Tests performed in labs are one thing, and by their design, they often
fail to expose fundamental problems. One of the most common is failure
of the Green wire to be properly bonded to the shielding enclosure of
the noise source. When this happens, the noise is ON the green wire, and
goes right past any traditional line filters.
There's also the issue of defining common mode. For EMC testing, it's
the voltage between neutral and Green, and traditional line filters are
DIFFERENTIAL filters -- they are not applied to the Green wire, they are
referenced to it! From an RF point of view, common mode is the algebraic
sum of current on all of the conductors of a cable, and it is that
algebraic sum that RADIATES. The only thing that a power line filter can
do, if properly applied at the noise source, is to bond the green wire
properly, and with what my EE profs called a "zero length lead."
As to the DC circuit -- all parts of it MUST be treated as a
transmission line run with twisted pair, as I noted in an earlier post.
It must be treated this way because it's NOT DC, it's DC modulated by a
rectangular wave, and, to minimize dissipation, usually has the fastest
rise and fall times practical, which maximizes harmonic content.
There is a better way -- mathematically shaping of the rise and fall
transitions like Elecraft introduced with their K3 transceiver to
minimize key clicks. Designer Wayne Burdick describes it as "sigmoidal"
or a "raised cosine function."
Bottom line -- this is a SYSTEM DESIGN problem that includes overall
system architecture, the equipment used, and its installation. And as
usual, the chain is as strong as its weakest link.
73, Jim K9YC
On 5/23/2021 6:03 PM, Carter, Michael wrote:
For those who would like to know what are the
conducted and radiated RFI limits in the EU and
U.S., this report (beginning on p. 11) provides
a nice summary:
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
|