Welcome to our Site for MG, Triumph and Austin-Healey Car Information.


MG parts spares and accessories are available for MG T Series (TA, MG TB, MG TC, MG TD, MG TF), Magnette, MGA, Twin cam, MGB, MGBGT, MGC, MGC GT, MG Midget, Sprite and other MG models from British car spares company LBCarCo.

MG MGB GT V8 Factory Originals Technical - Coolant Lose

My totally rebuilt (2001) 3.5 Rover looses cooling fluid. It always has.
I like to have the rad full and realize that the top of the rad is actually an expansion tank. I would fill the rad up and it would expand out.
To satisfy my idiosyncrasies, a couple of years ago I installed a overflow tank and the correct rad cap so I basically now have a closed system. Any excess water expands into the overflow tank and return to the rad when it cools.
There is no signs of fluid leaking from the engine, cooling hose’s, the heater core or from the exhaust pipes. The oil is not milky and the rad coolant looks fine.
But still the coolant fluid disappears.
Last year I did a leak down test and compression test which didn’t revel anything alarming.
Yesterday I headed out on a fall drive and realized I hadn’t check the rad level before I left. I turn around the block and back to the shop to pull the rad cap. As I pulled the cap there was a sound as the pressure equalized. The radiator cooling system had already built up pressure. (this was just around the block and the car had been sitting for over a week)

Any ideas”



Bruce Mills

are you sure that the water pump is O.K.?

It is difficult to see as it becomes wet at the bearing of its shaft, covered by the pulley adaptor and you can only feel the drops when the car has coold down.
An other ugly problem are leaking heater valves that are nearly invisuable and drain coolant to the cover underneath the intake manifold where it evaporates once the engine has heated up.



When I built my conversion which has a factory set up cooling system for the V8 the expansion tank would always settle at about half full and would always expel any additional coolant added to the system.

The header tank of the rad however always stayed full.

One answer to your query could be do you have the correct elbow on the inlet manifold, if you are using the type that points upwards instead of being straight the top hose is too high and air will get trapped into the system and you will then get air in the header tank.

To check, start car and allow it to warm up normally and wait for the thermostat to open, place a finger or two on the highest point of the rad header tank and one a bit lower you should feel the stat open as the header tank temp reacts very quickly, if you feel the temp rise at the lower position but not the very top then you have air trapped in the system.

To solve either fit the correct neck to the manifold, advised or bleed the system by parking the car on a steep slope with the rad cap off and the heater circuit open and allow engine to warm up, and then squeeze the top hose a few times to expel any air trapped in the system.

If you have the wrong neck on the manifold i would change it, and save a lot of hassle.


Kevin Jackson

A coolant system with a radiator with a header tank, i.e. as per 4-cylinder cars before 1977, shouldn't need an overflow bottle. If they are overflowing either it is being overfilled each time and simply expelling the excess, or there is a fault. Without the overflow bottle does the level in the radiator continue to fall if you don't keep topping it up? Or does it stabilise? This can be difficult to spot with the earlier radiator with the angled filler neck, much easier when the filler cap is directly above the tubes.

You don't say the rate of coolant loss in terms of volume vs mileage or running time, some rates are obviously more to be concerned about than others. How much goes into your overflow bottle, and how full does it get? Obviously if the cooling system is pushing more in than it can hold *it* will overflow and hence there will be less than before back in the rad when it all cools down again. Does it suck all the coolant back in evry time? Or leave some behind? It will only suck it back if you have the recovery-type radiator cap with the low-pressure suction valve below the main rubber valve, as well as the spring-steel shim above the main rubber valve which seals to the top of the radiator orifice. And that will only seal if the radiator orifice is smooth and flat. I'm guessing you *do* have the recovery type cap, but if you don't and it doesn't have the low-pressure suction valve, if any coolant is lost duringw arm up then when the engine cools you will have a slight vacuum in the system. The next time the system heats up again it will have to fill that vacuum before it starts to pressurise the coolant to resist boiling, and that can allow boiling and more coolant loss each time.

Were your leak-down and compression tests both on the cylinders? Or the cooling system? If not the cooling system then you should pressurise *that* with the engine cold and see if it loses any pressure over time. That test in itself may be enough to reveal a leak, although it may only leak when hot, and if it is small enough it may vapourise rather than being visibly wet. It could be in the heater which will simply wet the floor, and not be visible for some time under the car, although if you use anti-freeze you can usually smell it and it will steam up the inside of the screen if the heater is on.

The coolant will start to warm up, and hence expand, almost immediately the engine is started, and it takes very little to be noticeable as a 'gasp' when the pressure cap is removed.

With the factory V8 (and 77 and later) remote pressurised expansion tank system the radiator *should* always be full as the pipe between them comes from the top of the radiator and goes to the bottom of the expansion tank. This means that air is pushed into the remote tank on expansion and coolant is pulled back on cooling (assuming there is enough in the remote tank). There is a fault that can develop on these systems where air is pushed into the radiator, which displaces more coolant than it should, and eventually the remote tank can overflow but the radiator (and top of the engine) is full of air. This can be caused by the pump or bottom hose sucking in air. It took me absolutely ages to find this on my V8, and only stopped when I did an exploratory removal of the heads and changed the water pump (my hoses were tight)

Paul Hunt

I also had a vehicle that was losing coolant from no apparent source. Turned out to be a pinhole leak in the heater core, which was evaporating too quickly to form a puddle, or even a drip. Just one thought.
Joe Ullman

Hi Guys Thanks for your ideas Ralph. The water pump appears to be okay. Whether hot, or cold at normal temperature I cannot feel any leakage under the pump shaft, nor the thermostat. As I re-read this I see you said heater valve. I only have a few driving days left but will check that out this weekend. It’s suppose to nice here.

Kevin. The thermostat housing cover actually does point up. Not straight as your does. Where did you get that housing from? I usually burp the system but can see how air would be trapped there. Squeezing the hose is not an option as the top rad hose is a piece of heavy duty stainless steel braided line which came from a pulp mill. I will change the housing this winter if I can source the part. But I am not sure this is the problem. The system continues to loose water. I would think even with air trapped that the fluid would stabilize at a certain level.

Paul The level without an overflow bottle would continue to fall. On all of the long distance trips I have taken, each morning I have to fill up the rad. A days driving for me when on a trip is 4-500 miles. The next morning before I set out on my days drive I check the fluid levels and usually I can just see the fluid at the top of the cores.. In my case I have the 1963 Ford Falcon rad and can see directly to the cores. I contacted the BBS because the current lose, down to the cores in under 200 miles, car sitting about 10 days in the garage seemed excessive. ” How much goes into your overflow bottle, and how full does it get?” Sometime the amount of fluid in the overflow bottle seems equal to the amount the rad seems to be short. Sometime is doesn’t seem to match up. I add more water. Sometimes it levels out and there is excess in the overflow bottle when the engine cools down and sometime there isn’t but the rad is short. “Does it suck all the coolant back in every time?” Or leave some behind? Both, sometimes it does suck back all the fluid and sometimes it doesn’t. I do have the recovery type radiator cap (18#’s) “Were your leak-down and compression tests both on the cylinders?” Yes all test were done on the cylinders. But I will perform a pressure check on the cooling system. I had the heater core and radiator tested 2 years ago and I do use antifreeze and have not notice any antifreeze smell. I did try water wetter this year and it actually seemed to increase the temperature not reduce it. .What fluid level should be maintained in the overflow tank? I always fill the rad and let the system do as it will and fill the rad when needed. Perhaps I should also be maintaining a certain level in the overflow tank to avoid any air being inadvertently drawn back into the system??

I am pulling the engine this off season and feel somewhat relieved that nobody mention any type of internal problem. As I have report I just cannot see any evidence of cross contamination between the oil and cooling systems and the leak down test does not show a problem via bubbling through the rad.

Thanks Bruce

Bruce Mills

The pressurised remote expansion tank on V8s and late 4-cylinders should be kept at about 1/3rd to 1/2 full when cold, and this keeps the rad itself completely full. I had a Celica with *unpressurised* expansion bottle, and the marks on that were about the same. But that had a very low profile pressure cap and rad header tank and that was kept completely full. For earlier 4-cylinder rads without any remote expansion tank or catch bottle the header tank was supposed to contain all the expansion when correctly filled, which means there was always some air *space* in the top, less when hot than cold of course. It depends on what your rad is like and where the top of it is in relation to the engine. If it has a smaller header tank space then it may well overflow when hot even when filled to the correct level when cold, and so a catch-bottle would be desirable if not essential. If it is lower than other radiators for that engine, then you would need to keep the coolant level higher when cold, which would make it more likely to overflow when hot, making a catch-bottle desirable again. Unless the steel shim at the top of the cap makes a perfect seal with the radiator filler then it won't always draw back all the coolant it has expelled. If sometimes it does and sometimes it doesn't and you are frequently removing the rad cap to check the level, then maybe you are putting the cap back on sometimes one way and sometimes the other, and one way has a better seal than the other.

You might like to try completely filling the rad, putting a little coolant in the catch bottle when cold as well, and see if that still gives you coolant loss over time. If so then you must have a leak somewhere. I'd advise using some coolant additive to prevent corrosion if not freezing, rather than plain water. In fact anti-freeze being better at searching out leaks may be a useful diagnostic tool that would reveal a leak - from staining, smell or simply greater volume.
Paul Hunt

In order to avoid coolant overflow (loss) in my Rover SD1 (no overflow catch bottle) it was necessary to completely eliminate any air within the system. The highest point in my system is the upper radiator hose. I machined a fitting to insert at this point with a 1/8" NPT plug. I top-up there when I'm sure any other air pockets have been purged and do not loose coolant any longer running a 15# cap.

Edd Weninger


I got the thermostat housing from V8 conversions in Farnborough Kent UK whch is a bit distant for you to pop round and get one!.

I think it's basically the original MGBV8 one, it may be it was also used on other rover V8 engines, the one you have is the later range rover type.

If you really want one i can give Dave Vale at V8 conversions and see if he can send you one.

I can't find an e-mail address for him but you could fax him on 01689 861211 or Tel 01689 858716

Kevin Jackson

This thread was discussed between 19/10/2008 and 26/10/2008

MG MGB GT V8 Factory Originals Technical index

This thread is from the archives. Join the live MG MGB GT V8 Factory Originals Technical BBS now