Content-Transfer-Encoding: 8bit
Message-ID: <3.0.1.16.20010323151053.08afed50@mail.cc.kuleuven.ac.be>
Date: Fri, 23 Mar 2001 15:10:53
To: rsgb_lf_group@blacksheep.org
From: "Rik Strobbe" <rik.strobbe@fys.kuleuven.ac.be>
Subject: LF: QRS beta version (external timing)
In-reply-to: <65AECDF1F89AD411900400508BFC869F9C03FA@pdw-mail-1.dera.gov .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>

Hello group,

Recently there have been some requests to be able to let an external (and
accurate) timing source control the PC clock while running QRS.
I had a look at it and a basic 'external time control' was rather easy to
implement.
In version 3.06 beta 1 it is possible to connect a accurate timing source
to  COM port CTS pin (on a D9 connector : pin 8, on a D25 : pin 5). At this
moment it must be a 1 minute clock source (maybe more posibilies later).

You can download this version from : 
http://www.qsl.net/on7yd/QRS/qrs306b1.zip

In SETUP you can click on TIME to get to the external clock menue. There
you have 3 possibilies :
- DISABLE : no external clocking
- ABSOLUTE : the PC clock will increment 1 minute for every pulse at CTS
(compared to time at the previous pulse).
- ADJUSTING : a pulse at CTS will adjust the PC clock to the closest minute

May sound complicated, therefore an example :

1. Before starting QRS you have to set the PC clock as accurate as you can
(by hand using the default windows device for setting the clock or using
software that sets the PC clock based on DCF77 or similar).

2A. If you select ABSOLUTE then QRS will take a 'reference time' at the
first pulse on CTS and from then on will increase time by 1 minute for
every pulse. In the time between the pulses the PC clock will run at its
own pace.
eg. Assume 14:25:36 is taken as reference time. At the next CTS pulse the
clock will be set to 14:26:36, then to 14:27:36 etc ...

2B. If you select ADJUSTING then QRS will also take a 'reference time' at
the first CTS pulse but from then on will only adjust the clock on every
CTS pulse.
eg. Assume 14:25:36 is taken as reference time. At the next CTS pulse QRS
reads the PC clock, let's say 14:26:35 -> this will be adjusted to 14:26:36.

The difference between the 2 options is that if (for whatever reason) there
is a CTS pulse missed that in the ABSOLUTE mode the PC clock will be 1
minute wrong while in the ADJUSTING mode the clock will be set correct as
long as the in the time between 2 CTS pulses it hasn't gone wrong more than
30 seconds in its own pace. 

One could use an atomic clock or GPS to generate the 60 second timing, but
for the mere mortals a simple Xtal oscillator + divider train will do (keep
the error less than a second over a period of several days).
Keep in mind that the pulses must exactly come in a 60 second interval, but
that there is no need that the pulse comes just on a full minute.
Finaly : the pulse applied to CTS must be positive and have a width between
10us and 10ms.

If there is sufficient feedback I am willing to continue and make more
flexible external timing possibilies.

73, Rik  ON7YD