From owner-ibis-users@eda.org Fri Oct 19 14:08:06 2001 Received: by server.eda.org (8.12.0.Beta7/8.12.0.Beta7) id f9JL7Ybs020579; Fri, 19 Oct 2001 14:07:34 -0700 (PDT) content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Subject: RE: High frequency performance of IBIS model x-mimeole: Produced By Microsoft Exchange V6.0.4712.0 Date: Fri, 19 Oct 2001 17:07:24 -0400 Message-ID: X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: High frequency performance of IBIS model Thread-Index: AcFY359LFiH7BHVpQCuizNbIw6+esQAAMpzn From: "Ingraham, Andrew" To: X-OriginalArrivalTime: 19 Oct 2001 21:07:24.0439 (UTC) FILETIME=[0CD78A70:01C158E2] Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by server.eda.org id f9JL7T7n020576 Sender: owner-ibis-users@eda.org Precedence: bulk > IBIS models have only a constant capacitance to account for > the die capacitance, which means that there is no frequency > dependency in their impedance. No voltage dependency either. And various other second-order effects that are missing, which don't bite you until you push the frequencies high enough. Also, most IBIS models you will find, have just the simple package model which consists of simple RLC parameters, and no specified way in which they are connected, i.e., lumped vs. distributed, pi vs. tee vs. ell ... which may be OK for lower speeds but not so hot at higher frequencies. And no couplings either. Finally, many IBIS models are garbage simply because the people who made them, didn't really know what they were doing. It's easy to throw together an IBIS model using s2ibis, and not understand the first thing about IBIS, nothing about SPICE modeling, and very little about using SPICE. It's real hard to make a SPICE model without a solid grounding in both SPICE usage and SPICE modeling. Andy