Hello Dave,
Sorry for the setup problems - indeed, the calibration of sample rate
and frequency tuning is essential, and I am still looking for an easier
way to tune. The waterfall (and spectrum graph) will be pretty useless
when it comes to the really weak signals, since WOLF's decoding
potential exceeds the waterfall's weak-signal displaying capability by
far (especially when looking at a BPSK signal). There is a tuning mode
in the program which can produce a thin line on the receiver's sreen. It
must be exactly on the place which is marked with the larger tick in the
middle of the frequency scale (phew, I should mention that in a
manual... but no time yet).
About the crash: What compilation date (and time ! ) does the
"About"-box show ? I recently uploaded a new version in which a bug has
been fixed, which caused a crashed after quickly switching from RX to TX
and back to RX (the latest version was compiled today, April 24, which
was used by DF6NM and DF0WD).
But before downloading that version, wait for a few days until the next
one is out. I will add the option for a waterfall with selectable
frequency solution, which would simplify tuning if another station (with
precise TX) transmits the "Tune Tone" for a couple of minutes.
About the same message being decoded more than once, even if there is no
signal: Strong signals "accumulate" (or integrate) in a memory, causing
the same message to be displayed more than once. Since the message text
is spread all over the 960 bit message frame (which takes 96 seconds to
send at 10 bit/sec), the coding scheme makes it possible to decode a
strong signal which was only 24 seconds long.
To try this "off-line" with a reasonably fast PC (say >= 300 MHz), you
can let up to three instances of the WOLF GUI on the same machine, each
of them using its own configuration. With an AC97 audio system (which
most PCs use these days), let one instance transmit, and the other
receive. If your soundcard plays an ugly game with you, it uses
*different* sampling rates for in- and output (this can happen, but
fortunately not too often). If it happens, enter different sample rates
in the "Transmit" and "Receive" fields on the config tab. A digital
frequency counter connected to the soundcard output, and the "Tune Tone"
mode will quickly reveal such problems. After calibrating the sampling
rate (until 800.00 Hz get out as 800.00 Hz), you have a stable generator
to test the input ("Receive") sampling rate.
Note: Some digital frequency counters can be switched into "Period
measurement" mode, which gives a much higher resolution than the
"frequency measurement" mode, even at 10 second gate time.
About the WOLF-test today: Before calling Markus in WOLF BPSK after the
Formula 1 race today, I noticed a signal on the waterfall which I
couldn't decode - was that you Dave ?
Now back to the workbench...
73, Wolf DL4YHF
( Wolf = Wolfgang, that's me;
WOLF = BPSK by KK7KA, that's not me - all the flowers to Stewart
please ! :-)
Dave Pick schrieb:
Marcus and group (especially Wolf)
Thanks for your patience whilst I struggled to get Wolf GUI set up!
I had literally just unzipped it and ran it into the TX and I saw your
signal come back on the waterfall but couldn't make it decode for ages...
When I got the calibration sorted I could then decode fine (but the signal
was audible so it SHOULD have worked!). I had two copies running one on
Win98 and one on Win 2K and they both kept crashing after about 10mins, then
I couldn't make the TX start on the right frequency unless I re-started the
program several times.
Also I had the RX constantly repeating the last message it had received,
which was my TX message from the other computer, even though I was receiving
Marcus on the waterfall... Most odd.
Perhaps I need more practice?!
Thanks Wolf for making Stewart's mode accessible to the dim operators like
myself. It shows a lot of potential but maybe it's not quite ready as a QSO
mode yet? (Or maybe I'm not...?)
73 and thanks both again
Dave G3YXM.
|
