Richard Harrison wrote:
Terman says on page 96 of his 1955 opus:
"The reflected wave is identical with the incident wave except that it
is traveling toward the generator."

Gene needs to tell us how the TV modulation that
causes ghosting makes its predictable round trips
to the source and back without the aid of the
reverse traveling wave.
--
73, Cecil http://www.w5dxp.com

Cecil Moore wrote in news:7iYLh.73$Kd3.72
@newssvr27.news.prodigy.net:

Gene needs to tell us how the TV modulation that
causes ghosting makes its predictable round trips
to the source and back without the aid of the
reverse traveling wave.

Talk about echoes!

Cecil, at the time of writing, you have made 4 of the 7 responses to
Walt's posts, and there is nothing in what you have said that you haven't
said recently.

The language from various posters like:

"Some of the posters apparently are unable..."

"Some would say..."

"people obviously don't understand..."

"How can you possibly deny..."

doesn't seem to me the language of convicing arguments, much less proof,
from either side. They seem more a sign of the posters frustration, but
not otherwise convincing.

My guess is that this discussion will not converge on a convincing
outcome.

If the past is any indicator, just when agreement of two or three people
looks likely, someone will inject some noise like lets start dealing with
time domain and transient issues to prove that steady state analysis is
invalid in the practical sense, or this needs a photon explanation with
reference to a text no one is likely to have. It as though those posters
intended to wreck logical development and conclusion. Ah, but that is
USENET!

The basis of the assertion that a PA is naturally or magically conjugate
matched as a necessary consequence of adjustment or design for maximum
power output is based on an leveraging the Maximum Power Transfer Theorem
which depends on a linear source. I don't recall seeing experimental
results to convincingly demonstrate that the PA is a linear source,
though I have seen those that suggest otherwise. If the source cannot be
proven to be sufficiently close to a linear source, then the basis for
arguing the implicit conjugate match dissolves.

No one has yet come up with a quantitative proof that in the general case
PAs of all kinds have an equivalent source impedance the conjugate of
their load, nor convincing experiments that would place bounds on the
reflection coefficient looking into the PA for practical transmitters.

No one has demonstrated that using equivalent impedances etc is not a
valid analysis of the steady state behaviour.

Owen

On Tue, 20 Mar 2007 22:06:35 GMT, Owen Duffy wrote:

I don't recall seeing experimental
results to convincingly demonstrate that the PA is a linear source,
though I have seen those that suggest otherwise. If the source cannot be
proven to be sufficiently close to a linear source, then the basis for
arguing the implicit conjugate match dissolves.

Owen, despite our previous discussion, I have explained many times that even though the PA source upstream of
the tank circuit is non-linear (and no one's saying it isn't), the energy storage in the tank makes the output
of the tank a linear source, no matter what the shape of the current wave form may be at the input. The output
of the tank is proved linear because the voltage/current ratio at the output is non-varying and the shape of
the voltage and current wave forms are essentially sine waves. Consequently, the output circuit can be
represented by a Thevenin source that supports both a conjugate match and the maximum power transfer theorem.

Are you now denying that the output of a PA with the routine Q of 10 to 12 is not substantially a sine wave?
If you agree that it is a sine wave, then why are you arguing that there is no basis for a conjugate match?

However, none of the responses above respond to the issue of why the reflected power does not cause heating of
the amp, which is what my treatise was all about.

Walt, W2DU

On Tue, 20 Mar 2007 22:57:48 GMT, Walter Maxwell wrote:

On Tue, 20 Mar 2007 22:06:35 GMT, Owen Duffy wrote:

I don't recall seeing experimental
results to convincingly demonstrate that the PA is a linear source,
though I have seen those that suggest otherwise. If the source cannot be
proven to be sufficiently close to a linear source, then the basis for
arguing the implicit conjugate match dissolves.

Owen, despite our previous discussion, I have explained many times that even though the PA source upstream of
the tank circuit is non-linear (and no one's saying it isn't), the energy storage in the tank makes the output
of the tank a linear source, no matter what the shape of the current wave form may be at the input. The output
of the tank is proved linear because the voltage/current ratio at the output is non-varying and the shape of
the voltage and current wave forms are essentially sine waves. Consequently, the output circuit can be
represented by a Thevenin source that supports both a conjugate match and the maximum power transfer theorem.

Are you now denying that the output of a PA with the routine Q of 10 to 12 is not substantially a sine wave?
If you agree that it is a sine wave, then why are you arguing that there is no basis for a conjugate match?

However, none of the responses above respond to the issue of why the reflected power does not cause heating of
the amp, which is what my treatise was all about.

Walt, W2DU

In the fourth line in the first paragraph above the word 'time' was inadvertantly omitted. It should have read
....the output is non-time varying and the shape...

Sorry about that,

Walt, W2DU

"Walter Maxwell" wrote
... I have explained many times that even though the PA source
upstream of the tank circuit is non-linear (and no one's saying it isn't),
the energy storage in the tank makes the output of the tank a linear
source, no matter what the shape of the current wave form may be
at the input. The output of the tank is proved linear because the
voltage/current ratio at the output is non-varying and the shape of
the voltage and current wave forms are essentially sine waves.
Consequently, the output circuit can be represented by a Thevenin
source that supports both a conjugate match and the maximum
power transfer theorem.
______________

If this statement about the tank circuit being ~ a linear source is valid,
does that mean that any load-reflected power that appears across the output
terminals of the tx stops at the tank circuit, and never sees the
non-linear, non-matching Z of the active PA?

And if so, would that also mean that such a tx would not be prone to
producing r-f intermodulation components when external signals are fed back
into the tx from co-sited r-f systems?

Yet experience shows that this is not the case for ~closely spaced
interfering signals. The only mitigation for this for a PA with a tank
circuit is the rejection of that tank circuit to those off-freq, external
signals, and to the resulting IM products generated by mixing with the main
tx signal in the active (and non-linear) PA stage of that tx.

And the tank has VERY low rejection to load reflections of the signal
bandwidth to which it is tuned.

Also to be considered are the modern broadband (88-108MHz) FM broadcast
transmitters, which have no tank circuits, but except for some designs
incorporating balanced 3 dB hybrid combiners are affected by load
reflections about the same as a tx with a tuned tank circuit.

RF

On Tue, 20 Mar 2007 19:19:26 -0500, "Richard Fry" wrote:

"Walter Maxwell" wrote
... I have explained many times that even though the PA source
upstream of the tank circuit is non-linear (and no one's saying it isn't),
the energy storage in the tank makes the output of the tank a linear
source, no matter what the shape of the current wave form may be
at the input. The output of the tank is proved linear because the
voltage/current ratio at the output is non-varying and the shape of
the voltage and current wave forms are essentially sine waves.
Consequently, the output circuit can be represented by a Thevenin
source that supports both a conjugate match and the maximum
power transfer theorem.
______________

If this statement about the tank circuit being ~ a linear source is valid,
does that mean that any load-reflected power that appears across the output
terminals of the tx stops at the tank circuit, and never sees the
non-linear, non-matching Z of the active PA?

Richard, my earlier treatise considers only tube-type PA's with pi-network output coupling circuits used in
the Amateur Service, such as the Kenwood TS-830S on which my measurements were made. It was not intended to
consider PA's used in the tv service. Sorry, I didn't make this distinction earlier.

And if so, would that also mean that such a tx would not be prone to
producing r-f intermodulation components when external signals are fed back
into the tx from co-sited r-f systems?

This issue is irrelevant, because the signals arriving from a co-sited system would not be coherent with the
local source signals, while load-reflected signals are coherent. The destructive and constructive interference
that occurs at the output of a correctly loaded and tuned PA requires coherence of the source and reflected
waves to achieve the total re-reflection of the reflected waves back into the direction toward the load.

Yet experience shows that this is not the case for ~closely spaced
interfering signals. The only mitigation for this for a PA with a tank
circuit is the rejection of that tank circuit to those off-freq, external
signals, and to the resulting IM products generated by mixing with the main
tx signal in the active (and non-linear) PA stage of that tx.

Again, Richard, this condition is irrelevant to the re-reflection of the waves reflected by the load, because
the relevant signals are not coherent.

And the tank has VERY low rejection to load reflections of the signal
bandwidth to which it is tuned.

This may be true for PAs with bandwidths wider than those occurring in ham tx. However, the destructive and
constructive interference between the reflected and source waves in a correctly loaded and tuned ham tx
results in total re-reflection of the reflected waves.

Also to be considered are the modern broadband (88-108MHz) FM broadcast
transmitters, which have no tank circuits, but except for some designs
incorporating balanced 3 dB hybrid combiners are affected by load
reflections about the same as a tx with a tuned tank circuit.

And still further, Richard, the FM transmitters you refer to above are not in the same category as those used
in tube rigs used by hams.

Incidentally, Richard, have you really reviewed the report of my TS-830S experiment?

Walt

"Walter Maxwell" wrote
(RF): And if so, would that also mean that such a tx would not be prone
to producing r-f intermodulation components when external signals
are fed back into the tx from co-sited r-f systems?

This issue is irrelevant, because the signals arriving from a co-sited
system would not be coherent with the local source signals, while load-
reflected signals are coherent. The destructive and constructive
interference that occurs at the output of a correctly loaded and tuned
PA requires coherence of the source and reflected waves to achieve
the total re-reflection of the reflected waves back into the direction
toward the load.

But even for coherent reflections, if the PA tank circuit has very low loss
for incident power (which it does), why does it not have ~ equally low loss
for load reflections of that power? Such would mean that load reflections
would pass through the tank to appear at the output element of the PA, where
they can add to its normal power dissipation.

Also, does not the result of combining the incident and reflected waves in
the tx depend in large part on the r-f phase of the reflection there
relative to the r-f phase of the incident wave? And the r-f phase of the
reflection is governed mostly by the number of electrical wavelengths of
transmission line between the load reflection and the plane of
interest/concern -- which is independent of how the tx has been
tuned/loaded.

If the ham transmitter designs that your paper applies to produce a total
re-reflection of reverse power seen at their output tank circuits, then
there would be no particular need for "VSWR foldback" circuits to protect
them. Yet I believe these circuits are fairly common in ham transmitters,
aren't they? They certainly are universal in modern AM/FM/TV broadcast
transmitters, and are the result of early field experience where PA tubes,
tx output networks, and the transmission line between the tx and the antenna
could arc over and/or melt when reflected power was sufficiently high.

RF

On Wed, 21 Mar 2007 08:18:14 -0500, "Richard Fry" wrote:

"Walter Maxwell" wrote
(RF): And if so, would that also mean that such a tx would not be prone
to producing r-f intermodulation components when external signals
are fed back into the tx from co-sited r-f systems?

This issue is irrelevant, because the signals arriving from a co-sited
system would not be coherent with the local source signals, while load-
reflected signals are coherent. The destructive and constructive
interference that occurs at the output of a correctly loaded and tuned
PA requires coherence of the source and reflected waves to achieve
the total re-reflection of the reflected waves back into the direction
toward the load.

But even for coherent reflections, if the PA tank circuit has very low loss
for incident power (which it does), why does it not have ~ equally low loss
for load reflections of that power? Such would mean that load reflections
would pass through the tank to appear at the output element of the PA, where
they can add to its normal power dissipation.

Also, does not the result of combining the incident and reflected waves in
the tx depend in large part on the r-f phase of the reflection there
relative to the r-f phase of the incident wave? And the r-f phase of the
reflection is governed mostly by the number of electrical wavelengths of
transmission line between the load reflection and the plane of
interest/concern -- which is independent of how the tx has been
tuned/loaded.

If the ham transmitter designs that your paper applies to produce a total
re-reflection of reverse power seen at their output tank circuits, then
there would be no particular need for "VSWR foldback" circuits to protect
them. Yet I believe these circuits are fairly common in ham transmitters,
aren't they? They certainly are universal in modern AM/FM/TV broadcast
transmitters, and are the result of early field experience where PA tubes,
tx output networks, and the transmission line between the tx and the antenna
could arc over and/or melt when reflected power was sufficiently high.

RF
Richard, your statement above begs the question, "Are you aware of the phase relationships between forward and
reflected voltages and between forward and reflected currrents that accomplish the impedance-matching effect
at matching points such as with stub matching and also with antenna tuners?

When the matching is accomplished the phase relationship between the foward and reflected voltages can become
either 0° or 180°, resulting in a total re-reflection of the voltage. If the resultant voltage is 0°, then the
resultant current is 180°, thus voltage sees a virtual open circuit and the current sees a virtual short
circuit. The result is that the reflected voltage and current are totally re-reflected IN PHASE with the
source voltage and current. This is the reason the forward power in the line is greater than the source power
when the line is mismatched at the load, but where the matching device has re-reflected the reflected waves.

This phenomenon occurs in all tube transmitters in the ham world when the tank circuit is adjusted for
delivering all available power at a given drive level. When this condition occurs the adjustment of the
pi-network has caused the relationship between the forward and reflected voltages to be either 0° or 180° and
vice versa for currents, as explained above. When this condition occurs, destructive interference between the
forward and reflected voltages, as well as between the forward and reflected currents, causes the reflected
voltage and current to cancel. However, due to the conservation of energy, the reflected voltage and current
cannot just disappear, so the resulting constructive interference following immediately, causes the reflected
voltage and current to be reversed in direction, now going in the foward direction along with and in phase
with the forward voltage and current.

In transmitters with tubes and a pi-network output coupling circuit there is no 'fold back' circuitry to
protect the amp, because none is needed, due to the total re-reflection of the reflected power. It is only in
solid-state transmitters that have no circuitry to achieve destructive and constructive interference that
requires fold back to protect the output transistors.

This has been a quick and dirty explanation of the phase relations that accomplish impedance matching.
However, I have explained it in much more detail in my book "Reflections--Transmission Lines and Antennas."
Yes, I know the book has been sold out and now unavailable, but I have put several chapters on my web page
avaliable for downloading. The pertinent chapters covering this issue are Chapters 3, 4, and 23, available at
www.w2du.com. I hope that reviewing these chapters will be helpful in clearing up some of the
misunderstandings that are clearly evident in some of the postings on this thread.

Walt, W2DU

"Owen Duffy"
No one has demonstrated that using equivalent impedances etc
is not a valid analysis of the steady state behaviour.
_________

A reflection is a reflection. The reflection of a ~steady-state r-f source
may produce a different perceived/effective result than if that source
includes transients (modulation), but such does not negate the existence of
reverse/reflected power in the steady-state case.

Decades of experience with analog broadcast TV transmission systems
demonstrate that the reflected power from a mismatch at the transmit antenna
produces an amplitude variation (ripple) and other effects across the r-f
and demodulated video channel bandwidths that are directly related to the
magnitude of the antenna mismatch and the round-trip propagation time of the
transmission line between the tx and the antenna (period = 1 cycle per ~491
feet of air-dielectric line).

This is evident not only from accurate measurements made via a highly
directional r-f coupler sampling forward power at the tx end of the
transmission line, but also from results seen on the screen of TV sets
viewing those transmissions. I suspect, Owen, that you would agree that
this example originates from a "practical" system.

The r-f power supplied even by a CW source is subject to the same amount of
reflected power for a given antenna mismatch, which will have an appropriate
effect on system performance. Whether or not that reflected
power/performance effect is important (or even recognized as existent) is
the issue at hand.

RF

"Richard Fry wrote
(period = 1 cycle per ~491
feet of air-dielectric line).
_____________________

Sorry, make that 1 cycle == per MHz of bandwidth ==,
per ~491 feet of transmission line.

RF