[SI-LIST] time domain simulation of s-parameter
- From: "Madhavan Swaminathan" <madhavan.swaminathan@xxxxxxxxxxxxxx>
- To: si-list@xxxxxxxxxxxxx
- Date: Tue, 4 Dec 2007 18:10:25 -0500 (EST)
Dear Folks,
A lot of these issues such as passivity, causality, transfer function
causality etc are discussed in the book "Power Integrity Modeling and
Design for Semiconductors and Systems"
www.prenhallprofessional.com/title/0136152066
There is in fact an entire chapter on time domain simulation methods
dedicated to this topic.
There is also a website www.powerintegrity.net where you can download
software and reproduce some of the examples in the book.
Regards,
Swami
> Thank you all for responding to my query.
> Actually no tool is perfect in this world. Each one has its own plus and
> minus points.
> Power SI is limited with number of ports, As number of ports increases,
> its
> a complete mess..
> HFSS is good for higher frequencies, but not good for lower freq. so
> transient simulation of s-parameter from HFSS is certainly not good.
> Q3D is limited to lower frequencies and also on the size of structure.
> Ansoft claims that they have the best algorithms to solve such kind of
> problems, but they put the condition that one should use all the tools
> (HFSS, Q3D, Nexxim) only from ansoft. This seems to me just a marketing
> gimmicks.. nothing else..
> Now let me come to my problem statement.
> I was interested to use ADS for transient simulation after s-parameter
> extraction of the signal nets including power plane from Power SI. I got
> my
> simulation running after lot of experimentation in convolution tab. my
> colleague tried ADS on Power plane related transient simulation after
> s-parameter from Power SI. Initially, he extracted power plane model till
> 2GHz, ADS simulation did not converge. Later he extracted till 5Ghz, then
> he
> got the simulation going with some changes in convolution tab options.
> ADS is able to handle such problems, but we had to lot of manual labour..
> I
> really dont know about the robustness of algo.
>
> Thank you all once again
> --Sanjeev
>
>
> On 12/4/07, Michael Tsuk <mtsuk@xxxxxxxxxx> wrote:
>>
>> > Hello All
>> > I am trying to do time domain simulation using Agilent ADS after =
>> > extracting s-parameter model from Power SI.
>> > But, i am getting lot of convergence issue. Simulation gets terminated
>> > immediately. I do not get any time domain waveform.
>> > kindly suggest a solution.. which tool can solve such problem very
>> > efficiently?
>> > Thanks in advanve
>> > Sanjeev
>>
>> Sanjeev,
>>
>> Although I'm not an expert user of either Power SI or ADS, I have some
>> suggestions that might help you resolve your difficulty. I've been
>> leading
>> the research on transient simulation of S-parameters for the Nexxim
>> circuit
>> simulator team at Ansoft for the past several years. We've had
>> significant
>> success in creating algorithms for robust transient simulation, and the
>> work
>> has deepened my appreciation for the intricacies of this type of
>> problem.
>>
>> To me, it sounds very much like you have a passivity or causality
>> problem
>> with the 3D extraction from your EM tool, which is causing the non
>> convergence in your transient simulator. You may want to try turning on
>> passivity enforcement in ADS, and also switching to a non-convolution
>> based
>> transient algorithm, if that's available. Convolution tends to have
>> trouble
>> with power-distribution simulations, because interesting things happen
>> over
>> a wide dynamic range in frequency. Convolution can miss the
>> low-frequency
>> effects, due to interpolation onto a uniform grid for the inverse FFT.
>>
>> Adding a DC point if you don't have one is important for high-quality
>> transient results, but you'll need to make sure you add a good DC point.
>> I've seen cases recently where a simple "perfect short" was added, and
>> that
>> was non-causal and so messed up the transient simulation.
>>
>> However, there could be passivity or causality problems in the original
>> Power SI data, which might be causing the non-convergence issue.
>> Passivity
>> is relatively easy to check (the maximum singular value of the S-matrix
>> at
>> each frequency must be less than or equal to unity), but causality is
>> harder. This brings up one of the advantages of having both EM and
>> circuit
>> simulation tool coming from the same company: the finger pointing is all
>> done internally. :-)
>>
>> Let me know if you have any other questions. Thanks!
>>
>> --
>> Michael Tsuk, Ph.D.
>> Ansoft Corporation
>> (781) 229-8900, ext. 351
>>
>>
>>
>>
>> ------------------------------------------------------------------
>> 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
>
>
>
--
Madhavan Swaminathan
Joseph M. Petit Professor in Electronics
Deputy Director, Microsystems Packaging Research Center
School of Electrical and Computer Engg.
777, Atlantic Drive
Georgia Institute of Technology
Atlanta, GA 30332-0250
Ph: (404) 894-3340
------------------------------------------------------------------
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:
