Difference between revisions of "GRIFFIN"

From HPSDRwiki
Jump to: navigation, search
m (changed "...received signal had a 'chip'...." to "...received signal had a 'chirp'....")
(new ip)
Line 11: Line 11:
 
More details of Bill's work can be found here:
 
More details of Bill's work can be found here:
  
   svn://64.245.179.219/svn/repos_sdr_hpsdr/trunk/PennyWSPR  
+
   svn://184.81.170.126/svn/repos_sdr_hpsdr/trunk/PennyWSPR  
  
 
There's a README file at:
 
There's a README file at:
  
   svn://64.245.179.219/svn/repos_sdr_hpsdr/trunk/PennyWSPR/README.txt
+
   svn://184.81.170.126/svn/repos_sdr_hpsdr/trunk/PennyWSPR/README.txt
  
 
One of the advantages we have in the way that HPSDR has evolved is the
 
One of the advantages we have in the way that HPSDR has evolved is the

Revision as of 17:02, 30 December 2010

GRIFFIN - Stand alone beacon board

Block diagram as of 23 Aug 2010. Click for full size.

Many will be aware of the work undertaken by Bill, KD5TFD, to develop FPGA code that enables Penelope to operate as a stand alone WSPR beacon. Not only does his code allow a full implementation of the WSPR protocol, without the need to connect to a PC, it also enables multiple beacons to operate simultaneously on multiple bands. This is the dual of multiple independent receives in a Mercury - multiple independent transmitters in a Penny.

More details of Bill's work can be found here:

 svn://184.81.170.126/svn/repos_sdr_hpsdr/trunk/PennyWSPR 

There's a README file at:

 svn://184.81.170.126/svn/repos_sdr_hpsdr/trunk/PennyWSPR/README.txt

One of the advantages we have in the way that HPSDR has evolved is the ability to try new modes without the need to build or buy new hardware. Here's another example.

Many CW operators will remember the time when a CW report, say 559, also sometimes included a 'C' at the end (e.g. 559C) which indicated the received signal had a 'chirp' i.e. unstable when keyed, FM etc.

Chirp was regarded as an undesirable feature and best to let the operator know. There was also the practice of adding an 'H' to the report indicating Hum. The story goes that the difficulty in obtaining high voltage smoothing capacitors in Eastern Europe resulted in some operators applying un-smoothed rectified mains to the anode/plate of the PA stage giving it a distinctive err... note. However, a story for another day perhaps.

The introduction of digital/HDTV in many parts of the world has triggered the removal of the analogue TV broadcasters that are/were located adjacent to the 6m band. Magic Band operators often use these stations as an indicator of propagation conditions. Since they typically run much higher ERP than Amateur stations, 10's of kW, they were/are an ideal source of propagation beacons.

With the removal of these services in many parts of the world 6m operators loose a very valuable source of propagation information. It's just this problem that has lead Andrew Martin, VK30E, to develop an alternative that can be used in either a RADAR or beacon mode.

Andrew describes his technique fully in the 2/2010 edition of DUBUS

 http://www.marsport.org.uk/dubus/last.htm

but here's the basic idea.

We deliberately, repeatedly, linearly sweep the frequency of a carrier over 1kHz in one second. The nominal frequency of the carrier is derived from a GPS locked 10MHz reference. The start of the one second sweep is synchronised to the 1 PPS signal from a GPS.

At the receiver we use a matched filter, again triggered from a 1 PPS signal from a GPS, to detect the signal. The matched filter in this case runs on your PC and uses your sound card or VAC to digitise the received signal.

The effect is similar to WSPR. However, we trade RF bandwidth for time - in which case the output of the matched filter is available every second rather then every few minutes as in the case of WSPR. We also gain a signal processing advantage over WSPR which typically can decode at -26dB in a 2.7kHz bandwidth. 'Chirp' will decode at -36dB in the same bandwidth. If we integrate multiple chirp signals over say 1 minute then the processing gain increases to -54dB.

So if we run a 100W amateur beacon by applying a chirp to the signal we effectively have a 400kW signal - in line with the TV transmitters we are trying to replace.

Initial tests on 20m between Andrew and Don, VK6HK over a 2700Km path, have shown the technique to work such that a barely audible SSB signal gives a spike with a 45dB S/N ratio at the output of the matched filter.

This is chirp in the beacon mode.

Alternatively, one station can transmit a chirp signal, using full power and a beam, in a certain direction and another station listens beaming in the same direction. The stations need to be sufficiently far apart such that the transmitting station does not overload the receiver of the receiving station.

The receiving station will then receive echoes of the transmitted signal from the various mediums in the direction he is beaming. Since his matched filter is triggered at exactly the same time as the transmitting station he can plot of graph of signal strength against distance in miles/km. Such a graph can clearly show the presence of single and multi hop E's and their intensity. Andrew's DUBUS article has some very interesting examples of enhanced 6m propagation that chirp is able to identify.

The free sound card software Spectrum Lab:

  http://www.qsl.net/dl4yhf/spectra1.html 

will both generate and demodulate chirp signals.

What would be interesting would be to have a single board that generates a chirp modulated carrier and could be used as the replacement for the low level drive stage in a conventional 6m beacon. You guessed it - Penelope!

Bill kindly offered to take on this 'Weekender Project' (which did actually fit in a weekend for once!) and produced a version of his PennyWSPR code that included BOTH a chirp and WSPR beacon and would trigger on the 1 PPS from a GPS receiver. The nominal frequency of the beacons is at 50.3MHz (for initial testing) with the chirp extending from 300Hz to 1300Hz above this and the WSPR beacon 1500Hz above. The WSPR beacon is 6dB below the power level of the chirp beacon.

Since VHF beacons in VK are allocated on a 2kHz channel spacing the spectrum of the two signals fits nicely within an allocated channel.

Work is also in progress to look at including an aural CW ident on the allocated carrier frequency.

Testing is about to start and I'll report progress in due course.

If testing is successful then we intend to develop an FPGA based PCB that will provide a 6m/2m beacon exciter stage. The board is intended to be used as a replacement for the low level driver board in an existing beacon. As well as the FGPA, power supplies etc, it will include a GPS receiver and oven controlled 10MHz reference.

The project leaders for the board was Phil VK6APH and Kevin M0KHZ.