From: ARNOLD,PETER (HP-Cupertino,ex3) (firstname.lastname@example.org)
Date: Wed Jan 19 2000 - 11:21:27 PST
From: Jonathan Dowling [mailto:email@example.com]
Sent: Wednesday, January 19, 2000 10:27 AM
Subject: RE: [SI-LIST] : receiver jitter
> Argueably, the worst problem with behavioral simulators is that they do
not (or poorly)
> predict switching noise.
My department at HP uses XTK, and translates vendors' models from IBIS. I
agree that return-path / ground-bounce related effects are poorly (or not)
handled by current behavioral models, but I don't see this as a limitation
of the technique. It is surely possible for these to be included in a model
definition without resorting to Spice-level description. I anxiously await
tool and model support for simulating SSO noise in my designs.
My biggest problem with behavioral simulation is analysis of the results at
the receivers when signal quality is known to be less than perfect. At the
moment I get reports that give max flight time to a receiver threshold, and
indicate if multi-crossing errors or non-monotonicity occur. So I set a
certain standard for these and offset my thresholds by a certain amount and
call one side pass and the other fail. In reality I suspect some of my
failing waveforms would actually be correctly received by the real silicon -
those with a minor level of ringback well after the main transition,
perhaps, or a tiny kink. I can't examine every plot by eye to determine
I would prefer a more sophisticated receiver model that can indicate to me
whether the real silicon would have detected the level accurately in the
time allowed. I would get that from Spice-level models. Could an XTK-type
receiver model be made smart enough to do the same?
**** To unsubscribe from si-list: send e-mail to firstname.lastname@example.org. 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
This archive was generated by hypermail 2b29 : Thu Apr 20 2000 - 11:34:43 PDT