Re: [SI-LIST] : Moment Method/Ensemble/Touchstone/Errors

Uwe Keller ([email protected])
Tue, 21 Sep 1999 14:00:25 +0200

Hi Craig,

I am not familiar with Ansoft ensemble but since it basically is a MoM
which reduces a volumetric problem onto a surface one by using a Greens
function - which is basically something like the Dirac delta function in
3D -
I should guess that effects of the layered medium have not been
modelled properly. The typical MoM approaches for layered media are
using Greens function of free space (no dielectric substrate) and
the matching of the conditions on the dielectric substrate numerically
by discretizing the dielectric surface and so-called equivalent
(magnetic) surface currents; the other is using a Greens function for a
canonical structure as e.g. an infinitely extending dielectric boundary.
Both procedures somehow approximate the real object and should give
reasonable results. However the phase will be more sensitive to
potential errors.

E.g. if dieletrics is neglected in favour for simplified computing the
phase velocity might be off by a factor of something arround 2 (inverse
of square root of effective dieletric, which is some kind of average
between free space or air and the substrat).

The line thickness might also play a role but I would assume a minor one
to the perception that the phase velocity is influenced by an increase
capacitance and a decrease in inductance and both effects tend to
a bit when forming the product.

Kind regards


Uwe ________________________________________________________________________________ | | | | Uwe Keller | tel. : +49 5251-60-6181 | | Projectgroup FhG/IZM - Paderborn / | fax : +49 5251-60-6155 | | University - GH Paderborn | email: [email protected] | | Fuerstenallee 11, 33102 Paderborn, Germany || |==============================================================================|

**** To unsubscribe from si-list: send e-mail to [email protected] In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at ****