Hi Mike, Psk125r (and the other "r" psk modes" for that matter) are simply a standard bpsk mode with the following changes:
1. FEC with a rate of 1/2 which means that we send two bits for every bit of data. This redundancy is what provides most of it's extra robustness but of course at the cost of the effective data speed. So psk125r is about (more on this later) the same typing rate as psk63. 2. Convolutional encoder to spread the bits around so that noise has less effect 3. Soft-decoder which takes into account the phase and amplitude of the signal received to decide if it is a "strong" "1" or "0" or a "weak" one. Since we send two bits per data bit, the decoder on the receiver end can than decide which one is of better "quality" and has more chances to be the real thing rather than noise. 4. MFSK varicode for the simple reason that is has for some patterns of characters about 13% speed gain on the standard psk varicode. So in conclusion you may ask why bother with double the bandwidth (and therefore a 3dB power handicap)? The coding gain from all the above is theoretically 5dB and the tests performed in the "lab" show that this is pretty right and therefore +5 -3 = +2dB of advantage in white noise conditions. But also (and probably more importantly) the spreading of the bits in time allows for the impact of noise bursts to be reduced since it is less probable to have two noise bursts at exactly the same time for the first bit and it's redundant counterpart which is sent later on. This set of modes was created mainly for ARQ applications like Pskmail and Flarq as we wanted to close a gap between the MFSK/IFSK modes which are robust but slow (while remaining below 500Hz bandwidth) and the psk125,250 and 500 modes which are fast but can be more easily disturbed by QRM and need good s/n ratios. Remembering that for ARQ applications, one bad bit is most likely one bad frame that needs to be retransmitted. Regarding QPSK modes: they have also FEC but the spreading of bit is more limited and the 90 degree instead of 180 degree spreading of the phase changes make it only marginally better than bpsk in my experience. In some cases they can be worse in fact. Now for your third question: I doubt that this would be a good mode for EME due to it's level of sensitivity. I am no expert in EME but the little I know it that you need to have a mode that has a very low minimum s/n as the path losses are very high. The JT65A mode has a minimum s/n of -23dB versus something around -13 or -14dB for psk125r. Even the slowest psk mode that I know of, PSK10AM (in Patrick's Multipsk) has a -19.5 minimum s/n, so still a few dBs below JT65A. If these king of modes had any chance in EME then it would need to be slowed down even further, plus I don't know what the EME path is like regarding phase distortion which is a major negative for these modes since they rely and a phase change to encode a 1 or 0. Hope this helps. Best 73s, John (VK2ETA) --- In digitalradio@yahoogroups.com, Mike Lebo <mike-l...@...> wrote: > > How does pak-125r work? Does it use the same varicode? Does it have error > correcting code like QPSK-125? How many phases does it use? Would it work > well for EME? > > n6ief >