Re: [SI-LIST] : HSPICE W-element and skin effect

About this list Date view Thread view Subject view Author view

From: Ray Anderson (raymonda@ha1mpk-mail.eng.sun.com)
Date: Tue Sep 19 2000 - 11:57:28 PDT


The new version of the W element in the 2000.4 release
allows one to specify RLGC values in a table format wrt
to frequency. This allows any measured, simulated, or
imagined functional dependance of the parameters as
a function of frequency.

Earlier versions of the W element implementation assume
the classical physic's textbook sqrt(f) variation of
the R value (skin effect) and direct f variation of the
G parameter. You give the model a normalized R and G
number and it scales it with respect to frequency.
If the actual dependence in a particular geometry is
not sqrt(f) and f then there is a built in error.
The new table definition option allows you to input
arbritrary dependence functions.

Also, the new method also allows frequency dependent
L and C to be specified as well, which previously wasn't
possible.

I haven't had the opportunity to mess about with the
new release as of yet, but the above comments reflect
how I understand it is supposed to work.

I've found that I need to really look closely at each
new "improved" version of the W element as earlier
problematic behavior of the model which was fixed
in earlier releases have been known to resurface in
later releases.

-Ray

>
>
> Sunil,
>
> You are correct. For versions 2000.2 and before only one
> loss tangent, or conductance value is allowed. This does
> cause some error in simulating the 1 - 3 GHz region with
> lossy materials like FR4.
>
> About a week ago Dan Oh, a former developer
> with Avanti!, reported the following:
>
> "I would like to introduce the new feature of W element which
> allows you to model any kinds of frequency-dependent behaviors.
> Starting from 2000.4 (beta version is already available now),
> W element will support a tabular model to describe rlgc
> parameters."
>
> It seems that the capability you would like is available.
> However, you will have to develop these tables yourself.
>
> I have not tried this new feature, so cannot comment on
> it's accuracy.
>
>
> regards,
>
> scott
>
> Sunil Kumar wrote:
>
> > Scott,
> >
> > In HSPICE, I can give only one value of loss tangent. But loss tangent
> > varies with frequency. So, how does W-Element take care of that fact?
> > Ideally, one should be able to give a graph between frequency and the loss
> > tangent, which is not possible in HSPICE.
> >
> > Regards
> >
> > Sunil Kumar Ph: 91-80-2204616
> > Research Engineer Fax: 91-80-2263256
> > ATM Group Email: sunilb@cdotb.ernet.in
> > Switching BII
> > C-DOT
> > Bangalore-52
> > INDIA
> >
> > On Tue, 12 Sep 2000, Scott McMorrow wrote:
> >
> > > Ron,
> > >
> > > the w-element also takes into account frequency dependent
> > > dielectric loss as well as skin effects.
> > >
> > > scott
> > >
> > >
> > > --
> > > Scott McMorrow
> > > Principal Engineer
> > > SiQual, Signal Quality Engineering
> > > 18735 SW Boones Ferry Road
> > > Tualatin, OR 97062-3090
> > > (503) 885-1231
> > > http://www.siqual.com
> > >
> > > Ron Miller wrote:
> > >
> > > > Scin effect is not enough.
> > > >
> > > > Above 1 Ghz, Dielectric losses dominate.
> > > >
> > > > For a Data rate of 1.25 GBS, 625 Mhz, 3rd harmonic and 5th harmonic must be considered.
> > > >
> > > > So, you are talking components that cannot be simulated without a frequency domain
> > > > simulator that takes into account dielectric losses.
> > > >
> > > > Ron Miller
> > > >
> > > > > -----Original Message-----
> > > > > From: Scott McMorrow
> > > > > Sent: Monday, September 11, 2000 4:25 AM
> > > > > To: Sunil Kumar
> > > > > Cc: SI List (E-mail)
> > > > > Subject: Re: [SI-LIST] : HSPICE W-element and skin effect
> > > > >
> > > > > Sunil,
> > > > >
> > > > > I believe that the RISETIME option is no longer necessary in 2000.2.
> > > > > Skin effect calculations are based upon the frequency content of
> > > > > the waveform being propogated.
> > > > >
> > > > > regards,
> > > > >
> > > > > scott
> > > > >
> > > > > Sunil Kumar wrote:
> > > > >
> > > > > > Hello everybody.
> > > > > >
> > > > > > I am simulating a gigabit (1.25Gbps) link in HSPICE 2000.2 by modeling a
> > > > > > transmission line with W-element. I am using 18mil coupled
> > > > > > (differential) microstrip. According to the manual, the skin effect is
> > > > > > calculated by using RISETIME option. I have used different
> > > > > > values for RISETIME e.g. 50ps, 100ps, 600ps, but I am not seeing any
> > > > > > difference in the eye diagram. Can anybody explain this? Where am I making
> > > > > > a mistake? How exactly does HSPICE calculate the skin effect for
> > > > > > W-element?
> > > > > >
> > > > > > In case of U-element, skin resistance is calculated for
> > > > > > f=1/(15*risetime), and this value is added to the dc resistance to get the
> > > > > > effective resistance. The resulting equivalent resistance depends on one
> > > > > > frequency; it does not depend on various frequency components in the
> > > > > > edge. I beleive this not is not an accurate model for simulating lossy
> > > > > > lines. In the W-element it is not mentioned how it calculates skin effect
> > > > > > resistance. Any comments?
> > > > > >
> > > > > > Thanks in advance for any help.
> > > > > >
> > > > > > Sunil Kumar Ph: 91-80-2204616
> > > > > > Research Engineer Fax: 91-80-2263256
> > > > > > ATM Group Email: sunilb@cdotb.ernet.in
> > > > > > Switching BII
> > > > > > C-DOT
> > > > > > Sneha Complex
> > > > > > 71/1 Millers Road
> > > > > > Bangalore-560052
> > > > > > Karnataka
> > > > > > INDIA
> > > > > >
> > > > > > **** To unsubscribe from si-list or si-list-digest: send e-mail to
> > > > > > majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
> > > > > > si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
> > > > > > si-list archives are accessible at http://www.qsl.net/wb6tpu
> > > > > > ****
> > > > >
> > > > > --
> > > > > Scott McMorrow
> > > > > Principal Engineer
> > > > > SiQual, Signal Quality Engineering
> > > > > 18735 SW Boones Ferry Road
> > > > > Tualatin, OR 97062-3090
> > > > > (503) 885-1231
> > > > > http://www.siqual.com
> > > > >
> > > > >
> > > > >
> > > > > **** To unsubscribe from si-list or si-list-digest: send e-mail to
> > > > > majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
> > > > > si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
> > > > > si-list archives are accessible at http://www.qsl.net/wb6tpu
> > > > > ****
> > > > >
> > > >
> > > > **** To unsubscribe from si-list or si-list-digest: send e-mail to
> > > > majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
> > > > si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
> > > > si-list archives are accessible at http://www.qsl.net/wb6tpu
> > > > ****
> > >
> > >
> > >
> > >
> > > **** To unsubscribe from si-list or si-list-digest: send e-mail to
> > > majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
> > > si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
> > > si-list archives are accessible at http://www.qsl.net/wb6tpu
> > > ****
> > >
> > >
>
> --
> Scott McMorrow
> Principal Engineer
> SiQual, Signal Quality Engineering
> 18735 SW Boones Ferry Road
> Tualatin, OR 97062-3090
> (503) 885-1231
> http://www.siqual.com
>
>
>
> **** To unsubscribe from si-list or si-list-digest: send e-mail to
> majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
> si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
> si-list archives are accessible at http://www.qsl.net/wb6tpu
> ****
>

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


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Tue May 08 2001 - 14:29:32 PDT