"Richard Clark" wrote
To date in this matter, I have yet to see any concrete value of source
Z offered from those of the NOT 50 Ohms camp. Further, I have yet to
see any of them offer any experimental confirmation of their assertion
________________

Please see the following. In the quote there, note the text starting "The
transmitter's output source impedance must be low...", and the following
sentences.

+ + +

Below is a quote from a paper titled "A Study of RF Intermodulation Between
FM Broadcast Transmitters Sharing Filterplexed or Co-located Antenna
Systems," by Geoffrey Mendenhall. Mendenhall is a registered professional
engineer, and now a VP for Harris Broadcast Division in Mason, OH. He is
responsible for the engineering research and design of the entire broadcast
transmitter product line for Harris: AM, FM & TV. Harris is the world's
largest supplier of broadcast transmitters.

This paper and quote has to be read here with some interpretation, because
it is an analysis of what happens when an in-band signal from one
transmitter is coupled into another transmitter when their antennas are
close together and/or when adequate filtering of the external signal is not
provided. But it is strictly applicable also for single tx and antenna
systems, where an antenna mismatch produces reflections back toward the
transmitter. In this case the "interfering signal" is not external, but a
reflection of the incident power of that tx.

QUOTE: Output return loss is a measure of the interfering signal that is
coupled into the output circuit versus the amount that is reflected back
from the output circuit without interacting with the non-linear device. To
understand this concept more clearly, we must remember that although the
output circuit of the transmitter is designed to work into a fifty ohm load,
the output source impedance of the transmitter is not fifty ohms. If the
source impedance were equal to the fifty ohm line impedance, half of the
transmitter's output power would be dissipated in its internal output source
impedance. The transmitter's output source impedance must be low compared
to the load impedance in order to achieve good efficiency. The transmitter
therefore looks like a voltage source driving a fifty ohm load. While the
transmission line is correctly terminated looking toward the antenna (high
return loss), the transmission line is greatly mismatched looking toward the
output circuit of the transmitter (low return loss). This means that power
coming out of the transmitter is completely absorbed by the load while
interfering signals fed into the transmitter are almost completely reflected
by the output circuit. END QUOTE

The transmitter topology in this study was a single PA tube operating Class
C. For these designs, an on-carrier return loss value of 2 dB or less is
rather common. At 2 dB the reflection coefficient is over 79%.

PAs comprised of multiple devices combined by balanced methods (e.g. 3dB
hybrids, Wilkinsons) can provide a source impedance closer to 50 ohms
(higher return loss). In these cases, power that is reflected off the load
and NOT re-reflected by the tx mostly is dissipated in resistive networks in
the PA combiner. However these networks do not provide a load for the
forward power from the tx, only for power reflected by the output
termination.

RF

Visit http://rfry.org for FM broadcast RF system papers.

Sorry, it still isn't clear.

Richard Fry wrote:
"Roy Lewallen" wrote

But we seem to now have a "true SWR" as opposed to
some other kind of SWR. And "true SWR connected to
the tx output" doesn't have any meaning at all to me.


My "true SWR" term is used is an attempt to differentiate between the SWR of
the antenna system, and the inaccuracies associated with trying to measure
it with devices that cannot isolate the incident power in the system from
internal reflections of that power. For the conditions and reasoning
outlined in my earlier posts in this thread, and even though the system SWR
is a constant -- the normal SWR meter used in/with an operating transmitter
working into a mismatched load won't have the ability to give strictly
accurate measurement of that SWR. That is all I'm saying.

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


I also have no idea of what "sample points within the
transmitter" might be.


In broadcast gear, these are the directional couplers whose pickup probes
are inserted transversely into the coaxial line between the harmonic filter
output and the tx output connector. I haven't been a licensed ham for over
40 years (when I went into the broadcast field), but I expect some ham txs
might have the same setup. Otherwise it could be a Model 43 or the like
inserted between the output connector of the ham tx and the transmission
line to the antenna.

In your last posting, you said,

Just one sec, please. I didn't say that the true SWR connected to the
tx output connector was affected. I said that the RF power measured
at the sample point(s) in the transmitter can be affected by the
source and load impedances of the tx, for the reasons stated.

So replacing "sample point(s) in the transmitter" with "Model 43 or the
like inserted between the output connector of the ham tx and the
transmission line to the antenna", you've said that the RF power
measured by the (model 43) SWR meter can be affected by the source
impedance of the transmitter.

Obviously, if we have a voltage or current source of fixed value and
change the source impedance, the power delivered by the source changes,
and any means of measuring the power at the source, load, or in between
should show that change. That follows from elementary circuit theory,
and doesn't require any consideration or knowledge of transmission
lines, waves, or SWR. On the model 43, both the "forward" and "reverse"
powers will change, but by the same fraction. Perhaps that's what you
mean. But if you mean that the SWR reading or the ratio of "forward" to
"reverse" power changes as a result of changing the source impedance,
that's easily shown to be false by the simple experiment I proposed.

I hope this is understandable now.

Almost, but not quite.

"Roy Lewallen" wrote in message
...
This has been explained many times, to no avail.

So instead of one of us explaining it yet again, I suggest that you do
the following experiment. It requires only a transmitter, one or two
dummy loads, an SWR meter, and no more than five minutes of your time.

1. Connect the transmitter to either a dummy load or an antenna through
the SWR meter and measure the SWR.

2. Connect the transmitter in parallel with a dummy load by using a tee
connector. Connect this parallel combination to the input of the SWR
meter, and the output of the SWR meter to the same load as before (dummy
load or antenna).

If you don't have 2 dummy loads, there is a simple alternative. Connect the
TX to the meter through a 1/4 wave section of 75 Ohm line. Unless the TX
output was 75 Ohms, the equivalent TX output impedance seen by the meter has
changed.

Tam/WB2TT

Do you see any change in the SWR?

If you don't, then something is wrong with your theory -- since the
source impedance is clearly different for the two measurements --, and
you should take the effort of resolving it with your recent observations.

Roy Lewallen, W7EL

Richard Fry wrote:
"Ian White, G3SEK"wrote:

Richard Fry wrote:

"Ian White, G3SEK wrote

The meter measures nothing that involves the source, except
the level of RF that it supplies. It does not respond in any way
whatever to the source impedance.

Not that I said it did in my part of the thread, but nevertheless the

above

statement is not strictly true. In the case where the source Z of the
tx

PA

does not match its load Z (which is typical), power reflected from the

load

mismatch will at least partly be re-reflected from the PA -- which then
contributes to the power sensed by a "wattmeter" in the output path.

Sorry, that statement cannot be correct. It would mean that the
impedance you measure at the near end of a transmission line (terminated
by some arbitrary load at the far end) would depend on the internal
impedance of the device that's doing the measuring - and that is not
true, either in transmission-line theory or in the real world. It is a
function only of the line and the load. etc

____________

How, then, do you explain the "ghost image" that can occur* in
analog(ue) TV
transmission systems arising from reflections at/near the antenna end of
the
station's transmission line?

*with sufficient round-trip propagation time in the transmission line

RF

On Sat, 4 Sep 2004 17:57:26 -0500, "Richard Fry" wrote:

"Richard Clark" wrote
To date in this matter, I have yet to see any concrete value of source
Z offered from those of the NOT 50 Ohms camp. Further, I have yet to
see any of them offer any experimental confirmation of their assertion
________________

Please see the following. In the quote there, note the text starting "The
transmitter's output source impedance must be low...", and the following
sentences.

+ + +

Below is a quote from a paper titled "A Study of RF Intermodulation Between
FM Broadcast Transmitters Sharing Filterplexed or Co-located Antenna
Systems," by Geoffrey Mendenhall. Mendenhall is a registered professional
engineer, and now a VP for Harris Broadcast Division in Mason, OH. He is
responsible for the engineering research and design of the entire broadcast
transmitter product line for Harris: AM, FM & TV. Harris is the world's
largest supplier of broadcast transmitters.

This paper and quote has to be read here with some interpretation, because
it is an analysis of what happens when an in-band signal from one
transmitter is coupled into another transmitter when their antennas are
close together and/or when adequate filtering of the external signal is not
provided. But it is strictly applicable also for single tx and antenna
systems, where an antenna mismatch produces reflections back toward the
transmitter. In this case the "interfering signal" is not external, but a
reflection of the incident power of that tx.

A critical point made in the quote below is evidence of a serious
misunderstanding concerning the relationship between the source impedance of the
tx and the load impedance.

QUOTE: Output return loss is a measure of the interfering signal that is
coupled into the output circuit versus the amount that is reflected back
from the output circuit without interacting with the non-linear device. To
understand this concept more clearly, we must remember that although the
output circuit of the transmitter is designed to work into a fifty ohm load,
the output source impedance of the transmitter is not fifty ohms. If the
source impedance were equal to the fifty ohm line impedance, half of the
transmitter's output power would be dissipated in its internal output source
impedance.

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.

I have made appropriate measurements in a professional RF laboratory that prove
this point. The data from these measurements and the procedure used is available
for downloading from my web site at http://home.iag.net/~w2du under the title
"On the Nature of the Source of Power in Class B and C Amplifiers." This piece
is Chapter 19 in Reflections II, and also appears in QEX,, May/Jun 2001.

Unfortunately, like the statement made in the 'quote' above, there are all too
many RF engineers who fail to appreciate the true relationship between the two
separate resistances in the amp, the resistance resulting in dissipation and the
resistance responsible for delivering the power to the load. I guarantee the
reader of the piece referenced above will come away with something to think
about.

The transmitter's output source impedance must be low compared
to the load impedance in order to achieve good efficiency. The transmitter
therefore looks like a voltage source driving a fifty ohm load. While the
transmission line is correctly terminated looking toward the antenna (high
return loss), the transmission line is greatly mismatched looking toward the
output circuit of the transmitter (low return loss). This means that power
coming out of the transmitter is completely absorbed by the load while
interfering signals fed into the transmitter are almost completely reflected
by the output circuit. END QUOTE

The transmitter topology in this study was a single PA tube operating Class
C. For these designs, an on-carrier return loss value of 2 dB or less is
rather common. At 2 dB the reflection coefficient is over 79%.

snip
RF

73,

Walt, W2DU

On Sat, 4 Sep 2004 14:46:38 -0500, "Richard Fry"
wrote:

|"Roy Lewallen" wrote
| Let me suggest an additional exercise for Richard and anyone else that
| believes that source impedance affects the SWR. (etc)
|____________________
|
|Just one sec, please. I didn't say that the true SWR connected to the tx
|output connector was affected. I said that the RF power measured at the
|sample point(s) in the transmitter can be affected by the source and load
|impedances of the tx, for the reasons stated.

Not so fast yourself. You said, "The generic function of this meter is
to measure the degree of match between a source and a load."

There is no power mentioned in your statement. I, and others, stated
that your first statement was incorrect and since that time you have
been introducing prodigious amounts of verbiage in an attempt to
obfuscate and avoid the obvious error in your earlier statement.

Just slap your forehead and say, "Shucks, I blew it with that one" and
we can all forget about it. I do it all of the time.

|
|The true load SWR does not change under these conditions, but it cannot then
|be determined by such a meter. Attempting to do so will yield some value,
|but it will be wrong.

Oh please. If an SWR meter, direction bridge, TLI or whatever you
want to call it has decent directivity, i.e. the ability to discern
forward and reflected power, forward and reflected waves, reflection
coefficient, scattering parameters, or whatever you want to call them,
then the applied power is immaterial.

We are trying to measure a RATIO, not some absolute value of power.

"Roy Lewallen" wrote

If we connect a transmitter to an SWR meter, and then to a long piece of
lossless cable with the same Z0 as the SWR meter, and finally to a load,
the SWR meter reading will be the same as the VSWR on the cable, i.e.,
the ratio of maximum to minimum voltages on the line.

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

It is at this point where impressionable novices are led astray by old
wives, never again to return to logical thought on the subject.

They imagine that because the meter happens to indicate the swr on the line,
the meter is actually responding to the swr on it. Whereas the meter is
actually responding to the modulus of the reflection coefficient caused by
the line's input impedance regardless of what its Zo may be. The act of
making the line's Zo and the meter's resistance both equal to the
transmitter's designed-for load resistance, has put additional infomation
into the system. Cooking the books!

If there's an SWR to be indicated it is on a long line between meter and the
transmitter. In the absence of such a line the meter wastefully discards
half of the information it is presented with and indicates the modulus of
the reflection cofficient. A more appropriate name is TLI.
----
Reg, G4FGQ

On Sat, 4 Sep 2004 14:08:53 +0000 (UTC), "Reg Edwards"
wrote:

NOTE: In the above description and calculation there is no mention of Zo,
terminating impedance, source impedance, reflection coefficient, forward
power, reflected power, reflected volts, reflected current, Smith charts, or
conjugate matches. All these things are superflous to the determination. No
information other than the two voltage measurements is needed.

Hi All,

No mention merely means there is no offer of accuracy (not very
important, eh what?). These two measurements (repeated at intervals)
can reveal a SWR that varies along the length of the line like a snake
- UNLESS of course, you DO observe unmentionables like Load reflection
co-efficients and Source reflection co-efficients. As such, a
description of how not to measure SWR, but rather how to exhibit error
if you perchance have the misfortune of having a transmitter that is
unmatched to a 50 Ohm transmission system whose load is in fact
mismatched also.

Need I point out that if both ends are matched - what's the point in
measuring SWR? ;-)

73's
Richard Clark, KB7QHC

In message , Reg Edwards
writes

"Roy Lewallen" wrote

If we connect a transmitter to an SWR meter, and then to a long piece of
lossless cable with the same Z0 as the SWR meter, and finally to a load,
the SWR meter reading will be the same as the VSWR on the cable, i.e.,
the ratio of maximum to minimum voltages on the line.

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

It is at this point where impressionable novices are led astray by old
wives, never again to return to logical thought on the subject.

They imagine that because the meter happens to indicate the swr on the line,
the meter is actually responding to the swr on it. Whereas the meter is
actually responding to the modulus of the reflection coefficient caused by
the line's input impedance regardless of what its Zo may be. The act of
making the line's Zo and the meter's resistance both equal to the
transmitter's designed-for load resistance, has put additional infomation
into the system. Cooking the books!

If there's an SWR to be indicated it is on a long line between meter and the
transmitter. In the absence of such a line the meter wastefully discards
half of the information it is presented with and indicates the modulus of
the reflection cofficient. A more appropriate name is TLI.
----
Reg, G4FGQ



Would this help?

On the subject of whether the TX impedance affected the SWR reading, I
propose the following practical test:
Using standard CATV bits and pieces, connect up the following-
Signal source directional coupler #1 directional coupler #2 load.
DC#1 picks off the forward signal, DC#2 picks off the reverse.
Use a spectrum analyser to measure signal levels.
Beforehand check the DCs for go directivity, and chose a frequency where
it is best (at least 25dB). This will probably be around 20MHz.
With good load, measure forward and reverse signals.
Repeat with known load mismatch.
Screw up source impedance (eg add T-piece at source o/p, and
double-terminate).
Repeat the above.
Think about what the results mean.

Ian.



--

Richard Fry quoted Geoffrey Mendenhall who is responsible for the entire
Harris broadcast transmitter line design as saying:
"If the source impedance were equal to the 50 ohm line impedance, half
the transmitter`s output would be dissipated in its internal output
source impedance."

Mendenhall is wrong.

A Class C amplifier`s "internal output source impedance" is largely
"dissipationless resistance" produced by the non-conduction time during
its RF cycle. A matched Class C amplifier typically produces
efficiencies exceeding 50% by a comfortable margin. That`s why they are
used despite their harmonic generating nonlinearity.

Best regards, Richard Harrison, KB5WZI

"Wes Stewart" wrote:
Oh please. If an SWR meter, direction bridge, TLI or whatever you
want to call it has decent directivity, i.e. the ability to discern
forward and reflected power, forward and reflected waves, reflection
coefficient, scattering parameters, or whatever you want to call them,
then the applied power is immaterial.

We are trying to measure a RATIO, not some absolute value of power.
_______________

However an SWR meter, direction(al) bridge, TLI or whatever you want to call
it has NO ability to discern between two waves traveling in the same
direction, unless their detectors are operating in the time domain -- which
normally they do not. That is the "feature" causing the anomalies I have
been writing about.

RF