I think it does use that sensor. It’s not a simple on off at some temperature. It uses the settings of the AC control and some formula

I tested the hell out of my bad fan before replacing it several years ago. Toughest part as usual was separating internet fact from fiction.

Based on tests on my 2001 3.0i, with an at-the-time dead aux fan ...

Short answer is that yes, it uses that temp, but the engine must be running for it to turn on the fan due to a high temp there, and also just pulling the connector off the sensor does not work since it simulates -54*C. Putting a 100 Ohm resistor across the connector simulates hot enough to turn the fan on when the engine is running.

My notes:
Experiments with the temp sensor on the lower radiator hose
This has nothing to do with the temp gauge in the instrument panel (which uses a different sending unit). No warnings result when the temp varies to extremes.
Tests done with ignition ON, engine not running, battery charger maintaining voltage.
I used the Foxwell to monitor live data on that temperature. Measured about 71*F (ambient temp) when starting.
Disconnected the connector, simulating infinite resistance ==> -54*F
Measured resistance of the sender = 2.8 kOhms
Put the following resistors across the connector to simulate different temperatures:
temp [F] resistance [kOhms]
-54 infinity (unplugged)
71 2.8 (actual sensor)
94 1.48
122 0.80
159 0.40
213 0.151
241 0.100

These were done to confirm how I could simulate an overheated radiator hose, to hopefully trigger the fan to turn. Interweb lore says that you can just unplug the sensor, but that does not seem to work, which makes sense. The 100 Ohm resistor simulates 241F, and successfully triggers the fan to come on (the command comes, even though the fan does not spin).

Monitoring voltage on the fan control wire from the ECU:
Ignition ON, engine stopped, battery charger maintaining voltage at ~14V

Normal state - 14V (It is a square wave, but the DC Voltage meter showed this solid value; switching to AC gave numbers bouncing between almost that and zero; implying the DC reading shown is an integrated average)

Foxwell active test of the fan - almost instant drop to 1.6V, holding for 20 seconds. So 14V means fan is commanded off, 1.6V means it is commanded full speed on.

Putting the 100 Ohm resistor to simulate 241F at the lower radiator hose - no effect on fan control voltage. (but will have effect when engine is running)

Now started the engine, AC off.

100 Ohm resistor to simulate 241F at the lower radiator hose - control voltage gradually dropped to 2.6V over a period of about 30 seconds. This would gradually ramp up the speed of the fan to almost full speed. Repeating, this response did not happen with the engine stopped.

aux fan failure confirmation
 

That's some great detail! So with a working fan, rather than attempt to generate a square wave you can dial-a-temp using a potentiometer across the sensor leads. You could run in parallel just use a high enough Ω potentiometer that the effect is negligible when set high.

Fascinating :-)

Can you control the speed or it just goes max when hits the magic overheat value?

Which fan brand was installed? Curious minds want to know. :bmw:

It was a TYC branded fan. I'm hoping to get several years out of this at least.

Pulling the coolant or A/C high pressure connector, engine running, even with A/C off, works to get mine running. Pulling the A/C connector to test would be what I recommend.

Lifetime warranty...Keep your receipt and the warranty -which you can download/print off the Amazon website- where you can find them.

But i bet that fan will last longer than the car.

Congratulations
 

Good question, those notes are better than my memory at this point. Doing those tests, I was mainly trying to be sure the fan was bad before buying a replacement, rather than to figure out extra details. I think if the next lower temp had triggered the voltage to change, I would have noted that. Since I did not, it probably did not trigger.

Knowing how BMW likes to run these engines hot, and having a coolant thermostat, I would not be surprised if the fan just turns on at some high enough temp, and stays off below it (unless AC is on). So that's a guess.

I agree on the idea for this being an input for active fan control.

I used a TYC fan. Seems great, no problems.

I agree. When I did all those tests, it was partly because so much of the info on the intertubes was contradicting itself, so it was tough to believe anything.

Here are my notes on the AC in general, and the AC pressure sensor in particular. As above, the main thing I was looking for was confirmation that the ECU was working and the fan was broken.

AC turned on

Control voltage gradually dropped to 13 ... 12 ... 11V, holding there. I expect this would be spinning the fan at least a little, as soon as the AC came on, and would regulate speed based on something currently unknown (e.g., some AC system measurement, radiator hose temp)

AC Pressure sensor - in AC line very close to the lower radiator hose. Measured in Foxwell:
Ignition ON, engine stopped, AC off, cold: 1.25V
... connector removed: 5.00V
Engine running: 1.24V
Engine running, AC on: 1.76V, rising
... connector removed: 5.00V

When the AC was turned on, the fan control signal dropped to 12 .. 11V, as before.
When the AC pressure sensor connector was removed, it very quickly ramped down, at about one Volt per second, ending at about 2.0V, which should be almost top speed for the fan.