Fans at low speed, how to reduce electrical noise?!?

[Mudgey]

Hi,

I am trying to find a way to operate a 200cfm+ 120mm fan at a couple of speeds, so say full power, medium and low. With the low intended to be inaudible. I have used PWM to operate the different fan speeds, however the fans are very noisey at the low rpm, they give out a buzz, a high pitched whine and vibrate heavily.

Does anybody know if it is possible to make the fans smooth and also to reduce the electrical noises?

Many thanks.

P.S. The reason the post is in the WC section is that it is being fitted to a WC rad

[OldChap]

Have a read through Martin's fan testing thread http://www.xtremesystems.org/forums/...d.php?t=261778 for some more info on this ....the technical fix is still being worked on. meanwhile you could use a rheostat for full variable control or maybe try the 5v or 7v you could get from a psu molex connector

[MpG]

Sounds like a cheap PWM controller that's running at too low of a frequency. A quick-fix can be to throw a small (say, 0.3-1uF) capacitor across the power and a ground lines to smooth out the on/off transitions, but that kind of solution isn't for everyone.

[OldChap]

You might also have to play around some with that capacitor value .... the one that works best may depend on the specs of your fan. It goes across the +ve and -ve of the fan output of your controller

[Martinm210]

I have a 470uF cap on the PCBheaven 3 pin PWM controller.....

[OldChap]

A lower value might help, Martin ...if you have a few to choose from in your parts bin ...have a play...see if it makes a difference. There is also the ability to change a cap in that controller design to change the frequency of the pulses ....if you are still getting the "tick"

[Martinm210]

A lower value might help, Martin ...if you have a few to choose from in your parts bin ...have a play...see if it makes a difference. There is also the ability to change a cap in that controller design to change the frequency of the pulses ....if you are still getting the "tick"

I have the variety pack of caps to play with, maybe I'll try some others. The tick does still seem to be present the same as voltage.

I'm a bit of a greenhorn on electronics. On the 555 IC, which cap controls the PWM frequency. Is it the .1nF, or 10nF cap?

Image from the good folks at pcbheaven

I would assume it's the 10nF cap tying to the 6&2 pin Trigger??

[OldChap]

A quote from same site...

"Nice and is working, but with my Glacialtech fans i needed to replace the 10pF with a 47pF to get a large and smooth area of control."

"Yes indeed, different fans or loads will require different PWM frequency. If the load draws too much current, it could be that the frequency should be increased (decrease the cap value) and vice versa."

Then there was this one too

"The frequency will be calculated from the total potentiometer value (from side to side), but the duty cycle will be according to the position of the middle potentiometer pick-up."

So it seems there is room to play quite a bit to try to find values that might cover a broad range of fans.

I know nothing ...I just build 'em too but I saw this and thought "it is not written in stone, you can change the values some"

[MpG]

I have a 470uF cap on the PCBheaven 3 pin PWM controller.....

Wow. With that kind of value, especially if there's only one fan, you're almost turning it into a tiny switching-mode PSU, rather than simply reducing the skew-rate of the PWM pulses (which is all the little cap is supposed to do). Were you still getting the same speed control with that high of a cap value?

I have the variety pack of caps to play with, maybe I'll try some others. The tick does still seem to be present the same as voltage.

I'm a bit of a greenhorn on electronics. On the 555 IC, which cap controls the PWM frequency. Is it the .1nF, or 10nF cap?

I would assume it's the 10nF cap tying to the 6&2 pin Trigger??

Yeah, it's the 10nF affecting the total pulse period there, with the varistor shifting the relative length of the on and off periods. Although that's an odd setup in that circuit layout. Pin 3 is typically the output pin on the 555, whereas he's using 7 to trigger the FET, and using pin 3 (which should be an oscillating output) to power the period portion of the circuit. Maybe have to try that myself and see how it works.