I've done a bit of this where we created a pseudo coax environment in a
pcb with a plane on top, bottom and the sides of a trace. What you need is
a field solver (2d or 3d will work) and then define the planes top and
bottom, and the lines on either side of the trace. You must be able to
"ground" the traces on the sides in the field solver you use.
We were able to get very good correlation of modeled to actual measured
impedance. It is important to understand the pcb construction, i.e., all
the geometries involved (layer to layer thickness, line width, spacing,
etch factor, etc.) and material properties (Er) to get this correlation.
Regards,
Mike Freda
Sun Microsystems
------------- Begin Included Message -------------
Date: Wed, 6 May 1998 09:22:46 -0400 (EDT)
From: Robert K Huber <rhuber@redcom.com>
X-Sender: rhuber@enterprise
To: si-list@silab.eng.sun.com
Subject: [SI-LIST] : Guard Trace Effect on Impedence
Mime-Version: 1.0
I would like to alternate guard traces with signal traces in a stripline
configuration and would like to calculate the effective characteristic
impedence of my signal traces. Could somebody either tell me how to
perform this calculation or direct me to a source where this is discussed.
I have already looked in Howard Johnson's "High-Speed Digital Design a
Hankbook of Black Magic" and this was not discussed there.
Thank You Very Much
Bob Huber
------------- End Included Message -------------