I had posted some questions on baluns earlier. Thank you very much
for all then information. I should have been more specific. I have
done some more reading, and the baluns I'm talking uses lumped
components - the Lumped-constant 180 degrees splitter. I found it on
the IEEE site in a paper by Tetsuo Hirota and Masahiro Muraguchi. I
was wondering how the signals get split by 180 degrees, equations etc.
Transformers are the easiest way to get 180 splits. But you can't
always use a transformer. Generally 180 splitting is only occurring
over a range of frequencies with lumped components. At a single freq
a single lumped capacitor or inductor can give you 90 degrees of shift
between voltage and current. Together you can create 180, but only at
one frequency aka a resonance. Expanding the number and spacing
resonant frequencies gives a broaden response and range of *nearly*
180 splitting. A perfectly distributed (non-lumped) implementation
can give the best implementation because you can create
transformer-like circuit without strictly using a transformer as they
are familiarly known - the limit is that there is a (low end) cut-off
frequency.
You'd need to look at the particular paper (most likely it focuses on
economizing lumped component counts, which in quantity look
distributed, against bandwidth of splitting given a desired error
level with (fanatically) not using a wire-wound transformer), or look
into filter design techniques, especially all-pass filters and
impedance design. Since a balun is about matching impedances, and
filters have input and output impedances, then one way to think of
impedance matching circuits is as filter design problem. There isn't
a simple or single equation for filter design but if you look at
filter design you'll find the answer.
MM
