Fan control circuits

[Th3_uN1Qu3]

I was trying to drive a 24v DC fan with PWM, which didn't work. A friend googled it for me and found out that DC brushless fans cannot be driven by PWM as they have their own internal controller based on that hall sensor.

A bit more googling on my side revealed this: http://www.maxim-ic.com/app-notes/index.mvp/id/1784

I built the "DC-DC mode, low side" circuit presented there, as it only required an inductor, cap and diode added to my existing PWM circuitry. It works great. I used a 100uH inductor, 100uF capacitor and a random diode from my junk box. PWM frequency is 30kHz.

Hope this comes in useful.

[mockingbird]

Re: Fan control circuits

But I have seen 120mm DC Brushless case fans with three pin PWM connectors... I'm aware that on some fans, the third pin is just a dummy wire - and then it is usually colored white. Maybe I am confusing it with sleeve bearing.

[Th3_uN1Qu3]

Re: Fan control circuits

On 3 pin fans the 3rd wire is for rpm sensing, the fan can still be driven by DC voltage only. 4 pin fans are the ones you are thinking about, those are a different animal.

[lti]

Re: Fan control circuits

A friend googled it for me and found out that DC brushless fans cannot be driven by PWM as they have their own internal controller based on that hall sensor.

How do all of those PWM fan controllers work?

[tom66]

Re: Fan control circuits

You *can* drive brushless computer fans with PWM, but the mistake a lot of people make is trying to use a high frequency like a few kHz, which will usually cause the fan to twitch, and not spin up properly. You should use around 30 Hz, which gives the fan plenty of time to start on each cycle... the inertia takes care of the rest.

[Th3_uN1Qu3]

Re: Fan control circuits

You should use around 30 Hz, which gives the fan plenty of time to start on each cycle... the inertia takes care of the rest.

That would be unacceptable in my application. Try keeping a 30Hz square wave out of a power supply in an audio circuit. FYI, the PWM controller i built originally used something like 200Hz, and the fan didn't work with that.

[tom66]

Re: Fan control circuits

That would be unacceptable in my application. Try keeping a 30Hz square wave out of a power supply in an audio circuit. FYI, the PWM controller i built originally used something like 200Hz, and the fan didn't work with that.

Then use 5-10 Hz. And decouple it (>470uF) with a large series inductor (>100µH.) It will create a little beating effect, but barely audible. 30 Hz just gives essentially continuous movement for an 80mm fan. Or, place it on a separate tap on a negative rail, so the noise gets reflected into the negative rail, and not the ground. (This is done in a lot of PSUs.)

[mariushm]

Re: Fan control circuits

Why bother with all these improvisations when you can use something as simple as a MC34063 ...

http://uk.farnell.com/texas-instrume...ip8/dp/1053579

The rest of the parts, a couple of capacitors and resistors, a diode and an inductor are all less than 1$ in total.

[momaka]

Re: Fan control circuits

You *can* drive brushless computer fans with PWM, but the mistake a lot of people make is trying to use a high frequency like a few kHz, which will usually cause the fan to twitch, and not spin up properly. You should use around 30 Hz, which gives the fan plenty of time to start on each cycle... the inertia takes care of the rest.

Indeed you can drive a fan with PWM.

I was actually able to drive all of my fans more or less fine at 5kHz. However, the high frequency does indeed create some problems. Mainly, you won't be able to start the fan at low speed. Instead, you'll have to have the duty cycle cranked nearly 90% or more so that the fan spins, and only then you'll be able to turn the speed down. And if you are using a fairly low KHz audible frequency like I was, you'll get alwful screeching sounds from the fan.

I found a solution around this, though - just place an electrolytic cap, 100uF or more, in parallel with the fan. It removes the noise and allows the fan to start even at low speed. The bigger the capacitor, the lower the speed at which you can start the fan. With a 470uF general purpose cap, I was able to start my fans at around 6V (normally most of them would start at 5V if fed DC).

The PWM circuit I used looks exactly like this (with the only difference being that I used a FQP19N20C MOSFET):
http://www.dprg.org/tutorials/2005-11a/index.html

I built the "DC-DC mode, low side" circuit presented there, as it only required an inductor, cap and diode added to my existing PWM circuitry. It works great. I used a 100uH inductor, 100uF capacitor and a random diode from my junk box. PWM frequency is 30kHz.

Some old (and cheap) motherboards actually used to use this same method to get the CPU VCore voltages (except that the MOSFET is on the high side). I even have one like that - it's a crappy socket A motherbaord (ECS perhaps?) that I got with a dumpster computer. Came with a 750 MHz Duron. Works well enough for testing stuff.

Now if you just replace the diode with another MOSFET and have it turn on when the other MOSFET is off, you'll have a synchronous buck PWM DC-DC circuit .