# Which Value of R1 do I use for thermistor-controlled fan?

(source: electronicdesign.com)
The MOSFET I am using is IRLU014PBF-ND. I am using a 10k thermistor, and a 24VDC fan, instead of 12vdc.

It's hard to say exactly because the FET doesn't turn on at one fixed and known voltage. The gate threshold voltage is specified as 1V, but that is at 1V D-S and only 250µA drain current. At 4.5V on the gate you get 210 mΩ or less, so most of the off to on transition is somewhere between the two of these.

With 24V power supply it would take a 435 Ω pulldown to make 1 V on the gate when the thermistor is at 10 kΩ, so maybe start experimenting with a 470 Ω resistor for R1.

This whole control scheme is rather primitive and inefficient. You didn't say how much current the fan draws, but when the FET is half on (12V on the FET and 12V on the fan), it could get too hot for that package. It would be much better to use the thermistor to make a voltage into a microcontroller A/D, which then drives this FET with a 5V PWM signal to drive the fan. In that case lose the capacitor and replace it with a reverse diode to keep the inductive kickback from frying the FET.

You now say the fan takes 100 mA at 24 V. That means it will take at least 50 mA at 12 V, although probably more because motors are not linear like resistors. 12V x 50mA = 600mW, so plan on at least 1 Watt. That FET should be able to handle that if you solder it to a oversized pad.

You want the fan to cool the thermistor back to room temperature, then turn off. It probably won't work that way. You have a immediate smooth control system. It will probably find the equillibrium where the fan speed just cools the thermistor to maintain the speed. There may be some oscillation about that point, but it's not going to slam on and off. That would be better for both the fan and the FET, but your circuit has no provision for that. If you insist in doing this in analog, then a little positive feedback will provide some hysteresis.

• The Fan needs 100mA to run at full speed. I want it to cool the thermistor back to room temperature, and then turn off. Heres the site where I obtained this circuit [link] (electronicdesign.com/article/analog-and-mixed-signal/…) @Olin . Also~ I DO have a 12VDC fan available, along with a potentiometer, but I figured that a 24VDC would cool the thermistor more effectively Commented Nov 22, 2011 at 19:33
• @maraldz So, um why did you ask the question here if you already had an article with the answer in it? They gave you the formula, all you have to do is plug in the correct values.
– user3624
Commented Nov 22, 2011 at 19:57
• @olinlathrop I often think that microcontrollers are too frequenty advised when there are simpler solutions-- but in this case I agree with you. A small 8-pin microcontroller with a thermal resistor, PWMing a MOSFET is a simple solution that has a lot of capability.
– user3624
Commented Nov 22, 2011 at 19:59
• That is a circuit that I provided alink to. The discussion that went wityh it was useful. It's intention is to be a, more or less, proportional controller, hence the lack of hysteresis. Properly implemented it could do reasonably well. Commented Nov 22, 2011 at 23:18
• My circuit right now is technically acting as a proportional controller, however I cant find a good relationship between the fan speed rise with temperature. I think it's because I am using a blowdryer, and the temperature rise is too intense, so the fan just jumps to full speed. I want the fan speed to rise slowly with temperature. Also, the cooling aspect isn't effective because once the heat is removed, the thermistor cools itself within 3 seconds to turn off the fan. Commented Nov 23, 2011 at 22:45

You need 3 things.

1. Get the datasheet for your thermistor so that you can see the resistance equivalent with their temperature.
2. Choose Fan turn on temperature.
3. Based on the turn on temperature and knowing your gate threshold voltage you can then select the value of R1 that will give you the correct gate voltage when the thermistor resistance fall on that particular temperature.