[Fwd: LF Stability]

Bob Bruhns bbruhns@erols.com
Sun, 07 Feb 1999 22:46:39 -0500

This is a multi-part message in MIME format.
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Sorry, I goofed and sent this to the sender only...
Content-Type: message/rfc822
Content-Disposition: inline

Message-ID: <36BE5D26.8EC1CB9A@erols.com>
Date: Sun, 07 Feb 1999 22:42:30 -0500
From: Bob Bruhns <bbruhns@erols.com>
X-Mailer: Mozilla 4.06 [en]C-RR082798  (Win98; U)
MIME-Version: 1.0
To: Terry Fox <tfox@erols.com>
Subject: Re: LF Stability
References: <36BE14C3.85A68BB@erols.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

With single-signal conversion being done with SSB receiver techniques, I
think we have to input only real values into the FFT, but we need to
take the vector sum of the real and imaginary outputs of each bin to get
the actual signal reading for that frequency band.  Also, when
attempting to read data modulation, I think we would make the most of
the situation if our FFT sample period is equal to the bit period of the
signal, and is synchronized to it.  Don't know quite how to do that, but
it would be possible to run a full multi-kilosample FFT for every new
input sample, taking the new sample and the preceeding
multi-kilosample-minus-1 samples, and saving the results while looking
for sync...  then using the data from the appropriate sample period(s)
once we identify sync.

One of the European operators suggested that a little bit of keying yoop
made a signal stand out against steady birdies...  Maybe we can apply
some SS tricks to beat the birdies?

I only wish I wasn't (A) buried in overwork (likely for a lost cause, at
that), and (B) a victim of this killer flu - or I would be working on
this stuff right now.  As I expected, the job lined right up against
this great LF work.  Perfectly.

  Bob, WA3WDR