Hello Joe and all, > My attempts to get Linrad to lock onto a JT65 sync signal > have not been encouraging, so far. My impression is that it sometimes > gets confused by some data tones, which may be no more than a few Hz > from the sync tone, or by the gaps in the sync tone -- which can be up > to almost 3 seconds. For wsjt one has to set the fft1 (or fft2 if present) bandwidth to well below 5 Hz. I think the optimum is somewhere between 0.2 Hz and 1 Hz on 144 MHz. The averaging parameter in the AFC should be set to something that does not degrade the line width of the sync tone and therefore it may be optimum to set the average to 1 only. The number of transforms to fit a straight line to should correspond to a time of at least 15 seconds because of the characteristics of the EME path. A 3 second gap in the sync tone is VERY short in comparison.
I have made very few experiments because I do not have suitable test files. It would be interesting to try some cases where wsjt fails because of too unstable frequency near the detect limit and also at somewhat better S/N. > Linrad tries to lock onto the carrier phase. WSJT uses only frequency > locking. No, linrad just locks to the frequency. With typical parameters for cw at the detect limit (not much different from wsjt) the locking is accurate enough to provide a constant phase over pretty long times. > My experience so far seems to suggest that is the signal is > strong enough to make phase-locking over an appreciable portion of a > 46.8 s transmission, it is strong enough that you don't need to lock > phase. The WSJT decoders assume nothing about signal phase, except that > it not wander appreciably during a single tone interval, 0.372 s. My very limited experience points in another direction. For signals with reasonable stability the line width of the sync tone should be no worse than the line width of a good CW signal and that is about 0.2Hz. It means that the phase is constant over about 5 seconds provided that the optimum frequency and low order derivatives of it have been evaluated and that the phase relations between several sync tone intervals contribute to make the very narrow spectral line just as I can see happen in CW mode between dots and dashes. Now, this is another problem. I actually believe some advantage could be gained from using the phase information that one can get from the sync tone to decode the information carrying tones even at levels well below the current wsjt detect threshold but it is a quite different question whether the Linrad AFC is capable of compressing the bandwidth of a signal that is just a little too unstable for the wsjt AFC at signal levels where a detect should otherwise have been OK. > > There are two different ways of using Linrad for wsjt that I > > think would be optimal - but in different situations. > > > > 1) Extremely weak signals that have good frequency stability: > > ------------------------------------------------------------- > > In this case it will be a good idea to disable the AGC > > Do you mean disable the AFC? Sorry, of course AFC!!! (Never use AGC for any weak signal work) > > 2) Unstable signals that can currently not be decoded with wsjt. > > --------------------------------------------------------------- > > > > The AFC can be set to compress the bandwidth of wsjt signals down > > to very low S/N. The price to pay is a very long time delay. > > Something like 10 to 20 seconds. One would have to respond to > > the previous transmission, not the most recent one. I do not > > want to go into details since I have not verified this on signals > > that actually fail with the wsjt program. > > I do not see why there needs to be a large time delay. It seems to me > that this case is best addressed by doing the AFC after breaking the > received signal into a one-minute chunk. Ooooh! In CW mode - and that is all there is right now - there has to be a large delay. > AFC and decoding would both be > applied after the full minute (actually 53 s) of data are available, > just as WSJT does it now. The delay could be quite small if the > computer is fast. Yes. Absolutely. Applying the AFC to a 53 s chunk of data and computing the bandwidth copmressed down-sampled data to fit the decoding algorithms would be done in a couple of milliseconds if a "batch" mode were added for JT65 and other modes that use predetermined transmit/receive periods. Before I know for sure that the Linrad AFC would indeed give a significant improvement I do not find it meaningful to start thinking about implementing it for wsjt. I only have access to recordings that do decode well...... 73 leif / SM5BSZ > > -- Joe, K1JT > > ############################################################# > This message is sent to you because you are subscribed to > the mailing list <linrad@antennspecialisten.se>. > To unsubscribe, E-mail to: <[EMAIL PROTECTED]> > To switch to the DIGEST mode, E-mail to <[EMAIL PROTECTED]> > To switch to the INDEX mode, E-mail to <[EMAIL PROTECTED]> > Send administrative queries to <[EMAIL PROTECTED]> > ############################################################# This message is sent to you because you are subscribed to the mailing list <linrad@antennspecialisten.se>. To unsubscribe, E-mail to: <[EMAIL PROTECTED]> To switch to the DIGEST mode, E-mail to <[EMAIL PROTECTED]> To switch to the INDEX mode, E-mail to <[EMAIL PROTECTED]> Send administrative queries to <[EMAIL PROTECTED]>
