Content-Transfer-Encoding: 8bit
Message-ID: <5.0.2.1.2.20011117174327.00a1ea70@mail.pncl.co.uk>
Date: Sun, 18 Nov 2001 18:19:08 +0000
To: rsgb_lf_group@blacksheep.org
From: "Walter Blanchard" <blanch@pncl.co.uk>
Subject: LF: A modest proposal
In-reply-to: <3BF656EA.67F83CD3@virgin.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Sender: <majordom@post.thorcom.com>

I don't quite understand what this proposal is about.

Try this:
The remote Tx sends a steady carrier for an hour:
Both rxs "hear" him for a few moments now and again but at both sites he is 
too weak to be detectable.
The two rxs are added together in the hope that some of those moments will 
coincide and give a 3 dB gain which might be sufficient to lift him out of 
the noise.
Is that the idea?

Walter G3JKV.


Re phased rx:

Unless I mistake what you're trying to do, you will have to use some sort 
of phase-coherent detection.  That's the usual reason for time-locking and 
baseline compensation.  This means the incoming signal would have to hold 
constant phase over the integration period. At the tx itself no ham stuff 
is good enough for this even if you didn't switch it so would-be DX-er's 
would have to build a special high-stability driver using constant-phase 
switching, not just plug in the mfr's allegedly "high-stability" 
option.  Then, the path itself is unstable at long ranges 
and  signal  phase would constantly wander as the path length drifted 
because of iono changes.  At relatively short ranges (1000-2000 kms) that 
might on occasion have a predominant single skywave mode you could possibly 
get away with it for five or ten minutes under the most favourable 
circumstances but no longer. At greater ranges there are always several 
modes and you will get almost random phase changes.  To see what I mean 
take a look (if you have the right gear) at a Loran-C pulse coming from 
these ranges at night. Not at the pulse  envelope but phase-track the 100 
kHz signal itself in wide-band (20 kHz) mode.
Also, there is no chance of forming a single "beam" at these frequencies. 
At best, if all the signal delays were properly compensated between rx 
antennas you would get multiple lobes with deep nulls in between, 
interferometer-style. On a 3 lambda baseline, about 24 nulls each side. 
Longer baselines would produce more and more nulls.
Incidentally, without special timing rxs the accuracy of 
time-synchronisation via GPS is not likely to be better than 1 uS. At 136 
kHz that's nearly 50 degs of phase so you could be an eighth-cycle out 
right away.
Sorry to be a "wet"!

Walter G3JKV.