On Sat, 26 May 2007 20:51:37 -0400, Chuck
wrote:

Owen Duffy wrote:
Other works that I have read describe the loss mechanisms as quite
complex; magnetic loss described by a complex mu value that is
temperature, frequency and flux dependent, resistive loss in the core
material, dielectric loss in the core material, and resistive loss in the
conductors. Sevick seems to say that only one of these is relevant, or
that loss can be simplfied to a single equivalent loss, the dielectric
loss.

He is actually distinguishing between transformer type losses (as
Chuck notes later in material I have not quoted). In the conventional
transformer, loss is Ohmic due to the circulating eddy currents. This
explanation does not serve Ferrite, which is principally an insulator,
a dielectric with magnetic properties. Further, the operation of
transformation for conventional transformers is through magnetic
coupling, that is not so for tranmission line transformers (that no
doubt all parties here are already aware of, and which Sevick is
distinguishing to the naive readers of his paper). Hence, he is
responding to loss issues that are found in conventional designs, but
missing (or rather optimized for the Common Mode) in our chokes or
BalUns.

I guess it is appealing to equate loss that increases with frequency to
an equivalent dielectric effect, but the loss is flux dependent and in a
non-linear way, so it doesn't seem to fit well with a simple dielectric
equivalence.

Ferrites are not simple dielectrics, not even linear as a class (but
seemingly so for our limited applications as chokes).

Owen, I didn't see any reference in
Sevick's fourth edition to the
voltage-dependency of core losses.

I can observe it in the reference offered. However, Sevick couches
this dependency parenthetically to the dielectric property. As this
is specifically true, and he distinguishes other losses in
conventional transformers as Ohmic, his statement in its full context
is valid if perhaps too sparse. It is quite obvious that voltage and
current (and thus flux) are irrevocably inseparable and yet the loss
does not specifically arrive due to conduction in the Ferrite
material.

On the other hand, discussion of Ferromagnetics is couched in magnetic
fields and electron spin, not in voltaics.

Instead, the following quote seems to
characterize his philosophy in the book:

"With transmission lines, the flux is
effectively canceled out in the core and
extremely high efficiencies are possible
over large portions of the
passband--losses of only 0.02 to 0.04 dB
with certain core materials.

The flux of Owen's contention is not the flux of the differential
currents, but of the Common Mode current. [or so I read his query -
at least insofar as the application of the choke serves]

For student of the microwaves, Ferrites offer far more unique
properties than are made use as I've suggested above. The
nonlinearity is that Ferrite can be remarkably transparent to flux of
a given polarization, and with a shift in that polarization it becomes
quite opaque. In this service, it is also characterized as a
non-reciprocal attenuator. This attribute can be modulated,
literally, with an external DC (actually low frequency AC) field to
impart modulation to what would otherwise be a CW signal.

Ferrites employing this polarization characteristic are used in what
are called Faraday Isolators (another one of those devices that could
be used to separate forward and reverse waves; however, having only
two ports so as to not be confused with the rat-race, and to my
knowledge wholly unknown in this group).

73's
Richard Clark, KB7QHC