Coolant Pressure playing silly buggers!

[tencentlife]

What you are describing sounds like the beginnings of a fire ring (multi-layer steel combustion seal) leak. Combustion gases are being forced into the coolant envelope, but only under high cylinder pressures that are present at wide throttle openings. Gas bubbles will be entrained with the liquid flow to the radiator, where they will accumulate at the top, and system pressure will rise to above the pressure cap's break-open point (1 bar), causing liquid to be displaced to the reserve tank (top-up tank).

Once a certain volume of gases accumulate at the top of the radiator, radiator dynamic flow is interrupted; at that point you will see the coolant temp gauge go to the top of the scale and more severe engine damage is imminent, such as warped cylinder heads.

You can limp along with this condition for awhile by occasionally returning coolant from the reserve tank to the pressure tank manually, and releasing the trapped gases from the radiator top bleed screw (just crack it open while the engine is warm and coolant system pressure will force the gases out). The fire ring leak will get steadily worse, though, since the hot gases burn away material bit by bit when they are forced out the leak.

This type of leak at this early state will not be revealed by a compression or leakdown test. That's because at wide throttle openings where the leak originally occurs, pressure in the combustion chamber is well over 800psi, while compression test pressure will be well under 200psi and pressure used in a leakdown test is typically only 100psi. But as the leak worsens, it will take less and less pressure to force gases out, until you will see the cooling system pressurise as soon as the engine is started.

The reason you don't see coolant return to the pressure tank on cooldown is because the system has a volume of gases in the pressure tank and the radiator top. Gases can of course be compressed easily while liquids cannot, and normally there is nearly zero volume of gases inside the enclosed system. When you stop the hot engine, the gas pressure will probably be close to the cap's break pressure of 1 bar, then as the system cools the gas pressure falls but is still somewhat above atmosphere.

The only remedy is to reseal the top end, with new MLS fire rings, and checking the head and cylinder top for burn damage. I recommend using Hylomar on the fire rings, it is a non-hardening elastomer specifically designed for this type of high-temperature sealing requirement with differential rates of expansion between elements (Rolls Royce developed it for their large jet engines). It must be used according to instructions, applying it to one part, smearing that part against its mate to create a film on both parts, then the parts separated and the volatile vehicle allowed to gas off for a few minutes. Then the parts are assembled and brought to 1/2 to 3/4 of full torque, then wait 20-30 minutes for the elastomer to "creep" under pressure, and finally bringing the fasteners to full torque. This means both sides of the fire rings must be coated and smeared to the heads and cylinder tops. The head capnuts can be first assembled dry, then after all have been brought to final torque, taken off one by one and retorqued with sealant. Hylomar can also be used under the capnuts, which will allow retorquing if ever necessary without the need to drain coolant, remove the nuts and apply new sealant, since the Hylomar never hardens. There are several variants of Hylomar, get the Advanced Formula ("Hylomar AF") for its superior workability, or the Universal Blue spray if you are willing to mask your parts carefully . There is a special solvent available as well, "184", which makes clean-up easy, although I believe acetone can also be used. Hylomar products are easily found on eBay or Amazon.

[Smiffo]

Excellent write up TenCentLife.

Would one of those `sniff tests` I have heard about highlight whether that is the issue?

[itchyfeet]

Very interesting tencentlife,

I had always assumed the compression rings would fail suddenly and pop hoses or push coolant out but It's locical they could fail slowly and coolant is lost into the compression chamber.

Would this explain the erosion of aluminium between valve seats that you often find due to being driven with a coolant leak into the combustion chamber over a period of time, never seen a good explanation?

like this



or even this ( pic from MGP)

[tencentlife]

A viscous liquid at less than 15psi won't flow thru a leak space when the condition is such that only hot gases at over 40 times the pressure can force their way thru. Keep using the engine, though, and that equation will change.

Part of the gas-escape equation is also the extreme pressure acting to lift the heads off the cylinders, thereby lessening the clamping force on the fire rings. One bore is over 10 sq. inches area, so at 800 psi there's momentarily 4 tons of lift on a cylinder when it's firing at open throttle.

As I said, let this leak go on long enough and it will open to where gases will escape even at idling combustion pressures, causing the cooling system to pressurise immediately after startup on a cold engine. Once the leak is that bad, some coolant may force its way into the cylinder right after shutdown when there is residual pressure in the cooling system, and eventualy it will bleed in on the intake stroke, at which point the engine will be blowing quite a lot of steam. The vehicle has become practically unusable by that pojnt since it will overheat in mere minutes of running. But at the early phase of development, which the OP describes, this type of leak will normally show no evidence of coolant burning for the simple fact that none is getting into the cylinders. Yet.

Smiffo, a chemical test of the coolant can show dissolved CO, but that test can be unreliable until the problem has become fairly severe. Sampling gases from the pressure tank with an exhaust gas analyser can also show the presence of CO, probably on the whole it would find the problem earlier than a chemical test, but who has an old-school exhaust gas analyser around? The simple signs described by the OP are already sufficient for a diagnosis.

[Sir Brixalot]

'Fraid it sounds familiar. I had exactly the same symptoms a couple of years ago when i first got this van and engine needed rebuilding. It passed a sniff test by the AA as well but got progressively worse. It completely did my head in trying to fathom it out. Nicely explained by TCL. Fingers crossed though!

[Smiffo]

Nice one TCL...
`Hope for the best and plan for the worst` springs to mind here.

[300CE]

Many Thanks for taking the time out to provide an excellent write up TCL, t's much appreciated! I've forwarded this onto my pal for some Tuesday evening reading

My head was telling me that this issue could be the start of something expensive and, if I'm being honest, I'm getting tired of throwing money at the bloody thing even though I am lucky enough to have a mate who can do the work.

I'll see what he comes back with and we'll work out a plan.

Thanks again to all who have responded this is such a great forum!!

Cheers,

Justin

[300CE]

OK, just a quick update: Changed the thermostat today & re-bled the system. Wanted to replace the coolant sensor with another new one as the one I currently have is from a MK 1 Golf and temp gauge seemed to be reading slightly higher (3/4) when at temp. The new one I bought didn't seal properly and subsequently started to throw out coolant so put the old one back on!

I also added some of this:

http://www.steelseal.co.uk" onclick="window.open(this.href);return false;

Whether this will make any difference remains to be seen, but I'll provide an update as it's possible it might do a job for the time-being whilst I look into getting the job done properly.

[300CE]

Further update - looks like steel seal has done its job - no leaks at all when I came out this morning (had little leaks from any weak point due to the pressure build up, mainly the coolant hose to carb).

Took off the header tank cap and hardly any pressure release at all which allowed me to top up to the brim after bleeding the system yesterday. Took it for an hour's run (both motorway and country roads) and when I got back you can see the slight drop in coolant from the header tank and the increased coolant level in the top up tank (pic taken immediately after getting home):



I could hear the header tank cap making a noise which hasn't happened previously so am hoping this means the header tank will top back up as it cools.

I know this is just a temporary repair but may just give me a bit of time to get the bits together to get the job done properly.

[itchyfeet]

may as well get the sniff test done.

[300CE]

Yep, is next on the list Paul. Might get the kit and do it myself as there's obviously something wrong. Am thinking about the GTi route worse case scenario.

[kevtherev]

splitter!

[itchyfeet]

splitter!

x2

[Smiffo]

Subaru??

[kevtherev]

Subaru??

Subaruined