[Top] [All Lists]

Re: [RFI] Eliminating TVI

To: "Pete Smith" <n4zr@contesting.com>, <rfi@contesting.com>
Subject: Re: [RFI] Eliminating TVI
From: "Tom Rauch" <w8ji@contesting.com>
Reply-to: Tom Rauch <w8ji@contesting.com>
Date: Sat, 27 Dec 2003 12:21:08 -0500
List-post: <mailto:rfi@contesting.com>
> Do you have a recommendation for a GOOD commercial filter?


I generally build my own because the commercial filters I find either have
output and input bypass capacitors to common grounds making them poor or
useless for common mode rejection or they have bifilar windings with poor or
NO shunt caps between the lines making them poor at differential mode

I find by far the most important and profound TVI and/or RFI change comes
from tying all the lines entering and leaving the area of the device
together with low RF impedance to prevent RF from looping through the
sensitive device (or device transmitting RFI). A high impedance for common
and differential modes on the leads from that point to the device or device
cluster (if you have several devices interconnected) is second.

Just adding a few beads on a cable is often like "spitting in the wind".
Most devices and cables already have several hundred ohms of series
impedance in the RF loop, so  a few hundred more ohms often does very little
to change the system.

The same goes for cable shielding.  There is almost never any change by
shielding that can't be accomplished better with the simple addition of a
couple cheap capacitors and a few beads at the device input. People just
have to get away from the mindset that parallel windings on a core (like an
audio line case) reduce differential problems, and that TV deflection yoke
cores all make good HF cores.

Cores have a peak in permeability at a certain frequency, and you certainly
don't want to use a core that peaks at 100kHz to suppress a 7 MHz signal.

Certainly audio inputs or outputs are NOT very well balanced for RF and do
not have a certain differential or common mode impedance that relates at all
to the audio impedance. In my experience you need to hit the device with a
bypass (if possible) and  isolate it for both common and differential mode

I keep repeating this story because it is a good example. I was hired to
clean up an apartment complex that was located right over what used to be
the radial system of a high power AM and FM BC station. The tower was just
on the other side of a fence from one of the buildings. I could actually
draw a pretty good RF arc when the CATV cable was touched to the power line
ground in many apartments!

The CATV company had gone through all the TVI filtering and triple shielded
cables, but they totally missed the point. Every filter they installed had a
direct path for common mode from input to output, and the extra shields did
nothing at all over normal cables.

I had all the cable ripped out and new cables routed parallel with the power
lines using normal CATV cables. The cables went to bulkheads at the TV sets
where the shield was grounded to the power strip, and the lines were
bypassed to the safety grounds. The CATV cables were routed past breaker
boxes and grounded to the boxes. ALL the high pass filters were tossed in a
box, and all the triple shield was put in a dumpster.

Out of hundreds of TV sets, only a handful remained a problem. Out of those,
a line choke and antenna ground isolation cleaned all up except one. That
one TV had RF into the internal wiring, and could not be cured. Of course it
was in an apartment that had such strong fields the telephones and digital
clocks wouldn't work.

The real cure is providing a path between the conductors entering the area
of the device so the bulk of current can bypass the sensitive device, and
then (if that doesn't work) we need to choke off the differential and common
mode on any leads we can.

It is only when we have a lead we can't mess with (like a high impedance
very low level line) that things become a headache. There are devices that
are not correctable, but they are not common.

When you think about this, most of the things we do for lightning protection
also work for RFI.

73 Tom

RFI mailing list

<Prev in Thread] Current Thread [Next in Thread>