Algunos detalles mas: Hi James, Rik, Alberto and the group,
Yes an impressive distance.. even though the ERP was a bit above the usual value for amateur radio on 13 cm. AMSAT-DL has applied for, and been granted, a special permission for this experiment which expires at the end of this year. The *principle* was simple: - send a 5 minute carrier on 2.45x GHz, then - receive for 5 minutes, remove impulse noise (WLAN), run the signal through an FFT with ~11 Hz equivalent receiver bandwidth - add all received power spectra to calculate an average (*) - repeat the above until the sigma (standard deviation) in the frequency bins gets low enough to see the weak reflection in one of those bins. - while transmitting, keep the frequency constant to one or two Hz (not as easy as it sounds) - while receiving, permanently compensate the Doppler shift (which a marvellous piece of software by G3RUH did) - during the experiment, keep the dish pointed to Venus with an accuracy better than 0.1° (which another piece of software by G3RUH did) - hope that the 5-kW magnetron, and the cooling system (including a vaccum cleaner and a water cooler from a VW Polo) holds.. (*) A single FFT is not sufficient to lift the signal out of the noise. It was between 8 and 11 dB below the noise in a 10 Hz bandwidth; the ratio of <signal plus noise> to <noise> was between 0.3 and 0.4 dB; the latter measured today when it wasn't raining. This was sufficient to "see" the signal after 2 minutes, so we didn't have to average over consecutive 5-minute-receive intervals. Each dot of the "HI"-message lasted 5 minutes, because that's the approximate travelling time... in other words, "QRSS300" dictated by Venus ;-) Details, including the "Magnetron taming" circuit by DJ4ZC will appear in English language on www.amsat-dl.org hopefully soon. This vital part of the system includes 3 PLL circuits, a servo motor controller for the coarse frequency control (slow mechanical tuning, 6 MHz, on the output waveguide), and a fast controlling loop with about 1.5 MHz (iirc) loop bandwidth which involves 2 * 4CX1000 for the 'faster' frequency control, which also provides the SSB modulation. In fact, the energy not "used" for the SSB signal at a given instance is 'moved away' (spread out) by 15 kHz to both sides of the center frequency. Maybe some of these principles can be used in class-D amplifiers for other frequencies, which finally takes us back to LF / MF-related topics <g> . Cheers, Wolf DL4YHF .
