RE: [IBIS] RE: [IBIS-Users] hspice vs IBIS model correlation at different capacitive loads

From: Muranyi, Arpad <arpad.muranyi_at_.....>
Date: Wed Jan 04 2006 - 09:14:19 PST
Vinayak,

In addition to my previous reply, that IBIS is not
able to describe any feedback mechanisms inside
the buffer, let me explain what the fundamental
problem is with reactive loads with IBIS models.

1)  The concept of using I-V and V-t curves in the
models is based on the idea that the V-t curve can
be used as a time variant scaling coefficient on
the I-V curves.  In other words, a falling edge
waveform could mean that at its beginning the PD
transistor (or its I-V curve) is completely OFF,
and at its end it is completely ON.  This means
that you would multiply the PD I-V curve with 0
or 1, respectively.  All other points on the
waveform can be converted to numbers which are
usually between 0 and 1.

However, a falling edge could also mean that the
at its beginning the PU transistor is full ON, and
at the end the PU transistor is fully OFF.

2)  This is why an IBIS model which has a PU and
PD should have at least two rising and two falling
waveforms.  That way you have two unknowns (the
Kpu and Kpd scaling coefficient tables) and two
knowns (the two V-t tables) which can be solved
easily.

3)  The solution comes from writing a sum of the
currents equation:

            Ipu = Ipd + Iout

where Iout = (Vout-Vfixture)/Rfixture.  Vout is
basically your V-t tables, and the rest are the
conditions at which the waveform was generated.

4)  Theoretically the load doesn't have to be
resistive, you could have L and C, but if you
did that the equation would become a differential
equation including dI/dt and dV/dt.

5)  If tool vendors implement the above equation
without accounting for dV/dt and dI/dt, you will
start seeing significant miscorrelations when
your load is strongly reactive.

6)  This is not a problem when the buffer is loaded
with transmission lines, because they look like a
resistive load.  However, as soon as you start using
AC coupling (series capacitors between the buffer
and the T-line) or other reactive loads, you will
be in trouble.

I hope this helps.

Arpad
-----------------------------------------------------


-----Original Message-----
From: owner-ibis-users@eda.org [mailto:owner-ibis-users@eda.org] On Behalf Of Vinayak Ghatawade
Sent: Tuesday, January 03, 2006 11:21 PM
To: tom@teraspeed.com
Cc: ibis-users@eda.org; ibis@eda.org
Subject: Re: [IBIS] RE: [IBIS-Users] hspice vs IBIS model correlation at different capacitive loads

Tom,
Thanks for the reply.

Buffer for which IBIS is been generated will be used
with a shorter transmission line in which case c effect is more
pronounced than r-load effect. There is no termination in this case at the
recieving end since we are making sure the buffer will be used
with shorter trace.

Can we define transmission line load for V-T. The current V-T data that
we have generated has oscillations due to adaptive behaviour of the 
buffer since
this is not real case (might happen with longer trace).

Thanks again
vinayak

-----------------------------------------------------------------
|For help or to subscribe/unsubscribe, email majordomo@eda.org
|with the appropriate command message(s) in the body:
|
|  help
|  subscribe   ibis       <optional e-mail address, if different>
|  subscribe   ibis-users <optional e-mail address, if different>
|  unsubscribe ibis       <optional e-mail address, if different>
|  unsubscribe ibis-users <optional e-mail address, if different>
|
|or email a request to ibis-request@eda.org.
|
|IBIS reflector archives exist under:
|
|  http://www.eda.org/pub/ibis/email_archive/  Recent
|  http://www.eda.org/pub/ibis/users_archive/  Recent
|  http://www.eda.org/pub/ibis/email/          E-mail since 1993
Received on Wed Jan 4 09:14:24 2006

This archive was generated by hypermail 2.1.8 : Wed Jan 04 2006 - 09:15:23 PST