[SI-LIST] FWD: RE: Some issues related to spiral inductor modelling.
- From: Ray Anderson <Raymond.Anderson@xxxxxxx>
- To: si-list@xxxxxxxxxxxxx
- Date: Wed, 22 Jan 2003 12:06:08 -0800 (PST)
------------- Begin Forwarded Message -------------
From: "Dmitriev-Zdorov, Vladimir" <vladimir_dmitriev-zdorov@xxxxxxxxxxx>
To: "'mike_bihan@xxxxxxxxxxxx'" <mike_bihan@xxxxxxxxxxxx>,
List` Si <si-list@xxxxxxxxxxxxx>
Subject: RE: [SI-LIST] Some issues related to spiral inductor modelling.
Date: Wed, 22 Jan 2003 10:30:06 -0700
MIME-Version: 1.0
Hi Mike,
I's been working with S-parameters for a while and can tell you my point of
view. Sorry, if this is not exactly what you want to have as an answer. The
more sophisticated become the devices and their S-parameters, the more
problematic becomes and approach based on creating the equivalent circuit.
There are two main limitations inherent for this approach. First, you have
to solve the "synthesis" problem, to create a circuit that corresponds to a
given mathematical description. This problem is non-trivial and still have
no ideal solution (as most of other synthesis problems). Second, when you
incorporate your equivalent circuit into your design, as a subcircuit, this
adds an incredible number of additional internal variables; the size of
resulting Jacobian matrix "inside" simulator may even become prohibitive, if
you have many such models at a time. Even if not, you have considerable
slow-down of time-domain simulation.
The most promising way of solving such problems is avoiding sinthesis of
equivalent circuit. There are ways of creating a rational polynomial
approximation for your S-parameter matrix (sometimes called fitting), that
you can store as a table of frequency poles and their matrix coefficients.
[Equivalent circuits are often created from this table, not from original
frequency-domain dependencies ].Once such table is generated, it has all
the information that is contained in your S (or Y, Z) parameters. Then,
with this table, you can use very effective "indirect numerical integration",
"difference approximation" or "recursive convolution" approaches, something
similar to what some W-element model do [see for instance D. Kuznetsov, J.
Schutt-Aine "Optimal transient simulation of transmission lines, IEEE Trans
on CAS-I, v.43, No.2, where the "real-pole" equivalent of this approach is
presented].
In this case, you can avoid synthesis stage and also perform all
pole-handling operations "inside" the recursive model, not by placing the
expanded stuff into the Jacobian. It is not difficult to reach a convention
about format of such tables, containing poles and their coefficients
(something similar to Touchstone format). Then, these tables can be
effectively used as "models": recursive integration procedure is not a big
deal; it can handle all such models. There is no technical problem behind
this suggestion, there is a problem of convention and standard that is
missing.
Thanks, Vladimir
-----Original Message-----
From: Bi Han [mailto:mike_bihan@xxxxxxxxxxxx]
Sent: Wednesday, January 22, 2003 9:47 AM
To: List` Si
Subject: [SI-LIST] Some issues related to spiral inductor modelling.
Hi, all:
Recently I got some problem about 'spiral inductor modelling'.
After obtaining S-parameters from network analyzer, we had to translate =
it
into some kind of equivalent circuit to simulate it in hspice. Some =
methods
of annealing were used for this process. Since there will be many sets =
of
resolutions during the rocess of annealing, so what should be the final
criteria?
Also, since the value of the elements, such as resistor, cap, and =
inductor
in the equivelent circuit, are frequency dependent, how can we =
calculate the
exact value of these elements? Should I calculate every parameter at =
each
frequncy point?=20
=20
How to simulate this equivelent circuit is also a big issue, because =
the
parameters are frequency dependent, how to handle these problems? I =
wish
some experienced guies can help me.
Thanks!!
------------- End Forwarded Message -------------
------------------------------------------------------------------
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 archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
