How to design noise cancellation system for PCs cooling fan

Noise cancellation

How can I design a noise cancellation system. I want to design a circuit so that I may use it to remove the noise of PC cooling fans.

Re: Noise cancellation

You shoud use some kind of Digital signal processing. Look for example Active noise control. Such DSP based system can bring some improvment. Pure analog solution probably is can not give any good result...

Re: Noise cancellation

But what if I output picked noise at the 180 degrees out of phase. Will this analog design be helpful?

interface not ipad friendly, or vice versa, so please ignore typos, etc.

active noise cancellation is tricky and expensive. simpler solution is to lower rpm of fan and get rid of noisy turbulence. Will this over heat the supply?
There are two aspects to cooling electronics. one is gross heat removal and the second is hot spot cooling by inducing forced convection near a chip, etc.
So the rule is, if you feel the exhaust air, and it is hot, you shouldnt back down the rpm. this is rarely the case though. Usually the designers have a tiny fan cranked up to high rpm producing lots of noise. they do that because they want to stir up the air inside the power supply and have confused the turbulent transfer needed near a hot chip with an overall exhaust function.
So an ideal solution is to have a low rpm fan exhausting the supply and another low rpm fan stirring up air inside the supply space, preferably blowing directly on the hot areas.
you can check spot temperatures with a remote pistol type thermometer sold in home depot for 30$.
Finally, you can make a foam baffle on the output of the fan. Another mustake designers make is they choke off the air flow of the fan with a grill too close to the fan, causing turbulence and stalling. and noise.

i am in process of fixing tektronix equipment in ways mentioned above. I need to reduce voltage to fan, say to 8 v.a chip regulator?
comments?



this cooling noise mistake is common in electronics. I have Tektronix and HP equipment that makes a big roar with cranked up fans, and I am

---------- Post added at 09:55 ---------- Previous post was at 09:44 ----------

btw, design approaches above can reduce noise >90% and reduce dust intake greatly.

I think, your fan does not draw a huge amount of current. You can simply insert a variable resistor in series of the fan supply. That will slow down the fan.

It is nearly a 5 year old thread. I remember when I had noise problems. But once again, I am combating noise in my new PC unit. I have removed the fan on GPU of the graphics card and have added a bigger heat sink without fan. The cooling fan for the processor has also been replaced by a quieter one. It is much more silent now. I remember people submerging their whole PCs in mineral oil to dissipate heat and thus there was no need of any fan.

one solution is to use a resistor (about 1 or 2W)
the other solution is to use voltages that are already available in your computer, these are 0 5v 12v.
you can connect 0 and 5v or 5v and 12v =7v to slow down the fan (assuming it is a 12v fan).
Or you can buy a low RPM fan

Alex

thanks. i may try the resistor. i am modifying tektronix and not sure of bus v avail.
i am also considering a very low rpm fan internal to the piece of equipment and shutting off the exhaust fan.
probably better to have exhaust idling gently though. I just peeked into the grill and fan is 12v .37A in a tektronix
2021 function gen. In that case, if I want to reduce v to 5 v say, i am concerned a resistor will not allow the fan to
start reliably, before the back emf of the fan motor starts. when fan is stationary, it is a dead short and wont
have much voltage to start it. ive seen little voltage regulator cans...
i just turned on the unit, and the fan noise is atrocious. I put my hand near the grill, and there is minimal
air flow. BTW, grills are non-linear: their resistance to air flow goes way up as you try to force more through.
in any case, the experimenting with fans is safe and easy, since you can find and measure hot spots with the ir
thermometer from Home Depot I mentioned. Only thing to remember it needs a high emissivity surface, so you cant
measure clean bright metal temp with it. if you must measure a metallic surface, jyst put a piece of tape on it.

I have also tried reducing the duty cycle with a PWM to reduce the rpm of my graphics card fan, this circuit
Simple DC motor PWM speed control
has worked well but there was a noise in the fan coil created from the PWM frequency,
i wasn't able to get rid of the noise by increasing the PWM frequency or with a capacitor (https://ww1.microchip.com/downloads/en/AppNotes/00771b.pdf) so i have returned to the
resistor solution but you can try it,maybe it will suite you, it is a very simple circuit anyway.

Alex

Cancellation into a small space is one thing. Cancellation
at all positions in a large room I don't hold much hope for.

There are fans that spec low noise levels, what about a
simple swap-out? If you have a bunch of fixed eqpt,
consider a squirrel cage blower and ducting to feed
multiple ones? Squirrel cage blowers are almost silent,
run forever, but an inconvenient form factor.


Squirrel cage

dick_freebird said:

Cancellation into a small space is one thing. Cancellation
at all positions in a large room I don't hold much hope for.

There are fans that spec low noise levels, what about a
simple swap-out? If you have a bunch of fixed eqpt,
consider a squirrel cage blower and ducting to feed
multiple ones? Squirrel cage blowers are almost silent,
run forever, but an inconvenient form factor.


Squirrel cage

Click to expand...

a swap out would be ok, with low rpm fan. need to find a low noise fan. i think someone here or on eevblog swapped a fan in his Rigol scope. worth looking for hot spots while the case is off.
jb

Cancellation into a small space is one thing. Cancellation
at all positions in a large room I don't hold much hope for.

Click to expand...

You surely noticed, that the original poster has given up his plan to apply active noise cancellation. Nevertheless, there's an interesting application example for Analog DSPs dedicated to noise cancellation for a server fan. Analog Devices: Analog Dialogue: Adaptively Cancelling Server Fan Noise

thanks for reference. I looked at it and respectfully point out that this is an example of mis-placed technology, like
putting a bilge pump under a leaky faucet when you should just fix the faucet, or adding a second engine to a car
when the original just needs fixing. Fan noise happens because designers run them too fast through restrictive grills.
This creates turbulence and noise like a madman waving his arms and shouting, or a politician.

One solution is to use a resistor (about 1 or 2W). The other solution is to use voltages that are already available in your computer, these are 0 5v 12v. Like ceiling fan can connect 0 and 5v or 5v and 12v =7v to slow down the fan (assuming it is a 12v fan). Or you can buy a low RPM fan

router99 said:

One solution is to use a resistor (about 1 or 2W). The other solution is to use voltages that are already available in your computer, these are 0 5v 12v. Like ceiling fan can connect 0 and 5v or 5v and 12v =7v to slow down the fan (assuming it is a 12v fan). Or you can buy a low RPM fan

Click to expand...

Your answer seems to be an exact copy of the answer i gave in this thread about a month ago, answer #6

alexan_e said:

one solution is to use a resistor (about 1 or 2W)
the other solution is to use voltages that are already available in your computer, these are 0 5v 12v.
you can connect 0 and 5v or 5v and 12v =7v to slow down the fan (assuming it is a 12v fan).
Or you can buy a low RPM fan

Alex

Click to expand...

Alex