Hi,
If anyone is interested, I have the latest copy of my software for
arbitrary
transmission lines to Sourceforge. See
http://atlc.sourceforge.net
This allows one to find the characteristic impedance of any two conductor
transmission line of any cross section - with conductors being round, square,
elliptical, hexagonal or any other shape. I can not handle hollow waveguides or
optical fibres, but apart from that, it will do anything you can throw at it.
One needs to draw the transmission line's cross section in a graphics program,
save it as a bitmap and then let the program 'atlc' compute Zo from the bitmap.
In addition to computing the characteristic impedance Zo, the programme also
computes the inductance per metre, capacitance per metre and produces images of
the electric field profile, where the energy is concentrated etc. See
http://atlc.sourceforge.net/fileformat.html
and
http://atlc.sourceforge.net/examples.html
While using 31 different tests, based on 3 different analytical models (normal
coax, eccentric coax and a stripline between two infinitely wide conductors),
the maximum error found was under 1 % and the RMS error is less than 0.3 %.
The program works with both single dielectrics and multiple dielectrics,
although I must admit I've not checked the accuracy with multiple dielectrics.
The results seem reasonable, but I know of a few areas in the code that are not
100% accurate when there is a boundary between two dielectrics. Hence I might
expect the error to be slightly higher on multiple dielectrics. I will
eventually remove those problems and attempt measurements of accuracy, although
this is hampered by the lack of any obvious exact analytical models.
The software is distributed as C source code and needs to be compiled on the
target system. It should be no problem on any Unix or linux machine. The code,
being rather cpu intensive, has been written to support multiple processors. If
anyone needs to use the program, but does not have a linux machine, I'm happy to
run a few runs for you on my own computer, but this will of course have its
drawbacks.
I've only tested support for multiple processors on my Sun Ultra 60. There may
be some changes needed for linux or other machines, but it should be okay if the
pthreads library is installed. I'd be particularly interested to hear of anyone
using it with more than one cpu on any machine other than a Sun.
I've not currently produced an executable for a Windoze machine, but will
attempt this if anyone is interested. If anyone with a Mac C compiler can
produce a Mac executable, I would appreciate one. Someone is keen to use it on a
Mac and I have no way to create such a binary. While I don't have a C compiler
for my PC, I can get one for free, so will produce a Windoze binary if anyone
wants it.
The program is open-source and free to use by anyone for commerical or amateur
purposes, under the conditions of the GNU Public Licence.
--
Dr. David Kirkby PhD,
email: drkirkby@ntlworld.com
web page: http://www.david-kirkby.co.uk
Amateur radio callsign: G8WRB
|