RE: [SI-LIST] : 20-H Rule and Self-Resonant Frequency of PowerPlanes

About this list Date view Thread view Subject view Author view

From: Chris Rokusek ([email protected])
Date: Fri Apr 21 2000 - 10:07:26 PDT


Concerning radiation from the edge in absence of any other traces/metal
besides the pair of planes:

It sounds like you're saying that one of the launching antennae (and perhaps
the dominant when no other traces/metal are present) is really ALONG the
edges of the plates rather than ACROSS the edges of the plates. Is this
interpretation correct?

It seems the 20H rule would have little impact on this mechanism involving
transfer from ACROSS plates to ALONG edges. True or false?

If this discussion continues much more we should start attempting to name
each different mechansim at its point of introduction since its starting to
get complicated. E.g. Parrellel-Plate-Edge-Orthognal-Transfer (PPEOT).


Chris Rokusek

> <<Do you know if the fringing (NEAR) fields in a board NOT using 20H can
> radiate DIRECTLY into the far field significantly given typical excitation
> (say <500mv up to 1GHz) or must these NEAR fringing fields couple onto a
> more efficient far field radiator first?>>
> I'll bound your question so as to minimize possible misinterpretations of
> " a board..." If you are referring to radiation from traces on the
> surface of a PCB that are well inboard of the PCB edges, they
> will radiate
> the same independently of any power plane setback. (You may
> check out the
> IBM paper presented about three (four?) years ago at the IEEE EMC
> Conference
> in San Diego, CA, for the effects of trace spacing from PCB
> edges.) If you
> are (as I suspect) referring only to those emanations from the PCB plane
> edges (without any power plane setback), YES, direct radiation
> will result,
> and ALSO additional radiation will occur with coupling. When the
> interplanar
> TEM wave hits the open edge, it does not see an infinite
> impedance. Rather,
> the energy wave sees an impedance transformation from the low
> (transmission
> line) impedance of the planar structure to the 120*PI=377 ohms of
> free space.
> This is equivalent to a sizable impedance mismatch which will
> normally yield
> poor energy transfer; however, the fields (and currents) "spill
> over" on the
> copper edges of each/both planes and transfer energy (through continuing
> propagation) in both directions along the edges. Some energy
> also folds back
> on the outside surfaces of the planes. Henry Merkelo (Heads up
> EMC studies
> at U. of Illinois, international lecturer, knows his stuff) has produced
> several excellent papers over the last few years that uses software to
> illustrate these effects. This added effect increases the
> efficiency of the
> coupling to free space (as well as to other traces that may radiate more
> efficiently) and results in increased levels of radiation from
> the PCB. The
> worst case of radiation induced from this action will be when the TEM
> harmonic(s) coincide with the surface/outside resonance of the
> PCB structure.
> In that case, substantial radiation will result. This is yet another
> example of why potential EMI sources are best suppressed at the source,
> because there are multiple mechanisms of "leaking" RF energy from a PCB.
> As an added note, the common manufacturing practice of extending the PCB
> dielectric beyond ALL PCB copper layers captures more of the
> field lines at
> the abrupt impedance interface and reduces the radiated energy
> somewhat (but
> not nearly as much as the 20H setback).
> With consideration for all the preceding SI Reflector discussions
> on this and
> related topics, do not forget that much of this potential
> radiation can be
> suppressed or contained by good bypassing (and/or dissipative absorption)
> around the edges of the PCB. By forcing a voltage minimum (as
> best as good
> bypassing can) at the PCB edges, the PCB resonant waves (representing the
> most efficient radiating frequencies) are forced to voltage
> maximums on the
> more lossy interior of the PCB. Any fields that try to radiate
> more easily
> find termination somewhere on the PCB and are thus reduced. I
> extolled the
> virtues of this phenomenon in my presentation at the EMC '98
> Colloquium and
> Exhibition, "PCB Design Issues at the Macroscopic Level." The
> same concepts
> are carried forth in recent papers by Larry Smith and Istvan
> Novak, both at
> Sun Microsystems (although on different sides of the continent).
> *********************
> continue...
> <<As the planes move further apart from each other I expect the "far field
> direct radiation" to increase but I'm curious that if planes are very
> closely spaced that it is possible to radiate directly.>>
> Your expectation regarding further separation of the planes is valid.
> Referring to the discussion presented above, the interplanar transmission
> line impedance is raised by the further separation which provides a more
> efficient impedance match to the impedance of free space; hence,
> more energy
> transfer (through radiation) will occur. A good illustration of
> this effect
> is evidenced by the easily demonstrated (and measured) radiation from a
> single microstrip. Measure the radiation at any given frequency
> from a 50
> ohm line, then repeat the same measurement for a 100 ohm line. The Zo
> increase was 2:1 or 6 dB, but the measured radiation from the 100
> ohm line is
> 8 to 10 dB higher than the 50 ohm line. FYI, radiated emissions
> prediction
> software (such as EMCAD1 from CKC Labs) predict this same result.
> For the
> planar case, since the impedance mismatch is on the order of 200
> to 400 to
> one, I would expect a more linear correlation rather than an
> amplified one.
> That is, a one ohm planar Zo increasing to two ohms would
> probably give 6 dB
> increase in energy transfer.
> Hope this helps.
> Mike
> Michael L. Conn
> Owner/Principal Consultant
> Mikon Consulting
> *** Serving Your Needs with Technical Excellence ***
> **** To unsubscribe from si-list or si-list-digest: send e-mail to
> [email protected] In the BODY of message put: UNSUBSCRIBE
> si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
> si-list archives are accessible at
> ****

**** To unsubscribe from si-list or si-list-digest: send e-mail to
[email protected] In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at

About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Wed Nov 22 2000 - 10:50:04 PST