Hi all,

Jake's post is quite exciting; it would be great to try those techniques.
I'm presently in a very noisy location (Paris), but could possibly
contribute by writing some code to process recordings.

Under ideal conditions, if both Rx locations experienced the same S/N,
the signals could be coherently added, and the noise was uncorrelated,
the combined S/N would be improved by 3 dB.  Perhaps 2 dB is practically
realizable.  While that's nothing to sneeze at, it's not a dramatic
improvement.  LF is not like EME, where 2 dB of additional antenna gain
results in a big increase in the number of stations heard.  A wide range
of ERPs, distances, and propagation conditions contribute to a huge
variation in incoming S/N.  So, many stations can be heard with a single
antenna, but for most of those that can't, an extra 2 dB is not enough.
How often would it help?  Of course, we won't know until we try it, but
here is an attempt at an optimistic estimate.  Let's assume that the
S/N of incoming signals is normally distributed, with a standard deviation
of only 10 dB, and a mean 1 dB below our reception threshold.  In that
case, we could receive about 46% of the signals using one antenna, and
about 54% of the signals using two.  This means that only one in 12
experiments would be successful, if success is defined as decoding a
message which could not have been decoded using either station alone.

I believe that the above effect could be mitigated by making SSB bandwidth
recordings at each station, so that the entire LF band is captured.  Of
course, this requires receiving locations where there are no really strong
local signals to swamp the receiver.  One could then search the recordings
for signals which can be detected, but not decoded.  These would then be
further examined to see if decoding is possible by combining the signals.
If the original recordings had dozens of signals, the probability of a hit
should be pretty high.  If we don't see any, we can try again using three
or more stations.

Now for the good news: IMO it's possible, using suitable software, to
combine LF recordings made at different locations, with no need for GPS
locking or special references.  That's because a superb reference is
already present: LORAN.  Even if individual 'LORAN lines' are quite weak,
there are a couple of hundred such lines in SSB bandwidth, so from most
locations there should be no problem tracking the frequency and phase.
LORAN can also enable precise time alignment of the recordings.  While not
time standards in the normal sense, all North American and NELS chains are
locked to UTC.  This means that the master stations conceptually emitted
their first pulses on 1958/01/01 at 00:00 UTC.  If a recording contains
pulses from a dual rated station, and the time is known within a few
seconds, then precise time stamps with sub millisecond accuracy can be
easily derived.

Bob WA3WDR wrote "If we get a few transmitting stations running, we can also
try phase-synchronized transmitters beaming in a particular direction."
This has the possibility of delivering monstrous signals!  Because not only
do we increase our antenna gain, we are also launching more power into the
air.  Two stations of equal ERP, properly phased, will deliver 6 dB more
to a specified target than either station alone.  If four stations are
used, we can gain up to 12 dB.  I don't think that it is too hard to
achieve the required synchronization, if the transmitters are close enough
to hear each other via a reliable ground wave path.  One transmitter is
designated master.  The slaves stop transmitting for say, half a second
out of every ten.  During that time they measure the apparent phase of the
master with their local receiver, and compare it with the measured phase
of their own signal.  An adjustment is then made to maintain the desired
offset.  If you have an SSB transmitter and a separate receiver, connected
to a PC sound card, it should be possible to implement this system with no
special hardware.  Of course, transmitters locked to GPS would be even
better.

Finally, I believe that these experiments can result in many benefits
unrelated to long baseline.  For example, if you have a communication
link between two LF stations, one with a powerful transmitter and one
with a quiet receiving location, it's obvious that you can complete some
QSOs that would be impossible from either alone.  Or, with multiple
receiving locations, there is a greater chance of copying a particular
signal; one location might be favored with good propagation or low noise.
And, if your equipment is all locked to GPS or LORAN or whatever, then
the improved stability will enable narrower bandwidths, and hence,
reception of weaker signals.

All comments and flames welcome.

73,

Stewart KK7KA


_______________________________________________
lf mailing list
[email protected]
http://www.amrad.org/mailman/listinfo/lf