Hello Rick, >Note that the bandwidth is perhaps five times wider for 2 1/2 times more >throughput. Normally it would be possible to have a 2000 Hz ALE at 250 bauds instead of 125 bauds. For, I suppose, a reason of frequency diversity, the shift between two adjacent tones in ALE is twice the baudrate (when the minimum between two adjacent tones would be a shift in Hz equal to the baudrate).
However for big transfer speeds, there are the choice between several solutions, for example: * something close to the 110A solution: one carrier modulated at 2400 bauds 8ary PSK and different configurations (but with a necessary equalizer and some regular known data to supply the equalizer), * or many carriers modulated in BPSK (or QPSK) as with MT63, but a weak crest factor and no need for an equalizer. 73 Patrick ----- Original Message ----- From: Rick W. To: digitalradio@yahoogroups.com Sent: Friday, June 06, 2008 8:48 PM Subject: Re: [digitalradio] ALE 400 auto speed change With the ALE/FAE modes, the basic ALE 8FSK waveform is used, only slowed down in baud speed from 125 to 50, with the improved sensitivity but reduced throughput, but of course, a much narrower ham friendly bandwidth conserving mode when in ALE/FAE400 vs ALE/FAE 2000. Note that the bandwidth is perhaps five times wider for 2 1/2 times more throughput. This strikes me as one of the best features of Pactor 2 since it always stays close to a 500 Hz footprint, with a constant 100 baud rate, even when it moves to the higher speed constellations from DPSK and through QPSK, 8-PSK, and 16-PSK. Pactor 3 can expand and contract from over 2000 Hz down to around 1000 Hz when it drops to only 2 tones in its most robust mode. It only uses DPSK and QPSK and no higher constellations. At the slowest speed, it may only exceed P2 capability a modest amount (due to the wider spacing of the two tones), but has a net raw speed of only 77 bps. I would expect other modes to perform close to that with similar tone spacing. If you have larger set of tones, such as 8FSK, I wonder if you can adapt as much in terms of baud rate speed changes vs. keeping tone numbers smaller but with higher constellations? And then the bigger effect of P3 which can completely drop tones as needed for more robustness instead of speed. Dr. Rink (SCS) has said that PSK modes do require slightly less S/N ratio over FSK modes and perhaps with always on FEC coding may be a wise choice of modulation using two tones that are modulated with varying constellations? Most of the new soundcard modes have quite a few tones in them, at least 8 or more, and maybe reducing the number of tones might be more ham friendly and still have good throughput? Two tones effectively doubles the throughput compared with single tone PSK mode and yet allows for a relatively low crest factor in concentrating more energy into each of the tones rather than spreading them very thin across many tones. Pactor 3 is very similar to P2, but has the nearly five times wider footprint without being able to operate 5 times faster. I am not sure how often P2 can reach the highest speed level of around 700 bps plus compression compared to how often P3 can reach its highest speed level SL-6 at 2722 bps plus compression, but I suspect that P2 can generally outperform P3 when you take the bandwidth into consideration. But if we only need to change the baud rate of the 8FSK signal, we can do that now with switching between 8FSK50 and 8FSK125 with the attendant problem of drastically widening the footprint and the increased difficulty of finding a clear area to transmit. 73, Rick, KV9U So Andrew O'Brien wrote: > > Would it be possible for ALE 400 in multipsk to use the SNR > measurement of the slave station and signal the master to switch to a > slower/faster speed when indicated ? > >
