Message-ID: <41BC4158.9080808@freenet.de>
Date: Sun, 12 Dec 2004 14:02:16 +0100
From: Wolf DL4YHF <dl4yhf@freenet.de>
User-Agent: Mozilla Thunderbird 0.9 (Windows/20041103)
MIME-Version: 1.0
To: rsgb_lf_group@blacksheep.org
References: <82.1d33c60e.2eec2207@aol.com>
In-Reply-To: <82.1d33c60e.2eec2207@aol.com>
Subject: Re: LF: Frequency errors on PSK31
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org

Hello David and group,

In all PSK31 demodulators / decoders I know, the bit error rate 
increases gently with increasing frequency and/or symbol rate error.

Talking about "digimode" beacons... Do you use a DDS for the frequency 
generation, and -if so- does the DDS have TWO frequency registers which 
can be toggled via control pin (some chips have this option) ? If so, it 
would be great to have a low-symbol-rate MSK test signal on air... 
choice of symbol rate and codeset would be up to you. For simplicity, 
one could start with Peter's VARICODE without convolution. If you are 
interested in MSK, I can send more information to you (too early to put 
everything online yet).

When comparing PSK31 and PSK08 recently, the PSK08 signal gave perfect 
copy even when it was not audible with 100-Hz audio bandwidth, where 
PSK31 only gave faint traces on the spectrogram. So simply dividing all 
clock by 4 in your PIC firmware may be worth considering. A suitable 
decoder for PSK08 is available in SpecLab since a few years, it has been 
slightly optimized for lower symbol rates recently with an optimized set 
of coefficients in the FIR filters (see 
http://people.freenet.de/dl4yhf/spectra1.html ) . Of course, the 
frequency error tolerance in PSK08 is only 1/4th of the PSK31 value, 
something to be aware of.

Best regards,
73 Wolf DL4YHF (back to DF0WD now, hpe cul).