Comments on: How to make a temperature-based fan controller without a thermistor

By: nico

nico — Tue, 10 Jul 2007 18:50:51 +0000

That is right, the forward voltage drop would be easier to sense. The one saving factor is that I am operating in the 50-80C temperature range, so the reverse current is more appreciable than room temperature.

By: KPL

KPL — Tue, 10 Jul 2007 06:22:04 +0000

I’ve been using diodes for temperature sensing a lot of times, but using the forward voltage, which is much easier to sense.
Feed some tiny current (so it does not heat up it) thru the diode and measure the voltage drop, or compare it to the drop on another diode, which is located in different location.

By: nico

nico — Thu, 28 Jun 2007 16:22:38 +0000

Thanks for the comments everyone!

I don’t think that the 741’s input bias current requirement will work well with the present sensor design. The 741 won’t do rail to rail output either, but I am not sure how much of a problem it will be in this design.

Regarding the first op amp having gain, I just got lazy and didn’t want to put extra components (resistors) on the board. You have a very valid point with the anti-spiking diode since the load will be mostly inductive. The reason that I selected a LM317 (and 1N4148 diode) is that it what was available on hand. It would have been fairly straightforward to use a Darlington pair to drive the fan with appropriate biasing.

Lastly, the diode can be treated as a current source, but I am sensing voltage, so I am really measuring the voltage across the 10M resistor due to the reverse diode current. The operational amplifier employed has effectively infinite input impedance so it doesn’t load the sensor circuit appreciably.

By: Dave Bell

Dave Bell — Thu, 28 Jun 2007 14:24:36 +0000

Nice! Sweet and simple, very straight-forward proportional controller… The only thing that struck me was in the schematic: You appear to be buffering and comparing the diode’s reverse voltage, not the reverse current! Shouldn’t the diode/cap be switched with the 10Meg resistor?

Dave

By: don hersey

don hersey — Thu, 28 Jun 2007 13:56:03 +0000

Besides a resistor in series with the pot, the first op-amp could be configured to have gain, in the non-inverting configuration, Zin would be the same. An LM335 temp sensor would have been another alternative for the sensor.
I would consider shunting the load with a fast, reversed diode. Antispiking the load should help longevity.
Using the voltage regulator as an output device is clever, and has its appeal, but similar advantages could have been achieved by using a NatSemi temperature controlled transistor from the classic era. Another choice would have been a MOSFET.

By: Vinny

Vinny — Thu, 28 Jun 2007 12:23:30 +0000

Is it possible to use a 741 op amp? (Two, in fact)

By: MAKE: Blog

MAKE: Blog — Thu, 28 Jun 2007 10:32:41 +0000

HOW TO – Make a temperature-based fan controller without a thermistor…

Nick made a temperature-based computer fan controller that use a reverse-biased diode for temperature sensing – From a design standpoint, the only complaint I have about my Shuttle small form factor desktop is the extremely fragile on-board fan contro…