[SI-LIST] Re: AW: Re: Different Simulation Results
- From: "Yuriy Shlepnev" <shlepnev@xxxxxxxxxxxxx>
- To: "'Ing. Giancarlo Guida'" <gianguida@xxxxxxxx>, "'Havermann, Gert'" <Gert.Havermann@xxxxxxxxxxx>
- Date: Wed, 17 Mar 2010 10:13:08 -0700
The question is how to distinguish the errors due to the simulation setup from
the errors due to limitations for the simulator. It raises very important
subject of software benchmarking. There must be a set of problems solved with
known accuracy. Fifteen years ago James Rautio suggested to use simple
strip-line configurations with the characteristic impedance exactly 25, 50 and
100 Ohm - the test is still not popular and there are reasons for that. There
is exact formula for a strip-line with ideal strip and filled with vacuum.
S-parameters of a segment of such line can be calculated with very high
precision and can be used as a benchmark for numerical tools - see more on that
in:
J. C. Rautio, "An Ultra-High Precision Benchmark For Validation Of Planar
Electromagnetic Analyses," IEEE Tran. Microwave Theory Tech., v. 42, N11, 1994,
p. 2046-2050.
There are similar standards for the idealized coupled strip-lines based on
exact formulas and on computations with extremely high precision. Such
idealized tests show how well the strip edge singularities are treated by your
tool (important for the characteristic impedance computation) and how well it
approximates propagation constant of simple TEM-eave. If error is large, it is
either the accuracy setup or problems with the algorithm. Such tests should
pass both in static and in electromagnetic tools.
Obviously, there are no such "idealized" configurations in real life and we
need to use some numerical algorithms to compute parameters of traces on PCB.
Some structures on PLRD-1 board from Teraspeed can be effectively used as such
"benchmarks" - see the results of such investigation in
App Note #2009_06 at http://www.simberian.com/AppNotes.php
To solve the Mohamad's dilemma, I would suggest the following steps (in
addition to the benchmarking):
1) Try to increase simulation accuracy in both tools and see if the results are
converging. The tool that is not converging may produce questionable results.
Note that some tools may also converge to a wrong solution - it is so called
relative convergence phenomenon. Make sure that the problem statement is
similar in both tools (geometry, similar dielectric and conductor models, same
normalization of S-parameters).
2) Compare the results with the results obtained with a third tool. Make sure
that the "benchmarking" tool does not have major limitations for the given
configuration and frequency band.
3) Compare the simulation results with the measurements. This is the most
difficult step. Note that it might be particularly challenging to compare the
reflection loss for the transmission lines due to the coaxial to planar line
transition discontinuities. You have to simulate the transitions or de-embed
them. Both options are relatively difficult and error-prone. The simplest way
is to compare the reflection-less generalized modal S-parameters as suggested in
http://www.simberian.com/AppNotes/DesignCon2010_Paper2807.pdf
(see also
http://www.simberian.com/AppNotes/MaterialParameterExtractionWithSG_DesignCon2010_Final.pdf).
Best regards,
Yuriy Shlepnev
www.simberian.com
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of Ing. Giancarlo Guida
Sent: Wednesday, March 17, 2010 7:52 AM
To: Havermann, Gert
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: AW: Re: Different Simulation Results
Gert,
this is exactly my point...
I fully agree with your comments.
It is important to have a knowledge about the tool you are using and
about the phenomenon you want model
and it is also essential to recognize that each tool has its own field
of application, limitation, advantages and disadvantages.
Giancarlo
Havermann, Gert ha scritto:
> Giancarlo,
>
> it is not HFSS or SIwave making mistakes, its the user. But on the other
> hand, some tools make it easy to create errors, and some tools prevent
> errors. Therefore it is key to know the tool, and to know what you are doing.
> And even if you think you know what you are doing, verify that you did it
> right.
> Every simulation engine has some simplifications or approximations build in
> to speed up the simulation process (things like limited frequency range, or
> approximations of boundaries, material models with fixed parameters....).
> Knowing those will help to use the tool in the right way.
>
> Two good examples are:
> ADS: if you use the math function to calculate TDR impedance from
> S-Parameters, you have to specify: ref.Imp, time delay, Tstart, Tstop and the
> window function. Many users don't know that the Ref impedance is defined at
> the impedance at Tstart, wich means the plot os forced to the ref.Imp at
> Tstart. If you have S-Parameters of a 110 Ohm trace, refImp is 100 Ohm, and
> Tstart is a positive number, then you will see a flat 100 Ohm TDR plot.
> Chosing a negative number will result in 110 Ohm.
> CST: CST uses different mesh engines. if impedance matters, you must use the
> PBA mesher instead of the FPBA because the latter requires a much finer mesh
> (approx factor 3) for the same acuracy.
>
> For most of the problems (especially a simple trace) you will find a way to
> get the right answer, but as I said, some tools require more work to get
> there.
>
> BR
> Gert
>
>
>
>
>
> --------------------------------------------------------------------------
> Absender ist HARTING Electronics GmbH & Co. KG; Sitz der Gesellschaft:
> Espelkamp; Registergericht: Bad Oeynhausen; Register-Nr.: HRA 5596;
> persönlich haftende Gesellschafterin: HARTING Electronics Management GmbH;
> Sitz der Komplementär-GmbH: Espelkamp; Registergericht der Komplementär-GmbH:
> Bad Oeynhausen; Register-Nr. der Komplementär-GmbH: HRB 8808;
> Geschäftsführer: Torsten Ratzmann
> -----Ursprüngliche Nachricht-----
>
>
> Von: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] Im
> Auftrag von Ing. Giancarlo Guida
> Gesendet: Mittwoch, 17. März 2010 14:41
> An: Istvan Novak
> Cc: Mohamad Haghtalab; si-list@xxxxxxxxxxxxx
> Betreff: [SI-LIST] Re: Different Simulation Results
>
> Hi folks,
> I seriously doubt that software like HFSS and SI wave can make a mistake
> simulating a simple configuration like two coupled parallel microstrip lines.
>
> I would suspect that one of the model have something wrong; I bet that any
> Ansoft AE could easily help on this.
>
> By the way, with HFSS it is possible to simulate a transmission line
> structure of an arbitrary length by modeling a little piece and using the
> deembedding feature, provided that the structure is uniform.
>
> Regards
> Giancarlo
>
>
>
>
> Istvan Novak ha scritto:
>
>> At what frequency do you see this difference, and was the length the
>> same in the two simulators? In SiWave you can easily simulate very
>> long traces, whereas in HFSS your maximum length might be just a few
>> hundred mils.
>>
>> Regards,
>>
>> Istvan Novak
>> Oracle
>>
>>
>> Mohamad Haghtalab wrote:
>>
>>
>>> Hi all
>>>
>>> I've simulated two coupled parallel microstrip lines with both HFSS
>>> and SIwave(from
>>>
>>> Ansoft), there is good agreement between the SHAPE of Scattering
>>> Matrix results in two
>>>
>>> softwares,BUT in the LEVEL of them there is about 15 - 20 dB
>>> difference,what do you
>>>
>>> think about this?
>>>
>>> Regards
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>> ------------------------------------------------------------------
>> 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
>>
>> 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
>
> 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
>
> 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
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
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
