Odd BMW cooling problem

After about 10-15 minutes, following a cold start, this 2007 BMW X3 3.0 six does the following: 1) automatic cabin fan behaves erratically, jumping back and forth between low and high speeds; 2) cabin air does not heat; 3) electric radiator fans run on high; 4) coolant temp rises to normal, but then will suddenly jump up to full hot. At this point, driver of course pulls over and shuts off engine. The amazing part is this: the engine can be re-started IMMEDIATELY and all of the above problems vanish instantly. Radiator fans shut off, temp returns to normal, cabin fan works normally, and cabin heat comes on. I’m thinking sticking thermostat, but why does shutting off the car momentarily fix it? Change of water pressure when coolant pump stops running?

My first guess is a poor ground connection.
Check ground straps to the engine, body and harness to the interior.

If the temperature gauge immediately goes from, “full hot”, to normal just by shutting down & restarting the engine, I think that this is your clue to a probable electronic problem because, if the engine was actually overheating, it would take at least 15-20 minutes for it to return to normal temperature.

In any event, just to rule out the possibility of actual overheating, you first need to have a mechanic check the coolant temp at the point when the gauge tells you that the engine is overheating. Once you can rule out actual overheating, then you can focus on where the electrical gremlins are located that are causing these strange problems.

Unless you want to take out a second mortgage on your house, I suggest that you find an independent BMW specialist, rather than taking the car to the dealership.

Clean your battery connections and all your grounds.

While I hope it’s a bad connection (cheap) it might be a electronic gremlin that a dealer will need to diagnose (expensive).

Hate to be the bearer of bad news, but everything that is misbehaving is controlled by the cars computer. None of it is directly controlled by sensors and switches. This can be a real headache to troubleshoot.

It is possible that the ECTS (engine coolant temp sensor) is giving a problem, it is probably good itself, but maybe some water or corrosion in the connector or a ground issue. But since the cabin fan is also acting up, I would check the computer grounding as well. Worse case, it is the computer itself.

Just thinking, maybe the cabin fan is jumping to high because the engine thinks it is overheating and that is programmed to turn on the cabin fan in an effort to help cool the engine. If thats the case, a glitch in the computer becomes a stronger possibility, shutting the car down and restarting also reboots the computer.

I am thinking that an engine coolant temp sensor has something to do with it, indeed, since the cabin fan will change speeds according to the water temp, not wanting to blow hard until the water is warmed up. However, a stuck thermostat could be the primary cause and the temp sensor could be getting no good reading due to no good water flow. (Bear in mind we are talking cold climate here with am temps below freezing.) Now, the temp gauge in the dash doesn’t move instantaneously, either toward hot or back toward normal, but it doesn’t take 5 minutes, either. Maybe one minute each way, from normal up to hot, and then back down to normal. Also, if you happen to notice the antics of the cabin fan before the temp gauge “overheats,” you can cycle the engine off and on and everything goes back to normal before the temp gauge ever does read high.

That is an interesting set of symptoms, I have not seen that particular set of symptoms before, but my first suspect is air in the cooling system.

Turn the key on, heater temperature high as it will go. Open the bleed point and fill the expansion tank until bubbles stop coming out the bleed point. If this makes the problem better, that was your problem. If the problem comes back, and I suspect it will, then you are sucking in air when the car cools. That can be tough to find, but the first place to look is the expansion tank. An internal crack can allow the system to suck in air as it cools.

Manolito could be on to something. A little air in the system could cause the water pump to cavitate. If the thermostat and ECTS are in the same vicinity, then you could get a false reading from the ECTS. When you shut the engine down, the coolant submerges the ECTS as the air bubble rises.

It’s probably a worthwhile investment to have this looked at straight away. Hopefully it is just an electronics/gauge problem, but it may be some kind of cooling system trouble in brewing. Definitely have the thermostat tested, and discuss the possibility of testing for a leaking head gasket. If it is a cooling system problem and you catch it in time, you might can save a good deal of $$$ on repairs compared to waiting for it to get worse.

“Having it looked at right away” means me figuring out the problem. I live 6 hours from the closest dealer. I have a lift in my shop and full set of tools. None of you guys have hit on the fact (which I just learned myself) that the X3 3.0SI 2007-up has an ELECTRIC coolant pump. (Bizzare, eh?) And apparently it is known to fail. However, it obviously hasn’t failed completely. Since the thermostat and the electric coolant pump are near each other (and both blocked by the engine/suspension mounting beams under the engine), I am going to change both at the same time.

More data: we have had a cold spell lately, like most folks, and after the initial pull-over-shut-it-off-and-restart episodes, we drove down into a valley where the temp was below -10 degrees F. and the problem occurred again, after having driven 20-30 minutes with a fully functioning system. (This happened three times, so now the Bimmer is parked until I get finished fixing my Cayenne Turbo’s coolant pipes (another known problem) and get it out of my shop space.) My theory: the extreme cold caused the thermostat to shut momentarily, and when it did so, it stuck again, so that as the coolant warmed up it had a hard time re-opening. OR all of this could be electrical since the water pump is electric. OR the water pump could be failing intermittently.

Electric cooling pump? Basically a good idea, but I wonder why you would need a thermostat too, just program the pump to not run until the coolant comes up to temp. Between the electric pump and a thermostat, I’d bet on the pump, I have never seen a thermostat stick closed, but I know people who swear they have.

I too am going to vote for the pump. Or a connection to the pump.
I can understand having a T-stat with an electric pump. It’d probably allow the engine to reach operating temperature more quickly than if the path were completely open between the coolant in the radiator and the engine. I’m guessing, of course.

I think it is now pretty clear that the water pump has failed and it is now failed permanently and can’t be revived by the turn-off-turn-on method. (Perhaps it is related to the very cold temps we are having right now). I tried driving it today and the water never did circulate. It is now stranded 5 miles from town, cooling off. Water pump (and thermostat) on order and hopefully will arrive by end of week!

@keith wrote:
“Electric cooling pump? Basically a good idea, but I wonder why you would need a thermostat too”

Even when the thermostat is closed the coolant needs to circulate in order to cool hot spots and keep a fairly uniform temperature throughout the engine.

, I have never seen a thermostat stick closed, but I know people who swear they have.

I’ve seen them stuck open and closed. Open won’t kill your engine (just make it run poorly). Closed can destroy your engine.

The question is - which is more reliable…the mechanical thermostat or a sensor. If using a electric pump with no thermostat…then I’d use a 2 sensor system. One in case the other failed. You don’t want a state where the thermostat failed and tells the pump the engine is cold when it’s still hot.

circuitsmith, A multi-speed motor wouldn’t work?

Guys, I suspect that a T-stat allows the engine to reach operating temperature more quickly than an open line.

If you think about it, in truth, a T-stat’s function is not to allow the engine the cool, but rather to allow it to reach operating temperature and stay there.

My new Bavarian Technic computer tool confirmed water pump failure. Apparently it is a multi-speed pump as there are options for actuating it a various percentages of full speed (none of which produced any results because it is fried). Interestingly, the thermostat also has within it an electric pre-heating element, to add to the complexity, which is also involved in maintaining “just the right temperature” for the engine to minimize emissions. (That pre-heater threw a CEL over a year ago that subsequently went away on its own.)

Nice work.