The antenna is a Ground Dipole ( or probably better a Ground Loop ) as in the picture .
The antenna is of course very inefficient being a very small fraction of the WL ( Wave Lenght) that is around 26km .
Also the power supplied ( and dissipated , rather than radiated ) to the antenna system is small , in the range of 50 Watts .
Current is sensed with a resistor of 1 Ohm 5W .
The schematic of the I-V sensing was taken from Renato Romero article :
http://www.vlf.it/ed/earthprobes.html
Calculation shows a power rms input around 58 W .
The power amplifier is an analog class AB audio amplifier with Mosfets , powered with differential power supply +/- 25V that I found somewhere on a flea Ham market many years ago and used ,cutting the feedback , for tests in the 136 KHz Band .
I put a couple of fans on the heater for having 24h operation .
What is important is that the frequency of the carrier is very stable because generated from an HP double OCXO ( Oven Cristal Oscillator ) disciplined with GPS HP5850 A shown in the following picture together with the Arbitrary Waveform Generator Synthesizer .
The specs of the frequency reference :
https://www.sglabs.it/public/HP58503A.pdf
The overall schematic of the transmitting system is the the following picture :
This Frequency/Phase high stability was one of the key of success of this "Mission impossible" .
With "normal grabbers spectrograms " , we was even not able at the beginning to say surely that the signal was detectable in Cumiana that is around one WL from my QTH , till Paul Nicholson , a VLF guru , started to record and analyze the streaming of the signal coming in real time from Cumiana , using samples lasting some hours .
At this stage the signal popped out clearly from the noise at the right frequency as shown in the picture after 50hours in a BW of 5.3 uHz .
As you can see the " span " of the spectrum mean for 50 hours is 2 mHz and the RBW ( Resolution BandWidth is 5.3 uHz ( microhertz = 1 Hz / 1000000 ) .
Impossible to do without a very precise frequency/phase reference of both RX and TX sites .
At this stage , encouraged from the results , Paul decided to put on the web the possibility to postprocess in real time from everyone the signals recorded on air from the stations that are sending to his elaboration center the signals received .
Here you can use this marvelous tool from his extreme site : http://abelian.org/
The tool :
http://78.46.38.217/fbins3.html#r=vlf1&p=1591009200&b=000&s=sp&m=cardioid&w=r&h=62&z1=0.34&z2=0.64&c=1
Just for fun we tried to see the signals at Bielefeld ( Germany ) and being surprised that there was clear sign of the signal on the right frequency .
Again for fun we got some signs that maybe something was also received at the QTH of Paul in Todmorden at around 1200 km of distance from my QTH in Turin .
To be sure about , Paul pushed very hard on many algorithm of noise and interference suppression , compensate the different delay of propagation between night and day , cut out part of the signal where another transmission on the same frequency was present and patiently waited that the very very small signal increased at the right frequency during the time .
As much as you can wait time , you can decrease the bandwidth to better extract the signal from noise and interference .
After nearly 5.5 days of elaboration the following spectrum popped out very well also at Todmorden :
Please note the calibrated vertical linear scale of Field Strength : nV/m ( nanovolt/m ), this means some electrons being received ...., but especially the horizontal frequency scale with a total span of 1.2 mHz and 0.05 mHz/div = 50 uHz/div !
If you look how many signal you can isolate in one division , we can roughly estimate the RBW of the spectrum analysis in 0.0025 mHz , that is 2.5uHz .
Something very extreme , but at the end the satisfaction for everyone to have demonstrate how the processing techniques can hardly dig to make very low signals popping out from the noise .
What we have broadcasted ?
1bit in around 5.5 days ......
This is not the longer detection in this band .
Other have made even intercontinental TX-RX , but with much bigger EIRP and broadcasting at much higher bit rates .
Milestones in this field are reported here from Paul :
http://abelian.org/vlf/amateur-radio/
The "Mission Impossible" here was to use extreme postprocessing techniques to allow to receive/recognize with no doubt a far signal with extremely low EIRP ( Ultra QRP ... ) .
Here is the official announcement on the VLF list : https://groups.io/g/VLF/topics from Paul :
Da: paul@...
Data: 30-mag-2020 10.41
A: <VLF@groups.io>
Ogg: Re: [VLF] Transmitting at 8720.020 Hz with a new antenna
I have decided to claim detection of Claudio's ultra-QRP signal
at 8270.02 Hz.
http://abelian.org/vlf/tmp/200530-i1rfq.png
Frequency is exact;
Strongest peak in 24000 bins excluding the other known amateur
signals;
The S/N continues to improve as more data is added;
Azimuth is correct (GC is 139.8 deg, peak S/N 143 deg);
Partitioning the data gives the same peak;
Paul Nicholson
For future improvements I have placed a remote station at the QTH of Rodolfo Zucchetti that is around 700m from my QTH .
The RX antenna is a Wideband L101ES connected to the audio Sound card of a laptop , with HDSDR and Anydesk .
With an RBW of 0.1 Hz ( looks small , but it is 100000 uHz !) the S/N is around 20 dB ....
With a CW demodulator , event with the smallest filter , you can't listen the carrier .
