Message-ID: <013a01c11216$25afc020$0700000a@parissn2>
From: "Stewart Nelson" <sn@scgroup.com>
To: rsgb_lf_group@blacksheep.org, "LowFerQTH" <lowfer@qth.net>
Subject: LF: amateur VLF DX?
Date: Sat, 21 Jul 2001 20:51:25 +0200
Organization: SC Group
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Sender: <majordom@post.thorcom.com>

Hi all,

Although the recent amateur VLF contacts were fairly short range, I
believe that DX operation is possible, perhaps even across the pond.

Please don't laugh, but I think that the very high stability of
the path would permit operation at extremely low S/N, albeit at very
low data rates.

We all take for granted that when a signal is too weak, relative to
the noise received with it, communication is impossible.  Why can't
you just send the information more slowly?  Because you quickly reach
a point where variations caused by the propagation path or equipment
instability limit the narrowest bandwidth that can be used.

Consider a QRSS system on 20 meters using an Rx bandwidth of 1 Hz and
a 3 second dot time.  Assume that the received signal is spread over 1
Hz by the propagation path.  We need about +5 dB/Hz S/N for adequate
copy.  What to do if we only have -5 dB/Hz?  Of course, we can use
longer dots.  But, since the path prevents us from narrowing the
bandwidth, the dot time needs to be 300 seconds (with noncoherent
averaging, the S/N only improves with the square root of the averaging
time).  And this could fail, if the path closes before the message is
complete.  But even with a stable path, you are out of luck at, say,
-35 dB/Hz.  The dot time would need to be 300,000,000 seconds (about
ten years).

On LF, it's much better, because the path spread is typically only a
few millihertz.  I'll use 0.01 Hz to keep the numbers simple.  Our 300
seconds/dot will now work down to -15 dB per Hz, thanks to 0.01 Hz
bandwidth.  But below that level, averaging is again needed; -35 dB/Hz
would still require dots about a month long.  If you have an unusually
stable skywave path, and excellent equipment, you could use 0.001 Hz
BW, but you would still need dots lasting a few days.

But at VLF the path is so stable that phase jitter is small, relative
to the carrier period, except for shifts at sunrise and sunset.  I
don't understand the physics to know why this is so, but I'm pretty
sure that it *is* so, because the (now defunct) Omega navigation
system depended on it.  A signal sent from New York to London will
arrive with nearly the same phase delay on Tuesday morning, as it had
on Monday morning.  You could coherently integrate "forever".

I don't have any figures for path loss or noise levels at VLF, so I'll
start from the assumption that the signal from SAQ, using ordinary CW
at about 10 bps, can be fairly reliably received across the Atlantic.

It shouldn't be too hard for an amateur to generate a VLF signal 50 dB
weaker than SAQ.  The PA might be 200 W to 2 kW, compared with SAQ's
200 kW.  An "earth loop" antenna might be 0.001 to 0.01 as efficient
as SAQ's tower array.  Depending on the "gain" of our antenna, we
would use sufficient power to achieve the -50 dB relative level.
Using coherent detection of BPSK or m-ary FSK gives a 6 dB advantage
over OOK, and it should be easy to get another 4 dB from a suitable
low rate ECC.  Our overall system would then be "only" 40 dB too weak.

Now, I believe that with external synchronization, we can make up the
40 dB by simply sending 10,000 times slower!  0.001 bps is not that
bad.  We can send a 15-character message in a day, and complete a QSO
in three days.  We don't have to wait for good propagation, because
it's always about the same.  If there is bad QRN for a while, we can
just ignore that period.

A possible Tx rig would feed the output from a PC sound card, through
a high power audio amplifier and impedance matching device, to the
antenna.  The audio signal would also be fed back into one sound card
input channel, and the 1 PPS signal from a GPS receiver into the
other.  Software would periodically adjust the phase of the Tx signal
to keep it in step with GPS.

The receiver would consist of a tuned loop and preamp feeding one
sound card channel, with the 1 PPS from GPS feeding the other.
The software LO would be kept locked to GPS, and a phase adjustment to
compensate for diurnal variation would be added.  The result would be
used to demodulate the incoming signal.

In the US, I believe that AC below 9 kHz is not considered to be "RF"
at all, so it would be easy for this type of Tx to meet the Part 15
requirements.  The Tx would be classified as an "incidental radiator"
or, if the computer were included, an "unintentional radiator".  Those
radiated emission limits start at 30 MHz; it shouldn't be hard to keep
harmonics above the 3000th within the spec :) .  Of course, the device
must also not cause any harmful interference.

Is there any reason why this won't work?  Are my estimates reasonable?
I'll be back in Reno in about a week and would like to give this a
try.

73,

Stewart KK7KA