Question Frequencies of sounds emitted by PC components

[System64MC]

Hi,
I'm doing a project for school about analysis of background noise. In particular I've chosen to focus on the frequencies emitted. Part of what we have to do is record background noise; I recorded some noise with my PC on, since it's on most of the time when I'm home, and I noticed some lines at certain frequencies, indicating that there's some component of the PC that's emitting noise at that frequency. Can anyone help me identify what those components may be or what's causing the noise? I've heard about coil whine - could it be the culprit? Here's a picture of the spectrogram of a sample of sound taken from Audacity. The lines around 200-300 Hz are probably caused by the fans (I read somewhere that the formula to calculate the frequency was (RPM/60)*number of blades, is it correct?), but there are numerous other fainter lines around there (which will probably be harder to identify) and two strong lines at around 17.5 kHz and 20 kHz (which I'm particularly interested in) with other fainter lines around them. With another microphone I found lines at 90 Hz and 120 Hz, and they match perfectly the frequency that my 5400RPM and 7200RPM hard disks should be emitting. Sorry if this question is strange.
Here's the picture:

https://imgur.com/QTpLzII

View: https://imgur.com/QTpLzII

By the way, the repeated noises at ~9kHz are strange sounds one of my hard disks was making. I'm making sure everything from it is properly backed up. The other noises after that, the ones that cover a wide band of frequencies, are just clicks.
Thank you.

 

[alceryes]

A very low-tech way to find certain noises is to use the cardboard inner tube of a roll of paper towels.
Take the side panel off, hold one end of the tube up to your microphone, and move around the other end to the different components inside the case. You should be able to isolate the 17.5 and 20kHz noises this way and find out what's causing them.

 

[Lutfij]

Turbulence from your fans and the front panel of a chassis with the air being forced through slits or gratings in said chassis can also lead to sound being generated, hence a frequency. You would need to isolate each component into their own chambers so you can identify which component is emitting noise at that particular frequency. You will also need to rule out that the noise isn't being bounced off another material and reducing the effective frequency(dampening effect).

 

[System64MC]

A very low-tech way to find certain noises is to use the cardboard inner tube of a roll of paper towels.
Take the side panel off, hold one end of the tube up to your microphone, and move around the other end to the different components inside the case. You should be able to isolate the 17.5 and 20kHz noises this way and find out what's causing them.

Thank you! I'll try that.

Turbulence from your fans and the front panel of a chassis with the air being forced through slits or gratings in said chassis can also lead to sound being generated, hence a frequency. You would need to isolate each component into their own chambers so you can identify which component is emitting noise at that particular frequency. You will also need to rule out that the noise isn't being bounced off another material and reducing the effective frequency(dampening effect).

Okay, thank you. What about coil whine? What kind of frequencies would it produce?

 

[USAFRet]

In addition, you'll have harmonics between the devices.

Fan A + Fan B would be different than Fan A + Fan C, even if all 3 were very very close in RPM. Or even "identical". They are in different places in the case, producing different airflow.

 

[alceryes]

Okay, thank you. What about coil whine? What kind of frequencies would it produce?

Mostly high pitched but it'll be across thousands of Hz depending on the coil in question and the dampening effect described above.

 

[System64MC]

In addition, you'll have harmonics between the devices.

Fan A + Fan B would be different than Fan A + Fan C, even if all 3 were very very close in RPM. Or even "identical". They are in different places in the case, producing different airflow.

I guess this explains the many lines and general noise below 2000 Hz, right?
As for the noises at 17.5 kHz and 20 kHz, it seems they're coming from the hard disks. Is it from the motors that are spinning?

Mostly high pitched but it'll be across thousands of Hz depending on the coil in question and the dampening effect described above.

I guess this could possibly explain some lines/areas of noise at higher frequencies?

 

[USAFRet]

I guess this explains the many lines and general noise below 2000 Hz, right?
As for the noises at 17.5 kHz and 20 kHz, it seems they're coming from the hard disks. Is it from the motors that are spinning?

It doesn't necessarily explain. Just one more thing you have to take into consideration.

 

[ex_bubblehead]

I guess this explains the many lines and general noise below 2000 Hz, right?
As for the noises at 17.5 kHz and 20 kHz, it seems they're coming from the hard disks. Is it from the motors that are spinning?

I guess this could possibly explain some lines/areas of noise at higher frequencies?

You're getting into an extremely complex subject here. You don't just get the base frequencies from the various components but the (nearly) infinite number of harmonics and beat frequencies as the sound waves combine both additively and subtractively. The bottom line is that anything that vibrates or moves at all creates sound whether it be mechanically or electrically induced.

 

[System64MC]

You're getting into an extremely complex subject here. You don't just get the base frequencies from the various components but the (nearly) infinite number of harmonics and beat frequencies as the sound waves combine both additively and subtractively. The bottom line is that anything that vibrates or moves at all creates sound whether it be mechanically or electrically induced.

Ok, thank you!

 

[Paperdoc]

The use of a tube and your ear to locate sound sources is a common technique and easy to do. BUT here we're talking about frequencies of 17.5 and 20 KHz, and most human ears are poor in that range. On the other hand, you have demonstrated already that you do have a MIC that picks those up well. So use THAT detection device. Regarding the tube for concentrating on a narrow area, I suggest using a stiff metal rather than cardboard. Cardboard has enough non-uniformity and flexibility that can absorb (reduce the intensity) over a wide range of frequencies.

A thought about the 9 KHz sound you say comes from the HDD's. You've figured out the two sounds from rotational rates of the disks, and those units do rotate constantly. The other moving parts in an HDD are the arms carrying the heads. They move only is short spurts, not continuously. Further, although they might appear to be short continuous movements, they really are driven by stepping motors so the movement is a rapid series of small jerks. I don't know whether that matches the 9 KHz peak, but consider that, especially if this sound comes in brief pulses.

 

[System64MC]

The use of a tube and your ear to locate sound sources is a common technique and easy to do. BUT here we're talking about frequencies of 17.5 and 20 KHz, and most human ears are poor in that range. On the other hand, you have demonstrated already that you do have a MIC that picks those up well. So use THAT detection device. Regarding the tube for concentrating on a narrow area, I suggest using a stiff metal rather than cardboard. Cardboard has enough non-uniformity and flexibility that can absorb (reduce the intensity) over a wide range of frequencies.

A thought about the 9 KHz sound you say comes from the HDD's. You've figured out the two sounds from rotational rates of the disks, and those units do rotate constantly. The other moving parts in an HDD are the arms carrying the heads. They move only is short spurts, not continuously. Further, although they might appear to be short continuous movements, they really are driven by stepping motors so the movement is a rapid series of small jerks. I don't know whether that matches the 9 KHz peak, but consider that, especially if this sound comes in brief pulses.

The HDDs are actually the 17.5 kHz and 20 kHz sounds (they're louder the closer I get to the HDDs with the microphone, I can't hear the sound with my ear). I heard however some short noises coming from them, and they're at 9 kHz, but they're just short bursts; from the spectrogram there's some more faint noise that goes from 8 kHz to 11 kHz with some faint lines around those frequencies, but since they're faint I doubt I'm gonna discover much about them. What remains is all the noise below 1500 Hz; I've managed to figure out a few lines that are emitted by the fans, and I'm guessing the general noise under 1500 Hz is also fan noise or other similar stuff, not really anything specific. Thank you for your suggestions!

 

[Paperdoc]

SOME of the general noise under 1500 Hz will be simply from power systems - 60Hz plus several harmonics - harmonics tend to be of lesser amplitude as you get to higher harmonics. Similar for fan vibrations, of course. There MAY be some noise in there originating from the switching action of the PSU system.

 

[System64MC]

SOME of the general noise under 1500 Hz will be simply from power systems - 60Hz plus several harmonics - harmonics tend to be of lesser amplitude as you get to higher harmonics. Similar for fan vibrations, of course. There MAY be some noise in there originating from the switching action of the PSU system.

Ok, thank you!