Tony wrote: > Patrick, > > I get the same minimum SNR for Contestia but can squeeze -8db out of > MT63 when using DM780 and IZ8BLY.
Yesterday I had no luck with DM780 while monitoring Tony's QSO's on 14106. Of course, I have not calibrated DM780, so that is no surprise. Propagation was not good, but MultiPSK did fairly well. > The threshold difference shows on-air as well as under controlled > conditions and so it would seem that the best way to get the most out of > MT63 is to use software that decodes deeper into the noise. No doubt... > The 10-to-1 peak-to-average power ratio is an excellent point and > it's obvious that Contestia will put more RF into the air on average. > There's no doubt in my mind that the Contestia 16-1K will do better most > of the time. I have doubts on this department. If the peak amplitudes are the same, as may be happening with the audio tests, the decoding on those conditions should be equally valid. Of course, Vojtech points out that MT63 is more sensitive to distortions which are pretty common with some not so careful operators. Fast attack slow decay ALC could have a chance of correcting some of the overloads after the transmit IF, but the rest of the chain, from the audio input to the IF should receive proper audio levels. I have seen rebel cases of distorted PSK-31 when people closes the mic gain and distortion remains... because the early stages of the transceiver are already overloaded, and I had a real bad luck when explaining that to the other operator. People should know their way around... > On the other hand, it does not seem to recover from the complete > drop-outs that occur during deep fading or with lightning static the way > MT63 does. I was browsing my references this afternoon (local) and I I decided not to send a reply, since it seemed that Vojtech had a good point and was not worth arguing about it. Nevertheless, I wonder how the degradating effect of -30 to -20 dB IMD, the usually accepted values when adding that to the channel noise. Even more when I read that Contestia was devised with a flat envelope on mind (nonlinearity does not affect it) and uses about the same Walsh-Hadamard code. But it _might_ mean, conversely, that Contestia is more power greedy, an important consideration for emergency operation. > I've tested this theory by removing short 1-to-3 second segments of the > signal at random intervals and the mode continues to print despite the > missing 'chunks'. As you say, this could be due to the difference in > modulation speeds. Is there an alternative mode that I can test that > might have similar characteristics? I did not find any details on my references, but seemingly interleaving is done both in the time and frequency domains, so there is more chance for MT63 to get the message thru, specially with long interleave. On the other hand, if someone "pulls the carpet" (heavy doppler) there is a risk that MT63 will fail strepitously with bits falling on the wrong bins while a mode with less, wider frequency "bins" (like Contestia or Olivia) will really shine. I had really a low esteem for MT-63, but it had been hard to make a MT-63 QSO before Tony started the present tests campaign. It just happens that each mode should be used according to its most promiment strengths. I still have low steem for Chip-64... 8-) 73, Jose, CO2JA