[SI-LIST] Re: HFSS S-parameters... issues?
- From: "Ken Cantrell" <Ken.Cantrell@xxxxxxxxxxxxxxxx>
- To: "George Peterson" <george_peterson@xxxxxxxxx>, "Chung-Hsiao Wu" <wuchunghsiao@xxxxxxxxx>, "Ken Cantrell" <Ken.Cantrell@xxxxxxxxxxxxxxxx>, "Cristian Gozzi" <cristian.gozzi@xxxxxxxxx>
- Date: Fri, 11 Nov 2011 11:35:41 -0700
George,
You suspect right on the humor part. Scott and Steve have forgotten more
than most people will ever know. A previous post had called Steve "The
Steve", so I just added Scott to the list.
You can still use integration lines in terminal, it just orients the
E-field. If I don't like the results, I put the lines in and re-run.
Something to check.
Yes, that's what I meant by cannonical. For microstrip: port width: 10w, w
>= h, 5w, w < h, port height: 6h to 10 h. Striplne is different.
I'm not sure what's going on. With a 1 inch microstrip(or stripline), you
shouldn't have any issues. It's probably going to be some little thing that
you typed in wrong. Something simple that you overlooked. We've gone
complex, now I would suggest going simple.
The only time you would need to decrease the tetrahedron size, as The Scott
pointed out, is if you are running at a very high frequency relative to your
geometry, or have complex geometries. I usually keep the radiation boundary
at default, and run the metal/dielectric at the reduced size to keep the
solve time down.
Ken
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx
[mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of George Peterson
Sent: Friday, November 11, 2011 9:35 AM
To: Chung-Hsiao Wu; Ken Cantrell; Cristian Gozzi
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: HFSS S-parameters... issues?
Robert,
Regarding line length, I'm simulating a 1 inch line which solves in about 1
minute for me. I get the same S11 issue when trying 0.5 inch lines though.
I have no problem with using a 2D extractor, but I would like to understand
what is wrong with my current 3D setup. The results I'm seeing are MUCH
further off than they should be.
Ken,
The Ansoft rule of thumb for port size, that I'm aware of, is 10w by 8h. Is
this what you mean by canonical port size? What is the best determinant in
choosing optimum port size?
As for integration line, I'm using Driven Terminal solution, so no line
definition needed. I wouldn't expect Terminal or Modal solutions to be very
different here... but I'll take another look. The HFSS Help file is really
good, but as with all help files there's always that one-bit of information
that's missing
"The Scott", "The Steve" ? I suspect humor lol.
Cristian,
Yes, Driven Terminal solution type. The airbox does indeed touch the
dielectric on the two ends of the transmission line structure where the
Waveports are. The airbox on top-bottom and left-right has a 200 mil gap as
suggested by my Ansoft FAE. When exporting the Touchstone I am
re-normalizing to a fixed impedance. I'll be shocked if order of basis
function makes a difference here, but I'll look into it. I'm currently
using 1st order.
Will you please elaborate on what you mean by reducing the Z0 error
tolerance? Z0 of this simulated structure is about 55.7 Ohms from 1 to 10
GHz, plotted as re[Port Z0] in HFSS.
*Best Regards*,
George Peterson
Signal Integrity Engineer
Jabil Design Services
St. Petersburg, FL
On Fri, Nov 11, 2011 at 3:11 AM, Cristian Gozzi
<cristian.gozzi@xxxxxxxxx>wrote:
> Hi George
>
> I assume you are working with driven terminal solution type in HFSS,
right?
>
> does your airbox touch dielectric edge? or is it far away from
> PCB/dielectric edge?
>
> Which is the accuracy of Z0 in your main setup?
> since you are using waveport, I suggest to reduce the error tolerance of
> Z0 computation for waveport
>
> what do you do when you export touchstone format?
> do you re-normalize model to fixed reference impedance or not?
>
> Try to play with these above parameters!
>
> P.S. don't work with port solver only... otherwise you won't benefit so
> much of adaptive mesh!
> I suggest to have a small model, fix the mesh to highest freq. bandwidth
>
> Use also mixed order basis function!
>
> Hope you can solve your issue
> keep me updated
>
> Cheers
> Cristian
>
> SI & PI Specialist
> Technoprobe Spa
>
> On Thu, Nov 10, 2011 at 4:37 PM, Ken Cantrell <
Ken.Cantrell@xxxxxxxxxxxxxxxx> wrote:
> George,
> Changing the radiation boundary size/type does effect results, and
changing
> the port size effects results. There are three types of radiation
> boundary,
> and the port size effects the number of modes, wider lets in more lower
> frequency modes. I go with their cononical port sizes for a baseline, and
> select the appropriate boundary for the model I'm investigating. Search
on
> radiation boundary and port sizing in the index on the Help menu. Might
be
> a place to start. Did you put in integration lines on the ports? Other
> than that...what The Scott said...with apologies to The Steve.
>
> Ken
2011/11/10 Chung-Hsiao Wu <wuchunghsiao@xxxxxxxxx>
> Hi George,
>
> What is the trace length you are trying to solve in HFSS? The 3D mesh will
> have to refine further when the structure is long and skinny but it will
> takes longer time to solve. Usually 2D extractor is the best for
> transmission lines.
>
> You can try shorter trace in 3D, and cascade them to get to the length you
> want. It's faster and you can get reasonable accuracy if you insist using
> 3D.
>
> Good luck,
>
> Robert
>
>
> Sent from my iPhone
>
> 2011/11/10 George Peterson <george_peterson@xxxxxxxxx>
>
>> Hi Scott,
>> I have not tried adjusting mesh parameters, or order of basis function,
>> but
>> I'll give it a shot. For such a simple structure through only 10 GHz I
>> assumed that whatever defaults are present would suffice.
>>
>> My Ansoft FAE reproduced the problem and his solution was to increase the
>> size of the surrounding airbox. This worked for me on microstrip lines
but
>> appears to fall apart on stripline; that is, larger airboxes don't fix my
>> S11. I would hate to waste his time on a case of operator error if that's
>> what this is.
>>
>>
>> *Best Regards*,
>> George Peterson
>> Signal Integrity Engineer
>> Jabil Design Services
>> St. Petersburg, FL
>>
>>
>>
>> On Thu, Nov 10, 2011 at 1:54 PM, Scott McMorrow <scott@xxxxxxxxxxxxx>
>> wrote:
>>
>> > George
>> >
>> > I've experienced no problems in HFSS V13.0.2 or with V14 beta. If the
>> > port impedance and the impedance of the structure are well-matched, you
>> > might be suffering from insufficient mesh on the transmission lines,
>> > because there are no reflections (and therefore only minute delta S
>> changes
>> > with meshing). I suggest that you speak with your Ansys application
>> > engineer about the problem you are having.
>> >
>> > best regards,
>> >
>> > Scott
>> >
>> > Scott McMorrow
>> > Teraspeed Consulting Group LLC
>> > 121 North River Drive
>> > Narragansett, RI 02882
>> > (401) 284-1827 Business
>> > (401) 284-1840 Fax
>> >
>> > http://www.teraspeed.com
>> >
>> > Teraspeed® is the registered service mark of
>> > Teraspeed Consulting Group LLC
>> >
>> >
>> > On 11/10/2011 1:31 PM, George Peterson wrote:
>> >
>> >> Good Day All,
>> >> Is anybody else noticing unexpected behavior of S-parameters
>> (Touchstone
>> >> models) generated from transmission line structures in HFSS v13.0.2?
>> >>
>> >> I'm performing a linear port analysis in HSpice and when I use
>> Touchstone
>> >> models generated from transmission lines in HFSS I get wacky results;
>> >> specifically, S11 magnitudes vary dramatically\unrealistically and
>> >> expected
>> >> resonances change frequencies or are missing altogether.
>> >>
>> >> When I take that same structure in HFSS and export a W-element model,
>> the
>> >> HSpice results are as expected; for a 1 inch line, S11 peak ripples
>> occur
>> >> at about -20 dB and, using lossless calculations as a rough guide,
>> minima
>> >> match well at multiples of 1/2 lamdba (wavelength). I've also compared
>> to
>> >> models generated in Polar and all transmission lines models exhibit
>> >> roughly
>> >> the same S11 behavior... except Touchstone from HFSS.
>> >>
>> >> Within HFSS, adjusting the size of the surrounding airbox, changing
the
>> >> boundary conditions, adjusting the size of the waveports, and changing
>> the
>> >> technique for modeling dielectrics all effect the outcome that I'm
>> seeing
>> >> but I have yet to see a lot of rhyme and reason.
>> >>
>> >> Can I get a sanity check please?
>> >>
>> >>
>> >> *Best Regards*,
>> >> George Peterson
>> >> Signal Integrity Engineer
>> >> Jabil Design Services
>> >> St. Petersburg, FL
>> >>
>> >>
>> >> ------------------------------**------------------------------**------
>> >> 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<
>> //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<
>> //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:
