[Lowfer] PSK Update

[John Andrews]

NC, VD and TAG continue to play with the PSK02 found in Spectrum Laboratory.
As of this writing, VD and TAG can copy each other solid 24/7. NC has copied
fragments of TAG, and VD has copied one fragment of NC, all done at night.

Some reality checking:

1. These PSK modes are slow speed variants of PSK31, which was designed for
HF use. PSK31 has two primary virtues: the length of the code for each
letter is shorter for more frequently used letters, and the audio normally
is shaped to occupy a minimum of bandwidth. The audio shaping goes out the
window when we use our typical high-efficiency lowfer transmitters. The
resulting wide bandwidth isn't much of a problem at the 1 watt power level.
The variable character length is an asset for speedier communication, but a
liability for the decoding software, as it can't count on 7 or 8 bit
characters.

2. PSK31 is easy to tune, and the needed frequency stability is well within
the means of modern HF gear. For PSK02, you should determine your actual
sound card sampling rate, and know the actual audio frequency coming from
your receiver. You would be very lucky if you could just start PSK02 after
downloading SpecLab and decode much of the signal.

3. BPSK is a very viable mode for LF. WOLF is BPSK with a lot of error
correction. VE2IQ's DOS-based BPSK programs allow a variety of error
correction levels, including the GRAB function that will "build up" copy
like WOLF does. PSK31, 08 and 02 do not have this capability, and are just
chugging out and decoding the characters without error correction. For that
reason, don't expect WOLF or AFRICAM performance from these modes. Jay and I
are getting nice 24/7 copy, but we can do regular-speed CW during the
daytime, so the signals are pretty good.

All of that being said, these PSK variants could provide some good local or
"regional" communication. This weekend, I will try to work up a simple
calibration procedure using Loran signals. Stay tuned.

John Andrews