Hi all,I'm trying to get to the bottom of a possible head gasket issue on my 520i (M20 engine). Hoping some of you can help with the diagnosis, because the signs are conflicting.
Symptoms:
Constant stream of small bubbles visible in the coolant expansion tank
CO2 test (on coolant reservoir bubbles) turned out negative
Cooling system gets slightly pressurized over time
Engine tends to run hot in traffic or during slow driving (temp gauge goes up to ~¾). Otherwise stable at the halfway point.
Compression test results:
Cylinders 1, 2, 3, 4, and 6: 12–13 bar
Cylinder 5: 10 bar
After adding a oil into cyl #5: 11.4 bar
All spark plugs looked similar—no major fouling or signs of coolant entry
Any of you know other ways to confirm or rule out a head gasket leak beyond the CO2 test? Pointers are much appriciated!
That’s exactly the CO₂ test I used. The liquid didn’t change colour after five minutes. I also exposed the liquid directly to the exhaust fumes, and it changed colour within 30 seconds.
I'd make sure the fan belt is nice and tight. When loose, thay have been known to make the pump foam/cavitate, leading to overheating.
Here's an example: black shark
The cooling system is sopposed to pressurize; up to 15 psi when fully warmed up is normal.
Running hot in traffic or at low speed is indicative of a bad fan clutch or partially obstructed radiator.
If the cooling system is not building pressure to the point where it's losing coolant and the hoses are rock hard I wouldn't be too concerned. Bubbles are also not conclusive as there is enough turbulence to generate a certain amount during normal operation. A combustion gases test is fairly conclusive.
Do you have a radiator shroud installed? This is essential for the fan to work efficiently.
Federico wrote: Jul 16, 2025 9:32 AM
I'd make sure the fan belt is nice and tight. When loose, thay have been known to make the pump foam/cavitate, leading to overheating.
Here's an example: black shark
The cooling system is sopposed to pressurize; up to 15 psi when fully warmed up is normal.
Thanks for the pointers!
The belt seems and thight, and was recently replaced.
turbodan wrote: Jul 16, 2025 10:54 AM
Running hot in traffic or at low speed is indicative of a bad fan clutch or partially obstructed radiator.
If the cooling system is not building pressure to the point where it's losing coolant and the hoses are rock hard I wouldn't be too concerned. Bubbles are also not conclusive as there is enough turbulence to generate a certain amount during normal operation. A combustion gases test is fairly conclusive.
Do you have a radiator shroud installed? This is essential for the fan to work efficiently.
This might be it then.. thanks!
Hoses are not rock hard under normal running conditions.
No radiator shroud installed. So I'll replace the fan clutch and find myself a radiator shroud. Do the radiator shrouds normally come standard? It seems mine didn't.
A working fan clutch is usually obvious, if it's warm out and you're driving at moderate speeds you can hear the fan making whooshing sounds when it kicks in or out. There's also the 'newspaper test' where you take a rolled up section of newspaper (or some other piece of suitable paper) and hold it against the fan while the engine is running to see if the fan stops rotating.
I'm reasonably sure that non airconditioned vehicles dont have a fan shroud. I've owned two 520i, a 525e and a 528i and none of them had a fan shroud.
What does your exhaust look like? Any sign of condensation aka "white smoke"?
One would expect the temperature gauge to be at 10:30 not 12:00 when the engine is warmed up.
Likely issues are that the water pump is not working properly or the thermostat is failing.
Check the waterpump drive belt as it may be loose.
Have you checked radiator and thermosat housing temperature? They should be the same give or take, if thermostat is ok.
stuartinmn wrote: Jul 17, 2025 8:30 AM
A working fan clutch is usually obvious, if it's warm out and you're driving at moderate speeds you can hear the fan making whooshing sounds when it kicks in or out. There's also the 'newspaper test' where you take a rolled up section of newspaper (or some other piece of suitable paper) and hold it against the fan while the engine is running to see if the fan stops rotating.
This is very helpful thanks -- I've never heart the 'whoosing' sounds ever since I bought the car 3 years back. And the newspaper test turned out positive; super easy to stop the fan with the engine in idle. So we've got our suspect!
JohnH wrote: Jul 17, 2025 10:28 AM
I'm reasonably sure that non airconditioned vehicles dont have a fan shroud. I've owned two 520i, a 525e and a 528i and none of them had a fan shroud.
What does your exhaust look like? Any sign of condensation aka "white smoke"?
One would expect the temperature gauge to be at 10:30 not 12:00 when the engine is warmed up.
Likely issues are that the water pump is not working properly or the thermostat is failing.
Check the waterpump drive belt as it may be loose.
Have you checked radiator and thermosat housing temperature? They should be the same give or take, if thermostat is ok.
That's what I thought too regarding the fan shroud.
Exhaust looks normal. No white smoke. I'll check your other pointers in case a fresh fan clutch doesn't solve my problems sufficiently. Cheers!
With a fresh fan clutch I thought I had solved the problem. The temperature stayed stable, never really passing the "12:00" point, even on a hot day with some traffic. With the new whooshing sound added, I thought I had a winner.
But about a week ago I got stuck in traffic for around 30 minutes. I kept an eye on the temp gauge, and it crept just past the "12:00" point. Then my coolant started boiling over — right through the cap on the reservoir (which was replaced recently, so it should have held pressure just fine).
Any ideas what could have caused the coolant to boil over, with the temp gauge only just past halfway?
We're you able to find a fan shroud? The fan will run with or without a shroud but it will not run efficiently without the shroud. You'll probably double the amount of cooling with a shroud.
A good cap will raise the boiling point but will not stop coolant loss in the event of significant overheating. If you have pressure and coolant circulation, it overheated because the radiator wasn't exchanging enough heat. If it doesn't overheat at speed we can confirm that the radiator is functioning. The problem with idling is the lack of airflow, which must be generated by the fan and ducted with a shroud.
Plan B is to utilize a pusher type fan in front of the radiator. US spec e28s use the fan on the ac condenser for this. If the coolant temp reaches a certain threshold the fan will engage and run at high speed until temperatures drop to normal levels.
I'm not sure the combustion leak testers are all that reliable. Do shops in your area have emissions testing equipment? Might be worth a stop to see if they'll stick their sniffer in there to check for CO.
But... how old is the radiator? They aren't really lifetime devices and the car is ~40 years old. Have you checked the radiator for cold spots? Warm it up, good and warm, full normal operating temperature, shut it off and feel the radiator. It should feel uniformly hot everywhere. A cold spot or area means low flow of coolant thru that area.
I assume you have an expansion tank where you're checking for bubbles? Check it again, it can be very hard to tell the difference between bubbles and turbulence, but you should be able to tell. Also you can clamp off the small 8MM hose that goes to it, just clamp it in the middle somewhere and check for bubbles then. That should minimize the turbulence in there.
turbodan wrote: Sep 18, 2025 2:34 PM
We're you able to find a fan shroud? The fan will run with or without a shroud but it will not run efficiently without the shroud. You'll probably double the amount of cooling with a shroud.
A good cap will raise the boiling point but will not stop coolant loss in the event of significant overheating. If you have pressure and coolant circulation, it overheated because the radiator wasn't exchanging enough heat. If it doesn't overheat at speed we can confirm that the radiator is functioning. The problem with idling is the lack of airflow, which must be generated by the fan and ducted with a shroud.
Plan B is to utilize a pusher type fan in front of the radiator. US spec e28s use the fan on the ac condenser for this. If the coolant temp reaches a certain threshold the fan will engage and run at high speed until temperatures drop to normal levels.
They are available online (part #17111177834, 17111712240), but it’s difficult to determine if they’ll actually fit my radiator, since the radiator they’re meant for seems to be a different model. Has anyone here tried fitting one on a model that didn’t originally come with it? I think it’s definitely worth a shot.
Mike W." wrote: Sep 18, 2025 4:44 PM
I'm not sure the combustion leak testers are all that reliable. Do shops in your area have emissions testing equipment? Might be worth a stop to see if they'll stick their sniffer in there to check for CO.
But... how old is the radiator? They aren't really lifetime devices and the car is ~40 years old. Have you checked the radiator for cold spots? Warm it up, good and warm, full normal operating temperature, shut it off and feel the radiator. It should feel uniformly hot everywhere. A cold spot or area means low flow of coolant thru that area.
I assume you have an expansion tank where you're checking for bubbles? Check it again, it can be very hard to tell the difference between bubbles and turbulence, but you should be able to tell. Also you can clamp off the small 8MM hose that goes to it, just clamp it in the middle somewhere and check for bubbles then. That should minimize the turbulence in there.
Not sure exactly how old the radiator is, but there are no cold spots. The temperature at the entry side is definitely hotter than at the exit side (bottom left), so coolant seems to flow through nicely. I do have a radiator flush liquid that I’m considering using, to see if it might improve cooling capacity.The tiny bubbles in the expansion tank form a constant stream, which doesn’t seem to increase with higher RPM. That makes me think it’s just turbulence from the water pump. I can make a video if that helps. Also, the pressure in the cooling system only builds gradually as it heats up, which should indicate it’s not related to a gasket leak.
If you can find a professional radiator repair shop, they'll be able to do a better job of flushing the radiator than you can do at home. However, I don't know how it is where you live, but at least in the US they are getting harder and harder to find.
Like I mentioned before, my coolant boils over just past the “12:00” mark—right through the reservoir cap. The cap is a new OEM BMW (17112241974). Earlier this week I ordered a new radiator, thinking flow restrictions might be causing a sudden gush of super-hot coolant into the expansion tank and making it boil over. But yesterday I discovered the new cap doesn’t fit quite right. It sits at a slight angle and the thread just doesn’t match perfectly. The tank is 17111707541 (without level sensor). From what I can find, 17112241974 should be the correct cap. Is anyone familiar with caps not fitting this tank? If I can’t find the right cap, I’ll probably just order a new expansion tank.
So to wrap up this thread: I found a “new” used coolant tank with a tightly fitting cap, and that seems to have solved the problem. It’s late autumn here now, so no hot summer temperatures to fully stress-test it yet, but no more issues past the “12:00” mark. The new caps I bought also fit the new reservoir perfectly. It seems the threads on the old reservoir had warped just enough to prevent any cap from sealing properly. Quite a journey of replacing parts and troubleshooting for such a simple solution. Thanks for all the help guys!
