Re: [SI-LIST] : Ferrite Bead models

Ray Anderson (raymonda@radium.eng.sun.com)
Mon, 22 Mar 1999 20:10:56 -0800 (PST)

>
>
> Does any one out there have a model (usable in Spice) for a ferrite bead?

>
>

One particular bead that was specified to have 2Kohms Z at
30 MHz was reasonably modeled by the following ckt up to some freq.
(which I don't remember at the moment) where the frequency dependent
losses of the ferrite material really started to kick in. A more complex
model is needed to accurately model the effects of the freq. dependent
losses.

The bead models as:

----+------L-----+-----
| | L=18.47uH
+------C-----+ C=1.4pF
| | R=1.6K
+------R-----+

This was derived with the HP4194's built-in model firmware.


.subckt bead2k 1 2
Rpar 1 2 1.6k
Lpar 1 2 18.47e-6
Cpar 1 2 1.4e-12
.ends


The model is representative of the beads performance up to a
certain point and then runs out of gas.

A much better way to model a bead is to measure the s-parameters
of the device you are interested in and then create a "black-box" model
realized as a ratio of polynomial controlled voltage and current sources
(a Pade' approximation). A program such as Apsimtech's SPAR or HP's
ADS (MDS) with the Picosecond modeling option can assist in making such
a model.

You really need more than just a generic bead model if you
want your simulation to mean anything.

I believe this topic was discussed on si-list around April 98.
Also, the people at Georgia Tech's High Performance Packaging Lab have
done some interesting modeling of discrete components utilizing s-parameter
based macro models.

-Ray Anderson
Sun Microsystems Inc.

**** To unsubscribe from si-list: send e-mail to majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****