I have a 12vdc .5amp fan, and want to make a variable speed control unit for it. Something like this:



But it only needs to control one fan, and it doesn't need the fancy voltage switcher, or the LEDs...

I have a LAN party on new years eve, so i would like to have my controller done. Once it is done, i will post a thread on my sweet new computer mod!

__________________
There is no "I" in "team", unless Apple makes it... Then it would be iTeam.

 

4 ways of doing it really. Rheostat is simplest but most wasteful of power. Provides full range desired and creates no additional noise.

A potentiometer and pass transistor is a good simple way to do it, but the maximum speed is dropped slightly by the transistors voltage drop (probably not an issue).

A variable regulator can be used. This is quite simple but the maximum speed in hindered

The other way is PWM. this gives full range control (providing a mosfet switching element is used), but can make buzzing noises on lower speed settings; it is the most complex method but wastes virtually no power (probably not an issue with small fans). All methods are well documented online:

http://www.cpemma.co.uk/rheo.html
http://www.cpemma.co.uk/ef.html
http://www.cpemma.co.uk/reg.html
http://www.cpemma.co.uk/pwm.html

Have fun

__________________
Angry!? I'm absolutely electrolytic!

Will have to make do with myspace now I guess...

 

I would strongly recommend using PWM, that's how PC's themselves do it. You only need two things. Some way to generate a clock at a fixed frequency (preferably high, like 10-20khz) and some way to alter the duty cycle (on off times) If you have any, a basic 555 timer can be used, I can't provider refrences currently but you can create a variable duty cycle signal with a 555. a POT would control the duty cycle and the 555 would control the fan through a transistor. Most 555's can only source 200-300ma's of current.
Micro controllers are also a good sollution for this as most of them have facilities to generate variable duty cycle PWM signals. You really should learn the basics of micro controllers (even a simple one) as using them you can do an incredibly large variety of things with a single device with a few pieces of support circuitry.

__________________

 

Here is a suitable one:

http://www.nomad.ee/micros/pwm555.html

And there are some in my thread too:

http://www.electro-tech-online.com/e...tml#post182532

But personally I wouldn't really reccomend it. It is overly complicated for such a low powered device, heat dissapation from linear fan control is minimal and can often be dissapated from a to-92 package quite happily. I'd go with the pass transistor and pot myself, this will avoid the possiblility of growls/buzz at low RPM. You won't obtain full speed because of the voltage drop of the transistor but that really wouldn't be a problem (if it were, your computer would be running far too close to its thermal limit anyhow).

Also, when I tried a computer fan with high frequency PWM (to avoid any noise, 20k or so) I found the control very unsatisfactory. It would change in a very unpredicatable non-linear fashion and low speed settings were hard to achieve. Possibly something to do with the electronic systems in the brushless fans not liking the higher frequency?

__________________
Angry!? I'm absolutely electrolytic!

Will have to make do with myspace now I guess...

 

At 10 or 20khz (or higher) there will be no buzz at any RPM, it's outof hte range of human hearing and the power dissipation is so low it's not even audiable. If you pass the PWM output through an inductor (such as a motor) and a capacitor and/or resistor you end up with a low pass filter, so no mater what you do, you end up with an effectivly steady state voltage. It's called a 1 bit DAC =)

__________________

 

on paper that may be true, but in reality, it's not.

many electronically commutated (brushless) motors don't like PWM at all, and all though you might not hear the pwm itself, you get buzzing and growling in addition to poor fan performance.

sometimes fine-tuning the duty-cycle and frequecy while the fan is running helps to minimize this, but often times you're left without the option of freq.

the fan in my gaming machine makes a loud growl at 55% duty, but is much quieter at 51%, I've got no idea what the freq is, I guess whatever AMD and DFI came up with.

__________________
If you don't have a planet, what good are gold bars?

want to contact me directly? gmail gordonthree
check out my project website: http://projects.dimension-x.net
Favorite numbers:
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

 

Yeah, you can't PWM a pre-electronically controlled fan to control speed. What you have to do is tap directly into the fan. If you just PWM the input to the fan (ie. the power input to the commutating electronics) it's like PWMing a motor controller's power supply to control the motor.

The work-benefit ratio of tapping into a brushless fan with communication electronics built in is very high. It's considerably harder to commutate a brushless fan than a brushed one.

It's probably easier to turn it full off/on in 5 seconds bursts or something.

 

Well you can reasonably effectively but only at lower frequencies (normally under 200hz so pretty low). Like I said, when I tried at about 20k the performance was highly unsatisfactory. Pass transistor and pot is the cheapest and nearly most effective solution (rheostat being slightly better for full range control but more tricky to implement properly). Perhaps the last circuit shown here would be most suitable for your fan:

http://www.cpemma.co.uk/ef.html

__________________
Angry!? I'm absolutely electrolytic!

Will have to make do with myspace now I guess...

 

Ok... The fan was out of an old 350 watt computer PSU. It was connected to some sort of controller unit. I am going to look at it some more, and see if i could easily interface it... probably not.


Ok, now that the bickering has halted, what is recommended? I want simplicity, but i really want efficiency

__________________
There is no "I" in "team", unless Apple makes it... Then it would be iTeam.

 

why do you want efficiency? its not like you're doing this on a laptop

I've built a few of these, they work well:
http://www.heatsink-guide.com/control.shtml

__________________
If you don't have a planet, what good are gold bars?

want to contact me directly? gmail gordonthree
check out my project website: http://projects.dimension-x.net
Favorite numbers:
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

 

My electrical machines profs says "...but who gives a crap about power anyways when it's fractional horsepower."

 

Well, i am probably just paranoid, but i really don't want to give my power supply more work than it has to do. Besides, i like to conserve energy(ok, maybe not, but now is a good time to start...)


I just desoldered a bunch of stuff.

1) LM358N
1) 555 timer
1) TIP2955
1) IRFZ24
2) LT5530 diodes

Would any of this stuff be useful?

I also got a really geeky looking temperature sensor. It looks like a capacitor, and it is green. It looks neat.

__________________
There is no "I" in "team", unless Apple makes it... Then it would be iTeam.

 

I would try running it from a variable 0-12V power supply first to see how well it responds. If it works well, then a Pot and transistor is the simplest but it will waste a little power at the halfway setting (probably less than 3Watts, but it's hard to say exactly how much). I would try the simple method on a breadboard first and see if the heat is really a problem (You could put the heatsink in the airflow of the fan also). If it is, then the more complex PWM method is in order.

__________________
--- The days of the digital watch are numbered. ---

 

Mahah! I just looked at that link JustDIY left. I desoldered everything i need for that one! [sorry for the double post, i am lazy]

__________________
There is no "I" in "team", unless Apple makes it... Then it would be iTeam.

 

PWM'ing a PWM output is going to give you a beat frequency if the two clocks aren't in precise synchronization which could be very bad audiably.

__________________