Re: [IBIS-Users] How do simulators handle C_Comp?

From: Andrew Ingraham <a.ingraham_at_.....>
Date: Thu Mar 03 2005 - 07:33:32 PST
> In practice, I would like to know how simulators such as Hyperlynx account
> for this C_comp? Does it exclude double-counting of the C_comp when
> performing rise/fall time domain simulations?

Lengthy discussions about this have come up on these maillists a few times
before.

Honestly, I don't see what the problem is, why people think that there is
"double-counting" if a non-zero C_comp is included.

The IBIS model of an actively driving device is very simple.  It starts with
just a switch, that changes in an ideal way from one state to the other,
with some drive strength (the I-V curves) in each state.

The I-V curves do not contain any information about C_comp in them.  None
whatsoever.  I-V curves are static (DC) parameters, and are unaffected by
C_comp.  The I-V curves could be generated with C_comp=0, or C_comp=2pF,
or C_comp=10,000uF, and it would make no difference to the I-V curves.

C_comp is something that needs to be added to help make this model more
representative, more accurate.  The I-V curves represent the static drive
strengths of the output transistors.  C_comp represents the capacitance that
is really there on the chip, in those transistors and around them.  If the
simulator didn't include C_comp in the simulation, the IBIS model would
represent output transistors that had no capacitance, and they would switch
incorrectly.  You need C_comp to be there.

C_comp is even more important when the driver tri-states and the IC receives
signals from elsewhere.  But I don't think anyone has an issue with
"double-counting" C_comp when the device is a receiver.

If the simulator primarily uses the I-V curves to create the driven
waveforms from an output driver (and I believe that they do), then I think
there's no issue with double-counting.  The simulator needs to know about
C_comp in order to adjust the dynamic response from the "perfect" switch
that behaves according to those I-V curves.

IBIS data sheets can also have V-T curves, and those curves were measured
with the actual C_comp in the device (or in the SPICE model from which the
IBIS data was extracted).  What does the simulator do with the V-T data?  As
far as I know, they generally use it to adjust the simulator's transition
from one state to the other (from one I-V curve to the other), but they do
not use it to actually generate the driven waveforms themselves.  That is,
the simulator doesn't synthesize a time-dependent source whose value equals
the V-T curve, and then attach C_comp to it.  If it did, it would always be
wrong.  So simulators don't do that.

If a simulator tries to replicate the V-T curve, it must know that it can
only replicate it with C_comp present and with R_fixture present.  Including
them is not double-counting; it is providing exactly what the simulator
expects to be there.

Regards,
Andy


|------------------------------------------------------------------
|For help or to subscribe/unsubscribe, email majordomo@eda.org
|with just 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 written 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 Thu Mar 3 08:04:18 2005

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2005 - 08:05:11 PST