not the HG.


If the heads are lifting, and exhaust gases are getting out, will this show up on the HG? Is it possible for the HG to look fine while at the same time allowing air from the cylinders into the water and vice versa?


Or am I looking at a cracked head / block here?

It's looking like the latter, if no signs on gasket

What block has been used on your build? Ej25?

It's a used EJ207 from a JDM hatchback GRB STI, which was bored out to 92.5mm and fitted with a 79mm crank for a 2.1 stroker setup (also detailed in post #3 above).


I went this route because I needed an EJ20 setup that could work with my 2007 STI which has 32-bit ECU, e-throttle and single AVCS (I had the exhaust cam modified on the GRB engine so it would remain in a fixed position).


The strange thing is, if it is indeed a cracked block, it seems to have happened / got worse over time as I put more boost / revs on it, since there was a stage where it was holding boost, wasn't losing coolant and only had isolated bubbling. Now it is losing power, losing coolant, and bubbles constantly. All this happened over maybe 100km of hard driving.


Are Subaru blocks known to be prone to cracking if overheated? The engine did overheat a couple of times when it was just put together previously before the recent rebuild, due to the shop not properly burping the coolant system.


The shop is now telling me that the hatchback engines are not as good as the GD ones, due to revised water jacket layout and size... is this true? I know the water jacket is smaller compared to the EJ207 up to 2007, but still...

From what I know the hatchback 2009on JDM block is actually a thicker liner than the pre ej207 than a ej25 uk 2.5 block so I believe garage is incorrect.
It's a hard one to call tbh. Still a semi closed deck block though.
Most use a cdb for 2.1 and a ej22 for a 2.235 etc.
I'm not experienced enough to tell anymore tbh, but I would get that block pressure tested somehow for cracks etc.
There definitely seems to be some gases get in though.

Just to clarify, have the garage took the heads off the block to actually see the head gasket for signs?