Re: shield layer connection

Dr. Edward P. Sayre (esayre@nesa.com)
Tue, 9 Jan 1996 12:21:19 -0500

Fellow colleagues;

The concept of a "EMI shielding layer" has had a lot of play in designs and
literature because it intuitively seems to offer a way of containing
radiating fields of high performance systems. However, if the conditions for
radiation are present, namely that the E and H fields are orthogonal and
forms a Poynting vector ExH which is directed away from the sources, then
shielding layers are of limited value since they are not by definition
closed Faraday shields. Under those conditions, image currents are set up in
the "shielding layers" such that the board will radiate.

However, how much it radiates is set by the radiation resistance of the
signal harmonics which is set pridominately by the "electrical size" of the
source in wavelengths. It is the filtering action of shielding layers more
than any containment property which one is seeing.

The other down side of shielding layers is that they must be connected to
the chassis ground through a very low inductance connection which says it
has very low RF impedance. But, there are high potential isolation
requirements, especially in the telephone industry which make the presence
of shielding layers difficult to fabricate and sometimes results in hi-pot
breakdown.

It is NESA's considerable experience in backplane and board design and EMI
compliance engineering activites which leads me not to recommend the use of
shielding layers for EMI containmnet for either logic cards or backplanes.

Sincerely,

Edward P Sayre, Ph. D., P. E.