Roy Lewallen wrote in
treetonline:

....
But at RF, as Richard says, many ferrite cores are more resistive than
inductive. Not all are -- there are high frequency ferrites that
remain inductive at very high frequencies. However, for the HF range
and above, the best choice is usually "low frequency" ferrites which
are primarily resistive in that frequency range. They're best for two
reasons -- first, they provide much more impedance per turn squared
than high frequency ferrites, making the first two goals much easier
to achieve; and second, there are no resonance effects. An inductive
winding will resonate with stray capacitance at some frequency. The
impedance drops above that frequency, limiting the upper range of the
transformer. The winding and leakage inductances can also resonate
with external capacitance and cause ringing in some applications. A
properly designed transformer using an essentially resistive core can
operate very well over several decades of frequency.

Hams have a affinity for powdered iron cores in RF applications, mislead
by the thought that low loss material naturally produces a better
solution. Thing is that it is a huge leap from low loss *material* to a
low loss *solution*.

Here is a recent design for a Guanella 1:1 balun for 6m and 2m on a #61
ferrite co http://www.vk1od.net/balun/G1-1-FT140-61/index.htm .

The article contains a graph of the material characteristics, which
shows that at 2m, µ' has fallen to 10% or so of µi, and µ'' has risen,
so an inductor will be more resistive than inductive.

But, does that make it a poor TLT? Not at all, it has very high common
mode or choking impedance (a critical performance parameter that is
rarely reported for commercial baluns).

Would such a balun wound on a powdered iron core work better? Probably
not. I say probably because I have been unable to find manufacturer's
loss data for powdered iron materials above about 5MHz. The probability
is that a powdered iron core would yeild a balun with a choking
impedance that is lower and more inductive than resistive and may likely
be *more* lossy. (I have a prospective project to measure a couple of
powdered iron cores at 6m to further explore this, but they will not be
lab grade measurements.)

The key is in thinking about I^2*R. The lossy ferrite balun has very
high Z, so very low I, very high R, and the product of I^2*R is
relatively low. A powdered iron choke will have lower Z, so higher I,
low R, and the product I^2*R may be quite higher than the ferrite.

Higher choking impedance can reduce loss, even if the higher choking
impedance is by way of a lower Q inductor. The characteristic is a
curve, concave down and the trick is to choose a design (whether it is
powdered iron, air cored, or ferrite cored) for acceptable loss, and
that often means an operating point that is well on the right hand side
or the left hand of the curve maximum.

Another facet of ferrite TLTs operated in their lossy region, is that
they remain useful well above the device self resonance, whereas when
choking depends on a high Q impedance, it rapidly falls above resonance.
(Self resonance is ignored in most models of balun performance that I
have seen.)

Owen