On Wed, 3 Feb 2010 10:19:02 -0500, Mike Wetzel wrote:
>Does that sound like a solution that would work?
As usual, "it depends." You need to figure out where the (noise)
current is flowing and prevent it from radiating. If you could
force the noise current into a transmission line, you would
essentially kill radiation from it. Something as simple as twisted
pair would do it -- IF you could run the pair over the entire loop
that carries the current.
What often happens is that designers treat wiring like this as a
DC circuit, with individual "hot" conductors and a common return.
If hot and return are closely spaced, as in a transmission line,
their fields cancel and there's little or no radiation. If they're
separated, they create a magnetic field that's proportional both
to the loop area and to the current, AND the wires act as
antennas. At low frequencies, source and victim are more likely
to have a "near field" relationship (within wavelength/6), so the
coupling is magnetic. Beyond that distance, antenna action
It's also possible that some of this noise is coupling to the
power line, and on control wiring (thermostat, digital readiout,
etc), so I DO recommend a serious ferrite choke on those cables.
Follow the winding guidelines in my RFI tutorial to achieve high
choking impedance at the frequency(ies) where you have the
That's the wiring part of the solution. The other element is the
relative strength of the RF component, which is essentially driven
by the risetime of the pulse (and. like key clicks, can be
"shaped" to soften the edges of the pulse). Your RFI is harmonics
of square waves, so the second part of the solution is to reduce
the strength of the harmonics. It MIGHT be possible to do that
with some ferrite cores on a single conductor carrying the noise
current. Depending on the strength of the current, there might be
some saturation, but if you used a big enough core you might get
enough rounding to reduce the noise.
These are all "outside the box" solutions -- that is, no mods to
the electronics -- but you may need to get inside at least a bit
to get at the wiring.
RFI mailing list