[SI-LIST] Re: Why discontinuity must be far away Wave Port excitation?
- From: "Yuriy Shlepnev" <shlepnev@xxxxxxxxxxxxx>
- To: "'Tesla'" <emcesd@xxxxxxx>, <si-list@xxxxxxxxxxxxx>
- Date: Fri, 12 Aug 2011 09:57:50 -0700
Tesla,
In general, sufficient distance from a discontinuity to a wave port or to a
lumped port depends on discontinuity type as well as on the wave port type
or lumped port size. I will try to explain that.
Technically a discontinuity in a transmission line or wave-guiding structure
distorts fields of a dominant mode propagating in the line.
A local discontinuity forms near field with extra energy typically
associated with parasitic inductance of capacitance. Different
discontinuities may have larger or smaller near fields. Such field can be
de-composed into a set of evanescent (decaying) modes in addition to the
modes propagating in the line (similar to a Fourier expansion). A
multi-modal wave-port that includes multiple propagating and evanescent
modes can be technically placed as close to a discontinuity as you want.
Though, in general, the number of evanescent modes has to be larger as port
gets closer to the discontinuity. Set up a simple numerical experiment with
wave-ports at different distances from a discontinuity and compare computed
S-parameters. Convergence of a reflection parameter may be particularly
sensitive to presence of un-accounted modes. The distance is sufficient as
soon as the reflection magnitude becomes relatively independent of the
position. Typical distance for a localizable strip or micro-strip
discontinuity is 3-4 width of strip or substrate thickness (whatever is
larger).
You will not see the convergence in the reflection parameter if in addition
to the dominant modes in t-line there are propagating modes that are not
accounted for. It may happen if the line has propagating modes that are not
accounted for at wave-ports (case 1) or discontinuity excite modes that
cannot be associated with the transmission line at all (case 2). In the
first case, a transmission line with electrically large cross-section may
have propagating non-TEM modes for instance. Keeping cross-section small and
reference planes connected may reduce the problem, but always only up to
some frequency. Computation above that frequency with un-accounted
high-order propagating modes (or modes close to cut-off frequency) is the
common source of non-causal S-parameters. The second case is even more
interesting and often misinterpreted even in technical publications. This is
the case of non-localizable discontinuities such as via-holes without
stitching vias or with insufficient number of stitching vias. Technically,
any number of stitching vias or any differential via pair become
non-localizable at some frequency. Such vias excite waves of parallel plane
waveguides in addition to near fields associated with the capacitance and
inductance. Taking into account parallel-plane wave-guide modes at the
cross-section of a t-line approaching vias is not sufficient -
parallel-plane waveguide modes have to be accounted at all boundaries around
the vias (alternative radial wave-guide modes can be used for the same
purpose). A via model in that context will have multiple parallel-plane
waveguide ports in addition to the quasi-TEM ports for connected lines.
Connecting those additional ports with a model for PDN can theoretically
produce a system-level model for a non-localized vias that will account for
effect of board geometry on behavior of the vias up to microwave frequencies
(no tools that do that and thus the localization seems like much better idea
at this point).
Considering the lumped ports - they can be constructed in an area that is
much smaller than the wavelength. As long as it holds, the effect of
high-order modes is negligible (though the result of analysis may be
different from what you get with a wave port). A primary purpose of such
port is to be able to use a circuit theory model (voltage direction coincide
with the current direction) of a lumped element in the electromagnetic
analysis. It may be suitable for building discontinuity model for a
de-compositional analysis only in cases without the other possibilities.
Best regards,
Yuriy
Yuriy Shlepnev
www.simberian.com
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of Tesla
Sent: Friday, August 12, 2011 3:21 AM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Why discontinuity must be far away Wave Port excitation?
Hi, experts
In an application note of Ansoft hfss, it said geometrical discontinuity
must be far away Wave Port. I think High-order modes due to discontinuity
may affect the E and H field distribution leading to a wrong S-parameter.
But Ansoft said Lumped Port can be used in this situation. but i think the E
and H field of high-order modes also affect the current and voltage of
"Lumped Port". so i assume it also lead to a wrong result.
Thanks and Best Regards.
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List technical documents are available at:
http://www.si-list.net
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List technical documents are available at:
http://www.si-list.net
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
