The radiated field in which a feedline is immersed produces a common
mode, not differential, current on the feedline(*). The balun creates a
high impedance to the flow of common mode current. The result is that
much less common mode current is induced than would be the case if the
balun were absent. AM broadcasters put insulators periodically in tower
guy wires. Baluns have the same effect, although they're of course not
as perfect as the fully open circuit created by the insulator.

I'd also like to add that the induced current isn't unpredictable, as
you stated. It has to follow rules like all other physical phenomena, so
it's entirely predictable. I have to qualify this, though, by noting
that in many or most amateur installations, the path from the antenna
along the outside of the feedline to the rig and from there to the Earth
is often not well known. And precise predictions of feedline current
can't be made without knowing this path.

This effect is easily observed in models, as well as being
experimentally verifiable with a small amount of effort. For example,
make a simple wire Yagi with wire elements. Slip a few ferrite cores
over one of the elements and see what happens to the pattern. What
you've done is to reduce the current in the parasitic element, which is
immersed in the field from the driven and other elements, thereby
reducing its contribution to the total field. Cutting the element
(analogous to the broadcasters' insertion of an insulator) has the same
effect, although it'll be even more profound than the high but finite
impedance of the cores. Quanitative measurements aren't hard to make,
either. You can find details of current measurement devices in Chapter 8
of the ARRL Antenna book and other references. When placed over a coax
feedline, or over both conductors of a twinlead feedline, they'll
measure the common mode current.

(*) I'm assuming here that we're dealing either with coax, or with
twinlead whose wire spacing is small in terms of wavelength and whose
spacing from the antenna is large compared to the wire spacing. If you
put twinlead in a field such that the field seen by one conductor is
significantly different than the field seen by the other, you will get
an induced differential, as well as common mode current. This won't
happen with coax. Unlike common mode current, an induced differential
current will affect the properties of the line as a transmission line
(that is, change its apparent Z0 and velocity factor). But this would be
an unusual condition in an amateur installation, and I'm not considering
it here.

Roy Lewallen, W7EL

Richard Fry wrote:
A balun can reduce common mode feedline currents due to the power supplied
to the line by the transmitter, but if the feedline is immersed in an
asymmetric radiated field, how does the balun reduce/remove the resulting,
unpredictable differential current on the feedline, and its contribution
toward producing the net radiated pattern?

RF
_____________

"Roy Lewallen" wrote:

A "non-radiating" feedline is one on which there is no net current
(i.e., no common mode current). In the case of coax, this translates to
zero current on the outside of the shield; for twinlead feedlines, it
means that the currents in the two conductors are exactly equal in
magnitude and opposite in direction.


A transmission line placed symmetrically with respect to a dipole
won't have any current induced in it, although current can be conducted
via a direct connection.


The amount of common mode current flowing in a transmission
line can be reduced by introducing an impedance to the common
mode current. It's desirable to do this without disturbing the differential
mode transmission line operation. That's the function of a balun.