On Tue, 14 Jun 2011 14:18:53 -0700 (PDT), walt wrote:

Thanks again Richard, for your insightful response. And goodness no,
Richard, you won't hurt my feelings. I like being correct, but if I'm
not I surely want to be told about it. This situation is no different.

However, if you're not following me on the 'non-dissipative' path, I'd
like you to review the last portion of the TS-830S experiment, and
follow the numbers. If you don't have that material in front of you,
you can find it again on my web page at www.w2du.com. Click on
'Preview Chapters from Reflections 3', and then click on Chapter 19A,
which is a part of Chapter 19 in the 3rd edition.

Observe that when the load impedance is changed from 50+j0 to the
complex impedance 17.98 + j8.77 ohms, the plate current rose
expectedly from 260ma to 290ma. This change occurred because the amp
is now mismatched, and the pi-network is also detuned from resonance.
The unwashed would conclude that the reflected power caused the
increase in plate current. However, one will also observe that after
the pi-network has been retuned and adjusted to again deliver all the
available power into the otherwise 'mismatched' load (which is now
matched to the source), the plate current went back to it's original
value, 260ma. In addition, the amp returned to deliver 100w into the
complex impedance at the line input, with the 30.6w of reflected power
in the line, and adding to the 100w of source power, making 130.6w
incident on the mismatched load, absorbing 100w and reflecting 30.6w.

Hi Walt,

There is absolutely NOTHING that I can dispute in your numbers or
method. However, as to "dissipation" this says nothing.

As for what it says about the reflection coefficient, I would agree
with you. However, the math and the math experts you speak of - they
are not answering the model we are examining above. You distracted
them with stubs and reflecting waves.

If you cast this agreement in the reflection coefficient back into
effects purported to exist in the matching with stubs model, we may
yet argue.

IMHO, these data prove beyond doubt that the pi-network not only
didn't absorb any of the reflected power, but totally re-reflected it.
In my book this says the pi-network reflection coefficient rho = 1.0.
How can anyone disagree with this?

*** Nothing further useful about the topic is to be found below ***

There is a curious side bar to this found at:
http://www.w5big.com/purchase4170c.htm
Observe the second data screen that purports to examine a mismatch
through a length of both RG58 and RG59 terminated with 100 Ohms.

Of particular note is the distinct transition from the Zc of 50 Ohms
to the Zc of 75 Ohms and the similarly distinct transition from the Zc
of 75 Ohms to the termination R of 100 Ohms. Note that the
transitions both span a distance of 2 feet. 2 feet is not
insubstantial compared to connection technology that spans, probably,
no more than one tenth that distance. Whence the extra 11 inches on
both sides of the connection plane of either connection?

In the physics I've been studying for the past 10 years, near fields,
this would not be unusual. In fact it would be expected. There is a
transition zone (in waveguide design, there would be a taper or a
sweep section to anticipate this and they would be physically large in
terms of wavelength) not a transition cliff.

Similarly, at the connector to your TS-830S, there is a zone that
surrounds it that only approximates the 50 Ohm which occurs (in the
terms of this link's demonstration) some distance away, deep inside
your TS-830S.

To me, dissipation inhabits this zone (certainly large enough for the
tube and tank to occupy) and embraces the match with a complex
addition of phases that could result in loss or even gain. This is
also to say that I do not subscribe to dissipation being all about
loss - especially when the hand of man is on the tuning knob instead
of letting the chips fall where they may.

73's
Richard Clark, KB7QHC