Content-Transfer-Encoding: 8bit
Message-ID: <200102141551.KAA11032756@indyweb.cgocable.ca>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Date: Wed, 14 Feb 2001 10:52:00 -0500
To: rsgb_lf_group@blacksheep.org
From: "Bill de Carle" <bill1@cgocable.ca>
Subject: Re: LF: Re: GPS-Disciplined BPSK
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Sender: <majordom@post.thorcom.com>

At 02:53 PM 2/14/01 -0000, Andy wrote:
>> > in view. The 1PPS output is specified to be within 50ns of  "true
>time" (GPS time
>> > or UTC, selectable) with SA on. Now that SA is gone, the accuracy is
>even better
>> > (there is always a ~30ns sawtooth error due to finite  resolution in
>the software/hardware).
>> > 
>
>Surely, if the 1us jitter is a function of clock timing synchronisation
>then by averaging over many 1s periods, as the Brooks Shera design does,
>errors will be averaged out and high accuracy frequency sources may be
>properly locked to this 1PPS signal.    No GPS disciplined frequency
>source worth its name would, or should, ever try to correct output
>frequency absolutely on a pulse by pulse basis.

When removing jitter from a signal that's accurate over the long term,
we can do better than just averaging.  By letting a counter run (e.g.
from the oscillator we're trying to discipline) and latching it with the
1 PPS signal, we get the number of oscillations that have occurred from
one second to the next.  One approach would be to simply average up a
bunch of these and say that's the best answer.  But if we don't clear
the counter after each 1PPS latching signal - if we just let the counter
continue running upwards, then we can use least squares to solve for the
*slope* of the line.  Over any given time period that least squares slope
solution will give a significantly better result than averaging alone.
Try it if you don't believe it.

>>From memory, I think that design only claims about 10^-9 accuracy anyway
>locking with a time constant of many tens of minutes using a digital
>phase locked loop, still far less than GPS 'ought' to be capable of,
>although this was in the days when Selective Availability was still on.
>But accurate enough for most LF experiments - at the moment.    For bit
>timing purposes, I would have thought 1ms was good enough for 100ms or
>slower BPSK.

Absolutely.  What we have now is plenty good enough for BPSK.  There is
a problem which will limit us to MS100 or slower anyway.  It's that we
have considerable difficulty compensating for the number of *milliseconds*
delay through our transmitters and our receivers.  At MS100 or slower,
the existing GPS 1PPS outputs (even from "cheapie" GPS receivers) are
very well suited to our purposes.

Bill VE2IQ