[RFI] RFI Canceling RFI?

[Jim Brown]

On Fri, 10 Mar 2006 07:37:31 -0500, Pete Smith wrote:

>Never did figure out what the mechanism could have been to account
>for it.

Not speaking of your problem or the original post directly, but there 
are some power wiring errors that could account for the sort of 
results being described. Consider outlets or fixtures that are 
miswired so that the green wire (ground) and the neutral are 
interchanged, or the neutral and the hot are interchanged, or even the 
hot and the phase. 

And consider a fixture where the neutral is improperly bonded to the 
green wire. In a properly wired system, the neutral is bonded to 
ground at one, and ONLY one, point -- the place where the system is 
established. That point is either the service entrance, or, if there 
is a local transformer, at the transformer itself. 

When there is only ONE bond, as is required by code, all of the load 
current flows "out" on the phase and "returns" on the neutral. This 
causes the magnetic field to be almost entirely confined to the space 
between the two conductors, and external fields nearly cancel. If 
there is a second neutral bond, the "return" current divides between 
the paths permitted by every "grounded" object in the building. Now, 
the cancellation is quite poor, so the external field is both a lot 
larger in magnitude and spreads out over a wide area. 

Now, consider an electronic device that has a capacitor between 
neutral and ground, and between the phase and ground. In this 
condition, just as in the double-bonded neutral, the high frequency 
return currents are not equal (because in real equipment the 
capacitors, especially the strays, are rarely even close to being 
equal). And just and in the case of the double-bonded neutral, the 
magnetic field does not cancel, and spreads out over a wide area to 
cause interference. 

I am raising this issue to alert the EMC folk on this list to the 
issue, because I am seeing systems where the "willy-nilly" use of 
bypass capacitors on multiple circuits in large systems (like large 
lighting or motor control systems) is shoving huge currents onto the 
green wire and causing serious interference to audio systems. Last 
year, I measured 60A on the green wire in the distro panel of a 
theatrical lighting system in a "megachurch" in Maryland. When we 
asked the Electrical Contractor to find and fix the problem, he told 
the client the condition was "normal."  Huh?

Also think about these mechanisms in the context of RF. Bypassing RF 
to "ground" is not necessarily a good way to reduce radiation of noise 
if the "ground" is also an antenna! Remember that although WE call 
that green wire "ground," Mother Nature calls it an antennna. And she 
ALWAYS wins.

Jim Brown K9YC
http://audiosystemsgroup.com