PSU fan controller

[Pentium4]

Any of you geniuses out there know a simple way of making a heat controlled "fan controller" or a real fan controller for those PSU's that just have them running at full speed all the time? Is it as simple as just adding a little thermistor?

[lti]

Re: PSU fan controller

I made one with a TC648 IC.
http://www.microchip.com/wwwproducts...cName=en010720

There are a lot of other circuits you can use. I remember one on this forum that used a bunch of transistors.

[mariushm]

Re: PSU fan controller

It's usually not quite as simple as adding a little thermistor.

Usually, there's a voltage regulator or a current limiter which controls the speed using the resistance of the thermistor as if it was a potentiometer.

You can look for example at a datasheet for lm1117, which is a simple linear regulator that can output up to 0.8A which is enough for a fan:



Being a linear regulator, it means it needs a specific voltage above the output voltage to work properly, in this case of this voltage regulator the voltage is 1-1.2v.
So with this regulator, you could for example adjust the voltage of the fan between 5v and 11v, if you give it 12v at input.

If you scroll down to page 9, you can see a schematic for a simple adjustable regulator:



R1 is fixed, and recommended to be between 100 and 120 ohm, let's say you go with 100 ohm, so you're left with adjusting R2 to get 5-11v.
According to formula ...
R2 = (5v / 1.25v - 1 ) x 100 = 300 ohm (for 5v)
R2 = (11v / 1.25v - 1 ) x 100 = 780 ohm (for 11v)

So now you can pick a PTC (resistance goes up as temperature increases) so that when the psu is at 25-30c your ptc will give 300 ohm and when the temperature goes up to 60-80c your ptc will give 780-800 ohm or thereabouts.

Alternatively, you can also put a simple resistor in series with this ptc, for example in case you can only find a ptc that's 0 ohm when cold, 470 ohm when 90-100c then you could put a 330 ohm in series with this ptc.


Another example ... get a basic microcontroller and a mosfet/transistor and some capacitors. Use the ptc or ntc with another resistor to form a voltage divider and measure the voltage with the microcontroller. Based on that voltage, turn on and off the mosfet/transistor to adjust the power going to the fan.

[c_hegge]

Re: PSU fan controller

There's a very simple one here (http://www.heatsink-guide.com/control.gif). The only trouble with it is, that it has an extremely 'aggressive' curve. 1-2*C lower than the target temperature, and the fan turns off. 1-2 higher, and it's at full speed.

[Pentium4]

Re: PSU fan controller

Very cool, exactly what I was looking for!

mariushm, have you made one of those before?

[mariushm]

Re: PSU fan controller

I have made adjustable "power supplies" but i never had a need to make one with a ptc or ntc.

For basic experiments and small projects, it's good enough.

In real world products, you won't see something like this because linear regulators are not particularly efficient (the difference between input voltage and output voltage turns to heat) and there's that voltage drop which means you don't get the fan running at 12v.

Basically and for example, if you run a 12v 200mA fan at low speed by setting the voltage to 6v, the fan will run slow but the linear regulator will dissipate (12v-6v) x 0.2A = 1.2 watts which means the regulator will be very warm/ hot and will heat the insides of the psu.

Most fan controlling stuff out there relies on limiting current instead of adjusting voltage for that reason, and because it also sidesteps the issue of heat (that happens with linear regulators). That's done by using a npn transistor (or something similar) in linear region or a mosfet...

You can see these methods on this page: http://www.talkingelectronics.com/pr...ET/MOSFET.html

[goontron]

Re: PSU fan controller

why not try something with pulse duration modulation?

[domas]

Re: PSU fan controller

i have one in my box at home, it has 2 input wires for constant 12V, a 2 pin thermal sensor, and 2 output wires for 12V fan.
you can get it from me, just cover few dollars of shipping

pulled from working FSP p/s but with bloated caps, and i used it for spare parts

[Khron]

Re: PSU fan controller

A while ago, Th3_un1qu3 showed me a really simplified one (see attached picture).

The 10nF cap in parallel with the NTC is to give the fan a "kickstart", since it takes a higher voltage to start them, than to actually keep them spinning.

I'm guessing the "FAN" resistance is just its rated voltage divided by its rated current. Feel free to play around with the values in LTspice though

[ben7]

Re: PSU fan controller

Not sure if it would pose a problem but the PTC's and NTC's usually have a terrible temp/resistance curve.

I could probably cobble one together sometime and see.

I would probably use a ramp generator circuit, and then compare the voltage from the PTC with the ramp, basically getting a PWM output.

[Pentium4]

Re: PSU fan controller

Getting some good ideas, will have to round up some parts!

In real world products, you won't see something like this because linear regulators are not particularly efficient

Couldn't be more than a watt or two though??

I found this fan controller from an old SF unit. interesting, although it has a button on the back to select "Auto, Slow, Fast" That's quite a row of BJT's.

[ben7]

Re: PSU fan controller

Getting some good ideas, will have to round up some parts!

Couldn't be more than a watt or two though??

I found this fan controller from an old SF unit. interesting, although it has a button on the back to select "Auto, Slow, Fast" That's quite a row of BJT's.

Looks like the typical 4017 decade counter with decoded outputs.
https://cdn.badcaps-static.com/pdfs/...26b16388c2.pdf

I guess it is doing some sort of PWM thing, and selecting different outputs of the 4017 depending on the button on the back. Is there any sort of signal generator/timer/oscillator on there?

[Behemot]

Re: PSU fan controller

Friend of mine is working on PWM regulator with Arduinos which will be able to regulate PSU fan with sensor attached to heatsink and fan connected to it…

[larrymoencurly]

Re: PSU fan controller

A while ago, Th3_un1qu3 showed me a really simplified one (see attached picture).

The 10nF cap in parallel with the NTC is to give the fan a "kickstart", since it takes a higher voltage to start them, than to actually keep them spinning.

Wintech forgot to include that cap in my Ultra V-series PSU, and the fan sometimes wouldn't start when it was cold.

Here's another fan controller that prevents the fan voltage from going below a minimum:

[Khron]

Re: PSU fan controller

I think you'll find that's "cheaped out on", not "forgot"...

[TELVM]

Re: PSU fan controller

Found this one here:


"... The thermistor is connected in a voltage divider, so that as its resistance changes, the voltage at the input of the op-amp changes. The non-inverting input is connected to a potentiometer to set the reference voltage, which effectively sets the lower limit on the fan speed. The output voltage follows the difference in voltage at the two inputs. The output is connected to a small transistor which is darlington-connected to a power transistor that acts as a current amplifier, which drives the fans.

This circuit is a difference amplifier with negative feedback. As temperature increases, the difference in voltage between the op-amp's inputs increases. This causes the op-amp's output voltage to increase, causing the transistor output voltage to increase. This makes the fans spin faster, which hopefully results in the temperature decreasing. As the temperature goes down, the fans slow down. Use the potentiometer to set the low fan speed to suit your setup, I prefer 6-8V output at room temperature (i.e. when the computer is cold) ..."