On May 18, 6:11*am, Cecil Moore wrote:
On May 18, 12:33 am, K7ITM wrote:

I'm _still_ trying to figure out _why_ anyone would care about the
output impedance of a PA of the sort used at HF to drive antennas.
Nobody has ever convinced me that it matters at all, except perhaps as
academic interest.

Nobody is questioning the efficacy of design methods. Whatever works,
works. What we are discussing is indeed only of academic interest.
Knowing whether my IC-706 is conjugately matched or not does not
affect its operation at all.

From the time (t0) that a PA first outputs a Zg signal to the time
(t1) that the PA senses its load impedance is NOT zero time. How does
the PA know what its load impedance really is when it is not Zg?
Einstein's spooky action at a distance? No, feedback from the load.

Obviously, the PA receives some sort of feedback in real time. What is
the nature of that feedback? What can it be besides feedback energy
reflected from the load? (not in zero time, but at the speed of
light). In the real world, it takes measurable time for the forward
energy to reach the load and measurable time for the reflected
feedback (if any) to arrive back at the PA. The load seen at the PA
source is always an E/I ratio, i.e. a lossless image impedance that
always experiences a delay if it is not equal to Zg, i.e. it usually
contains reflected energy.
--
73, Cecil, w5dxp.com

Cecil, just HOW do you propose to MEASURE the effect you describe, as
seen at the transmitter's output port, using only our HF ham
transmitter/PA that transmits a signal with a maximum bandwidth of
perhaps 10kHz? If it is going to have the dire consequences you
suggested a few postings ago, then it must be trivial to
measure...unfortunately, I don't see how, and that bugs me, as one who
strives to provide accurate, sophisticated measurement technology to
engineers.

Perhaps you missed it, but nobody is disagreeing that the mechanism
for establishing a load impedance is reflections in the system of
lines and lumped loads attached to the PA output. We are simply
saying that, for the bandwidth signals involved, you'll be extremely
hard-pressed to distinguish between a load consisting of any number of
transmission line segments, along with one or many lumped loads
wherever you want along those line segments, and a simple equivalent
series RLC.

I suppose it will be lost on most of the lurkers, but it's a bit of a
bad joke to deny that one can make valid PA output impedance
measurements with a signal very slightly off-frequency (less than
1kHz), and then claim that reflections in a system of maybe a couple
hundred feet of coax makes a major difference in how the load behaves
as compared with a lumped RLC.

Cheers,
Tom