PWM Trouble with PC Fan

[tony48]

Hello,

I'm trying to use a PWM controller to control a 12v PC fan. When I wire it up, it controls the fan as expected but there is a loud squealing noise the changes pitch when I change the fan speed. Can anyone help me understand why this is happening and hopefully fix it?

Here's the PWM unit I have:
http://www.amazon.com/LEDwholesaler..._1?ie=UTF8&qid=1393714784&sr=8-1&keywords=Pwm

Thanks very much!

 

[shumifan50]

It seems from the reviews it is a product best avoided if you don't want to start a fire.

The noise is most likely the frequency of the PWM. The fix would be to increase the frequency until it is not audible. But the better option is a better quality PWM unit. It seems this one does not cope too well with higher currents.

 

[tony48]

It seems from the reviews it is a product best avoided if you don't want to start a fire.

The noise is most likely the frequency of the PWM. The fix would be to increase the frequency until it is not audible. But the better option is a better quality PWM unit. It seems this one does not cope too well with higher currents.

Thank you very much for reply! Unfortunately I've already purchased this unit...I was actually recycling it for this project. Luckily I'm using it for this small fan which draws less than an amp.

I should have been more clear when I mentioned the high pitched sound. It is coming from the fan when I use the PWM circuit. It doesn't happen when I bypass the Pwm circuit and run the fan directly from 12v.

 

[shumifan50]

My answer still applies. It is the voltage switching on and off that causes the fan to squeal. If the freq is high enough it should not happen. You could open the unit and trace the circuit and post it here then someone should be able to advise on component changes.

 

[tony48]

Here's the guts. It's the best I can do. Hopefully it's easy to follow the circuit.

 

[(*steve*)]

I don't think it will cause a fire, but if the switching is in an audio range and the motor has some resonance near that frequency then you'll get audible artifacts.

This is a long winded way of saying "it will scream its head off".

 

[(*steve*)]

Can you tell us what is written on each of the multi-legged black devices. Your photo obscures one and does not allow the others to be read.

 

[tony48]

Can you tell us what is written on each of the multi-legged black devices. Your photo obscures one and does not allow the others to be read.

Here you go. Let me know what else I can do

 

[(*steve*)]

The answer is probably to increase the speed of the 555 oscillator (I was hoping it was a 555)

It looks like C1, R1, and R2 are the components setting the speed of the oscillator.

R1 and R2 look like 1k and 10k respectively.

My guess is that C1 is between 0.1 and 0.01uF. Replacing it with something between 4n7 and 2n2 would raise the frequency well above audible. The question is whether the rest of the circuit would operate correctly at the higher frequency.

You'd have to try.

edit: Oh, and I'd also solder a 0.01uF capacitor in as C2

 

[tony48]

The answer is probably to increase the speed of the 555 oscillator (I was hoping it was a 555)

It looks like C1, R1, and R2 are the components setting the speed of the oscillator.

R1 and R2 look like 1k and 10k respectively.

My guess is that C1 is between 0.1 and 0.01uF. Replacing it with something between 4n7 and 2n2 would raise the frequency well above audible. The question is whether the rest of the circuit would operate correctly at the higher frequency.

You'd have to try.

edit: Oh, and I'd also solder a 0.01uF capacitor in as C2

Thank you so much! I'll give it a try

 

[tony48]

Oops, repost. Please ignore!

 

[tony48]

Steve, Thank you so much! I soldered it all up and it worked beautifully. I can't thank you enough. It's love to be able to understand the math and reasoning behind this situation

 

[(*steve*)]

Gee, you were quick. Glad I was able to help.

Just be careful that the mosfet (Q1) doesn't get hot. If you set the speed to about 90% and leave it running for 5 minutes or so then put your finger in the mosfet. If you can hold your finger on the mosfet for at least a couple of seconds that's fine. (even better if it's cooler)

 

[(*steve*)]

I'd (?) love to be able to understand the math and reasoning behind this situation

I think that's what you meant.

I figured out the important parts of the circuit by following traces that were (fortunately) visible.

I then went here (the first google hit, it's nothing special) and plugged in the values for your R1 and R2 (which are the other way around on your circuit).

Then I tried a few values out for C and figured that from your description the existing value was between 0.01 uF and 0.1 uF. (that would make them in an audible and annoying range)

I then tried a few values to get frequencies above audible and not too high.

Nothing really tricky about it

Then you did the hard work of replacing the components and testing it.