Hi Igor,Earlier this month you made suggestions for a possible new protocol for minimal weak-signal QSOs. I have been away on vacation since that time, so have not had a chance to respond.
Of course there are many possible ways to make design trade-offs involving message size, duration of transmissions, sensitivity, bandwidth, undetected error rate, coding and modulation scheme, and so on. In the modes implemented in WSJT, WSPR, and WSJT-X we have made choices that I believe provide close to optimum solutions for a wide variety of different Amateur Radio activities.
By all means, if you think your ideas have merit you should proceed to develop a new mode along the lines you outline. But I should point out that the particular goal your message seems to favor is close to that of the "WSPR QSO Mode" that I introduced in May, 2008. A brief description of this mode was published in the proceedings of the 13th International EME Conference, held in Florence, Italy, later that year. A link to this article may be found as Reference 11 here:
http://physics.princeton.edu/pulsar/k1jt/refs.html"WSPR QSO Mode" reached a sensitivity of -29 dB and the required bandwidth was only 6 Hz, but the mode never became popular. One reason is surely that the supported message types (in 50-bit packets) were too restrictive. Another reason is that the 2-minute T/R sequence length is too long: QSOs were necessarily verrrry sloooooow. Scaling the transmissions to 15 s sequences would make the bandwidth about 47 Hz and sensitivity -20 dB. FT8 has about the same bandwidth, better sensitivity, and a much wider range of supported message content. WSPR QSO MOde was an interesting idea, but FT8 is far better.
-- 73, Joe, K1JT
On 8/8/2018 2:29 AM, Игорь Ч via wsjt-devel wrote:
Hello Joe and all, .We all have been missing JT65 mode sensitivity and proposed WSJT-X 2.0 new FT8 approach with 0.2 dB sensitivity penalty can make things even worse..I would like to ask you to consider a new protocol where callsign hash would be used instead of the real callsign in all messages but CQ and incoming call, this way we can get back to -25..26dB SNR sensitivity although will get more limited with the free message length.. CQ message: 28 bit callsign1 + i5bit + 12 bit CRC = 45 bitincoming call: 10 bit call1 hash + 28 bit callsign2 + i5bit + 12bit CRC = 55 bit report message: 10 bit call2 hash + 10 bit call1 hash + i5bit + (10 bit call3 hash for DXpedition) + 6 bit report + 12 bit CRC = 43(53) bit roger+report message: 10 bit call1 hash + 10 bit call2 hash + 6 bit report+ i5bit + 12bit CRC = 43 bit 73 message: 10 bit call2 hash + 10 bit call1 hash + 15 bit GRID + i5bit + 12bit CRC = 55 bit RR73 message: 10 bit call1 hash + 10 bit call2 hash + 15 bit GRID + i5bit + 12bit CRC= 52 bit.Spare bits can be used for nonstandard(special) callsign transmission in CQ message. call1 hash could be omitted in the incoming call message if this message is originated by the nonstandard(special) callsign..Probably we can optimize protocol even better while a main idea is to transmit a full callsign only once per each QSO and to transmit not more than one full callsign in the message.. 73 Igor UA3DJY
------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ wsjt-devel mailing list wsjt-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/wsjt-devel
