Message-ID: <200208160048.g7G0m7F08953@block.iamanidiot.com>
To: njqrp@njqrp.org,  rsgb_lf_group@blacksheep.org, 
 lowfer@mailman.qth.net,  lf@amrad.org,  psk31@bipt106.bi.ehu.es
Subject: LF: PSK31 no longer audio beacon -- update
Date: Thu, 15 Aug 2002 20:48:07 -0400
From: "David Willmore" <davidwillmore@iamanidiot.com>
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Sender: <majordom@post.thorcom.com>
Content-type: text/plain; charset=windows-1252; format=flowed
MIME-Version: 1.0
Content-transfer-encoding: 8bit


Just a quick email to let everyone know that the enhancements to 
the NJQRP PSK31 audio beacon kit are going well.  With the help
of Peter G3PLX, I have been able to get the PSK31 modulation
code working.  The code is in such a state that it can send a
pre-varicoded message from a table in ROM.  This is as far as
I can go without messing with the serial port and building the
EEPROM daughter card--which are the next steps.

I've had to modify the name of the project a little as it's
no longer an 'audio' beacon.  As it uses DDS to generate the
waveform and because it samples at 500KHz, it can easily produce
output up to 200KHz.  At 1KHz, run directly into a PC sound 
card, most programs claim that its IMD is around -52 to -56 db.
I have no way of measuring it at RF as I don't even have a
scope, but when received through an FT-817, and piped into a PC
sound card, it comes out at -31 db IMD--which is receiver limited.

I did PSK31 first as it's the hardest of the modes I hope to
generate.  It takes two DDS operations and a multiply while the
others take two DDS operations and an add(MSK16) or just one DDS 
operation (CW, QRSs, Dif-CW, Jason).

Now that I know I can fit the complicated mode in, the others
will be a lot easier.  It even looks like it will be possible
to take 'live' input from a serial connection for all modes--PSK31
might not make it, but I will try.

Several people mentioned wanting PSK31 receive.  Well, I didn't
think it was possible, at first, but I'm starting to think that
it might not be all that far fetched.  At 50 8 bit MIPS, I can
do 1.5M 8x8 multiplies/sec.  At a 4KHz sampling rate, that's
375 multiplies/sample.  That might just be enough for a few
simple FIR filters, etc.  Memory will be the killer for that,
though.  If anyone has any suggestions, I'm open to them.  It's
starting to look like it *might* be possible.  No promises, 
though.

Thanks, Peter.

Cheers,
David N0YMV