VLF Amateur News 2016

Dec 26th 2016

Paul Nicholson has written a paper on forward error correction.This is worth looking at if, like me, you never found maths that easy. It is on FEC, written so we can understand! Paul is famous for his DX reception of amateur VLF and ULF signals. I am not sure if he is involved in VLF professionally. He certainly knows his stuff.

See http://abelian.org/fec/pn161218.pdf .

Dec 24th 2016

Paul Nicholson continues working with DK7FC to decode his 6470 Hz EbNaut messages.

Dec 19th 2016

Stefan, is yet again pushing the boundaries:

Hi ULF,

In addition to the informations from older mails about this experiment, here are the results of the post-processing:

First, a spectrogram in 424 uHz FFT bin width (after sferic blanking) showing the whole transmission: https://dl.dropboxusercontent.com/u/19882028/ULF/65km_424uHz.jpg

The SNR reached levels up to 22 dB. Even during the (relatively) strong QRN periods at night, the transmissions are still visible. Even the EbNaut transmissions were visible. All this is about 10 dB stronger than i expected. The low winter noise levels and the quiet location certainly plays a big role here. The 'short' DFCW-10800 (3h per element) message "FC" is O copy as the experts say. Inbetween the "F" and "C" there is an EbNaut message (#2) with the content "73". One more example showing the power of EbNaut.

In the last experiments i've shown reference spactrograms beeing generated in low distance (tree grabber) just to give a better overview of what was transmitted, but this is not necessary here :-)

3 EbNaut messages were transmitted. All of them have been decoded with a strong SNR:

1st message: https://dl.dropboxusercontent.com/u/19882028/ULF/65km_EbNaut_2970Hz_Message1.png

2nd message: https://dl.dropboxusercontent.com/u/19882028/ULF/65_EbNaut_2970Hz_Message2.png

3rd message: https://dl.dropboxusercontent.com/u/19882028/ULF/65km_EbNaut_2970Hz_Message3.png

Last but not least a spectrogram showing the recorded range from 0...12 kHz, without a noise blanker: https://dl.dropboxusercontent.com/u/19882028/ULF/65km_wide.jpg

A new experiment in 140 km distance is in the pipeline. The plan is to run it in 2016.

73, Stefan DK7FC

Dec 14th 2016

Stefan has had success o ULF:

Hi ULF,

My post-processing of the 4 day recording is still running, almost 50% of the data is shown on the spectrograms now.

I'm happy to confirm the successful decode of a 5 character EbNaut message over 65 km on the 101 km band, with a BIG SNR! It's the one which started on 08.DEC, 7 UTC. First there was no decode until i found that i made a mistake, the actual start of the transmission is 6:58:40 UTC, an offset of -1 symbol.

Capture in attachment. I also attach the wav file, if someone wants to play with it, the size is very small.

73, Stefan

Dec 12th 2016

I missed this from Stefan DK7FC as it was in my SPAM mailbox!

"After doing a lot of modifications here on the TX side, the effect still exists. I can't get rid of the unstable phase shown on the VLF tree grabber.

One thing to mention is that the phase seems to change slowly, no sudden phase jumps. Thus the peak is 40 dB S/N in 47 uHz. So obviously there were much less uncorrected phase glitches than expected in the first moment when i saw the effect.

Just to be sure i downgraded to the last SpecLab version, the one i used since the last month. But the effect still exists.

Next to note is the i never saw the effect in my last tests. I still have captures showing a very clear trace on 2970 Hz during the 31 km experiment. OK that was in September. The QRN levels may have been a bit higher and the signal was 7 dB weaker. So maybe the effect was just hidden in the noise?

Next i found that some SpecLab configurations seem to be corrupted by forth and back installing old (V2.90b2) and new (V2.92b2) version. So i am now using the old version and a config file that was never opened in the new version. It all seems to work fine on the TX side but i can still see the effect. BTW i even disabled the NB and checked if this makes a change. But thill the same result.

The TX schedule for the experiment is:

07.DEC: TX 2970.000 Hz (antenna current between 105...110 mA all the time) since 13:30 UTC. The recording has started about 15 UTC.

08.DEC: EbNaut 5ch, 80s, 8K19A @ 2970.000 Hz starting 7 UTC, ending 18:22:40 UTC. Then, carrier at 2970.000 Hz. Some interruptions happened in the late carrier.

09.DEC: 0 UTC starting DFCW-10800 (3h per element), 'dit' at 2969.9975 Hz, 'dah' at 2970.000 Hz. The message is 'FC'. Between F and C there is a break of 7 hours. In this time, starting 15 UTC: EbNaut 2ch, 68s, 8K19A @2970.000 Hz.

10.DEC: End of DFCW message 'FC' at 13 UTC. Then, EbNaut 5ch, 80s, 8K19A @ 2970.000 Hz starting 13 UTC (repetition of the message of 08.DEC.)

11.DEC: End of EbNaut message at 0:22:40 UTC. Then, carrier on 2970.000 Hz until 20 UTC. The recording will stopp in the late morning, due to space limitations of the 32 GB USB-stick.

From the local recording i tried to decode the 5ch EbNaut message. But it didn't decode. I just got a '*****' message and a false decode. This is strange. The dots are quite concentrated in the two horizontal lines which indicates a stable phase. See attachment. Also the signal peak is strong and narrow. But in the red field there is nothing to see... This EbNaut message was transmitted with the new PSK config in SL. It is the reason why i will repeat the message in the end of the experiment again, using the old settings and version.

I should have done more pre-tests before starting the experiment. But now it is running. Well, the main goal is to detect a clear trace of my signal in a spectrogram running at 300 uHz or so. The signals do not look bad on the local grabber, not at all, despite the strange (propagation?) effect. I'm still optimistic...

73, Stefan"

Dec 5th 2016

More on ULF tests from DK7FC:

ULF,

It was a clear, cold and sunny day again and finally i was testing the new large 2.97 kHz coil on the roof: https://dl.dropboxusercontent.com/u/19882028/ULF/20161204_153126.jpg

Not a good idea on MF or LF, but on ULF is can be a good idea to put a resistor in series to the coil. I put 500 Ohm in series. This helps to attenuate eventual HV transients, makes the resonance curve a bit flatter, so the peak is easier to hold.

500 Ohm and 0.1 A, that just a loss of 5W, quite affordable. https://dl.dropboxusercontent.com/u/19882028/ULF/20161204_153113.jpg

Unfortunetely i found that the actual resonance region is adjustable from 2993...3130 Hz. So the inductance is still to low :-( I've been quite sure that my antenna replacement capacity for resonance tests in the shack has the right value. So maybe other effects like an additional capacitive load by the things standing arround in the shack? Or a different behaviour of the residual antenna + frozen ground capacity on 2970 Hz?

At least it was possible to get 100 mA anyway, just by applying some more voltage to the antenna. The PA is now running at 20V and less than 1A, most of that are PA- and transformer losses. Anyway i have 12 kV on the antenna! That's ULF!

I was planning to bring the coil down to the shack anyway, after the 65km experiment which will start in a few days i think. Then i can add more turns to come closer to 2970 Hz.

With 70 uOhm radiation resistance the resulting ERP is now entering the microwatt range: 1.2 uW on the 101km band.

A stable carrier at 100 mA was sent from 12:50 UTC to 14:24 UTC.

I tried to measure the antenna current with a true rms digital multimeter, shielded by aluminium foil, measuring on high potential. But the measurement was wrong anyway. The permanent installed analog amperemeter and SpecLab gave a correct measurement. Tomorrow i will check the antenna current measurement on high potential with a calibrated analog amperemeter.

For the 65km experiment i want to transmit some DFCW-12000 again as well as an EbNaut message. The RX site will be in JN39VH :-)

73, Stefan

Nov 5th 2016

Stefan DK7FC has been doing some tests with Paul Nicholson (UK) on 6470 Hz using EbNaut messages with partial success. Very narrow bandwidths have been used with a lot of post processing.

Oct 16th 2016

Hi ULF,

This night i have completed the 4th stack of my large ULF loading coil.

An image: https://dl.dropboxusercontent.com/u/19882028/ULF/20161008_ULF_TX_coil.jpg

The exciting question was, what is the new resonance frequency together with my antenna. In the lab i put an equivalent C in parallel and found the resonance peak at 3.67 kHz. So the L seems to be 3.8 H now. This is still not enough. The l/d of the coil dimension becomes larger now, so i need more and more wire to reach the final resonance frequency of 2.97 kHz at about 500 pF.

I will do a next test on the actual resonance frequency in some days, when the WX permits. I will also check where the resonance can be found when switching the small coil in series that i use for 6.47 kHz now. It could be quite close to 2.97 kHz then! So maybe this allows a next QRO step on 2.97 kHz: 100 mA antenna current must be reached for the planned experiment in 65 km distance.

BTW the 4 stack coil is wound out of almost 4 kg of enameled Cu wire with 0.4 mm diameter. The layers are isolated by Kapton tape. My assumption is that it will hold 20 kV now.

Ah and the DC resistance is 445 Ohm now. When connecting 275 V AC / 50 Hz, i can measure 269 mA. So the actual inductance is 2.93 H (beeing farer away from the self resonance frequency)

I can tell you it's quite fascinating to operate down on these low frequencies :-) To be continued...

73, Stefan/DK7FC

Sept 25th 2016

New record on 2.97kHz by DK7FC

Hi ULF,

Now the results of the post-processing of my 3 day recording are available. It shows very clearly that my 250 nW ERP (average 46 mA antenna current), 2970 Hz signal was copied in 31.3 km distance. The recording location was in a more or less quiet location, in http://no.nonsense.ee/qth/map.html?qth=JN49IQ37AM. The transmission path is shown in http://no.nonsense.ee/qth/map.html?qth=JN49IQ37AM&from=jn49ik00wd

In this second attempt, everything worked fine, stable GPS reception, accurate orientation of the vertical, single turn receive loop and the recording of a single 30 GB .raw file at 32 kS/s (covering 0...16 kHz). However the QRN could have been lower. But there was no QRM by electrical fences...

Here's a spectrogram showing the complete, unfiltered recording: https://dl.dropboxusercontent.com/u/19882028/ULF/31km_wide_3days.jpg

Of course there is nothing visible of the transmission in this resolution, it is just an overview to get an impression about QRN and QRM in that location. QRM from trains and mains harmonics is the critical parameter on ULF.

The transmission consists of 3 parts: A long dash on 2970.000 Hz, then the message "73" in DFCW-12000 (2.5 mHz frequency shift) with the lower dash sent on 2970.000 Hz and then another long dash on 2969.990 Hz

After filtering and noise blanking, here are the resulting spectrograms in 476 uHz and 238 uHz:

https://dl.dropboxusercontent.com/u/19882028/ULF/31km_476uHz.jpg

https://dl.dropboxusercontent.com/u/19882028/ULF/31km_238uHz.jpg

The experienced visual human spectrogram interpreter experts may find a significant part of the message :-)

Just as a reference, here are spectrograms showing the same transmission received in 3.5 km distance (http://www.iup.uni-heidelberg.de/schaefer_vlf/DK7FC_VLF_Grabber2.html).

https://dl.dropboxusercontent.com/u/19882028/ULF/ref_424uHz.jpg

https://dl.dropboxusercontent.com/u/19882028/ULF/ref_212uHz.jpg

I'm not sure if this is the highest distance ever crossed with a message by radio amateurs on ULF (0.3...3 kHz), but it is certainly one of them :-)

A new experiment in about 65 km distance is in preparation. Probably it will be done in a few weeks. The location will be quieter (the Pfälzer Wald, JN39VH) and there will be less QRN!

73, Stefan/DK7FC

Sept 12th 2016

More from Stefan DK7FC about his 2.97kHz (101km band) tests:

Hi ULF,

Since yesterday (10.Sep), 11 UTC, i'm running a new experiment on the 101 km band. It is in 31.2 km distance to my transmitter-site.

The last experiment did not produce satisfying results due to various reasons.

Now i improved the loop cross section area, the loop orientation, GPS reception (for the 1 PPS + NMEA reference signal). Furthermore the recording sample rate was reduced to 32 kS/s which allows about 3 days of stereo recording.

I'm still using the Raspberry Pi2 for therecording. The system consumes just 150 mA from a 12V battery. 3x 7 Ah lead acid batteries are in use now.

The transmission contains stable carriers and a short DFCW-12000 message, "73". The transmission can be followed on my 3.5 km distant forest grabber at http://www.iup.uni-heidelberg.de/schaefer_vlf/DK7FC_VLF_Grabber2.html , as usual.

Relative to the last experiment, the average QRN has dropped significantly now, another advantage. So i'm quite optimistic to get a good result!

The next experiment in about 60 km distance is already in preparation. Then i will use my new coil which is still in preparation. The last measurement, after completing the 3rd stack, showed a resonance frequency of 4.27 kHz. So it should have 2.8 H now...

In the next experiment, i will use 100 mA antenna current, again a about 6 dB stronger signal.

The current experiment will stop on Tuesday afternoon. Results can be expected in the end of the next week.

73, Stefan/DK7FC

Sept 2nd 2016

Joe VO1NA has built this loading coil for 8.277kHz. Sadly the coil arced and had to be rewound.