The author was prompted to work on this aspect of amateur radio after reading an article by Ray Scrivens, G3LNM. Download Reflections.pdf (3.7MB). The work has been rendered more timely by an article in the March 2006 issue of RadCom by Walter Blanchard, G3JKV.
If you observe a suitable VHF beacon on a waterfall display using a suitable program such as Spectran you are likely to see additional 'versions' of the main signal but different in frequency. The additional 'versions' are due to aircraft reflections and the difference in frequency is due to the Doppler Effect causing the signal to be observed higher or lower in frequency depending on the movement of the aircraft relative to the observer. A good illustration is a picture taken by Ray which shows the signal from GB3ANG in Dundee as observed in Cantley near Norwich, a distance of 330 miles. The direct signal may be seen and is much weaker than the reflection which traverses it moving from lower to higher in frequency.
Another example shows a reflection from GB3SSS which beams west from Poldhu in Cornwall. This was taken in Truro by G3PPT and, not surprisingly, the reflection is much weaker than the main signal. Cornwall sits under one of the main air routes into the UK and reflections from this beacon are very frequent.
Practical observation of aircraft reflections show them to be relatively weak since the reflection process is very lossy. They can last for several minutes and can shift in frequency by tens of Hz in the duration of the reflection. In general, the reflections are reasonably 'clean' and free from the distortions that can make the use of digital modes difficult on HF especially at night.
The use of any of the current digital modes to communicate via these reflections is fraught with difficulty. Most of the modes have an inbuilt frequency tracking capability but nowhere near the wide frequency shift encountered with many reflections. The sensitivity of digital modes is very high but not high enough to cope with all but the peak of most reflections. Lastly, multiple reflections will cause havoc with modes that use a number of tones.
A formidable problem indeed. However there is one digital mode that is robust and requires no synchronization whatsoever and that is Hellschreiber which is basically a form of FAX and was invented by Dr Rudolf Hell in about 1930. In its classic form it runs at about 2 characters/second and has been successfully adapted to run on PC's with soundcard. A good source of information on this and other so called 'Fuzzy Modes' is Murray Greenman's site:
At the conventional speed, Hellschreiber is not sufficiently sensitive for our purpose, but in 1999 the author developed a very slow form of the method for use where QRSS CW is used. This was called SlowFeld and was, at some three characters per minute, very slow but the sensitivity (readable at a level approaching -30dB below AWGN) adequate for this application. Instead of using conventional digital filters, the method plotted bins from the waterfall display of a Fast Fourier Transform (FFT). To get over the problem of tuning in to just one bin, a number of Hellschreiber lines was plotted in parallel with the desired signal appearing on at least one line and, if the signal drifted, the signal would move to an adjacent line. The original SlowFeld was written under Windows95 in the very early days of PC modes using the sound card and had ceased to function under modern versions of Windows. A complete rewrite of Slowfeld was long overdue and with the aid of Borland C++ Builder V6 was duly carried out.
Slowfeld was first tried on the path between G3SMW in Marlow and G3LNM in Cantley. G3SMW transmitted Slowfeld at 3 characters/minute on 2m using 30 Watts to a halo. For reception, G3LNM used an FT857 and the antenna was a 4 ekement J Beam Quad horizontally polarised.
The initial results are shown below. Individual characters can be clearly seen in different lines but despite valiant efforts it is not possible to manually track the signal with the mouse. Note the multiple reflections on the waterfall screen.
Then came the breakthrough. A modified version of SlowFeld was written whereby instead of attempting to track the moving signals, the passband was widened a little and then the strongest FFT bin in the passband chosen. For signals moving greatly in frequency the operator can manually reposition the passband to accomodate the shift. A recording of G3SMW's signals of 24-02-2006 was played through the new program and the results were a revelation. See below.
Two-way trials then took place between G3LNM and PA0OCD with signals being received both ways. Note that PA0OCD was using 400 Watts since this power was available but the power was reduced to 100 Watts in later trials.
During the trials G3LNM and PA0OCD spent some time beaming towards G3PPT in Cornwall whose reception facilities are a trifle basic consisting of a loft mounted 7 element ZL special feeding an FT-817. Nothing was received from PA0OCD and only the odd character was received from G3LNM as shown below:
This was disappointing but perhaps not that surprising if you look at a map of the locations of the participating stations:
The path from mid-Cornwall beaming ENE crosses the high ground of Dartmoor. As ranges increase, the time that a given aircraft is in a suitable position will decrease. Many of the aircraft will be descending to or ascending from airports, also aircraft on short-haul flights will use lower altitudes. On the other hand, the plots obtained by G3LNM from GB3ANG over a path that passes over the sea for a considerable length indicate that ranges of 300 miles plus should be possible there. The aviation sky over the UK and N. Europe is a busy place as evidenced by the following picture:
It took only slight modifications to produce a version of SlowFeld dedicated to aircraft scatter use, which was called SlowFeldXPAS. This program has 3 speeds: 3 character/minute, which was the speed used for the initial trials, and then 6 and 12 character/minute. Increasing the speed inherently reduces the sensitivity and it remains to be seen in practice if the faster speeds can be used to get more information through at the peak of a reflection or whether 'slow but more sure' is the better approach. Much will depend on the path and the station capabilities at each end of the link.
SlowFeldXPAS is available for download at:
Note that the zip file contains library files some or all of which may be required to allow the program to run under Windows versions earlier than XP. These should reside in the same directory as the program.
There are two .mp3 files of 3 char/min signals recorded off-air by G3LNM available for download. These may be played through the program to give a feel for how the system works.
Since starting this work it has come to our attention that a group of Australian amateurs have been working on a different approach to the subject in that they are deploying conventional digital modes to transfer data at a higher speed when the strength of the reflection is sufficiently high and enjoying some success. In our experience for northern Europe this has some merit but tends to fail when there are multiple reflections due to a multiplicity of aircraft in our crowded airspace. See:
Lionel Sear, G3PPT, March 2006.
With much help and support from:
Bernard Spencer, G3SMW.
Ray Scrivens, G3LNM.
Paulus W. Straks, PA0OCD.
Ko Versteeg, NL9222.
THIS SOFTWARE IS PROVIDED ''AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR AND CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Namings for products in the software and this manual, that are registered trademarks, are not separately marked. The same applies to copyrighted material. Therefore the missing (r) or (c) character does not implicate, that the naming is a free trade name. Furthermore the used names do not indicate patent rights or anything similar.