[SI-LIST] Re: 2D vs 3D EM based signal integrity simulators
- From: Hany Fahmy <hfahmy@xxxxxxxxxx>
- To: "'bradb@xxxxxxxxxxx'" <bradb@xxxxxxxxxxx>, "'si-list@xxxxxxxxxxxxx'" <si-list@xxxxxxxxxxxxx>
- Date: Fri, 15 Jan 2010 14:25:34 -0800
Momentum does not to simulate voids only on an infinite plane.
Moment-method is not new as a full-wave technique: the litreature is full of
analysis in IEEE-MTT and AP for the methodology.
Moment method can still model the surface current on finite planes while still
model the voids on the same plane.
U can still model the plane with electric-surface current elements and also
model the voids with magnetic currents to get maximum accuracy.
I understand that many tools does not model the voids, I will share with u guys
S-parameter correlation for HFSS, Momentum and CST all correlate very well to
VNA measurements: such voids along with planes, vias, via-stubs, surpentines,
are all need to be modelled using full-wave well-known techniques such as
Moment-method, FEM, BEM, FDTD, and such full-wave techniques.
I agree that best way to find out whether the method is good or bad is to
correlate with lab data: for passive interconnects, best way is to use VNA to
get fast correlation.
Just my 2-cents.
----- Original Message -----
From: Brad Brim <bradb@xxxxxxxxxxx>
To: Hany Fahmy; si-list@xxxxxxxxxxxxx <si-list@xxxxxxxxxxxxx>
Sent: Fri Jan 15 12:26:43 2010
Subject: RE: [SI-LIST] Re: 2D vs 3D EM based signal integrity simulators
hello Hany,
Want to provide a note of caution concerning the application of "planar" EM
simulators in the manner you mentioned. If, as you suggest, you define a
layer as an infinite plane and mesh only the voids in that plane for
magnetic current flow then you may not be simulating the design you intend.
This is especially the case for broadband Power Integrity analysis with more
than one plane.
Such planes are infinite and therefore have an infinite capacitance between
them at DC. This implies a virtual DC connection amongst all planes - think
of it as a DC global ground. To see the effects of infinite planes you can
build a simple design with 4-layer stackup; microstrip on top/bottom with
two infinite planes between. On the top and bottom define a length of
microstrip feedline with a port on one end and the other end connecting to a
via that goes from the top to bottom layer thought the two planes. Assure
you define an antipad in each plane that are meshed with resulting magnetic
current flow in them. Simulate from a low frequency to a high frequency. You
will find at DC there is connectivity between the top and bottom (i.e.
mag(S11)=mag(S22)=0, mag(S21)=1 at DC). This is likely not your intent. A
real circuit does not have infinite extent therefore you only have finite
capacitance at DC and you should observe a series capacitance at low
frequency for this 2-port circuit with mag(s11)=MAG(s22)=1, MAG(s21)=0 at
DC. The infinite plane result is not "wrong": it is a correct solution to
the infinite plane design you defined. The potential issue is that you may
not be defining the design you really wish to model.
You can view the above simple example (as a circuit simulator does) as a
thru via with relatively small localized inductive and capacitive
parasitics, but you must augment this view the plane capacitance as a series
element in the return path. Infinite planes have an infinite DC capacitance
and yield a thru connection at DC. A finite sized circuit has only finite
capacitance and yields an open at DC.
I believe John Dunn from AWR covered this point briefly in this forum last
year at DesignCon concerning "ground" for EM analysis. If you don't have the
time to setup the planar EM analysis or peruse John's workshop materials I
may be able to find and share with you offline the slides I shared with John
for his forum last year.
If you are doing SI/PI analyses with planar EM simulators you must keep this
subtle issue in mind. Else you may be solving a different design than what
you intended. The simple workaround is to not use magnetic currents
(infinite planes). Of course, for some RF/microwave circuits the infinite
plane approximation is perfectly reasonable and saves a lot of simulation
time that would be required to mesh all [finite] planes.
Momentum is a fine tool, as are the many other planar EM simulators
available. I am somewhat biased, since I helped create Momentum many yeas
ago :-)
cheers,
-Brad Brim (Sigrity)
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx
> [mailto:si-list-bounce@xxxxxxxxxxxxx] On Behalf Of Hany Fahmy
> Sent: Friday, January 15, 2010 10:38 AM
> To: Istvan Nagy; Carrier, Patrick; si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: 2D vs 3D EM based signal integrity simulators
>
> I used a lot Momentum ( a part of Agilent-ADS) tool to
> accurately model the impact of layer-changes along with
> stitching-vias impact and the proximity of these vias to
> signal transitions.
>
> Momentum is a full-wave technique that depends on
> discretizing the metals with surface electric currents, also
> the voids with magnetic currents and solve for these
> currents, therefore, it can capture the impact of signal
> transitions along with the impact of gnd-stitching and the
> proximity of it.
>
> Hany Fahmy (Nvidia Corporation)
-----------------------------------------------------------------------------------
This email message is for the sole use of the intended recipient(s) and may
contain
confidential information. Any unauthorized review, use, disclosure or
distribution
is prohibited. If you are not the intended recipient, please contact the
sender by
reply email and destroy all copies of the original message.
-----------------------------------------------------------------------------------
------------------------------------------------------------------
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:
