[SI-LIST] Re: What are power integrity tools available out there? Anything inexpensive that works?
- From: Jory McKinley <jory_mckinley@xxxxxxxxx>
- To: Joel Brown <joel@xxxxxxxxxx>, Istvan Novak <istvan.novak@xxxxxxx>, Aleksandr Oysgelt <aleks@xxxxxxxxxxxxxxxxxx>
- Date: Mon, 9 Feb 2009 10:27:56 -0800 (PST)
Hello Joel,
What we have done is model this behavior through the knowledge of the design.
For example, if we want to understand the effects of a PDN de-coupling strategy
for a DDR3 IO bank, we could model the power supply bank and associated DDR3
data signals with an SIWAVE or PowerSI or your favorite. The model would
include the section of the chip package (including any bond wire) and board.
We typically can get the Spice model for the IO of the IC and through design
knowledge can input the switching behavior of the DDR3 data. What is missing
is the on-chip power/ground model. We have been able to get these models (at
least the last couple of projects I have worked on) but I understand they may
not be available to everyone....depending. What should be available is the
on-die decoupling model which coupled with the other models will get you real
close to your PDN goals.
Regards,
-Jory
________________________________
From: Joel Brown <joel@xxxxxxxxxx>
To: Istvan Novak <istvan.novak@xxxxxxx>; Aleksandr Oysgelt
<aleks@xxxxxxxxxxxxxxxxxx>
Cc: si-list@xxxxxxxxxxxxx
Sent: Monday, February 9, 2009 11:49:01 AM
Subject: [SI-LIST] Re: What are power integrity tools available out there?
Anything inexpensive that works?
How is it even possible to "design" a PDN when IC manufacturers do not
provide any information about the power characteristics of their parts such
and current draw vs. frequency. There are no models as to what is inside the
ICs at the power pins. So all the design tools in the world are not going to
solve that problem. It seems that IC manufacturers could provide this
information if they had the motivation to do so.
Joel
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of Istvan Novak
Sent: Saturday, February 07, 2009 4:10 PM
To: Aleksandr Oysgelt
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: What are power integrity tools available out there?
Anything inexpensive that works?
Aleks,
You raise valid and good questions, but I am afraid the answers may not
make you happy.
If you look around the tool market, there are several good commercial
tools on the market;
you named two of them, and we can add some more, for instance the
Ansoft/Ansys SiWave
and Cadence PI tools. These are the most sophisticated publicly
available tools at the moment,
and even these wont give you a design. Some tools have optimization
features built in, but
one could argue that the outcome of the optimization depends largely on
the initial constraints
you put in, so unless the user has some idea of what is going on, it
wont tell you what to do,
it just tells you the performance of the design you have come up with
(or within the constraints
you set in the optimizer).
It would be great to have both pre and post layout simulation
capabilities in the same package
at a low cost. When you look at the commercial tools, you will find
that they are all better
or more convenient only in one or the other. To get the cost down and
get you something
beyond lumped spreadsheet and rectangular-plane cavity-resonance
formulas, your best bet
is to look around among pre-layout tools. For instance, you might want
to try EZ PowerPlane
from http://www.ems-plus.com/ezpowerplane.html. When it comes to DC
drop, you can
do what-ifs with a pre-layout tool, but you really want and need a
post-layout check, because
pre-layout tools will not easily capture all the fine details (via
antipad dimensions, via barrel
resistance) that you eventually need for a reasonable answer.
We know that there are less demanding and more demanding designs in
terms of how much
work you may need to put into the power distribution network design to
get a satisfactory
product. If you deal with the category where you can stay competitive
only if you squeeze
out every drop of extra fat from your system, the possible alternatives
are: either you
outsource this aspect of the design (and beyond cost, this solution has
its own extra challenges),
or the company gradually gains the knowledge and expertise (and the
necessary tools) to
handle these tasks. So in this respect I view the last wish-list item
as not being compatible
with the rest of the list items. And finally, when you go through the
exercise of obtaining
the suitable tools and playing with them and a few designs have been
completed successfully,
you may find that you do not need the tools for certain tasks, because
you already gained
the knowledge so that you know what to do without simulations.
I hope this helps.
Regards,
Istvan Novak
SUN Microsystems
Aleksandr Oysgelt wrote:
> I am starting to look at power integrity tools and would like to know what
> is available and whether there are any "slimmed down" versions that can do
> what I need at lower price than full-blown Sigrity-like packages.
> Currently we use "spreadsheet" approach where we analytically compute
plane
> impedance based on self+mounting impedances of various types of caps plus
> plane impedance. This is fairly accurate to 100MHz or so. However, it
does
> not help us at all in figuring out where the plane resonances are and
where
> we should place caps to help with them.
>
> So, here is my wishlist. I don't require all to be addressed by same
tool.
> Depending on price - I'd be happy to get just some of the bullet points
and
> not all.
>
> - Figure out plane impedance curve
> - Figure out plane resonance - where should we place caps
> - Figure out resistive losses plus current distribution
> - Be able to figure out tradeoffs between using higher capacitance
> plane/ground structure vs using large number of caps vs using different
> types of caps.
> - Do pre and post layout simulation
> - Have non-SI expert be able to run the tools
>
> So far Sigrity and HyperLynx PI (just announced) seem to do the job but at
> quite substantial cost. I keep asking myself a question - what problem do
> they solve vs cost (in license fees and training/simulation time) vs using
> spreadsheet/simple simulation. So far I don't have a clear answer - thus
I
> am asking you for other ideas on approaching the power integrity problem.
>
> Thank you
> Aleks
>
>
>
------------------------------------------------------------------
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 technical documents are available at:
http://www.si-list.net
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 technical documents are available at:
http://www.si-list.net
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 technical documents are available at:
http://www.si-list.net
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:
