[SI-LIST] Re: Help: how to create IBIS model for a chip which containstwo dies

  • From: Lu@xxxxxxxxxx
  • To: "Crain, Dan S" <dan.s.crain@xxxxxxxxx>
  • Date: Thu, 24 Apr 2003 10:15:04 -0400


Hello Dan,

May I ask one more question about how to model two dies in single package?

Let's say I set RLC to 0 in individual ibs file, and in the ebd file I can
represent traces among PIN(user accessible pins of the package) and
NODEs(pins of dies).

                          PIN
+--------------------------------*--------+
|                                |        |
|         +------------------+---+        |
|         |                  |            |
|         | NODE             | NODE       |
|   +-----*--+        +------*----+       |
|   |        |        |           |       |
|   |        |        |           |       |
|   | die 1  |        |  die 2    |       |
|   |        |        |           |       |
|   +-----*--+        +------*----+       |
|                                         |
|                                         |
|                                         |
|              chip                       |
+-----------------------------------------+


My question is: where do you include package parasitics for the NODEs which
only have connection between two dies, since we set RLC to 0 in individual
.ibs file, and .ebd file does not include these traces?

+--------------------------------*--------+
|                                         |
|                                         |
|                                         |
|                                         |
|   +-----*--+        +------*----+       |
|   |        | NODE1  |           |       |
|   |        *--------*           |       |
|   | die 1  |  NODE2 |  die 2    |       |
|   |        |        |           |       |
|   +-----*--+        +------*----+       |
|                                         |
|                                         |
|                                         |
|              chip                       |
+-----------------------------------------+

Thanks

Lixin

*************************
 Lixin Lu
 IC CAD Specialist
 MOSAID Technologies Inc.
 11 Hines Road,
 Kanata, Ontario
 Canada K2K 2X1
 (613) 599 9539 ext 1638
*************************



|---------+---------------------------->
|         |           "Crain, Dan S"   |
|         |           <dan.s.crain@inte|
|         |           l.com>           |
|         |                            |
|         |           04/17/03 12:11 PM|
|         |                            |
|---------+---------------------------->
  
>--------------------------------------------------------------------------------------------------------------|
  |                                                                             
                                 |
  |       To:       <Lu@xxxxxxxxxx>                                             
                                 |
  |       cc:       <si-list@xxxxxxxxxxxxx>, <si-list-bounce@xxxxxxxxxxxxx>     
                                 |
  |       Subject:  RE: [SI-LIST] Re: Help: how to create IBIS model for a chip 
which contains two dies          |
  
>--------------------------------------------------------------------------------------------------------------|




My pleasure Lixin.  I am happy for the opportunity to give a little back to
members of this terrific forum.  Best of luck.

Rgds,

Dan

-----Original Message-----
From: Lu@xxxxxxxxxx [mailto:Lu@xxxxxxxxxx]
Sent: Thursday, April 17, 2003 8:33 AM
To: Crain, Dan S
Cc: si-list@xxxxxxxxxxxxx; si-list-bounce@xxxxxxxxxxxxx
Subject: Re: [SI-LIST] Re: Help: how to create IBIS model for a chip
which contains two dies



Hi Dan,

Thanks a lot! your email is very helpful. I found some IBIS files from
intel which have same setup as you mentioned, maybe you create them:) I'll
have a look at the IBIS section about EBD, and ask the package company to
see if I may get the EBD file from them, otherwise, I need create it by
myself.

Again, thanks.

Lixin




|---------+---------------------------->
|         |           "Crain, Dan S"   |
|         |           <dan.s.crain@inte|
|         |           l.com>           |
|         |           Sent by:         |
|         |           si-list-bounce@fr|
|         |           eelists.org      |
|         |                            |
|         |                            |
|         |           04/16/03 07:40 PM|
|         |           Please respond to|
|         |           dan.s.crain      |
|         |                            |
|---------+---------------------------->
  >
--------------------------------------------------------------------------------------------------------------|

  |
|
  |       To:       <Lu@xxxxxxxxxx>, <si-list@xxxxxxxxxxxxx>
|
  |       cc:
|
  |       Subject:  [SI-LIST] Re: Help: how to create IBIS model for a chip
which contains two dies              |
  >
--------------------------------------------------------------------------------------------------------------|






Hi Lu,

I have had success using IBIS version 3.2 and the EBD (electrical board =
description  --  in section 8) for multiple die/single package products.

For my case, I had a CPU and two memory devices vertically stacked and =
connected together via a package substrate.  The bottom side of the =
package substrate had a ball array for attachment/connection to a PCB.  =
My on-chip topologies looked very similar to your depiction.

My final IBIS setup contained the following items:
1) IBIS model for the CPU with package parameters zeroed out
2) IBIS models for the memory devices (also with package params zeroed)
3) An EBD file that described: the package R, L, and C parameters for =
each trace; trace lengths; bond-wire RLC params from each die connection =
to the substrate; and pin-to-pin-to-pin-to-ball connection mapping =
between the stacked devices and the substrate ball.  Note, each mini =
topology points to the appropriate driver/receiver model in the IBIS =
files for the dies.
4) Readme.txt file summarizing how everything fits together for future =
users

The hardest part of this effort was generating accurate electrical =
models for the substrate traces.

A good example of this type of IBIS setup can be found on Micron's web =
page under SDRAM modules.  I learned from this list that several widely =
used SI simulators work fine with IBIS/EBD models.

Best of luck!

Rgds,

Dan

-----Original Message-----
From: Lu@xxxxxxxxxx [mailto:Lu@xxxxxxxxxx]
Sent: Wednesday, April 16, 2003 2:50 PM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Help: how to create IBIS model for a chip which
contains two dies



Hello Guys,

I am new to the IBIS model creation. I have some questions about how to
create IBIS model for a chip which contains two dies. especially in the
following case:

There is an I/O buffer in die 1, which connects to another I/O buffer in
die 2, then to the pin.

+--------------------------------*--------+
|                                |        |
|         +------------------+---+        |
|         |                  |            |
|         |                  |            |
|   +-----*--+        +------*----+       |
|   |        |        |           |       |
|   |        |        |           |       |
|   | die 1  |        |  die 2    |       |
|   |        |        |           |       |
|   +-----*--+        +------*----+       |
|                                         |
|                                         |
|                                         |
|              chip                       |
+-----------------------------------------+

1. How can I choose the model for this pin, say I already have models =
for
each I/O buffers?
2. Is R_pkg, L_pkg and C_pkg is sufficient to model the trace among 2 =
I/Os
and pin?
3. which IBIS version should I use: 2.1 or 3.2?

I'll appreciate if you give me any opinions or suggestions. Thanks.


*************************
 Lixin Lu
 IC CAD Specialist
 MOSAID Technologies Inc.
 11 Hines Road,
 Kanata, Ontario
 Canada K2K 2X1
 (613) 599 9539 ext 1638
*************************



------------------------------------------------------------------
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 archives are viewable at:    =20
                         //www.freelists.org/archives/si-list
or at our remote archives:
                         http://groups.yahoo.com/group/si-list/messages=20
Old (prior to June 6, 2001) list archives are viewable at:
                         http://www.qsl.net/wb6tpu
 =20
------------------------------------------------------------------
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 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











------------------------------------------------------------------
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 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: