Jim, Roger, and All:
Jim and Roger have made excellent points in their respectuve postings. I
would just like to add my comments regarding overall RFI issues with
I have spent the past 11 years working in military and aerospace EMI/EMC
design and test. In my case, most of the emphasis was on equipment design
with the goal of passing the required EMC tests on the furst try. That is
because mil spec testing (and even commercial aerospace testing to RTCA
DO-160 specs) is very expensive. You must keep in mind that even the best
designs WILL have radiated emissions. The issue is: at what level and in
which frequency ranges? Because so much equipment is frequently jammed
together in aircraft cockpits or the radio ops areas of military vehicles
or base stations, the equipment must not interfere either with itself or
with other co-located equipment. (If interference occurs between radio
systems because of antenna-to-antenna coupling, then costly co-location
filter networks may be required. These are similar in concept to the
cavities used for repeater systems.)
Commercial gear for normal industrial and/or consumer use has no such
requirements imposed upon it. From roughly 26 MHz on up, the radiated
emissions allowed by FCC Part 15, Class B equipmant will be on the order of
20 to 30 dB greater than allowed for military or commercial aerospace
applications. That's a LOT of extra signal. As has been stated on this
reflector many times in the past, most radiated emissions come from
connecting cables - be they power cords, Ethernet, discrete signal &
control, audio, or whatever. Good equipment design will minimize those
emissions, but so will proper cable design and installation. The postings
already made go into enough detail involving chokes, by-passing, shielding,
and so forth.
My point is that equipment design and cable treatment can only go so far in
reducing emissions that, in many instances, are basic to the product. (A
high speed Ethernet cable WILL radiate harmonic energy - it's part of the
game. The issue is to reduce the radiation as much as possible.) One BIG
part of the EMC control triad that is overlooked all too often is: what is
the separation distance between the offending source (which can be both a
piece of equipment AND all of the cabling associated with it) and the
victim? In most cases, we regard the victim as being our radio receiver.
Now, think about it- is the victim the receiver or the antenna to which the
receiver is connected? I think you will find that in all too many
instances, the real problem is that the antenna (or parts of it) are too
close to the offending source.
Here's some food for thought: Mil spec EMC testing is done with a
measurement distance of 1 meter from the unit under test; FCC tests are
done at 3 meters from the unit under test, and the allowed limits are
higher, as well (for equivalent frequency ranges). That difference alone
is good reason to make certain that you do not locate equipment that is
potentially an RFI source too close to your antennas. Note that the
antennas might very well be outdoors and the equipment indoors, but unless
your house is built like a Faraday cage, having the end of your 80 meter
dipole located just over the roof above your shack might result significant
unwanted signal coupling. The same thing can happen when routers or cable
boxes are located in attic areas or second stories of houses and there is a
40 foot tower adjacent to the house with a tri-band yagi and a couple of
inverted-Vs on it. Sorry - you might lose the RFI battle on that one.
I agree with Jim's comments: if you must do something, do it to the cables
and interconnects. If that is not enough, try getting more separation
distance between the source and your antennas. It might not be easy, but
that's still better than trying to modify a piece of equipment that is
actually designed just fine (met its specs when tested), but you are trying
to make it work to a requirement that is more stringent than the design
There is one instance in which taking action against the equipment might be
practical: high noise power supplies, expecially wall warts. I have
personally seen wall wart supplies (using switchmode technology) in which
there are parts missing from the pc card within. You know the story: the
parts are all there for the required FCC or CE test, then the production
runs leave out those EMC-suppressing parts to save costs. That's very
common with some 3rd party "no name" warts from the Far East. Since you
probably can not get information as to just what the values of those
missing parts are, the most practical approach seems to be to find either a
linear power supply (bigger, heavier, less efficient) with the required
output, or a switcher that actually has all of the parts and does meet
specs. Good luck. I endorse Jim's "bucket treatment" for the offending
pieces of trash.
73 and Best New Year Wishes to everyone,
> [Original Message]
> From: Jim Brown <firstname.lastname@example.org>
> To: <email@example.com>
> Date: 12/30/2010 2:18:14
> Subject: Re: [RFI] Do we neglect problems and fix symptoms?
> On 12/30/2010 11:36 AM, Roger (K8RI) wrote:
> > Chokes work very well in many instances, but I think
> > we some times over look the problem and fix the symptoms.
> My published advice on the topic of RFI is two important premises.
> First, is the assumption that with any commercial product, the best
> solution is generally one that does not require modification of the
> equipment. In other words, the most practical solution is one that can
> be applied "outside the box." There are three good reasons for this.
> Staying outside the box is usually less expensive, it avoids warranty
> problems, and it avoids changes in internal wiring that could result in
> instability or other malfunction. That is, if the designer was dumb
> enough to build a product with poor EMC properties, changing the way he
> has implemented circuit common might make the product unstable.
> The second premise is that we need to understand the fundamental
> mechanisms that are coupling RFI so that we can not only diagnose which
> one(s) are coupling RFI in a given system and apply suitable fixes, but
> also so that designers of equipment can avoid those design errors in
> future products.
> It is generally well known that the most common causes of RFI are Pin
> One Problems, poor internal shielding (and/or poor layout of circuit
> boards and internal wiring), poor interconnect wiring (poorly shielded
> cable, or paired cable that is not twisted), and inadequate filtering of
> inputs and outputs. In some cases, it takes two errors to excite the RFI
> -- poor interconnect wiring PLUS poor filtering.
> While few would argue that eliminating a Pin One Problem is a more
> direct way of eliminating it, most products are built in such a way that
> it is not practical to do so without major surgery. The same is true of
> shielding and circuit layout errors, and the addition of filtering
> components. With most real world products, the only practical fix for
> ALL of these design errors is at the design stage.
> Thus, the most practical "in the field" fixes for most RFI are 1) good
> common mode chokes on all wiring that could act as an antenna, 2) robust
> braid shields on all unbalanced wiring, and 3) twisted pair for anything
> else. And, of course, good engineering practice includes
> chassis-to-chassis bonding of interconnected equipment in close
> proximity, especially when the connections are unbalanced.
> 73, Jim Brown K9YC
> RFI mailing list
RFI mailing list