Re: [SI-LIST] : 20-H Rule for Power Planes

Michael E. Vrbanac (mvrbanac@cisco.com)
Wed, 26 May 1999 12:46:26 -0700

Bill,

re: test data
I cannot divulge specific test data for proprietary or confidential matters.
Most everyone who works for a corporation knows this. My apologies.

re: conclusively relate 20H design to EMC compliance
Absolutely, although it might actually be xH depending on what I am doing
and what I need to accomplish. In some cases, it simply doesn't matter.
Perhaps that is the case with the equipment that you are dealing with.
Perhaps the rule is too conservative for what you need to do. As I have
said before, it truly depends on the design and what you are trying to
achieve.

re: EMC compliance ... in general
Any astute person knows that EMC compliance is a system issue and
many factors go into a failing/non-failing condition. It doesn't take a rocket
scientist to know that the spectral profile will be different when one system
uses a "Faraday cage" approach for an enclosure and a "non-Faraday cage"
approach and anything in between. I never try to prove anything by a
compliance test or set of those types of tests because there are too many
variables and too many things cloud the data output. I look for achieving
the "performance" that I need for the designs I do. A compliance test is the
last thing I would do to do that with this sort of thing. To me, that's like doing
a "go-no go" test on a very critical test requiring a large amount of data...
its incompatible with my goals.

One last thing on EMC compliance... if you mean meeting the FCC/CE mark
requirements, etc. consider this... There are EMC requirements that are not
found in regulations.... they are the inherent requirements of the circuitry being
designed. Bottom line... no matter how anybody feels about it, if the "circuit
doesn't work", there is no need to do the regulatory tests....

re: your being a naysayer... until you see an EMC difference...
Ok, that's fair. You must be convinced within your own sphere of influence
to do what you feel is appropriate. Within mine, I'm convinced. Physics,
experience, and many years of success are ample proof enough for me.

re: running 1 GHz edge rates...
Hmmm .... that's pretty slow.... and sounds like digital only, also <grin>

So don't let anyone make you follow the 20H rule if you aren't convinced of
it... just be sure to recognize that someday you'll may run up against it
and need to do something about it and be prepared for it. In the end, I hope
that the discussion will have assisted someone out of a "jam" perhaps even
one of the naysayers... which is the best I can hope for, I suppose.

Regards,

Michael E. Vrbanac

At 12:10 PM 5/26/99 -0500, Bill Dempsey wrote:
>Michael,
>
> What test data do have from either personal or company experiments
>to substantiate the 20H rule that you stated so eloquently? Have you
>been able to conclusively relatea 20H design to EMC compliance and a
>non-20H design failing?
>
> Until then I must agree with the nay sayers as I have never been able
>to "see" an EMC difference. Of course I'm only running 1 GHz edge
>rates and am not looking above that point...
>
>Regards,
>
>Bill Dempsey
>DNA Enterprises, Inc.
>
>Michael E Vrbanac wrote:
>>
>> Mark,
>>
>> Essentially the 20H rule derives from Gauss' Law for the Electric
>> Field which simply states "the net electric flux density leaving an
>> object is determined by its volumetric electric free charge density on
>> it surface". I am sure that more "proof" is unnecessary to verify
>> this as a physical phenomenon.
>>
>> The point that I believe applies to this question is "how much" is
>> important in "the design situation" not whether the situation exists
>> or not. The point of the 20H rule is to provide an all-around
>> recommendation for a wide variety of designs which can be encountered.
>> Obviously, if the power plane is closely coupled to the ground plane,
>> the shorter the actual "20H" distance will be as I believe has been
>> pointed out in another response to your question. It should be
>> important to note that while the actual distance depends on the
>> "conditions" present and the susceptibility of components/circuitry to
>> those conditions, this principle should not be ignored. There can be
>> cases made for 10H (or even 100H!) depending on the situation.
>>
>> I am puzzled that you have not been able to find any "data" on this.
>> Perhaps you are being too literal... Here's a example of what I mean.
>> Check Dr. Johnson's book "High Speed Digital Design" in Chapter 5, pg
>> 190 and 191 I believe or thereabouts... While it doesn't specifically
>> state the 20H rule, he gives an excellent equation on estimating
>> crosstalk levels from a certain distance away from a conductor (he
>> even supplies a very useful equation!). Now while that is not
>> specifically a 20H rule dissertation, the fundamentals are quite the
>> same when trying to estimate stimulus levels at the plane edges for
>> re-radiation. From there, some meaningful outcomes with regard to this
>> question can be fairly easily obtained. Again, the evidence and
>> analytical work is out there for those who can see it. I would
>> encourage anyone trying to work this through to consider taking a bit
>> of time to think through the ramifications of what these men who have
>> "explored" this territory before us were trying to tell us and perhaps
>> take what they have done and expand the knowledge further (i.e. sounds
>> like a great paper to write, too!)
>>
>> re: fringing fields too small... GHz range signals
>> Well, some of us work with stuff like that.... and with digital, too.
>> And we need to pay attention to electromagnetic emissions and immunity
>> as well.... Would I need to apply that technique to what I do....
>> you bet. Do you? Only you can answer that....
>>
>> Regards,
>>
>> Michael E. Vrbanac
>> mvrbanac@cisco.com
>>
>> >
>> > Now and again I come across references to the "20-H Rule" for reducing =
>> > radiation from power planes. This rule states that the power plane =
>> > should be smaller than the ground plane; The power plane edges should be =
>> > back from the power plane a distance of 20-times the plane spacing. =
>> > This reduces fringing fields from the power plane and reduces coupling =
>> > to adjacent planes and free space.
>> >
>> > Best I can tell, this rule originated with Mike King. The earliest =
>> > reference I found is Mark Montrose's "Printed Circuit Board Design =
>> > Techniques for EMC Compliance," pg. 26. I have not found any numbers - =
>> > analytical, simulation or measurement - which indicate the effectiveness =
>> > of this technique over frequency. Intuition (a dangerous thing for this =
>> > digital designer to rely upon) tells me that the dimensions of the =
>> > fringing fields are small, thus only affecting GHz-range signals. Is =
>> > this technique currently only of interest to cell 'phone designers, or =
>> > do we need to begin applying this technique to digital PBW design?
>> >
>> >
>> > Mark Freeman
>> > msf@stratos.com
>> > Stratos Product Development, LLC
>> >
>> >
>> >
>> > **** To unsubscribe from si-list: send e-mail to majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****
>> >
>> >
>>
>> **** To unsubscribe from si-list: send e-mail to majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****
>
>**** To unsubscribe from si-list: send e-mail to majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****
>

**** To unsubscribe from si-list: send e-mail to majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****