Tuned Sound Amplifier/Receiver for the VLF Christmas Eve transmission from SAQ, Grimeton, Sweden

Jump to Latest
#463
bucks bunny said:
Removing the noise h-field this is the noise floor remaining
Click to expand...
Oh yes at first glance I thought it was noise but its signals maybe electronic noise or real signals? I'm not used to looking at such a wide spectrum after concentrating on SAQ. The signals ofter seem to be at multiples of 10 KHz! I had another look at your circuit, it looks to me as if Q2 is a constant current generator whereas Q1 is providing the gain. If it works well its a very minimal low cost circuit. 2N2222 is very popular npn and the pnp is quite common too.

Understand this is intended to be a flat response - for non selective operation from 10 to 100 kHz so that's an interesting range. My original thinking for a strongly resonant circuit (for SAQ a long distance away) was to do part of the filtering as analog components and the rest digitally. Doing all the filtering digitally runs the risk of strange aliasing effects - digital filtering isn't perfect - the best possible digital filter is (Sin x)/x but unfortunately it has infinite number of coefficients! But digital filtering is pretty good.

Have you tried comparing the circuit noise generated by this transistor circuit and an equivalent OpAmp circuit? Have you tried feeding your home-brew coil from a signal generator and checking how linear the response is?

It seems to me that if an antenna is connected a good earth must also be connected, in all the tests I did signal strength shot up when a good earth was connected suggesting that there is strong current flow between antenna and earth. I also tried laying a 25 metre insulated wire along the ground (along a concrete path) also with an earth connected and the signal strength was very strong but unfortunately it picked up a lot of noise signals as well.
 
Last edited:
#466
fubar3 said:
VLF wizards might be impressed by the 3-axis receiver at http://abelian.org/todmorden-vlf/
I am no wizard but it seems they have a complete receiving system for H-field and another for E-field. The data is merged with GPS time signals using http://abelian.org/vlfrx-tools/
Click to expand...
The Todmorden VLF spectrum looks really weird, I don't understand it. The circuits on the abelian website are interesting much like what we have been working on. The example signals from Todmorden VLF they give are well beyond weird - to me they really so sound like the dawn chorus of bird song.

Impressive find, its amazing what's on the internet, people have taken a lot of time to do those websites.

To be honest I knew about VLF before especially the 60kHz (time standards) but learnt a lot more recently.

I'm surprised a lot of people use short (1-2m) vertical rod antenna, my small amount of experience suggests that long horizontal wire above ground (together with a good earth) is the best for VLF.
 
#467
[ I'm surprised a lot of people use short (1-2m) vertical rod antenna, my small amount of experience suggests that long horizontal wire above ground (together with a good earth) is the best for VLF.]

Is the long wire directional. Perhaps with a vertical antenna you can get more signals .. including the ones you don't want.
 
#470
richgwilliams said:
Oh yes at first glance I thought it was noise but its signals maybe electronic noise or real signals? I'm not used to looking at such a wide spectrum after concentrating on SAQ. The signals ofter seem to be at multiples of 10 KHz! I had another look at your circuit, it looks to me as if Q2 is a constant current generator whereas Q1 is providing the gain. If it works well its a very minimal low cost circuit. 2N2222 is very popular npn and the pnp is quite common too.

Understand this is intended to be a flat response - for non selective operation from 10 to 100 kHz so that's an interesting range. My original thinking for a strongly resonant circuit (for SAQ a long distance away) was to do part of the filtering as analog components and the rest digitally. Doing all the filtering digitally runs the risk of strange aliasing effects - digital filtering isn't perfect - the best possible digital filter is (Sin x)/x but unfortunately it has infinite number of coefficients! But digital filtering is pretty good.

Have you tried comparing the circuit noise generated by this transistor circuit and an equivalent OpAmp circuit? Have you tried feeding your home-brew coil from a signal generator and checking how linear the response is?

It seems to me that if an antenna is connected a good earth must also be connected, in all the tests I did signal strength shot up when a good earth was connected suggesting that there is strong current flow between antenna and earth. I also tried laying a 25 metre insulated wire along the ground (along a concrete path) also with an earth connected and the signal strength was very strong but unfortunately it picked up a lot of noise signals as well.
Click to expand...
Most maritime mobile VLF stations are using a protocol optimizing data rate vs bandwidth, which shows a characteristic spectrum. See
https://en.wikipedia.org/wiki/Minimum-shift_keying
Hence the FFT of the received VLF range should reflect that.
1705202338241.png

1705202468969.png
 
#471
Aridace said:
Most maritime mobile VLF stations are using a protocol optimizing data rate vs bandwidth, which shows a characteristic spectrum.
Click to expand...
I have seen that shape many times on VLF, it looks like some form of truncated sinc (sinx/x) function.

Mathematically its possible to prove which protocol is going to give the best s/n ratio at the fastest data rate at a given transmitter power and so on. I thinks its a lot like ADSL and ADSL2 works ?? Or maybe how older design modems worked.

I have seen some stations on there where they have two carrier frequencies close together and they switch between them.
 
#472
I have seen some stations on there where they have two carrier frequencies close together and they switch between them.
Click to expand...
That could be FT8, each transmission has a carrier that varies by only a few hertz, changing about twice a second. I think it's actually eight frequencies it uses within its narrow band (thus FT8). On most HF bands there's an established frequency where a dozen or more transmissions occur simultaneously within a SSB voice band (2.7kHz). It's generated and decoded on a computer connected to the radio's mic and audio out connections, and uses a lot of DSP. You can send and receive messages around the world with a 100 watt radio.
https://en.wikipedia.org/wiki/FT8

ETA: Here's the technical stuff on FT8 - notably (see graphs pp. 14-15), an FT8 signal can be decoded 20 dB BELOW the noise floor.
https://wsjt.sourceforge.io/FT4_FT8_QEX.pdf
 
Last edited:
#473
richgwilliams said:
I have seen that shape many times on VLF, it looks like some form of truncated sinc (sinx/x) function.

Mathematically its possible to prove which protocol is going to give the best s/n ratio at the fastest data rate at a given transmitter power and so on. I thinks its a lot like ADSL and ADSL2 works ?? Or maybe how older design modems worked.

I have seen some stations on there where they have two carrier frequencies close together and they switch between them.
Click to expand...
Due to my distance from most VLF transmissions I can see the sidebands of just a few. Long ago I treated FSK as two ASK signals which worked better than any FM demodulator (including PLL). It even worked well for MSK and similar modulations, up to 200 bps. I think a major reason to optimize modulation for maximum speed is the use of a coding protocol to hide the information for those who're not intended to be reading along. While with plain text it's easy to correct errors, that's different with a coded signal. So it's likely some form of FEC is used and the modulation is tweaked for maximum depth of the submarines. With an FM demodulator at the previous source, the rise and fall time are obvious.
1705256238565.png
 
#475
benb said:
That could be FT8, each transmission has a carrier that varies by only a few hertz, changing about twice a second. I think it's actually eight frequencies it uses within its narrow band (thus FT8). On most HF bands there's an established frequency where a dozen or more transmissions occur simultaneously within a SSB voice band (2.7kHz). It's generated and decoded on a computer connected to the radio's mic and audio out connections, and uses a lot of DSP. You can send and receive messages around the world with a 100 watt radio.
https://en.wikipedia.org/wiki/FT8

ETA: Here's the technical stuff on FT8 - notably (see graphs pp. 14-15), an FT8 signal can be decoded 20 dB BELOW the noise floor.
https://wsjt.sourceforge.io/FT4_FT8_QEX.pdf
Click to expand...

That above is the CW switching I referred to.

The FT4 and FT8 are amazing, the graphs and the detailed theory (did not try to understand it).
 
You must log in or register to reply here.
Top Bottom