At 9:01 PM -0700 6/25/03, Scott Townley wrote:
>Good stuff Chuck.
>One very important thing to remember is that the common-mode current
>"sees" a transmission line of unknown impedance. One conductor is
>the outside surface of the coaxial line, and the "other" surface
>(could be multiple) is whatever conductor is nearby...side of a
>house, fence, earth surface, etc.
>So you can get all sorts of interesting impedance transformation
>issues. For example, the "common" common-mode choke consisting of
>ferrite beads surrounding the outer conductor of a coaxial line
>would "like" to see a short circuit "behind" it...so that you have a
>large-ish series impedance followed by a shunt short
>circuit...maximum rejection of signal. If the coax is
>grounded/bonded 1/2 wavelength from the choke position, that will do
>the trick...at that frequency (and all multiples). But at a
>frequency where the grounded coax shield point is 1/4 wavelength
>away from the choke point...the choke is much less useful....
I agree. Because the impedance of a common-mode choke approaches
zero as the frequency approaches zero, it tends to be difficult to
achieve effective common-mode choking at low frequencies. Because
the lowest-frequency band that I use is 80 meters, I placed two
common-mode chokes 70 feet apart -- about one-quarter wavelength for
this band -- on my coaxial line. In effect, the current node at one
choke is transformed (potentially) to a current maximum at the other
choke, which maximizes the effect of the other choke. This
arrangement is probably responsible for the superior, rather than
inferior as one might expect, common-mode isolation that I measured
on 80 meters. On the even-harmonically-related upper bands, the
spacing of these two chokes is the worst-possible, of course. But
another trick, described below, is effective at higher frequencies.
(Did you notice that my best common-mode isolation occurs at the
highest and lowest frequencies?)
>Back in my DoD days when doing wideband HF engineering/testing, we
>would always bond the coaxial shields to ground (where we could) at
>least every 1/10wl (at the highest frequency of interest) between
>exciter and antenna. This kept the common-mode impedance low and
>greatly aided the effectiveness of common-mode chokes used.
I wish I could do this, but my feedline is at least twenty feet above
ground all the way from my antenna to my upstairs shack.
So I inserted a common-mode, L-C-L, T-network, low-pass filter at the
interface between the open-wire line and the 70-ft. coaxial line, as
The 4:1 balun transformer between the open-wire line and coax is a
"current" balun, i.e., a common-mode choke. It is the first "L" of
the common-mode, L-C-L, T-network. The shield of the short coax
jumper between this balun and the antenna tuner is connected to a
two-square-meter sheet of copper on the floor of my attic. This
sheet has capacitance (C) to "ground," or infinity. The second "L"
of the common-mode "T" network is provided by the coaxial common-mode
choke on the other side of the antenna tuner.
In effect, the C to ground provides the low source-impedance that is
required for a choke to be effective.
A NEC-4 simulation of this L-C-L T-network scheme showed that it
provided substantially more common-mode isolation than a single choke
(L), or the two L's in series, would provide. In this simulation I
modeled my whole three-dimensional universe, including the antenna,
the open-wire line, the metal boxes containing the balun and the
tuner, the copper-sheet counterpoise, the coax, the c-m choke in the
shack, metal boxes for my transmitter etc., continuing with the power
wiring (and the c-m chokes in it, which I haven't mentioned
previously) to the basement entrance panel, its connection to the
buried metal water pipe to the street, the buried power cable to the
street, and the water and power mains under the street.
This simulation showed that more RF current would flow into my shack
via the power line from the street, where the buried mains parallel
my antenna, than via the coaxial line from the antenna. So I added
73 -Chuck, W1HIS