I have an interesting problem on my VW Golf MK3 TDI.

Basically, sometimes the pressure cap on the coolant reservoir lets out some coolant.

Here's what I know:

  1. It happens when the car is under torque for a while (driving fully loaded on a motorway, or going up a small mountain).
  2. It doesn't happen when the car is not under torque (we drove about 1500km with light luggage, 2 people and a baby in the car, and no water was lost)
  3. The pressure is still there when the car has cooled down.
  4. There is no oil in the coolant
  5. There is no coolant in the oil.
  6. Thermostat works, I tested in a pan of water with a thermometer.
  7. Coolant reservoir (expansion tank) and cap are fine. I replaced them with new ones and the problem persisted
  8. The engine shows no signs of overheating. The thermostat remains at 90°C except of course if I idle the car for a while.
  9. The problem has remained the same for at least15000km
  10. The car is now at 227000km
  11. The water pump was replaced 2 years / 25000km ago

And here's what I think:

  • It can't be the cylinder head gasket because of 4, 5, and 9.
  • It can't be bubbles in the circuit because of 3.
  • It can't be the reservoir or pressure cap because of 7.
  • It can't be the thermostat because of 6, 8, and possibly 1, 2 and 3

I am at a complete loss at this point, as well as everybody I know.

  • 1
    I find points one and two to be the most interesting. Does your car exchange heat between the oil and the coolant? E.g., on mine, the oil and coolant pass in close proximity in a heat exchanger. I'm wondering if you're seeing much higher coolant temperatures (and therefore pressures) than normal due to prolonged periods under boost.
    – Bob Cross
    Jun 18, 2014 at 11:34
  • I was thinking along the lines of what @BobCross said. Also, don't necessarily discount a cylinder head gasket leak because of 4, 5, & 9. Since the intake is always under pressure, you could be pumping air/exhaust into the coolant without burning coolant or getting an oil/water mix. Unlikely, but not out of the question. Jun 18, 2014 at 11:51
  • Well i considered both those possibilities but although there is a coolant/oil heat exchanger, the engine temp indicator (known to be working) is not showing any rise. I have to idle the car for 15 minutes to get the temperature above 90 and the radiator fan to turn on. And if it was a cylinder head gasket leak, that would have to get worse over time IMHO. Jun 18, 2014 at 11:57
  • Just to clarify: when you say "pressure cap", you're talking about the cap on the radiator itself, right? We aren't discussing the expansion / overflow tank? That's the separate plastic container where the level raises and lowers based on the current temperature of the coolant.
    – Bob Cross
    Jun 18, 2014 at 12:12
  • 1
    In some cars, like my Astra Coupe, the overflow tank is pressurized and the only "radiator cap" is the cap on the reservoir. When I had a problem with the cooling system last year, it turned out to be caused by a crack in one of the nipples, which allowed pressurised gas to escape and subsequently diminished the efficiency of the cooling system, causing boiling of coolant which would normally stay liquid under pressure. Basically it created a vicious cycle. My suggestion would be to perform a pressure test on the cooling system (not the engine). Jun 18, 2014 at 14:10

3 Answers 3


tl;dr: Your radiator cap might be working as advertised and releasing a bit of overpressure.

The difference is points one and two is telling. In point two, you're explicitly staying out of boost on the turbo. Steady state cruising really doesn't require the turbo to spin up. In point one, you're explicitly relying on the turbo quite a bit. This is going to lead to:

  1. Hotter intake: air under compression heats up
  2. Hotter turbo-lubricating oil: the turbo gets hot compressing that air
  3. Hotter coolant: due to the oil->coolant heat exchanger

I'm going to assume that your engine temperature gauge is reading degrees C. That means that your indicated temperature of 90 is already approaching the boiling point of unmodified water at sea level pressure (like where I am right now - see there's the sea right over there! ;-). Your coolant isn't actually about to boil for a few reasons:

  1. It's not just water - it also has chemical additives that raise the boiling point.
  2. It's under pressue due to the radiator cap - this also raises the boiling point.

That pressurization isn't a magic spell, however. Eventually, the rising temperatures will lead to a pressure that exceeds the rating of the radiator cap, causing it to start to vent to atmosphere. This is on purpose: your cooling system isn't magic and you want any excess release to be controlled rather than explosive.

So, at this point, I think your vehicle is working approximately as intended. You have several options:

  1. You could decide this is not a problem. If the amount of coolant lost isn't a large amount, you could accept that every now and then you need to add coolant. This wouldn't be an unreasonable choice as you're driving an older vehicle.
  2. You could change your coolant. There are additives that you could introduce to the coolant that increase its ability to transfer heat.
  3. You could change your radiator cap. A radiator cap rated for a higher pressure will leak later.
  • I don't think this is entirely right. Althoughi agree that the points 1 and 2 correspond to turbo usage, there is simply no evidence of any overheating. Also, this being VW, the reservoir cap is rated higher than needed as they also use it on bigger (or hotter) engines. So there would need to be a lot of heat to push up the cap, and my thermo says otherwise. Jun 18, 2014 at 20:59
  • @TomMacdonald, I might be completely off base. Hopefully, you'll find out the correct answer and tell us here.
    – Bob Cross
    Jun 18, 2014 at 22:22
  • I have been thinking about this, and point 3 might be explained by the cooling system being underpressured when it has cooled down, having released water when overheating. In which case, this could well be the right answer. Sep 26, 2014 at 7:41

Old topic, but just had the exact same problem on a 1.9tdi 110. The problem occured only under load, and certain driving conditions "pushing" the engine harder. We changed everything, cap, thermostat, resevoir, radiator!(someone suggested i could be clogged), oil heat exchanger, pipework, watter pump, and the heater matrix was eliminated at one point because it started to leak ( caused by the excessive pressure). Finaly i have decide to check the engine, and a very tiny leak was found between cylinder 2 and 3. A new headgasket and a new set of headbolts fix it permanently.

  • I don't have the car anymore, but I think this may have been the right answer. Jan 11, 2016 at 9:39
  • I have owned may VW diesels, and Pedro has a highly probable cause. The Mk1 and Mk2 diesels were prone to headgasket leaks. The Mk4 were not. The A3 were in the middle. Compression tests, and cylinder leakdown tests are your friends, even when diagnosing heating issues.
    – mongo
    May 3, 2017 at 15:22

What is Your thermostat opening temperature? If it is 90 °C and if you live in country where temperature does not goes under negative Celsius, You can change thermostat with one that opens (for example) at 85 °C. If in winter temperature drops under 0 °C, then You can put that thermostat only in summer. You can also try thermostat with lower opening temperature and then test if that helps. I also do not know, if is available thermostat with lower opening temperature for Your car.

  • For this car, this engine, an 87C thermostat provides the best compromise between proper temperature for operation and keeping things cool enough to not have intermittent overheating. There is no need to switch thermostats with the seasons. I own and have owned many VW diesels and worked on many more.
    – mongo
    May 3, 2017 at 15:24

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .