Another new direction in radio for me; weak-signal work. Well, I've been involved with low-power radio (QRP) since my return to the amateur scene in 2005, but this is a new departure. I'm currently operating a manned experimental propagation transmitter (MEPT) in the 30m band. There are lots of people doing this now, and the body of knowledge we are corporately and severally building is impressive. The idea is to shift the energy density away from power, and expand the time instead. The net energy expended in communication is identical, but my 50mW signal takes many times longer to transmit than when I operate at 15WPM with a hand key at 5W. To put it into time terms, each dit takes three seconds to transmit.
So, did I make the MEPT, build it from a kit or buy it? There seemed so little to it, so I brewed my own. I've recently got up to speed on how to program and use Microchip PICs, and this was one project where a PIC is ideal. I use the PIC to control the keyline and frequency of the MEPT; four lines from one of the ports feed a simple R2R digital-to-analogue converter, which in turn drives a LED in the oscillator, pulling the crystal LF to give the 'waveform' structure of the signal. My signal is shown above, in the centre of the picture. It alternates between conventional dit-dah Morse code, giving my callsign and locator, and my callsign again in 'slant' CW. This form of encoding is very time-efficient; I can send my callsign in the time most stations take to send a zero in dit-dah Morse. The downward slanted lines are dahs, and the upward one are dits.
The picture above shows the oscillator, housed in a small steel tin. The crystal is a bit of a special; it was originally a 10.15MHz item, but with the aid of Hans Summers' excellent article on crystal penning (do a search for G0UPL), I was able to push it down ten whole kHz with a CD marker. It had to be 'aged', or dried-out after this treatment to drive off the volatiles in the ink but it eventually settled down. It is thermostatically controlled by heating it with a 47R resistor regulated by a simple linear FET / thermistor circuit, and the whole oscillator is enclosed in 20mm of closed-cell foam. The PIC controller, 74HC02 driver / keyer and the VN10KM PA are built up outside in the main enclosure.
I have found that two levels of voltage regulation are required for stability, and in spite of the thermostatic control holding the oscillator at 50 degrees centigrade, the MEPT still drifts slowly up and down during the day. An exercise for the coming weeks.