Using Diodes for Fan Control.

This in alternative to using rheostats and chips for controlling fan speed.

The idea for this came from http://www.cpemma.co.uk/sdiodes.html NOW http://www.pcsilencioso.com/cpemma/sdiodes.html and I wanted to explain it a little more and build a few of them. Basically a diode type 1N4001 drops around 0.75 volts with a typical fan. When you link them together then the voltage can be varied.

This is taken from the site
Using a chain of diodes mounted on a rotary switch, a series of voltage steps can be produced giving adequate control for fans.

This control method has several advantages;

• Cost is less than that of a half-decent rheostat, and much less than the price of a good one.
• number of fans wired in parallel (or to a fanbus) can be controlled from the unit without danger of overload. Cheap rotary switches will handle 1amp at low voltage, rheostats to handle such currents are big and expensive.
• The full 12V can be used if required. Basic semi-conductor voltage regulators lose about 2V and would need a by-passing switch to run the fans at full speed.
• The heat produced is spread over a number of diodes. Each one will only generate about a watt at the maximum 1amp load, and less than 0.1W per diode with a typical single fan.
• Unlike resistors, voltage drop is about the same at all fan currents, so fan start-up at low voltages is more assured and a wide range of fans can be used on the same unit.
• Last but not least, adding display lights to form a voltage bargraph or show switch position is easy.

This is my adaption to the original idea. I will use the pictures from the site to help explain.

I did one of the switches as a speed up video.

Step 1 Options For LEDs

Their are several ways that LEDs can be used for this to show what setting the switch is on.

• Bar graph
• single Indicator
• No LEDs

The pictures help explain.

Comments

[demsony]

how many fans can be controlled using only one controller?

[Rob K]

With the rotary switches that I used, it should be able to do 3 of the low power 120 MM fans. Just add up the fans current draw and then look for the switches current ratting.

[Gepetto Father]

What are those diodes just before the LEDs in the positions 11 and 12 of the SW1b in the first schematic? do you need to drop the voltage or something like that?

[Rob K]

Those are Schottky diodes. Each position puts out 12 volts but to have every LED on at the highest setting would require a higher voltage to keep it as a bar graph. So instead when a higher position is reached it turns on a second series of LEDs.

[Gepetto Father]

Ah, so the diodes are there just to avoid the current to kinda make a "loop" and running the LEDs of the first chain in reverse-biased? (i think i got it now... the thing of the parallel connection for lowering the power requirements of the bargraph i've got from start =D)

btw, is nice to see speed controlers that dont require some IC. those are pretty hard to find for me D=

[Rob K]

I have not tried to bar graph the LEDs that way. I used the option with the Schottky diode per LED. I was also using 3.6 fv aqua LEDs and another circuit that was suppose to fade them that messed up.

[sdallesasse]

Why not use zeners to regulate the voltages?

[Rob K]

I asked your question Cpemma responded with.

In theory you could use zeners, but it's hard to find low values. A 2.4V 1.3W zener in place of the first pair of 1N4001 would be worth doing to give a bigger starting step (12V to 9.6V) and extend the bottom of the range to 6.6V

[Rob K]

I'm not sure on the zener diodes. I know you can use schottky diodes.

[Phil B]

I assume you are utilizing the property of a diode that drops the voltage across it by about 0.6 volts per diode.

[Rob K]

Yes it is, I forgot to add that in there. I will update the instructible.