[RFI] ethernet EMI revisited - questions for the pros
jim at audiosystemsgroup.com
Sat Nov 29 22:07:26 EST 2008
On Fri, 28 Nov 2008 20:22:47 -0500, Mike Tindor AA8IA wrote:
>I'll let you all know what the outcome is when I'm finished. I'll also
>purchase some type 31 snapons to see how those help - they seem to be
>useful and something we all ought to have a few of lying around.
Before you buy anything, study my RFI tutorial if you haven't already.
The issue of 10MB or 100MB may not be as important as the the mechanisms
that are radiating the trash into your antennas. At HF, the primary
mechanism (that is, the loudest) is usually the Ethernet cables acting as a
long wire antenna. That's common mode current, and you must choke it at the
source (each end of an Ethernet circuit is a source, every piece of gear
with a clock is a source, each switching power supply is a source). A few
clamp-ons will NOT kill that common mode current unless you have enough
turns to move the resonance down to the HF spectrum.
Killing the common mode will ONLY kill what's on the cables. Usually some
of the trash (lower levels) is radiated directly by the wiring/circuit
traces inside the boxes. For that, you've got to either shield the box
(usually very impractical), or use the bucket and replace the box. But good
luck finding something well shielded.
The bucket is filled with water. Put the offending device in twice. Take it
As to shielded cables -- that can only help if the shields are bonded at
both ends to the shielding enclosure of the box, or to the common point of
circuit common and the power supply green wire. Lotsa luck with most boxes
in that regard. If you connect the shield to the wrong point, a pin 1
problem in the box is likely to shove RF out on that shield, so you'll need
the chokes to kill it.
More information about the RFI