I already have this guy with a (USA) broadcast band loop. I wonder if it wouldn't be easier to just temporarily try to add some parallel capacitance to bring it down to VLF?
The Si5351 synthesizer in my VFO is specified to 8 KHz, my RF Amp and mixer are specified to DC, so unless there is a programming problem with the Si5351, in theory I already have a receiver in place to try to receive this, but not optimistic about coming up with an antenna.
Assuming I can get away with playing radio at the right time ...
Win W5JAG
The Si5351 synthesizer in my VFO is specified to 8 KHz, my RF Amp and mixer are specified to DC, so unless there is a programming problem with the Si5351, in theory I already have a receiver in place to try to receive this, but not optimistic about coming up with an antenna.
Assuming I can get away with playing radio at the right time ...
Win W5JAG
Attachments
In the presence of pulse noise, filtering has the disadvantage of stretching a pulse, with the effect that it buries (part of) the wanted signal. So pulses have to be processed before filtering at the highest BW possible, and "close to clipping" opamps can do that rather well w/o additional hardware. Did you already connect the tuned ferrite coil to a mike input of the PC soundcard, and ran an FFT with sufficient resolution? When able to run an FFT, displaying it as waterfall, the SAQ morse can be decoded visually. It will look like this (15m ham band CW)
The parallel C for a MW coil to resonate at 17.2k is rather high but it can work. Induced voltage will be low which explains the need for low noise amp despite the very high noise level.
It's got a single stage bipolar amp in the loop base, but I don't know much about it. I bought this base and a couple of loops about twenty years ago, and can probably count on one hand the number of times I've used it.
It does work well on the MW band.
Win W5JAG
It does work well on the MW band.
Win W5JAG
Dec 24 is coming close. It will be interesting to hear how the recording goes!
Myself I live in Sweden, less than 500 km from the SAQ transmitter. I have found it possible to pick up the signal in the past using a loop antenna coupled directly to the microphone input of a laptop running e.g. Audacity.
Listening to a recording made with e.g. Audacity doesn't work for me - I am too old to hear the 17200 Hz tone.
Playing it back at, say, half speed should work in principle. However, there may be other VLF transmitters that drown out the SAQ message. These other transmitters may be as close as 1000 Hz away from 17200 Hz, so an antenna resonant at 17200 Hz will help. And bandpass filtering may improve the audibility further - an audio bandwidth of 200 Hz or less should be ideal.
Playing at reduced speed may also make it easier to work out the message. The telegraphist at SAQ sends at around 15 WPM (words per minute) - a bit fast for me; I can decipher it, but only by playing it over and over many times.
I have made a synthetic sound clip to give an idea of what the half-speed message could sound like: https://privat.bahnhof.se/wb748077/radio/vlf/saq-synthetic-8600.wav
The tone frequency is 8600 Hz, and the rate of sending is 7.5 WPM. The message is the SAQ call: "cq cq cq de saq saq saq". With my aged ears I can hear it, but only when playing it really loud.
Myself I live in Sweden, less than 500 km from the SAQ transmitter. I have found it possible to pick up the signal in the past using a loop antenna coupled directly to the microphone input of a laptop running e.g. Audacity.
Listening to a recording made with e.g. Audacity doesn't work for me - I am too old to hear the 17200 Hz tone.
Playing it back at, say, half speed should work in principle. However, there may be other VLF transmitters that drown out the SAQ message. These other transmitters may be as close as 1000 Hz away from 17200 Hz, so an antenna resonant at 17200 Hz will help. And bandpass filtering may improve the audibility further - an audio bandwidth of 200 Hz or less should be ideal.
Playing at reduced speed may also make it easier to work out the message. The telegraphist at SAQ sends at around 15 WPM (words per minute) - a bit fast for me; I can decipher it, but only by playing it over and over many times.
I have made a synthetic sound clip to give an idea of what the half-speed message could sound like: https://privat.bahnhof.se/wb748077/radio/vlf/saq-synthetic-8600.wav
The tone frequency is 8600 Hz, and the rate of sending is 7.5 WPM. The message is the SAQ call: "cq cq cq de saq saq saq". With my aged ears I can hear it, but only when playing it really loud.
Another thing, if a recording is made there is a program that can decode the audio to plain text. Find it here: https://github.com/fowlay/gerke-decoder
It can be run on Linux or Windows. You would have to specify option -F 17000,17400 to indicate that the signal should be looked for in that frequency range. If you encounter problems, please get back to me and I will try to help.
It can be run on Linux or Windows. You would have to specify option -F 17000,17400 to indicate that the signal should be looked for in that frequency range. If you encounter problems, please get back to me and I will try to help.
Theoretically it is necessary to find the spectral components of a noise pulse by use of Fourier transform. Then a perfect "notch" filter "bandpass" filter will only pass frequency components from the noise pulse that are in its pass band.
I agree that a noise pulse can be stretched by digital filters if the digital filter is too simple or sampling rate is insufficiently high and in this instance even 48,000 samples per second by the Laptop PC sound input is not high enough for comfort.
But noise pulse stretching does not happen with well designed analogue filters. It is possible to construct a very strong analogue band pass filter that is very sharp and normally these filters have a low Q but attenuate very strongly below or above the centre frequency. It is possible to construct an analogue filter with a gain of 1 at 17200 Hz that falls off sharply above and below this frequency at 48dB per octave or more. Such a filter would reject noise pulses, it would only see 17200 Hz.
Constructing such a strong bandpass filter is somewhat beyond the scope of what's reasonable at the moment. A reasonable thing is to make the Ferrite Rod a resonant front end. High Q is a good idea but I think noise pulses may cause this simple device to ring somewhat at high Q or allow noise to pass through it at low Q.
Assuming a Morse SAQ speed of 15 WPM (as @fowlay mentions) and 50 dits per word (dit being shortest event eg dot) then each dit is 80mSec long, that's 17200 x 0.08 = 1376 carrier cycles. So assuming we use the program Audacity (as mentioned) to display the recorded signal as a graph then a dot should be 1376 cycles of waveform, the smallest gap should be 1376 cycles of nothing and so on. Point being Ferrite rod ringing won't last 1376 cycles, sharp noise pulses hopefully won't last 1376 cycles etc.
I see what you are saying about making the front end wide band then allowing an OpAmp to clip noise pulses, often a pair of diodes in the OpAmp feedback loop are used to clip a signal like this.
That's an interesting post, thank you its nice to have some confirmation that the methods some of us are proposing are likely to work within the constraints you mention. I listened to the synthetic sound clip so yes at that speed or less it gives me a chance to read the morse code.
My approach is to record the 17200 Hz carrier as best as I can (using Audacity or something like it) and then work out how to process/listen to it at leisure afterwards. I think that Audacity can also be used in real time to tune in the carrier in the first place and make any gain adjustments.
I have a signal generator and a frequency counter so my plan is to tune the ferrite rod for as close to 17200 Hz as I can get and use the frequency counter to check that I am exactly at 17200 Hz. To get signal into the receiver Ferrite Rod I plan to have a second resonant coil a few metres away that is "transmitting" my test 17200 Hz signal.
No doubt a few things will be quickly improvised between now and tomorrow morning !
It is possible to select the strongest signal while blocking out most of the noise. Use devices such as LM339, LM358. Boost the signals with the LM358 and feed it to LM339 adjustable comparator which will indicate when the carrier turns on and off. Use the LM339 pulse to blink a LED and activate an oscillator to produce an audio tone. I live in the noisy Toronto area so I can't test these ideas. Good luck.
According to https://en.wikipedia.org/wiki/List_of_VLF-transmitters there are transmitters close to 17.2k and one of them is the Norwegian navy at 16.4k. with equal ERP as SAQ. It's possible the info is outdated but if so, the navy will produce the strongest signal. There are many others but here in Panama only 2 stations are adjacent to SAQ, the one to the left, JXN at 16.4k, to the right, INS at 18.2k. When wondering what the "straight lines" around NWC at 19.8k are, they are 60 Hz harmonics.
What also might succeed in a day is using HDSR as it can use direct input from soundcard, and will provide audible CW tones.
https://www.hdsdr.de
https://www.hdsdr.de
So I have built and partly tested a receiver as shown by the photo and the circuit diagram below. As you can see it is very simple, very basic. It peaks strongly at 17.2 kHz and picks up my test coil some distance away.
Unfortunately I don't think it has enough gain as is - ideally needing another tuned circuit and more gain, but I will connect a long wire and an earth to it tomorrow morning and see what happens. I did look for a better OpAmp but could only find some LM358's
-
Unfortunately I don't think it has enough gain as is - ideally needing another tuned circuit and more gain, but I will connect a long wire and an earth to it tomorrow morning and see what happens. I did look for a better OpAmp but could only find some LM358's
-
Last edited:
If the noise floor is low enough, you can always apply volume normalization (or extra filtering and volume normalization) afterwards in an audio editing program like Goldwave or Audacity.
If it's sensitive enough, noise has to be visible, centered around 17.2k so gain has to be increased till it's visible. What to expect is visible on this multi-user SDR: http://websdr.ewi.utwente.nl:8901/ which uses a mini-whip with near-perfect ground.
With a tuned ferrite antenna, op-amp gain of 40dB I feed into the Mic input of my soundcard. With max mic gain (ca 60dB) I see my resonant noise peak at -40~30 dB FS. The incoming antenna signal hence is in the ballpark of several uV.
At the time I use the RTA inside REW for monitoring. Tascam DR-40X for recording.
At the time I use the RTA inside REW for monitoring. Tascam DR-40X for recording.