[SI-LIST] Re: EMI simulaiton using speed2K

  • From: "Brad Brim" <bradbrim@xxxxxxx>
  • To: "'Rahul G'" <skg212@xxxxxxxxx>, <si-list@xxxxxxxxxxxxx>
  • Date: Fri, 24 Apr 2009 02:07:33 -0700

hello Rahul,

Speed2000 is commonly applied for SI and PI characterization of entire
board/package systems. The systems characterized can be quite large, e.g.
server or router boards with thousands of nets and multiple FPGAs/ASICs.
Just as SPICE results include node voltages and branch currents, the results
of Speed2000 SI/PI analyses include the current flow in nets and edge
voltages of the planes throughout the entire system. Since most EMI analysis
is done in the frequency domain, these time domain EMC "sources" are
converted to the frequency domain. The appropriate Green's functions are
applied to the top/bottom microstrip net currents and equivalent magnetic
edge currents to form the near and/or far fields of the system. The near
fields can be applied in a 3D EM simulator as so-called "compact sources"
(such as Microwave Studio or Microstripes) to characterize enclosures or
other environmental effects external to the system. You might find useful a
paper from the 2008 EMC Symposium by Jin Shi and other authors titled
"Equivalent Radiation Source Extraction Method for System Level EMI and RFI
Prediction". You can also review a presentation available online
'sigrity.com/papers/2008/How to reduce EMI using CAE (LG).pdf'.

There are some things the "hybrid" EM/circuit analysis of Speed2000 does not
address - e.g. heat sinks or other sizeable 3D portions of the system. A 3D
EM simulator such as those mentioned above can be very effectively applied
to characterize these effects. However, 3D EM simulators don't have the
capacity to characterize full boards/packages for typical SI or PI
applications. There are numerous "planar" EM or PEEC based simulators that
can work well for EMC applications with small boards with low/moderate net
count. No single analysis tool is likely to be adequate for an EMC guru.

I've heard many refer to EMC the "dB world" and circuit design as the
"percent world". Where you might be happy with a circuit analysis prediction
within a few percent you are similarly happy with an EMC analysis prediction
of a few dB. There are too many details and potential variations to
characterize a real system for EMC better than this. Even if you could model
the multi-resolution of fine geometric detail in combination with the
electrically large volumes of the entire EMC system, you would still need to
know all the signals in the entire system. This last part is what many EMC
engineers forget about.

Despite the uncertainties and complexities of the system, you can still
apply system level "hybrid" solvers such as Speed2000 subsequent to their
familiar SI/PI analyses to gain significant understanding of the EMC
behavior of your board/package designs. The earlier in the product design
you realize and fix your noisy boards the faster you will make it through
compliance test. Someone in your organization is likely already doing
board/package system level SI/PI analyses - why not leverage the same tools
for a quick prediction of emissions since most of the setup work is already
done :-)


P.S.  You can also quite effectively apply frequency domain "hybrid"
analysis for EMC characterization of board/package systems. For example, a
frequency domain analogy of Speed2000 would be PowerSI. To explore behavior
of the operational system you may need to know (or estimate) the spectral
content of your signals to weight the frequency domain results, since you
can only apply periodic signals for frequency domain analysis of the
board/package system.

> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx
> [mailto:si-list-bounce@xxxxxxxxxxxxx] On Behalf Of Rahul G
> Sent: Thursday, April 23, 2009 11:47 PM
> To: si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] EMI simulaiton using speed2K
> Hello EMI experts
> I need your feedback on using speed2k for EMI simulation for
> board/package.
> is it a complete EMI tool? is there any other better tool in
> industry for EMI simulation?
> Thanks
> Rahul...

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:

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List technical documents are available at:

List archives are viewable at:     
or at our remote archives:
Old (prior to June 6, 2001) list archives are viewable at:

Other related posts: