On May 17, 12:36*pm, Wimpie wrote:
On 17 mayo, 17:22, walt wrote:









On May 17, 9:10*am, Wimpie wrote:

On 17 mayo, 14:29, Cecil Moore wrote:

On May 17, 4:49*am, Wimpie wrote:

If you show up with a relevant
quest, maybe I am willing to dive into it.

Wim, here is why my questions for you are more than just relevant. It
is imperative that someone lecturing us on happenings inside that PA
RF volcano be able to understand what is occurring during a passive
event involving forward and reflected EM fields and waves occurring at
an impedance discontinuity outside of a PA.

Two of the physical quantities that must be conserved are energy and
momentum. EM RF fields and waves contain both energy and momentum
which must be conserved. I have asked you to tell us exactly what laws
of physics govern the reversal of the momentum and direction of energy
flow at a Z0-match at a passive impedance discontinuity in a
transmission line. You have refused to do so and asserted that such is
irrelevant. I contend that I could not have asked a more relevant
question - thus the reluctance to provide an answer.

The answer to the question is contained in my energy analysis article
at:http://www.w5dxp.com/energy.htm
A passive Z0-match relies on superposition of waves accompanied by
interference effects to explain the reversal of reflected wave energy
direction and momentum. Walter Maxwell has called the process a
"virtual open-circuit" or a "virtual short". In my article, I explain
how it is a two-step process involving normal reflections and
interference patterns at the impedance discontinuity. It works exactly
like non-reflective glass covering a picture with its 1/4WL thin-film
coating where two sets of reflected light waves undergo destructive
interference toward the viewer and, honoring the conservation of
energy and momentum, reverse their direction and momentum and flow in
the opposite direction toward the picture. This is a well-understood
phenomenon from sophomore physics 201. Why most RF engineers don't
understand this simple physical process involving EM wave interference
is beyond belief. Here's the Florida State University web page again:

micro.magnet.fsu.edu/primer/java/scienceopticsu/interference/
waveinteractions/index.html

Set the java application for opposite phase and when the result is
zero, scroll down to the bottom of the page to find out what happens
to the energy components in the two waves that cancel to zero. Those
energy components "are redistributed to regions that permit
constructive interference" just as they are at a Z0-match in an RF
transmission line where there are only two possible directions for RF
energy flow. For every destructive interference event in one
direction, there will be an equal magnitude constructive interference
event in the opposite direction. At Walt's "virtual short", total
destructive interference energy toward the source is redistributed as
constructive interference energy back toward the load.

I studied this subject in my EE courses at Texas A&M during the
1950's. The textbook was: "Fields and Waves in Modern Radio", by Ramo
and Whinnery, (c) 1944, 1953. The subject is covered under "Quarter-
Wave Coating for Eliminating Reflections" in the chapter titled:
"Propagation and Reflection of Electromagnetic Waves".
--
73, Cecil, w5dxp.com
"Halitosis is better than no breath at all.", Don, KE6AJH/SK

Hello Cecil,

I am familiar with quarter wave (and multi layer) coatings to reduce
reflection. I am not waiting for a lecture on (un)bounded wave
propagation. *If I don't have something present in my mind, I know
where to find it.

As mentioned earlier, you can convert all the wave phenomena in the
coaxial feed line to impedance as seen by the PA. You are unnecessary
complicating things, hence loosing more public that may have interest
in this topic.

Maybe you (and other people) should carry out the experiments I
suggested in this thread (looking to forward power, net power and DC
input power versus small load mismatch [normally referenced to 50
Ohms] ).

With kind regards,

Wim
PA3DJSwww.tetech.nl

Wim, *I'm amazed that you don't find the more-detailed explanation of
how impedance matching occurs via wave interference of any value. Many
RF engineers have traditionally believed that a PHYSICAL open or short
circuit is required to produce a reflection. As a professional antenna
engineer with RCA in 1973 I discovered and published the wave
mechanics that produces the VIRTUAL open and short circuits that are
required for achieving an impedance match. I took bashings from those
traditional engineers, who said reflections cannot be engendered by
wave interference, until they studied my data more carefully and
finally agreed I'm right. Remember, James Clerk Maxwell also had his
detractors until they finally saw the light.

Walt

Hello Walt,

It is not that I don't see the importance of reflections / wave
interference, as I use transmission line theory on an almost weekly
basis. *However one don't need to complicate the matter by using
transmission line theory for a HF PA.

When you open your rig, you will very likely not find a 10" long
100uh inductor in the output section of your PA. Also your capacitors
have very small size w.r.t. wavelength. *A lumped circuit model and a
load specified as an impedance is therefore more than good enough to
discuss PA complex output impedance and what CAN happen when you apply
mismatch.

You can't tell what happens exactly, because then you need to dive
into the circuit diagram of the PA to evaluate current and voltage
waveforms at the active device.

With kind regards,

Wim
PA3DJSwww.tetech.nl

For what it's worth, in my work I design a lot of filters and matching
networks. I regularly model the designs before I build them, using
the level of detail I feel is appropriate. When I build the physical
filter or network I've modeled, I compare the measured response with
the response predicted by my model. I do that in some detail. I
almost always modify the model as necessary, adding detail so it
matches the measured performance. When I say I add detail, I don't
mean that I do it haphazardly, but rather that I look closely at the
physical realization and add pieces to the model that match pieces of
the physical realization. Because I've gone through this design cycle
many times, with many different topologies and for a variety of
frequency ranges from below 1MHz to above 1GHz, I have a pretty good
idea before I start a new design just what level of detail I'll need.

What I find from this is that, just as Wim says about PA matching
networks, I seldom need anything beyond representations of the lumped
components when I'm dealing with low frequency filters that don't have
high loaded-Q resonators. Up to 30MHz, I don't recall ever having to
use transmission lines in my models to get excellent agreement between
the model and the physical implementation (except in the rare cases
where I've included transmission lines in the physical implementation
of a low frequency network, of course). I do often have to add
parasitic elements--sometimes even for relatively low frequency
filters. (I remember having a technician build a 1MHz filter for me;
he couldn't understand why his didn't work anywhere near as well as
the first one I had built, until I showed him where he'd allowed other
parasitics to creep in--short lengths of wire with currents shared
between two high-Q resonators...) I'm _FAR_ more likely to need to
include mutual inductance between two coils in a model, than I am to
need to include a transmission line.

On the other hand, when I'm dealing with filters above 100MHz, it's
not unusual to include transmission line sections and stubs in my
models. Above 200MHz or so, the models generally do benefit from
including transmission lines. Mind you, there aren't any hard and
fast rules; there is no magic transition frequency. But when you've
built models that match reality as closely as I commonly do, you learn
to just smile blandly at those who tell you that you "must" consider a
coil to be a transmission line or you'll be "wrong."

Finally, even when I do include transmission lines in my models, I
don't worry in detail about reflections, or about a time-domain
analysis of the situation. Just as there's an equivalence between
time-domain waveforms and spectral analysis, a frequency sweep of a
system (including phase as well as amplitude response) tells me
everything I need to know, in the domain I'm already interested in.

I hope Wim (and others?) will excuse the off-topic drift here. And
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.

Cheers,
Tom