Overclock.net banner

[Guide] Convert 3 Pin fan to PWM (56k Warning)(Courtesy of Lazzer408)

38K views 45 replies 10 participants last post by  Eastrider 
#1 ·
Courtesy of Lazzer408 from Techpowerup!!

http://www.techpowerup.com/forums/threads/so-you-want-pwm-control-of-your-3-pin-fan.115752/page-3#post-3143383

I personally don't want to spend 20 USD on fans that include PWM and are minutely quieter. So, I just went with Koolance fans. 9 Bucks, can't beat that. I bought 10 of them. http://koolance.com/fan-120x25mm-108cfm

For this project you will need soldering iron, flux, and solder, solder wick and decent soldering skills. I recommend helping-hands and a vice, and a magnifying glass with LED lighting. (But that's overkill)

This guide will show you the simplest way to PWM control up to 10 fans safely, (or about 3 amps). 11 amps on 18 gauge wire is the recommended maximum for molex connectors. Proceed at your own risk. (Looking at you Delta Fan owners and lovers)
tongue.gif
:p:p:p:p

Here are some references before we get started:

Pin Outs and Power






Schematic




Parts List and Links to Purchase

5.1v Zener Diode: http://www.mouser.com/ProductDetail/Vishay-Semiconductors/TZX5V1A-TR/?qs=sGAEpiMZZMtQ8nqTKtFS%2fC7KvLOwmmQCtCYT1%2fs1PtE%3d
10k Ohm Resistor: http://www.mouser.com/ProductDetail/Vishay-BC-Components/PR01000101002JA100/?qs=sGAEpiMZZMsPqMdJzcrNwiCREInfUjhCku0gD%2fN6gQU%3d
N Type MOSFET: https://www.radioshack.com/product/index.jsp?productId=13339909 (Best, I'd recommend a heat sink....
TO-220 heatsink for nte2985: http://www.mouser.com/ProductDetail/Wakefield/274-1AB/?qs=sGAEpiMZZMttgyDkZ5Wiuita4PD18Ap7e0ZyfdsUPiE%3d
Solder Board: http://www.amazon.com/gp/product/B0073XD90K/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1
4 Pin Fan: http://www.moddiy.com/products/4%252dPin-Male-Fan-Connector-%252d-Black.html
3 Pin Fan: http://www.moddiy.com/products/3%252dPin-Male-Fan-Connector-%252d-Black.html
Molex Female: http://www.moddiy.com/products/4%252dPin-Molex-Female-Header-Socket-Connector-for-PCB-Mounting.html

When making your runs, please use 22-24 gauge wire. Solder runs are fine, but may not hold up to the amperage. Solder joints also tend to crack very easily.

So, the first thing I did was space out my components and connectors in a nice, semi-tight area. If this is your first time ever using a soldering iron, I would use the full board to get used to routing wire and give yourself some room to play. Components are cheap, buy a few of them.

The schematic is as simple as it gets. Be sure to route gate between the zener and resistor. This is for PWM.













So, all completed. If PWM IS NOT CONNECTED YOU SHOULD RUN AT FULL SPEED. (12v)


 
See less See more
20
#2 ·
This is exactlly what I was searching (want to plug two 3pin Fans(CPU-Heatsink) into my one PWM-CPU-Header.
But I never did something like that and I'm still at school. Could you explain a little bit more precise how to exactlly do it (and how to do it for 2Fans)?
Would be really AWESOME!!!!!
DrBlock42

PS: Thank you a lot for posting awesome stuff like that.
 
#3 ·
No problem, for two fans just parallel two headers instead of the ten I have. Or use one header, and a fan splitter cable. With something like that you could make the circuit very very small, and may not even require perfboard. This design is capable of driving from a single fan to multiple fans.
 
#6 ·
Design this as a product (good looking, small and scalable) and get it on Kickstarter!

I will be backing this!
 
#8 ·
Quote:
Originally Posted by DrBlock42 View Post

So my one header ( on the asus p8 z77-v lx2 ) can power two corsair sp120 q.e. fans?
And does it controll the speed smoothly(so exactlly like the mainpoard pwm output)? (not like:its on 30% and if itbgets to hot it jumps to 100%
You can control and power two SP120 fans directly from the motherboard. Also, I forgot to mention that the fans will run at a minimum 50%, and slowly ramp up to 100% depending on temperature.
 
#10 ·
What's this fascination that lotsa people have with Corsair SP120 QE fans? ..... as if they have some magical properties.

For starters, they are bad fans for HeatSinks and radiators.. no matter how much hype is sold
And there is some issue with the PWM SP120s that make them misbehave as soon as you have 4 or more of them on the PWM channel

But aside that... Kudos to lazzer408
 
#13 ·
Quote:
Originally Posted by RnRollie View Post

What's this fascination that lotsa people have with Corsair SP120 QE fans? ..... as if they have some magical properties.

For starters, they are bad fans for HeatSinks and radiators.. no matter how much hype is sold
And there is some issue with the PWM SP120s that make them misbehave as soon as you have 4 or more of them on the PWM channel

But aside that... Kudos to lazzer408
I have a remedy for that issue.

Just remove the original circuit from this schematic, and connect the pwm out to the 4 pin fans. Corsair SP are not spec'd for intel.



Edit: that is a 556 timer. Remove the mosfet and resistor and diode from the original circuit. I'd still modify the headers to use proprietary power.
 
#16 ·
I'm currently deployed so I can't make a more detailed video / pictures. I'll see what I can do.

For calculating resistance all I did was take what I know. I know that 50% duty cycle is 10k resistor. So how do we get to 0% duty cycle? Double the resistor size. If you remove the resistor you will run at a minimum 100% duty cycle. It could be sporadic and jumpy in RPM.

To calculate your percentage, just double the size of the resistor as the base 50%.

10k @ 50%
20k @ 0 % (a resistor this size negates all continuity)

So we need 30% duty cycle (power) MINIMUM. So the higher we raise the resistance, the smaller the percentage duty cycle.

70% of 20k = .7 X 20,000 = 14k (30% Left over)
60% of 20k = .6 x 20,000 = 12k (40% left over)

For a 30% duty cycle, 70% of the power will be resisted, leaving 30% to be used while the MOSFET is in its normal state. This is why we chose an N-type MOSFET. Electricity will take the path of least resistance. So when there isn't a PWM signal, the fan runs at maximum voltage. When the PWM signal pulses, the MOSFET turns off the fan, but the resistor limits how low of the voltage it will go, but never completely turning it off and it woudn't, because it would just run max RPM unless we completely negated its continuty with a much larger resistor.
 
#17 ·
Decided to make some sort of shroud.....

out of wood :O

I sealed the wood with a mix of two-part epoxy and alcohol and brushed it on. The alcohol thins the epoxy and soaks into the wood, instead of layering on top and chipping off.

I used two machine screws to mount the perfboard. (Drilled some guide holes... m4 size drill bit. YEAH! I have an m4 bit!!)

I sealed the metal contacts underneath the perfboard where all the mess is with dow corning 3145 (very expensive, but worth it. No acetic acid bi product when curing and a non etchant)

Nice glossy finish. I'll stick it inside the carbide 540 mounted with some super industrial awesome double sided foam tape.



 
#24 ·
Quote:
This will work great, maybe find one that is N channel, and a few amps lower. Unless you're going to be powering A LOT of fans.

Here you go, this will work. , too

http://de.farnell.com/multicomp/1n5231b/zener-diode-500mw-5-1v-do-35-full/dp/2413123
http://de.farnell.com/vishay-bc-components/pr01000101202jr500/resistor-metal-film-12-kohm-1w/dp/1767765

booooom :3
 
#25 ·
I am thinking about building two of these, one for my videocard and one for my CPU.
thumb.gif
Thanks for the info and thanks to Lazzer408. This is my next project after I finish re-capping a mobo.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top