Extremely Low Frequency radio receiver update
After all these years, I finally have a good recording of the Extremely Low Frequency (ELF) radio signals.
I use a 100' tall pine tree in front of my house as an antenna, by wrapping it with a coil of wire. Unfortunately, there is a power pole right across the street from it. So I mainly see 60Hz power signals with some extra wiggles in it.
I have a very good way of removing the power frequency and get a signal that is around 1/10th of the raw signal.
The Schumann frequencies are approximately: 7.8, 15,6, 22 and 28 Hz. I can see these frequencies, but when I look at the spectrograph, I was expecting to see continuous bars at these frequencies, but they come and go.
To get a better feel of the signal, I tried to listen to the filtered recording, but it was much too low in frequency, even with headphones. If I speed up the playback It makes a very interesting sound, because frequencies of the signals are increased to the hearing range. Of course it loses its timing. I had taken an online Stanford class, actually from a Spanish university, Universitat Pompeu Fabra, Barcelona Music Technology Group, where they provided Python code that allows shifting the frequency of a signal to higher frequencies using a sinusoidal model. I use these tools to upscale the frequency 8 times. So a 5 Hz tone would become 40 Hz, etc.
During the first atmospheric river storm, we had a short blackout for ~30 seconds. My house has a solar battery power backup, I recorded the event.
I always wondered if the signals I was recording were just noises from the power line, because they are quite chaotic. But now I have a signal with greatly reduced power line noise, still some, because I had power.
In the attached a 2 minute recording, it starts with the power on, after 30 seconds the power is off, until 54 seconds, when power is restored.
The top plot shows the raw signal in green with the 60 Hz filtered signal in blue. The blue signal, while reduced, is still there with the power off.
The middle plot shows a small zoomed 2 second section of 2 the minute recording, which is the filtered signal.
Under it is a frequency spectrum for that 2 seconds.
Here is the Sonification of power failure with 8 times frequency shift, 3 octaves, as a playable 2 minute wave format file.
I had to increase the sample rate from 240 sps to 8000 sps in the sonified version to support the higher frequencies There was quite a bit of noise as the power glitched before going off.
Update
Found a python package that uses FFT to shift the pitch of the signal. I've switched to doing 16 times, 4 octave shift: Sonified 16X audio
