[5th_bike]

Aha, I think I figured it out.

You see, R1 is only rated 1/8 (=0.125) watt. And the current through R and R2 is about 12.6 V /(330+22) ohms = 0.0341 A. The wattage through R1 and R2 is then 0.0341 A x 12.6 V = 0.45W. Which is more than what R1 is rated for (0.125W), therefore R1 will run hot. And R1 will heat up the NTC much more when it is not submerged, and then the LED will go on.

So, it is not the NTC that heats up by itself, but R1; and R1 heats up the NTC.

I was kind of on the right track in my previous post, but I hadn't figured the function of R1 and R2 out. You see, once I used a thermistor to measure the temperature somewhere, with a 24V PLC; it didn't work and wondering why, I almost burned my fingers on it, because the current going through it was too much. duh. So that's why I thought the NTC would heat up by itself.

Yes R4 is to keep the LED off, to have a low voltage across the LED until the NTC conducts more.

And you can swap R3 and the (LED+R4), they are in series, R3 can also go between 12V and the (LED+R4). R3 is the current limiter for the LED.

Lastly, I'm not afraid the tank would explode, because there is no oxygen in there, just gas vapors. Still, I wouldn't like a hot resistor in my tank...