|
Hi Joe,
very good. Yes I guess that from your ~10^33
emitted photons, some 10^23 may have found their way into my RX
antenna ;-)
Regarding Opera, coherence is not an
issue for the standard decoder which basically decides each symbol
individually. But with correlation detection, it seems to
improve sensitivity by about 6 or 7 dB compared
to random-phase or chirpy dashes. On the other hand, in winter
conditions your signal had been strong enough for easy incoherent
decodes, but the extra dBs may be valuable now to overcome summertime QRN.
No need to worry about DDS rounding errors as the exact absolute frequency is
not relevant for detection.
The Manchester-coded Op-32 sequence consists
of 239 half-symbols (ie. on or off), each nominally 8.192 seconds
long, making up 32 minutes and 38 seconds total (don't be tempted
to round down to 8.0 seconds per dash as the accumulated timing error would
become much too large). According to the coding
scheme uncovered in the PE1NNZ paper, the sequence for VO1NA should
be
11011010100101010110100101100110100101011001100110101010101010100110011001100101100101101001011001011010011010101010101010101010100110101010101010010101011010010110101010011001100110011010101010101001011001010110010110010110100101101001101
It's all up to your preference whether you'd
generate that by a computer, a microcontroller, or even a mechanical
device based on conducting pencil marks on a paper strip ;-) As mentioned by
Graham, you could either use EA5HVK's software directly to generate keying,
or extract the sequence and program it into a PIC. For my own
transmissions, I have converted my 1010 sequence to a
16-bit IQ file, which is played out by DL4YHF's SndOutpt tool,
slowed down by a properly selected decimation factor. The audio
from the PC soundcard then activates a simple VOX keying
circuit.
Best 73,
Markus (DF6NM)
----- Original Message -----
Sent: Wednesday, June 19, 2013 8:57
PM
Subject: Re: LF: VO1NA 137777.0000
Hi Markus and group,
Thank-you very much for
the progress report and critique of the many
2200m photons that have been
eminating from my back yard over the past
week.
The AD9851 is clocked
by a DOCXO that's calibrated against WWV to within
0.1 Hz. Initially, there
were rounding errors and an error due to the
resolution (~0.01 Hz).
The former was eliminated using higher precision
arithmetic the latter
by setting the dds to 16x and dividing the
output enroute to keyer the
Decca. It isn't anything fancy or exotic and
perhaps a just bit of
good luck.
So it seems the photons are now last reasonably stable,
accurate,
monochromatic and exhibit phase coherence. The next steps
are to
make them frequency agile (137.555 kHz) and to key them in a
manner
that preserves these properties. Perhaps I can code the OP32
directly
in the microprocessor? How long would the minumim message be
or would
I be better off using a computer to take care of the
keying?
73 to all
Joe VO1NA
On Tue, 18 Jun 2013, Markus
Vester wrote:
> Hi Joe,
>
> it looks like you've moved
your QRG to an exact integer Hz a couple of nights ago. Just curious how you
managed to circumvent the 2^-32 DDS clock raster - did you just tweak the
reference frequency, or employ some sophisticated fractional-N frequency
toggling scheme?
>
> And how about serving some freshly cooked,
perfectly coherent Opera-32 to the world tonight? The famous international
reception commitee is awaiting your signal ;-)
>
> Best 73,
>
Markus (DF6NM)
|
