Gentlemen, excuse me butting in.

I suspect all these learned articles to be invalid because the authors
incorrectly assume the source impedance to be constant and the load
impedance to be the variable in any analyses or sets of measurements.

Whereas the internal or source impedance is actually a function of the load
(and many other factors).

I have not read these papers or articles but base my comments on what I have
gleaned from newsgroup conversations over the years.

In brief, how can you have a conjugate match with the source impedance
hopping about trying to follow the load? ;o)
----
Reg, G4FGQ

===================================

You assume he refers to the source impedance of/at output of the
amplifier.
More likely he is following convention and stating the load impedance that
the amplifier was designed to work into.

The source impedance of most transmitters is not published even today. If
it was, probably we wouldn't be having all of this confusion about it, and
its effects.

"Richard Harrison" wrote
The way to avoid intermodulation is to keep the foreign signals
out of the electronics so they don`t mix. Well designed and
adjusted pass/reject fikters in the transmission circuits of
KPRC and KXYZ saw to that.
_____________

Thanks. You report more evidence that the source impedance did not match the
load impedance of these txs. If they did, each of these txs would absorb
the coupled signal of the other -- and neither of them would generate mixing
products.

RF

On Sun, 5 Sep 2004 14:13:36 -0500, "Richard Fry"
wrote:

The thesis has been proven.

Hi OM,

It's always fun to find in the heat of discussion an opponent who
impeaches his own witnesses. This is like how Reggie abandons Lords
Kelvinator and Plushbottom to their graves when a troll is so much
more entertaining with that glass of wine. Such are the vagaries of
esteem so lightly held by gossamer minds.

It has been quite amusing. :-)

73's
Richard Clark, KB7QHC

Richard Fry wrote:
"How do (you) find it justified to assign a literal quote to me that I
did not write in the first place?"

A mistake?

Sorry if I misquoted Richard Fry on peculair intermodulation products.

The power hungry RCA 5 C`s with their linear amplifiers were later
replaced with RCA BTA5F`s which had Class C final amplifiers. Despite a
phony ceremonial switchover to the new equipment in which pads were
pulled from the audio lines to the new transmitters, the stations really
sounded the same before and after the equipment change.

Best regards, Richard Harrison, KB5WZI

Reg, G4FGQ wrote:
"In brief, how can you have a conjugate match with the source impedance
hopping about trying to follow the load?"

I think you take control of the process and tune for maximum smoke.

Best regards, Richard Harrison, KB5WZI

On Sun, 5 Sep 2004 07:49:06 -0500, "Richard Fry" wrote:
snip
Walter Maxwell wrote
The last sentence in the paragraph above is incorrect. This shows that
the writer of the quote is in the unbelievably large group that still
believes
incorrectly that half of the tx power would be lost if if it were
conjugately
matched. But we all know that efficiencies greater than 80% is achieved
by Class C amps, and greater than 60% is achieved by Class B amps
when the source impedance of the tx is 50 ohms resistive and the load
impedance is also 50 ohms resistive.
_______________

To Walter Maxwell:

1. You may be interested in reading Mendenhall's complete paper, which I
will email to you. The lab measurements reported in it used two, operating,
high-power FM broadcast transmitters -- and support his statements about
amplifier source impedance and its consequences.

2. I will ask again, if transmitters have a 50 ohm source impedance, what
accounts for the fact that TV ghosts are produced by an antenna system
reflection having a sufficient delay time? Calculations and measured data
show that the energy that produced the ghost originated by re-reflection off
the TV transmitter output stage of far-end reflections in the antenna
system. If the tx source impedance was 50 ohms, it would absorb the far-end
reflection, which would be incapable of producing a ghost image.

Further, if the tx source impedance was 50 ohms, then the RF intermodulation
measured and reported in Mendenhall's paper -- and verified in real-world
installations by the radiated interference those IM products produced --
would not occur.

RF
Thank you Rick, I've received the Mendenhall paper, which I'll review and
comment later. However, before reading it I have one comment. On the condition
that the tx has tubes (and I assume it does) with some sort of LC output
coupling network, then if the source impedance of the tx is 50 ohms it will not
absorb the far-end reflection, because the source impedance of this type of tx
is not absortive. The source impedance of a resonant tank circuit is a
resistance determined by the voltage-current ratio in the tank--high resistance
at the tank input and low resistance at the output. The resistance being
proportional to the load line there is no dissipating resistor involved. The
only dissipative resistance in the system is the cathode-plate resistance, which
is separate from the output resistance. And contrary to what I've skimmed in the
Mendenhall paper, the output circuit of the tx is linear, not non-linear as
Mendenhall says, because the energy storage of the tank isolates the non-linear
input from the linear output. Remember, the tx output is a nearly perfect sine
wave.

I'll not comment further on this point, Rich, until I've reviewed the Mendenhall
paper.

"Walter Maxwell"
Thank you Rick, I've received the Mendenhall paper, which
I'll review and comment later. (etc)
______________

Thanks for your very civil response. I was on Geoff Mendenhall's staff for
some years before I retired, and will contact him to ask for his response to
whatever of your points seem appropriate for that. He's a very busy person
these days and may not have the time to get involved, but there's no harm in
trying.

RF

On Sun, 5 Sep 2004 14:53:47 -0500, "Richard Fry"
wrote:

|"Richard Harrison" wrote
| The way to avoid intermodulation is to keep the foreign signals
| out of the electronics so they don`t mix. Well designed and
| adjusted pass/reject fikters in the transmission circuits of
| KPRC and KXYZ saw to that.
|_____________
|
|Thanks. You report more evidence that the source impedance did not match the
|load impedance of these txs.

|If they did, each of these txs would absorb
|the coupled signal of the other -- and neither of them would generate mixing
|products.

Why would they? One is tuned to 950 KHz, the other to 1320 KHz, a
good part of an octave difference.

This is the same fuzzy logic that says that you can measure the output
Z of an amplifier at one frequency by injecting another signal at a
different frequency.

If it was this easy, the bother of load pull techniques wouldn't have
become popular and the norm.

On Sun, 5 Sep 2004 19:24:02 +0000 (UTC), "Reg Edwards"
wrote:

|Gentlemen, excuse me butting in.

Not at all, after all, you started this mess.
|
|I suspect all these learned articles to be invalid because the authors
|incorrectly assume the source impedance to be constant and the load
|impedance to be the variable in any analyses or sets of measurements.
|
|Whereas the internal or source impedance is actually a function of the load
|(and many other factors).
|
|I have not read these papers or articles but base my comments on what I have
|gleaned from newsgroup conversations over the years.

What a hoot. You constantly battle old wives and then say that you
base your comments on newsgroup conversation!
|
|In brief, how can you have a conjugate match with the source impedance
|hopping about trying to follow the load? ;o)

Iteration.

Richard Fry wrote:
"Roy Lewallen" wrote:

Sorry, it still isn't clear.

What, then, is "system SWR"? How do you define it?


System SWR is the net SWR of a component assembly present at its input
terminals. "Antenna system SWR" then is comprised of the net SWR of
everything in the RF path from the output of the SWR meter to and including
the antenna. In a transmitter, the antenna system begins electrically at
the output of the SWR meter -- physically close to the output connector of
the tx.

You've still lost me. Let's say the "component assembly" is a half
wavelength of 75 ohm transmission line terminated with a 75 ohm
resistor. What is its "net SWR"? How about a half wavelength of 75 ohm
line terminated with 50 ohms? Or a plain 75 ohm resistor? You surely
have an equation you use to calculate "system SWR" or "net SWR" -- can
you share it with us?

This is getting more complicated rather than simpler. We now have "true
SWR", "antenna system SWR", and "net SWR". Quite a step from the ratio
of maximum to minimum voltages on a transmission line.

. . .

Roy Lewallen, W7EL