All,
The following contains 1 status of where we are, a discussion of s12p pre
layout SerDes Models. I am enclosing the working memory package IBIS file,
and a DIMM EMD file updated from the originally posted parameter tree
format to be more compatible with the package modeling formats being
discussed.
1. I think the open Use Model issues are:
a. How does the IC Vendor want the User and/or EDA Vendor to use
"Swathing" coupled models. The IC Vendors supplying these models need to
step up and explain. There are now two exemplars of this type of model:
i. An
s12p representing a SerDes victim and two SerDes aggressors (all three
being either Tx or Rx). This model to be used for all the SerDes Tx (or
Rx) pins as victims.
ii. A
subckt representing one DQ victim, two DQ aggressors and one DQS
aggressor. There is one unique instance of this subckt for each DQ, with
each instance having a single value Length parameter, all using the same
subckt.
b. Need to review Use Models and Modeling Scenarios of some of the
examples I added to the end of memory_package.ibs file.
c. Need to ask if there are other possible package models or
modeling scenarios that IC Vendors want to deliver, or EDA Vendors think
IC Vendors should deliver that cannot be expressed in the "SiSoft" or
"Mentor" package format.
d. Need to summarize what "Use Models" the "Required IBIS Packaging
Modeling Solution" needs to support.
2. This committee of four needs to come up with an IBIS Syntax that
supports the "Required IBIS Packaging Modeling Solution" determined in
1.d.
3. This committee should discuss an EMD strategy. I agree with Randy
that the new EMD format I sent out as an exemplar will be much more
palatable to the IBIS-ATM and IBIS community. I still like parameter tree
format, and I think it is still a good way to introduce a concept because
it is a self-documenting syntax, but what Kumar wanted (and I tried to
support) is all new IBIS sections be implemented in parameter tree is just
not going to happen. A fundamental decision is whether to create a new EMD
section or enhance EBD.
I have done some research on s12p package model that we have received from
IC Vendors (Intel and others).
In Post Layout we just use the s4p of the victim extracted from the s12p
on each channel. In Pre-Layout we us the s12p along with a channel with
one victim and two aggressors (either another s12p or a channel with
coupled w-lines). I think by distinguishing Rx_Xtalk as strictly pre
layout eliminates the difficulties we have of documenting swathing.
So taking the example we discussed in early February:
1. Rx_Xtalk could be used in both Pre and Post layout.
2. Rx_Using_S12p could be used in Post layout, but requires care in how
to terminate the unused ports.
3. Rx_Using_S4p is like Rx_Using_S12p, except the IC Vendor does the
extraction of the s4p from the s12p.
4. Rx_Using_S4p_C0_D0 to Rx_Using_S4p_C15_D15 is a more verbose way of
doing Rx_Using_S4p.
[Begin Package Model] Rx_Xtalk
Language Touchstone
File PDF .2 Rx_Min.s12p .6 Rx_Typ.s12p .2 Rx_Max.s12p
Ports Pin_Mod+.Rx.1 Pin_Mod-.Rx.1 Buf_Mod+.Rx.1 Buf_Mod-.Rx.1
Ports V_Pin_Mod+.Rx.2 V_Pin_Mod-.Rx.2 V_Buf_Mod+.Rx.2 V_Buf_Mod-.Rx.2
Ports Pin_Mod+.Rx.3 Pin_Mod-.Rx.3 Buf_Mod+.Rx.3 Buf_Mod-.Rx.3
[End Package Model]
[Begin Package Model] Rx_Using_S12p
Language Touchstone
File PDF .2 Rx_Min.s12p .6 Rx_Typ.s12p .2 Rx_Max.s12p
Ports 0 0 0 0
Ports Pin_Mod+.Rx Pin_Mod-.Rx Buf_Mod+.Rx Buf_Mod-.Rx
Ports 0 0 0 0
[End Package Model]
[Begin Package Model] Rx_Using_S4p
Language Touchstone
File PDF .2 Rx_Min.s4p .6 Rx_Typ.s4p .2 Rx_Max.s4p
Ports Pin_Mod+.Rx Pin_Mod-.Rx Buf_Mod+.Rx Buf_Mod-.Rx
[End Package Model]
[Begin Package Model] Rx_Using_S4p_C0_D0
Language Touchstone
File PDF .2 Rx_Min1.s4p .6 Rx_Typ1.s4p .2 Rx_Max1.s4p
Ports Pin.C0 Pin.D0 Buf.C0 Buf.D0
[End Package Model]
[Begin Package Model] Rx_Using_S4p_C1_D1
Language Touchstone
File PDF .2 Rx_Min.s4p .6 Rx_Typ.s4p .2 Rx_Max.s4p
Ports Pin.C1 Pin.D1 Buf.C1 Buf.D1
[End Package Model]
...
[Begin Package Model] Rx_Using_S4p_C15_D15
Language Touchstone
File PDF .2 Rx_Min.s4p .6 Rx_Typ.s4p .2 Rx_Max.s4p
Ports Pin.C15 Pin.D15 Buf.C15 Buf.D15
[End Package Model]
Walter
Walter Katz
<mailto:wkatz@xxxxxxxxxx> wkatz@xxxxxxxxxx
Phone 303.449-2308
Mobile 303.335-6156
Attachment:
memory_package.ibs
Description: Binary data
Attachment:
dimm.emd
Description: Binary data
