John Popelish wrote:
Richard Clark wrote:
. . .

In fact, nothing of that sort happens - at least not by your
description. The ferrite is simply bulk resistance inserted into the
common mode path.

Make that, "impedance (mostly inductive, if the ferrite is well suited
to the frequency)" and I agree.

"Low frequency" ferrites are well suited to applications as HF chokes
and broadband transformers. What you don't want to do is use them for a
high Q inductor in a filter, tuned circuit, or similar application.

That is why common mode current is suppressed. The
same thing occurs in the coiled transmission line choke, but the
resistance is replaced by reactance. Again, common mode current is
snubbed by encountering this too.

I agree with this, except that the purpose of the ferrite is to increase
the common mode inductance of the section of coax passing through it,
not add resistance. Some resistance is inevitable, because no ferrite
is lossless, but the intention is for inductance.

Baluns work fine with a resistive impedance, with the exception of
applications involving large power. In fact, resistive impedance is
desirable because the impedance changes little with frequency, and is
relatively free of resonance effects. (More below.)

The transformer property is in the isolation of the balanced circuit
from the unbalanced circuit through this resistive characteristic.

Try transmitting through such a resistance and you are going to lose a
lot of your power.

You're not "transmitting through" a balun's impedance. Only the common
mode current effectively flows through it, and the power dissipated by
the balun is Icm^2 * R, where Icm is the common mode current and R is
the resistive part of the balun's common mode impedance. If R is small,
dissipation is low. But if R is large, that makes Icm small, so
dissipation is also low. It's really an impedance matching problem when
the balun is resistive -- a very low or very high balun R results in low
dissipation. Dissipation is maximum at some intermediate value of R and
decreases on each side.

A typical balun made with "low frequency" ferrite (e.g. Fair-Rite 70
series) and operating at HF or above (and therefore primarily resistive)
having a common mode impedance of 500 ohms or greater generally won't
dissipate any significant fraction of the transmitted power. However, if
you're running high power, even a fraction of a dB dissipated in the
balun will cause it to overheat. Consequently, people running high power
often resort to type 43 ferrite (a Fair-Rite designation; or its
equivalent from other manufacturers), which is less resistive than lower
frequency ferrites. In extreme cases, high frequency (60 series) ferrite
is necessary. The problem is that it's increasingly difficult to get
adequate impedance with the higher frequency ferrites. Type 43 is often
a good compromise, and it's widely available in many core sizes.

. . .
I think you mean by this that a normal unbalanced signal in a coax has
no magnetic field external to the shield. It is all between the center
conductor and the shield. And I agree that this is what you are trying
to accomplish by adding this two conductor choke between the coax and
the balanced antenna. Without it, there would be some magnetic field
from a net (uncanceled) current and voltage on the outside of the shield
that would cause the coax to radiate. And the voltages and currents
fed to the balanced antenna would not be equal and opposite (balanced)
but somewhat unbalanced. There would also be non equal currents in the
center conductor and shield. I think we agree on all that, but have a
different picture of how a choke balun corrects these problems.
. . .

Common and differential mode currents are physically separated in a coax
cable, and so are the fields from the two components, providing that the
shield is at least several skin depths thick. The differential mode
current and its fields are entirely inside the coax, decaying rapidly as
you go outward from the inner boundary of the shield. By the time you
reach the outer boundary of the shield the fields from the differential
current is negligibly small. So any core you put over the coax doesn't
see or interact with the common mode current or its fields at all, and
you can completely ignore it when analyzing balun action. Similarly, you
can ignore the core when analyzing the differential mode properties of
the system. The common mode current resides in a thin layer on the
outside of the shield, it and its fields never reaching the inside. The
balun provides an impedance to this current just as it would to any
current on the outside of a conductor.

When bifilar wound, the fields from the differential mode current are
primarily between the turns, although some relatively small amount
extends beyond to interact with the core. The net result is nearly the same.

. . .

Roy Lewallen, W7EL