Content-Transfer-Encoding: 8bit
Message-ID: <5.1.0.14.0.20020107102830.00b0ecb0@gemini.herts.ac.uk>
Date: Mon, 07 Jan 2002 12:04:14 +0000
To: rsgb_lf_group@blacksheep.org
From: "James Moritz" <j.r.moritz@herts.ac.uk>
Subject: LF: Re: GPS Coherent PSK Transmission
In-reply-to: <7D653C9C42F5D411A27C00508BF8803D55C689@pdw-mail-r1.dstl.go v.uk>
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>

Dear Andy, LF Group,

I copied the G4JNT PSK beacon last night; assuming I have the start 
times/polarity correct the sequence was:

011 010 111 001 011

I found two methods of demodulating the signal using general-purpose 
equipment - the first is simply to produce a spectrogram, using parameters 
for 30s dots. The "key clicks" produced by the phase transitions are 
visible as a spreading of the trace. Setting Spectrum Lab to produce 30s 
time markers made it fairly easy to identify the sequence. I noticed a 
period around 2250 when there were no phase transitions for several 
minutes, but then the sequence returned to normal.

The second method was to look at the phase directly - to do this I fed the 
RX audio into one channel of a 'scope, and fed a stable audio frequency 
equal to the BFO offset into the other channel. The frequency error of the 
RX and locally generated tone was small enough so that the phase difference 
between the two traces only changed a fraction of a cycle in several 
minutes, so it was easy to see an abrupt phase change when it occurred. Due 
to QRM/QRN, I found that to make the beacon signal clearly visible on the 
scope, a very narrow bandwidth audio filter (<10Hz) was required; this was 
also implemented using Spectrum Lab. Even then, the apparent phase of the 
signal bounced around by some 10s of degrees due to noise, and noise spikes 
could cause a temporary phase inversion, so a lot of concentration was 
required to get the sequence. With a strong signal, the traces are quite 
stable, so you could use this method to detect an abrupt phase change, or 
measure frequency to millihertz resolution quite easily.

Although good fun, neither of these methods is really viable for DX 
reception, since they don't integrate the signal over the whole bit period, 
let alone over several sequences - a bit like decoding QRSS by ear. Andy's 
signal was about S3 at my QTH, but hard to actually hear due to a permanent 
S6 carrier I get on about 136.65kHz, a real nuisance for manual CW 
operation - does anyone else hear this?

Cheers, Jim Moritz
73 de M0BMU