Help with detecting 25-50Hz audio signal

[eyAyXGhF]

Hey all,
I'm wanting to detect an audio signal that's somewhere between 25-50Hz, and logging it's amplitude over time for several days at 1 second intervals.

My only existing knowledge is a bit of arduino and some linear electronics.

I know you can do FFT with an arduino, but I don't know enough about it to know how much accuracy you can get. I also don't even know if this would be the best approach. I'm also guessing some filtering on the microphone input would help? I've used digikey part #P9925-ND before, which is an electret microphone with a response as low as 20Hz.

Any ideas if this is possible?

Mike

 

[alec_t]

Do you need to know the actual frequency of the detected signal? Would, say, using a simple low-pass analogue filter to suppress (not necessarily reject entirely) frequencies above ~ 60Hz be adequate, or is narrow-band-pass filtering essential? What amplitude variation do you anticipate? How accurately do you need to measure amplitude?

 

[eyAyXGhF]

Hey Alec,
As of right now, I do not know the exact frequency of the signal. I'm hoping to make a recording of it in a few weeks, which should answer most of the questions. If you're interested, this is what I'm trying to detect:

https://www.windsorstar.com/technology/Windsor+Essex+rumblings+reminiscent+Kokomo/4631510/story.html

Ignore the crappy journalism. Basically it's a low frequency sound heard throughout the city and surrounding areas, with no explanation or source identified. I thought it sounded like some geek ingenuity could at least help. My follow up questions will eventually be about detecting directionality of the noise, but one step at a time I suppose

 

[carbonzit]

Interesting.

So in the spirit of scientific inquiry, wouldn't it be better for your experiment if you started with a more open-ended approach, rather than making a priori assumptions about the frequency of these purported signals? Your guess of 25-50Hz might be accurate or it might not. Set your instruments to pick up sound in the low-frequency range, then you can sort them out later.

I'm certainly no expert on this, but it sounds as if FFT might be the way to go here (but I'm not sure about using it to work on such a low frequency: is that easier or harder to do?).

 

[Reloadron]

Interesting read. Where you mention "low frequency sound heard throughout the city" others I believe from the article seem to "feel" the sound. If I place my ear close to for example a power type transformer I can hear the 60 Hz hum and if I place a finger tip on the transformer I can feel the hum. So where you are do you hear or feel these low frequency vibrations? This makes a very big difference in how these sounds or vibrations are sensed. Now I have no clue what you have for an electronics knowledge base or what tools you have available?

I guess if I were you I would go to my local Lowes and buy a Grounding Rod and drive the sucker about 7 of its 8 foot length into mother earth. I would drill and tap the top of the rod and mount an accelerometer something along these lines. Even if you could find a way to mount a guitar pickup. Then amplify the output and monitor it on a scope. The same sensor could be mounted on a large plate of metal and suspended in air to see what you get.

I would think before you worry about recording it it may be wise to try and see what it looks like?

Ron

 

[eyAyXGhF]

Hey Reloadron,
My wife and I actually drove all around the city listening for it one night when it was very active. It's not 60Hz hum, I'm very familiar with that (common problem with guitar/audio setups) it's definitely lower than that. I've never -felt- it, and we've been in various parks with our ears to the ground literally listening for it.

The problem is that it always sounds "off in the horizon". I can guess an approximate direction, but low frequencies are very hard for us to pinpoint (I'm guessing because of the distance between our ears?).

Other people are reporting having actually felt it though, so maybe they're closer to the source.

Here's a more recent article about it: https://www.windsorstar.com/technology/real/5231824/story.html

Either way, it's a fun little problem to think about.

 

[carbonzit]

Fun, indeed.

If you really want to investigate this, and if you can scrounge the gear somehow, it might be worthwhile getting your hands on the best microphone you can, at least so far as low-frequency response goes. The lower the better, since the phenomenon may actually be subsonic. (And you really need to take almost all specs on mikes with a huuuge grain of salt; many are wildly exaggerated.) Then try to capture a recording of it; a minute or two would probably be enough.

Then you could try to analyze it at your leisure, maybe trying analog techniques first (get a 'scope and look at it).

 

[Mikebits]

Maybe what you need is a geo phone, they are used in Seismic indicators. They are good for detecting low freq vibrations (10Hz-240Hz). I picked up a couple cheap on Ebay.
https://en.wikipedia.org/wiki/Geophone
https://www.electro-tech-online.com/custompdfs/2011/08/SM-2420ST20Brochure.pdf
https://www.geophone.com/Category.asp?CatID=1&SubCatID=3

 

[Reloadron]

I would be starting looking seismic. I would also be interested in what the current seismic devices actually record. Back to driving a stake in the ground. I liked Mike's links.

Ron

 

[Mr RB]

You can use a large diameter speaker as a subsonic microphone. A woofer speaker of 10" or better 12" should put out a pretty decent signal from 25-50Hz subsonics, although it will be a bit directional (which might even be a bonus).

 

[Externet]

What you actually could use is a low pass filter and greatly amplify and record it.

You are not alone; this is going on for years, and was discussed in other forum.

http://news.bbc.co.uk/2/hi/8056284.stm

https://www.electro-tech-online.com/custompdfs/2011/08/jse_18_4_deming.pdf

http://sandaura.wordpress.com/2011/03/23/a-constant-non-stop-low-frequency-hum-heard-in-my-home/

**broken link removed**

http://en.wikipedia.org/wiki/The_Hum

 

[squishy36]

Interesting problem. I've read that low frequency and infrasonic sounds are quite difficult for people to locate. If this is something that is amenable to study, I would first get out there with some general purpose measuring equipment, leaning towards seismic-type equipment that is capable at these lower frequencies. I'd also spend some time tracking down the technical reports made by others who have investigated it. You'll learn a lot to help you in your approach should you want to continue it.

You could set up multiple recording stations that are synchronized in time and look for dispersion or other behaviors that might give you a clue as to the direction or nature of the source(s). I'd first want to see a waterfall-type frequency spectrum of the noise. It also wouldn't hurt to get some Navy sonar techs involved too.

 

[Mikebits]

You can use a large diameter speaker as a subsonic microphone. A woofer speaker of 10" or better 12" should put out a pretty decent signal from 25-50Hz subsonics, although it will be a bit directional (which might even be a bonus).

And he can wear a tin foil hat on his head and look really special.

 

[alec_t]

Interesting topic. I've been toying with doing a similar project to locate a local industrial noise source (~100-500 Hz at a guess). Keep us posted on results.

 

[4pyros]

Look here. Speaker with low pass and amp.
http://amasci.com/hum/lennart2.gif

 

[4pyros]

And here is some talk about it.
**broken link removed**

 

[4pyros]

OK its an "ELF-receiver" Extremely Low Frequency Receiver. Lots of stuff on the net about it. Andy

 

[KeepItSimpleStupid]

Just pondering. Has anyone thought about underground fires? e.g a fire hits a large pocket of air or fuel such as a gas pocket and causes an avalanche of sorts. Just a thought.

 

[Mr RB]

And he can wear a tin foil hat on his head and look really special.

I don't get it? Are you implying that my suggestion to use a large diameter speaker as the sensing element was less credible than your suggestion to use a geo phone?

I suggested that as the lower cost and better availability of a large speaker might make it attractive as a sensor for a hobby-level infrasonic detector.

 

[Externet]

Disembowel an unwanted speaker from its cone, dust dome and frame very carefully with a saw/razor blade; leaving the magnet, coil, flex wires and spider intact.

With proper glue, attach the rim of the coil to a large as possible glass window centre, supporting it/hang it steady overnight.

Now you have a large area microphone to play with your low-pass filtered amplifier. Bonus: it works great as subwoofer speaker too.