Author Archives: DK8OK

ALE [MIL-STD-188-141A]: Which one is the best Decoder?

This is an update from my post two days ago. I have expanded the number of test signals and added some hints.

Does your decoder read this track? Buried in noise and plagued by multipath fading, the recordings below will separate the wheat from the whaff.

Often I am asked – and sometimes even asking myself! – “Which one is the best decoder for ALE?” This means: Which one delivers the best decoding under demanding conditions?

To test this, I made a recording of twelve stations “on the air” plus one weak signal, buried in Additive White Gaussian Noise, AWGN. All signals are correctly tuned, no one invers. All were read by at leastby one of my decoders “in a row”.

To test your decoders, you should download this WAV file of 131 seconds length and play it. It can be either directly opened by some decoder, or feed it via virtual audio cable (VAC) into a decoder. I used Audacity for this.

I am as interested in the results as you are – so please drop me a line to dk8ok [at] gmx.net. I like to encourage you to try all ALE decoders you have at hand – the more, the better.

Already the first results were surprising. This concerned both, the decoding ability of the decoders and the repeatability of the test. So far, the following decoders had participated: go2monitor, Krypto500, MARS-ALE, MultiPSK, Sorcerer, and W-Code. Steve, N2CKH, had written some valuable hints to optimize his MARS-ALE software for SIGINT purposes – please see his comment.


This WAV file contains the calls of thirteen ALE stations. Download and save this file (point to the icon, press right mouse button …). Then feed it to your decoders. Copy the results and send them to me. Have fun!

Ghosts in the Air Glow: HAARP on March 26th, 2019

Just after the spring equinox, interdisciplinary artist Amanda Dawn Christie did another performance of her ionospheric transmission art project “Ghosts in the Air Glow” via the High Frequency Active Auroral Project HAARP near Gakona/Alaska. I took an HF recording of a range, covering all frequencies and times scheduled – see here. At my location, on March 26th, 2019, reception was possible only on 5.100 kHz (best), 6.900 kHz, 7.900 kHz and 8.000 kHz. Signal strength was too low to hear any modulation, but the characteristics of the signals did exactly match the schedule – see screenshots and captions below.

Receiver: Elad’s FDM-S2, Antenna: Active Dipole MD-300DX (2 x 2.5 m), Software: V3 from Simon Brown


5.100 kHz, 01:16 to 01:26 UTC, West beam, gave the best signal.
The signal on 6.900 kHz on the East beam from 01:16 to 01:26 UTC was considerably lower.
On 7.900 kHz, the signal was transmitted by an electronically rotated beam, one rotation per minute, from 01:03 to 01:09 UTC on the West beam. This is clearly seen on the the Signal window below the spectrogram.
On 7.900 kHz, the signal was transmitted by an electronically rotated beam, two rotations per minute, from 01:09:30 to 01:15:30 UTC on the West beam. This is clearly seen on the the Signal window below the spectrogram.
On 8.000 kHz, the signal was transmitted by an electronically rotated beam, one rotation per minute, from 01:03 to 01:09 UTC on the East beam. This is clearly seen on the the Signal window below the spectrogram.
On 8.000 kHz, the signal was transmitted by an electronically rotated beam, two rotations per minute, from 01:09:30 to 01:15:30 UTC on the East beam. This is clearly seen on the the Signal window below the spectrogram.

Visualizing HF Networks

G-VIFT

A flight of G-VFIT from Atlanta to Heathrow, and its HFDL communications – visualized by GEPHI.

Gephi does it

Each communications has a structure. Visualization reveals this structure. This is also valid for HF communications with its networks with different stations, (even moving) locations, hierarchy …

Recently, I made my first steps with free visualization software GEPHI to get a deeper look into some aeronautical networks. The graph at the top shows the gephi’ed result of 125+ HFDL messages, transmitted by a flight of G-VFIT from Atlanta to Heathrow. I monitored six HFDL channels in parallel, one channel from New York, two from Shannon, and three from Reykjavik.

Each point represents a message, tagged by its time in UTC. The positions of the points are geo-referenced, as I used HFDL messages containing these information.

Wheras the longitude’s positions are to scale, their latitude’s positions had been spread for better reading.

This visualization shows that Reykjavik on 6.712 kHz did the main work. But it is surprising that the first contact just leaving the U.S. coast was made with Shannon, and not with nearby New York.

GEPHI also helps in visualizing the hiearchy of networks, see screenshot below:


In the Russian Aero Net on 4.712 kHz, Rostov plays a pivotal role.

Here the strength of the connecting lines and the position of the city’s names represents to hierarchy of this network, i.e. who is calling whom, and how often. There are several strongholds like Rostov, Yekaterinburg and Samara, but also some mere outposts like Novosibirsk and Syktyvkar.

This picture isn’t geo-, but social-referenced, to say so: you know that e.g. Chelyabinsk on the left is geographically situated east of Rostov. You may also geo-reference these data, turn it into a kml file and see it in Google Maps of Open Street Map. If the co-ordinates are correct, zooming will take you exactly to the feeding point of each of their Nadenenko dipole, see below …

The above visualized hierarchy has been changed here into a geo-referenced kml file, opened in Google Earth.

There are many more applications of such a great tool for visualization which will further enrich monitoring.

GMDSS & Display Launcher: Monitoring seven Channels in parallel

DiplayLauncher_4

GMDSS-Display reading decoded data streams from seven MultiPSK’s instances in parallel, presenting all information neatly in one database.

GMDSS is a system of ship-coast and coast-ship digital communications on six main HF channels. At an average location in Germany, you will receive about 5000 messages altogether during 24 hours.

In the past, I mostly used the excellent and free YaDD software to decode all channels in parallel (yes, YaDD can be opened in multi instances, each one in a separate folder).

During HFDL monitoring, I came across Mike Simpson’s free software Display Launcher which neatly collects now up to 24 different data streams, coming from up to 24 HFDL channels in a clear database format.
Mike’s software also contains a module called “GMDSS-Display” which now works similar in collecting datastream from up to seven GMDSS data streams, decoded by MultiPSK software.

Yes, also MultiPSK can be opened in many instances, each one in a sperate folder. By this way, it accepts e.g. the audio input of seven different GMDSS channels from an SDR via each different VACs, and decodes each of them.
To do so, the decoded data of each MultiSPK instance has to be backed up regularly:
Configuration -> Regular back-up -> 20 sec
Then, decoded data is automatically written into the appropriate QSO.txt file. This, in turn, is read by GMDSS-Display. Of course you first have to set the paths to guide the software to the appropriate sources.

It takes a bit time of setting it all up, but then you may run this combination until a Windows’ update forces the PC to re-boot 😉

With Mike’s development, you have a unique and mighty tool at hand for a 360° view now also in the field of GMDSS – thank you very much!

Please find below the results of a 24 hours’ session on all seven GMDSS HF channels – coast stations only, automatically drawn onto DX Atlas. All stations received in Germany with SDR FDM-S2 and MD300DX, an active vertical Megadipole of just 2 x 2.5 m of stunning performance.

DXAtlas_5

Received coastal stations on all GMDSS channel/HF during 24 hours in Germany world-wide and …

DXAtlas_6

… those with a focus onto Europe.

TDoA Direction Finding: First Experiences on the KiwiSDR Net

6465_5

With some iteration, as described in the PDF, the former unknown site of a CIS-12 transmission on 6.465 kHz has been disclosed as the Russian Navy from Baltysk, Kaliningrad.

The stunning direction finding tool on the KiwiSDR net has hit the community. Most people are enthusiastic about the new horizons, some some smart people had opened for free.

A few people, however, reported some disappointment as they couldn’t pinpoint each and every transmitter with expected high precision.

To avoid this disappointment, you have to know what you are doing. The TDoA tool for direction finding indeed delivers automatically stunning results. But you have to think a bit about the setup, and also do some iteration.

I wrapped up my first experiences with TDoA in this PDF. You may simply download it by double-clicking the link, and open it in a PDF reader. It consists of 22 pages and 37 instructive figures. I greatly stressed the practical part of direction finding with this tool – with 13 explicit case studies from 2,6 MHz to 15,6 MHz.

The idea is to have more fun by getting the most reliable results.

TDoA on KiwSDR Net: Direction Finding for All!

TDoA7_Saissac

Four receivers are nailing Saissac as transmitter site of STANAG 4285 transmission on 12.666,5 kHz on July 16th, 2018, 15:45 UTC.

A dream has become true: direction finding (DF) in the range of 0 to 30 MHz for all! The method is “time difference of arrival”. This is a speaking term, and it works exactly like it says: measuring the different times of arrival of a signal at at least three different receivers and calculating the position of the transmitter.

Base is the net of KiwiSDRs. If connected to GPS, each of this SDRs is able to deliver an I/Q stream with time stamps of a very high resolution. If you compare the recordings of different receivers tuned to the same station at the same time, you will find as slight time difference from receiver to receiver. This is due to the “time of flight” from transmitter to receiver. HF is traveling 300 km per 1 millisecond – so time is distance.

However, you don’t know the absolute transmitting time on transmitters other than time signal stations. So “Time Difference on Arrival”, or TDoA, measures the time differences of arrival. The first step is to compare the I/Q streams to find at which time difference their content does correlate. From these time differences, there are calculated curves onto a map. They cross in that region, where most likely the transmitter is located – see the screenshot at the top which I did to locate the STANAG 4285 transmitter on 12.666,5 kHz.

Due to concept, most reliable results (and, hence: sharp regions where the curves meet) are achieved only with the same propagation mode prevailing at all receiving stations. This is most strictly the case on VLF and long wave, where we mostly have one and stable waveguide-like propagation. But you can achieve also stunning results on HF if you carefully choose the SDRs – see next picture.

Isso_2a

Here the broadcast station at Issoudun on 15.320 kHz has been pinpointed by G0EVX, OE5EAN and OZ1BFM receivers at 08:35 UTC. The tip of the arrow points to the actual location, just one kilometer south of the RDoA result [see scale!].


As soon as we approach skywave propagation, you must care for more or less the same propagation mode, of which a one-hop propagation (e.g. 1 x F2) should be preferred. Often even a bit bigger heat map of correlation gives a significant clue from where the transmitter is operating  see next picture.

TDoA8_RussianOFDM

An even weak and fading CIS12 signal 11.836 kHz (spectrogram at the bottom) has been reliably located on Crimea, Sevastopol area. Hence, it is most likely associated with the Russian Navy there.

The software has been developed by some smart people around Christoph Mayer who also provides detailed information on the concept of this approach on his website. It quickly has become an extension of the KiwiSDR net. This makes it very comfortable to use – if you take care of what is said above.

 

SDR Console V3: New and indispensable Software

V3_Dimtsi

“Living Sonagram”: On the right window, you see a part of a 24 h recording at 6,1 MHz bandwidth (ca. 2 TB) with 1 line/second. Tagged is the sign on of Dimtsi Hafash which is received by the undocked “Receive” panel of V3’s GUI. At the bottom: signal strength on 7180 kHz over 24 hours reveals e.g. s/on, s/off and fade in.

Just a small note on a real real big event: Simon Brown, G4ELI, has published V3 of his indispensable SDR Console software on June 18th, 2018 – after three and a half years of heavy coding. Download it here and donate. Or vice versa.

V3 is a quite universal software for most SDRs on the market. For all, it provides the same graphical user interface (GUI) and the same functions (plus those specific to some devices).

All

DXer’s delight: On top the sonagram to visually catch signals (here: JDG from Diego Garcia; tagged). Bottom, from left to right: receive GUI for fine tuning, decoder W-Code showing “JDG”, below this “Playback” panel for controlling the recording (back/forward, e.g.), and on the right a database.

There are many unique functions and modules which will take DXing with SDRs to the next step. For now, let me mention just two of them:

  • 24 parallel demodulators within the SDR’s bandwidth – fully independent in e.g. mode, bandwidth and AGC to receive, record and decode 24 signals/channels in parallel.
  • a sophisticated File Analyser  which presents a recorded band as “living sonagram” – whre you see and click to a signal which then is played via the basic GUI

6pane

Up to six parallel demodulators can be seen on the main screen (from up to 24 possible).

 

1520

1520 kHz from 18:00 to 05:00 UTC (local SR/SS: 19:43/02:58 UTC) with 100 Hz bandwidth and 0,0031 Hz resolution (= +65 dB over 10 kHz!) reveals at least 27 stations and their offsets.

Each of these just two features mentioned will open new worlds for DXing and even serious professional monitoring. I will be happy to come back to some applications of V3 in more detail.

Thank you very much, Simon, for providing this excellent tool for free!

4800

4’800 kHz: First CNR1 with sign on at 20:15 UTC and fade out, then AIR Hyderabad with the same, but s/on around 00:06 UTC.

 

7435kHz

You may export levels over time on one frequency or level over frequencies at a given time. This graph visualizes the activity on 7435 kHz with 86’400 levels (on per second over 24 h). The data had been exported to QtiPlot for further investigation.

« Older Entries