[SI-LIST] Re: issue in IBIS modelling of LVDS buffer
- From: "Muranyi, Arpad" <arpad.muranyi@xxxxxxxxx>
- To: <si-list@xxxxxxxxxxxxx>
- Date: Fri, 27 Aug 2004 09:01:39 -0700
Salainithya,
p. 16 of the presentation shows the Matlab code for reading the file
that contains the data and generating the two surface plots shown on p.
8.
(The file being read generated by the Perl code, which basically
extracts
the raw data from the HSPICE .LIS file and formats it to make it Matlab
readable).
p. 17 shows the code that generates the top two plots on p. 12. This is
also described verbally on p. 10.
p. 18 shows the code which does the normalization described on p. 11,
and
the code which generates the two surface plots on the bottom of p. 12.
p. 19 shows the code which does a surface fitting to the equation
Idiff =3D a0 + a1*xp + a2*xn + a3*xp*xn + a4*(xp)^2 + a5*(xn)^2
which was done because at the time when I wrote this HSPICE didn't have
the
IBIS [Series MOSFET] keyword implemented in the B-element. This
equation was
used for implemented the same effect in HSPICE. It is not necessary any
more
if you can use the [Series MOSFET] keyword. The bottom of this page
plots
the fitted surface, which is not shown anywhere in the presentation.
p. 20 shows the code which saves the data in IBIS format. The top half
saves the data for the [Series MOSFET] keyword containing the
differential
current (surface plot), the bottom half saves the common mode currents
which are shown on the top half of p. 12. The first is a bit tricky,
because
you have to slice the surface diagonally to get the data in the format
IBIS
needs it. The second is nothing but the diagonal of the raw data, as
described
on p. 10. Note that this code will not reduce the number of points to
100
(if there is more than that) which is an IBIS requirement.
I hope this will help you to get your work done. I don't say that you
must
have Matlab, but I am not sure whether you can do all this in HSPICE
alone.
You may need some sort of post processing to make it happen. Before you
go
though the whole process I would suggest that you look at the normalized
differential surface first, because there are many differential buffers
which have practically no differential current flow, in which case this
whole exercise becomes a mute point, and you are perfectly OK to make an
IBIS model containing only the common mode models.
Good luck,
Arpad
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
=3D=3D
-----Original Message-----
From: chandrasekaran.nithya@xxxxxxxxx
[mailto:chandrasekaran.nithya@xxxxxxxxx]=20
Sent: Friday, August 27, 2004 3:03 AM
To: Muranyi, Arpad; si-list@xxxxxxxxxxxxx
Subject: RE: [SI-LIST] Re: issue in IBIS modelling of LVDS buffer
Hi Muranyi,
Thanks for the info.Actually I don't know matlab usage.The document you
had referred uses matlab coding to find out Idiff value.
Can you please suggest me any manual method or any other method using
spice simulation with which I can find the Idiff.Or can you exactly
define me what is the way to calculate
Idiff so that I can implement the logic through spice simulation.
Regards
salainithya
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx]
On Behalf Of Muranyi, Arpad
Sent: Thursday, August 26, 2004 10:14 PM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: issue in IBIS modelling of LVDS buffer
Salainithya,
Look at the presentation(s) on modeling differential buffers
if you haven't done so yet:
http://www.eda.org/pub/ibis/summits/oct03/muranyi.pdf
Arpad Muranyi
Intel Corporation
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx]
On Behalf Of chandrasekaran.nithya@xxxxxxxxx
Sent: Wednesday, August 25, 2004 8:39 PM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] issue in IBIS modelling of LVDS buffer
.
Hi,
Iam facing a problem in IBIS modelling of a non-terminated LVDS
buffer.
Actually the spec indicates that for proper functioning a terminated
resistor of 100ohms is to be connected between pad and padn.
Hence this led me to conclude that the buffer is a non-terminated one.
my doubt is
1.whether we need to connect the 100ohm resistor between pad and padn
for generating the ibis model?
Also we tried both - connecting the resistor and without connecting the
resistor while ibis model generation.
2.The ibis generated VT waveforms are not matching with the spice
waveform in verification(spice vs hspice (used IBIS model in
hspice)).The voltage levels itself are not matching in both the cases. A
difference of about (1.0-1.2)V. Hence not able to figure out what
exactly causes the mismatch.
3.Please can anyone clarify as to what setup I should give in case of VI
and VT waveforms for generating and verifying IBIS file for LVDS
buffer(terminated and non-terminated).It would be more helpful if anyone
could give the setup in diagramatic form which will ensure accurate
modelling.Also if possible could anyone think on a generic sense and
tell us what might be the possible causes for the mismatch.
regards
salainithya
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List FAQ wiki page is located at:
http://si-list.org/wiki/wiki.pl?Si-List_FAQ
List technical documents are available at:
http://www.si-list.org
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
