On Mon, 29 Dec 2003 03:01:38 +0000, email@example.com wrote:
>Earth-grounding the shields when they leave the shack helps that
Well, maybe, and maybe not. Consider that many shacks are located such
that a "ground" conductor will have considerable inductance and be some
significant fraction of a wavelength at many frequencies of interest,
thus the ground conductor can act as much like an antenna as anything
else in the system.
I think Tom's comments re: estabilishing a very low impedance bond
between the common reference for all filters and the common for the
equipment to be protected are right on the money. In essence, the
filters need to minimize the potential between the lines and the
equipment, and prevent the flow of noise current into the equipment.
But also, it is critical to avoid exciting a pin 1 problem in the
connection to the equipment ground. In other words, the equipment
ground must be to the chassis, not so some common point on the circuit
board that excites common impedance coupling.
I also strongly concur with Tom's comments re: the impedance reference
for the effective attenuation of filters and how they work with
termination to real equipment. The industry generally specifies a 50
ohm termination impedance for such measurements, and the only reasons I
can see for are 1) you have to pick some number; and : 2) all of the
common test equipment is built around 50 or 75 ohms. But the relevance
of these measurements to real world attenuation is tenuous at best. I
find it comical (in the tragic sense) that the noise attenuation of
large power isolation transformers rated for thousands of VA is
generally specified for a 50 ohm termination.
I also concur with the double edged sword Charlie mentions, whereby
noise current is dumped into the "ground" conductor with the Alice in
Wonderland belief that this will somehow make it go away. In fact, one
problem we frequently encounter in audio and video systems is that so
many EMI filters are dumping noise into the "ground" conductor that a
voltmeter or scope will quickly reveal several volts of noise between
any two "grounded" points. Folks forget that the connections between
these "grounds" have moderate resistance and considerable inductance.
They also forget that a typical ground electrode (that is, some big
honking object driven into earth) may have an impedance to mother earth
of several ohms at best and often tens of ohms.
Given this reality, dumping noise into "ground" can cause more problems
than it solves. A prime example of this is a dumb MOV wired as a surge
suppressor. It dumps a transient into "ground," raising the potential
of the "ground" by Ohm's law. Any equipment connected to that "ground"
will also have its potential raised, and if it's raised enough, may
fail catastrophically. In the pro audio world, we've been using a
family of series-mode surge suppressors that avoid this problem by
storing the transient in a high voltage LC network, then discharging it
slowly back into the line. http://www.surgex.com. No, I have no
connection to them and don't sell them, but I specify them for all my
systems and use them in my office, home, studio, and ham station.
Jim Brown K9YC
RFI mailing list