Re: [SI-LIST] : skin effect and transmission line simulators

J. Eric Bracken (
Thu, 04 Sep 1997 15:55:11 -0400


The new release (3.0) of Maxwell 2D Extractor from Ansoft can do what
you want. It's currently in beta.

We analyze the frequency dependence of resistance and inductance
using a full finite element model of the conductor cross-section.
This lets us compute the actual current distribution within the
wires, and how it changes with frequency.

The frequency sweep is done using AWE-like acceleration techniques,
so it's much faster than doing one frequency point at a time.

We can create Spice models that represent the frequency dependence
of R and L. These models can be analyzed in the time or frequency domain,
with nonlinear transistor-level driver and receiver models, in our
Spice or your own. Of course, you can excite with any kind of
waveform you choose--periodic or not.

This type of modeling handles dispersion, attenuation, and the
effects of skin effect losses on crosstalk; you can also model the
skin effect in your ground plane as well as in your signal line.


--Eric Bracken

>>>>> "Chris" == Chris Simon <> writes:

Chris> I am looking for a time domain lossy transmission line
Chris> simulator for digital signals with bit rates of 1Gbit/sec
Chris> and higher. I would like to get information on simulators
Chris> that you have experince with.

Chris> The simulator I'm looking for should include all of the
Chris> following: 1. The results can be viewed in the time
Chris> domain. 2. Transmission line models that include series
Chris> resistance and inductance which are functions of frequency
Chris> to ACCURATELY model skin effect. (i.e., a frequency
Chris> dependent line model that is used to get time domain
Chris> simulation results) 3. SPICE CMOS models and/or IBIS type
Chris> non-linear models can be used as drivers and receivers 4.
Chris> Arbitrary driver data patterns can be used (i.e., not
Chris> limited to repetative signals)

Chris> There is one other criteria, but I'm not certain that this
Chris> one is required for signals with 150 to 200 pSec rise and
Chris> fall times. Any opinions? 5. Includes the effect of
Chris> dielectric losses.

Chris> I am aware of several products that claim to meet these
Chris> critria. Upon close inspection several of the lossy line
Chris> models didn't really meet number 2. This seems to be the
Chris> tough one. I am going to take a detailed look at Microwave
Chris> Design System (MDS) from Hewlett-Packard, although someone
Chris> told me that the current version doesn't meet number 3
Chris> above. I would be interested in hearing which of the above
Chris> criteria are met by other tools, or whether a future
Chris> revision will incorporate these features. If you know how
Chris> the following tools stack up, especially with regard to
Chris> number 2, I would appreciate hearing from you. Quad
Chris> Design, Quantic, Interconnectix/Mentor Graphics, Pacific
Chris> Numerix, Cadence SigNoise, ApSim/Contec, EMC-workbench,
Chris> others?

Chris> Chris Simon

J. Eric Bracken, Ph.D.                          Tel: 1.412.261.3200 x135
Group Leader, Signal Integrity R&D              Fax: 1.412.471.9427
Ansoft Corp., Four Station Square, Suite 660
Pittsburgh, PA USA 15219-1119