# RE: [SI-LIST] : 20-H Rule and Self-Resonant Frequency of Power Planes

From: Chris Rokusek ([email protected])
Date: Wed Apr 19 2000 - 16:37:58 PDT

Mike,

I have a additional question limited only to the first "FIRST" mechanism you
described below--where fringing (NEAR) fields are less "bloated" when using
20-H Rule:

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?

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.

(Of course this question is moot since there's ALWAYS something to couple
to.)

Best Regards,

Chris Rokusek
Innoveda

> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]]On Behalf Of [email protected]
> Sent: Wednesday, April 19, 2000 3:49 PM
> To: [email protected]
> Subject: Re: [SI-LIST] : 20-H Rule and Self-Resonant Frequency of Power
> Planes
>
>
> Mary and Jim:
>
> I was waiting to see some further questions about the
> "clarification" on this
> topic presented a few days back. There are several statements in the
> clarification that add to the confusion. There are perhaps THREE
> fundamental
> effects going on that are being erroneously integrated and (IMHO)
> more confusion.
>
> FIRST, the setback of a power plane from the edge of a nearby
> ground plane is
> an intuitively obvious method to contain EM energy propagating
> between the
> two planes. When propagating waves (picture TEM) hit an unfettered edge,
> electric fields will literally spray into the surrounding ether
> and attempt
> to seek out a return reference. Although an inefficient process for this
> condition, significant (relative to allowable limits) radiation is the
> typical result. Setting the edge of the power plane (selected
> because it is
> typically the noisiest and most isolated from chassis or earth reference)
> back by a respectable ratio to the dielectric thickness between
> the planes
> (20H is arbitrary, but easy to remember) allows the electric
> fields hitting
> the plane edge a considerably better picture. Specifically, the sprayed
> fields immediately see a nearby ground reference plane on which
> to terminate.
> Not only that, but the concentrating factor of the relative dielectric
> constant of the dielectric (which must also be continued beyond the power
> plane edge for greatest effectivity) literally sucks up the potentially
> radiating field lines and captures/contains them. This method of field
> interception (and therefore containment) is particularly
> effective at higher
> frequencies. An excellent verification of this effect was
> published in the
> Motorola MECL System Design Handbook (one of the true "bibles" of signal
> integrity engineering) way back in 1980 where it was demonstrated
> that the Zo
> of a microstrip line over a limited width ground trace was essentially
> identical to using an infinite ground plane if the ground
> reference trace was
> >3 times the signal trace width. (For completeness, this finding only
> applies for signal trace widths greater than the dielectric
> thickness between
> the two traces.)
>
> Therefore, the recessed edge of the power plane stands on its own
> merit as a
> simple, viable method of reducing the EM radiation launched from a
> non-recessed planar construction. It's a good thing.
>
> SECOND, the INTERPLANAR SRF (self-resonant frequency) of the
> dielectric-loaded structure will be (because of distributed
> bypass capacitors
> and component pin parasitic capacitances) much lower than the simple
> calculation using 1/(sqrt Er) and the physical dimensions of the printed
> circuit board (PCB). No matter what one does with a few tens of
> mils around
> the outer edge of the power plane, it will have little affect on the
> interplanar SRF. Enough said.
>
> THIRD, invariably there will be some EM energy coupling to the
> SURFACE of a
> reasonably stuffed PCB structure. This effect is accomplished by
> parasitic
> leakage from component leads, connector structures that rise from the PCB
> (either sockets or pins), surface traces, etc. This energy will
> propagate
> associated currents (including harmonics) on the planes nearest to the
> surface of the PCB. Because the surface dielectric material
> in some cases) is thin, and interfaces directly to air (not
> another plane or
> conducting medium), the surface dielectric has negligible affect on the
> magnetic fields that are created in the air adjacent to the PCB
> surface.
> Excitation of these resonances reasonably accurately based on the
> physical
> dimensions of the PCB excluding dielectric effects, and
> efficiencies can be achieved. Note that this SURFACE SRF will NOT be
> directly loaded/lowered by bypass capacitors and may behave more
> like a solid
> planar conductor referenced to earth ground. This condition is
> similar to
> surface currents flowing on cable shields, which are typically excited by
> just such a mechanism... we call it common mode excitation... and
> we pay the
> penalty for it.
>
> BOTTOM LINE: EMC/EMI is a messy, complicated science subject to much
> interpretation (and confusion). The above is my partial take on the
> apparently mixed subjects discussed under this one topic.
>
> I hope everyone has a great three-day weekend.
>
> Respectfully,
>
> 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 http://www.qsl.net/wb6tpu
> ****
>
>

**** 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 http://www.qsl.net/wb6tpu
****

This archive was generated by hypermail 2b29 : Thu Apr 20 2000 - 11:36:15 PDT