Joe, Yes, I have determined that the residual error is due to clock frequency error in my sound device.
I see what you mean --- it would make more sense to accept separate calibration parameters for the input and output devices. But I also understand that most people won’t care about this level of accuracy, so it’s probably not worth doing at this point. There are bigger fish to fry. I have addressed the problem by switching to a different audio output device (EMU 0202) that has a smaller clock error that is well under 50 millihertz and therefore too small to detect with the 0.1 Hz resolution of the WSPR decoder. This is providing me with some motivation to shell out for a Hermes or Angelia board so that I can get rid of this sound-clock error. It also makes me want to add a precise frequency estimation routine to the decoder so that I can easily measure errors smaller than 0.1 Hz using “real” signals. I’ll look at the hopping code some more. I still like the idea of using a version of your tx-table to control all transmissions. This is the only way that I can think of that will enable us to (via iterative refinement) to come up with a table (like your tx(10,6) table) that (nearly) satisfies all constraints (tx_percentage, number of transmissions per band per hour, and max number of successive transmissions). Steve k9an > On May 29, 2015, at 7:44 PM, Joe Taylor <j...@princeton.edu> wrote: > > Steve -- > > One more thing that I forgot to mention. So far I'm not enthusiastic > about making WSPR's Tx frequency spinner a double instead of an int. We > could do that, of course; but if we really need to do something about > correcting a fractional-Hz error there might be better ways. > > We could, for example, add another input field to the Frequency > Calibration group on the Settings | Frequencies tab. > > Have you actually established that the error you observe is the result > of an incorrect sampling rate in your audio output device? > > -- Joe > > ------------------------------------------------------------------------------ > _______________________________________________ > wsjt-devel mailing list > wsjt-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/wsjt-devel ------------------------------------------------------------------------------ _______________________________________________ wsjt-devel mailing list wsjt-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/wsjt-devel
