From: "Dr. David Kirkby"
X-Mailer: Mozilla 4.78 [en] (X11; U; SunOS 5.9 sun4u)
Subject: Software of interest to RF engineers.
Date: Tue, 17 Dec 2002 12:02:30 +0000
NNTP-Posting-Date: Tue, 17 Dec 2002 12:02:31 GMT
Organization: ntlworld News Service
The following may be of interest to professional RF engineers,
students and radio amateurs.
I have in the last few days placed on the web the latest version of my
free/open-source programme suite 'atlc' which was originally designed
for finding the impedance of RF transmission lines (like coax) but
with an arbitrary cross-section and arbitrary number of different
dielectrics. The errors have always been less than 0.5% for any
structure I've analysed that has a known exact analytical solution.
About 7 months ago I extended 'atlc' somewhat to have the ability to
analyse directional couplers (giving the odd and even mode
impedance's) of an arbitrary cross section and number of different
dielectric. In this case, the errors creep up a bit, to a maximum of
0.8% for the one structure I've compared it against that has an exact
It has now been extended the suite of programs 'atlc' to a include a
program 'design_coupler' which can *design* you a directional coupler
(this time of only one particular cross section), given the frequency,
coupling factor and optionally the length of the coupler (otherwise
the default length is a quarter wave at the centre frequency).
The program is written primarily for a Unix system, not Windoze, so
while it was developed on a Sun workstation under Solaris, it has been
checked on a few GNU/Linux PCs, a Dec Alpha, a Sun running GNU/Linux,
a Macintosh running OSX. i.e. it should run under any Unix system.
There is no graphical user interface, so for analysing arbitrary cross
sections you need to use a graphics package (Photoshop, Gimp,
CorelDraw etc) to draw your cross section, save it as a bitmap, then
run 'atlc' to analyse it. There are some programs for rapidly
generating certain cross sections, such as one rectangular conductor
inside another, coax (since I could quickly test a case for which
there's an exact solution) and several others.
The main analysis program 'atlc' is CPU intensive and hence has been
developed to work with multiple processors if available, although
support for multiple processors has not been checked much on any
system other that Sun's Solaris operating system.
The C source code is freely available issue under the GNU public
If interested, see the homepage at:
IF ANYONE WITH A SYSTEM (OTHER THAN A SUN) WITH MULTIPLE CPUS, CAN LET
ME KNOW HOW IT WORKS I'D BE INTERESTED.
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408