On Feb 2, 12:42*am, "Antonio Vernucci" wrote:
Tom,

the story began when, a few days ago, I was going to replace a trap of my HF
yagi. Not to make mistakes, I consulted the antenna assembly manual where I
found a big banner: do not invert traps otherwise the antenna will not work.

So, I thought, this is a case in which a bipole cannot be inverted.

This is clearly due to the fact that the external body of the trap (an aluminuim
can about 2 feet long), which contains two coils resonated at different
frequencies by means of built-in capacitors, is effectively part of the antenna
radiating element. So, the trap is a bipole not only comprising lumped elements,
and that is the reason why it cannot be inverted.

So, as K1TTT has pointed out, a bipole can be inverted without consequences only
if it has only 2 ports, has only passive linear components, and is small enough
to be considered a lumped element.

73

Tony I0JX
Rome, Italy

As I see it, it has nothing to do with being small enough to be
considered a lumped element. Instead, it has to do with coupling
between the network elements and the outside world. Clearly your
trap is coupled to the outside world. Clearly the series LC in my
example is coupled to its surroundings. In such cases, you're NOT
dealing with a two-terminal network: there is a path for current
other than the two terminals. You can put as many transmission lines
in your network as you wish, and as long as they don't have coupling
to the rest of the universe except through the two closely-spaced
terminals, there will be no difference in behaviour if you reverse the
terminals. If you can show a valid counter-example, you've proven a
whole lot of textbooks wrong...

I suppose equivalently, if you decouple your measurement from the rest
of the universe, you'll get the same result. Typical broadband
directional couplers (of the sort used in S-parameter test sets) have
good decoupling, to be able to separate excitation from return power.

Cheers,
Tom