I recommend http://owenduffy.net/blog/?p=5442#more-5442, a new posting
by Owen Duffy (who in days of yore when rec.radio.amateur.antenna had a
number of people like him posting regularly here, was one of the best of
them).

Often the correctness of ideas can be tested by pushing their
applicability to extremes. He considers a two-wire air-insulated line
operating at 10 MHz, wonderfully short, just 30 mm long. Using it, he
shows that predictions made by some popular models for transmission line
loss cannot possibly be correct. To understand his argument you do not
need partial differential equations or Smith Charts or anything much
more complicated than Ohm's Law. This is what is nice about really short
lines, where for all intents and purposes the current and voltage do not
change along the line.

Caution: His argument clearly shows that loss is less in a mismatched
line with high load impedance than in a matched line, and more in a
mismatched line with low load impedance than in a matched line, for very
short lines. Do not apply his reasoning to longer lines. But his
argument does demolish the theory that additional loss depends only on
SWR.

He used to have a very nice line loss program on his old website, but
it's gone now. For a good one, one that does not just enter in the SWR
but instead uses the actual complex load impedance directly, see a Java
based program at http://fermi.la.asu.edu/w9cf/tran/ from Kevin
Schmidt, W9CF. Java can be dangerous to your computer's health, but his
program is OK. However opening it up on-line using Java will expose your
computer to evil things from other people while Java is running. You can
instead download his program, and then remove your computer from the
Internet while you run it.

David, VE7EZM and AF7BZ

--
David Ryeburn

To send e-mail, change "netz" to "net"

On 10/7/2015 1:45 AM, Jeff wrote:
On 07/10/2015 04:50, David Ryeburn wrote:
I recommend http://owenduffy.net/blog/?p=5442#more-5442, a new posting
by Owen Duffy (who in days of yore when rec.radio.amateur.antenna had a
number of people like him posting regularly here, was one of the best of
them).

Often the correctness of ideas can be tested by pushing their
applicability to extremes. He considers a two-wire air-insulated line
operating at 10 MHz, wonderfully short, just 30 mm long. Using it, he
shows that predictions made by some popular models for transmission line
loss cannot possibly be correct. To understand his argument you do not
need partial differential equations or Smith Charts or anything much
more complicated than Ohm's Law. This is what is nice about really short
lines, where for all intents and purposes the current and voltage do not
change along the line.

Caution: His argument clearly shows that loss is less in a mismatched
line with high load impedance than in a matched line, and more in a
mismatched line with low load impedance than in a matched line, for very
short lines. Do not apply his reasoning to longer lines. But his
argument does demolish the theory that additional loss depends only on
SWR.

He used to have a very nice line loss program on his old website, but
it's gone now. For a good one, one that does not just enter in the SWR
but instead uses the actual complex load impedance directly, see a Java
based program at http://fermi.la.asu.edu/w9cf/tran/ from Kevin
Schmidt, W9CF. Java can be dangerous to your computer's health, but his
program is OK. However opening it up on-line using Java will expose your
computer to evil things from other people while Java is running. You can
instead download his program, and then remove your computer from the
Internet while you run it.

David, VE7EZM and AF7BZ


I post is pretty much hogwash, it is implying that his results for a
very short line can be generalized and apply to any length of line. I
have not had time to go through his workings to see what other flaws
there are in them, BUT trying to draw conclusions from the analysis of a
short line is flawed anyway. A spice analysis may show the correct results.

Consider a very short discontinuity in the middle of a long 50 ohm line,
say a PL259/239 connector pair. The impedance of a PL259 is not 50 ohms
and usually somewhere nearer 100ohms. Do you see a 2:1 vswr when you use
it at 7MHz? No of course you don't, but you may well at 23cms. The
reason is the length of the discontinuity compared to the wavelength.

Jeff

I think I disagree. But we should wait until you read the article
entirely and think it over objectively.

On 10/7/2015 1:45 AM, Jeff wrote:
On 07/10/2015 04:50, David Ryeburn wrote:
I recommend http://owenduffy.net/blog/?p=5442#more-5442, a new posting
by Owen Duffy (who in days of yore when rec.radio.amateur.antenna had a
number of people like him posting regularly here, was one of the best of
them).

Often the correctness of ideas can be tested by pushing their
applicability to extremes. He considers a two-wire air-insulated line
operating at 10 MHz, wonderfully short, just 30 mm long. Using it, he
shows that predictions made by some popular models for transmission line
loss cannot possibly be correct. To understand his argument you do not
need partial differential equations or Smith Charts or anything much
more complicated than Ohm's Law. This is what is nice about really short
lines, where for all intents and purposes the current and voltage do not
change along the line.

Caution: His argument clearly shows that loss is less in a mismatched
line with high load impedance than in a matched line, and more in a
mismatched line with low load impedance than in a matched line, for very
short lines. Do not apply his reasoning to longer lines. But his
argument does demolish the theory that additional loss depends only on
SWR.

He used to have a very nice line loss program on his old website, but
it's gone now. For a good one, one that does not just enter in the SWR
but instead uses the actual complex load impedance directly, see a Java
based program at http://fermi.la.asu.edu/w9cf/tran/ from Kevin
Schmidt, W9CF. Java can be dangerous to your computer's health, but his
program is OK. However opening it up on-line using Java will expose your
computer to evil things from other people while Java is running. You can
instead download his program, and then remove your computer from the
Internet while you run it.

David, VE7EZM and AF7BZ


I post is pretty much hogwash,


This conclusion seems a bit strong for you, Jeff.


it is implying that his results for a
very short line can be generalized and apply to any length of line.


I cannot find where he implies that. Can you point to the particular
section where he alludes to this?


I
have not had time to go through his workings to see what other flaws
there are in them, BUT trying to draw conclusions from the analysis of a
short line is flawed anyway. A spice analysis may show the correct results.


Why is a short line analysis flawed? I kinda thought physics was the
same everywhere. Yes, please check it out and let us know your findings.


Consider a very short discontinuity in the middle of a long 50 ohm line,
say a PL259/239 connector pair. The impedance of a PL259 is not 50 ohms
and usually somewhere nearer 100ohms. Do you see a 2:1 vswr when you use
it at 7MHz? No of course you don't, but you may well at 23cms. The
reason is the length of the discontinuity compared to the wavelength.


But, that is not part of the analysis. Can you provide a disagreement
with his analysis under the same conditions?


Jeff

Jeff wrote:
On 07/10/2015 04:50, David Ryeburn wrote:
I recommend http://owenduffy.net/blog/?p=5442#more-5442, a new posting
by Owen Duffy (who in days of yore when rec.radio.amateur.antenna had a
number of people like him posting regularly here, was one of the best of
them).

Often the correctness of ideas can be tested by pushing their
applicability to extremes. He considers a two-wire air-insulated line
operating at 10 MHz, wonderfully short, just 30 mm long. Using it, he
shows that predictions made by some popular models for transmission line
loss cannot possibly be correct. To understand his argument you do not
need partial differential equations or Smith Charts or anything much
more complicated than Ohm's Law. This is what is nice about really short
lines, where for all intents and purposes the current and voltage do not
change along the line.

Caution: His argument clearly shows that loss is less in a mismatched
line with high load impedance than in a matched line, and more in a
mismatched line with low load impedance than in a matched line, for very
short lines. Do not apply his reasoning to longer lines. But his
argument does demolish the theory that additional loss depends only on
SWR.

He used to have a very nice line loss program on his old website, but
it's gone now. For a good one, one that does not just enter in the SWR
but instead uses the actual complex load impedance directly, see a Java
based program at http://fermi.la.asu.edu/w9cf/tran/ from Kevin
Schmidt, W9CF. Java can be dangerous to your computer's health, but his
program is OK. However opening it up on-line using Java will expose your
computer to evil things from other people while Java is running. You can
instead download his program, and then remove your computer from the
Internet while you run it.

David, VE7EZM and AF7BZ


I post is pretty much hogwash, it is implying that his results for a
very short line can be generalized and apply to any length of line. I
have not had time to go through his workings to see what other flaws
there are in them, BUT trying to draw conclusions from the analysis of a
short line is flawed anyway. A spice analysis may show the correct results.

Consider a very short discontinuity in the middle of a long 50 ohm line,
say a PL259/239 connector pair. The impedance of a PL259 is not 50 ohms
and usually somewhere nearer 100ohms. Do you see a 2:1 vswr when you use
it at 7MHz? No of course you don't, but you may well at 23cms. The
reason is the length of the discontinuity compared to the wavelength.

Jeff

The rule of thumb is that a line less than a tenth of wavelength long
shows negligible transmission line effects and can be viewed simply
as a shielded wire.


--
Jim Pennino

On 10/7/2015 1:12 PM, Jeff wrote:
On 07/10/2015 14:33, John S wrote:
On 10/7/2015 1:45 AM, Jeff wrote:
On 07/10/2015 04:50, David Ryeburn wrote:
I recommend http://owenduffy.net/blog/?p=5442#more-5442, a new
posting
by Owen Duffy (who in days of yore when rec.radio.amateur.antenna had a
number of people like him posting regularly here, was one of the
best of
them).

Often the correctness of ideas can be tested by pushing their
applicability to extremes. He considers a two-wire air-insulated line
operating at 10 MHz, wonderfully short, just 30 mm long. Using it, he
shows that predictions made by some popular models for transmission
line
loss cannot possibly be correct. To understand his argument you do not
need partial differential equations or Smith Charts or anything much
more complicated than Ohm's Law. This is what is nice about really
short
lines, where for all intents and purposes the current and voltage do
not
change along the line.

Caution: His argument clearly shows that loss is less in a mismatched
line with high load impedance than in a matched line, and more in a
mismatched line with low load impedance than in a matched line, for
very
short lines. Do not apply his reasoning to longer lines. But his
argument does demolish the theory that additional loss depends only on
SWR.

He used to have a very nice line loss program on his old website, but
it's gone now. For a good one, one that does not just enter in the SWR
but instead uses the actual complex load impedance directly, see a Java
based program at http://fermi.la.asu.edu/w9cf/tran/ from Kevin
Schmidt, W9CF. Java can be dangerous to your computer's health, but his
program is OK. However opening it up on-line using Java will expose
your
computer to evil things from other people while Java is running. You
can
instead download his program, and then remove your computer from the
Internet while you run it.

David, VE7EZM and AF7BZ


I post is pretty much hogwash,


This conclusion seems a bit strong for you, Jeff.


it is implying that his results for a
very short line can be generalized and apply to any length of line.


I cannot find where he implies that. Can you point to the particular
section where he alludes to this?


I
have not had time to go through his workings to see what other flaws
there are in them, BUT trying to draw conclusions from the analysis of a
short line is flawed anyway. A spice analysis may show the correct
results.


Why is a short line analysis flawed? I kinda thought physics was the
same everywhere. Yes, please check it out and let us know your findings.


Consider a very short discontinuity in the middle of a long 50 ohm line,
say a PL259/239 connector pair. The impedance of a PL259 is not 50 ohms
and usually somewhere nearer 100ohms. Do you see a 2:1 vswr when you use
it at 7MHz? No of course you don't, but you may well at 23cms. The
reason is the length of the discontinuity compared to the wavelength.


But, that is not part of the analysis. Can you provide a disagreement
with his analysis under the same conditions?



The analysis relies on taking a line that is 1/1000th of a wavelength
long which has pretty much a constant voltage on the line.

In the case of a line that short the analysis is correct because the
load impedance dominates over the characteristic impedance of the line.
However, when a source is connected to a load via a “long” transmission
line, the line’s own characteristic impedance dominates over load
impedance in determining circuit behaviour. In other words, an
electrically “long” line acts as the principal component in the circuit,
its own characteristics overshadowing the load’s.

Because of this the claim in the Conclusions that "A very simple
transmission line scenario that could be solved accurately using basic
linear circuit analysis was designed as a basis for evaluation of some
published techniques for predicting TL loss under mismatch" is untrue as
the analysis is ONLY valid for a very short line.

His statement does not imply his results and analysis will apply to a
long line. He is talking about analyzing a short line and comparing to
the results obtained with the other analysis method for a short line.


And of course the conclusion that "Reflections II did not reconcile with
the linear circuit analysis solution, and showed gross error" is correct
because of the flawed analysis.

I haven't read the article yet, but your statement of a flawed analysis
seems premature and certainly not supported by your ideas.

--

Rick

On 10/7/2015 12:12 PM, wrote:
Jeff wrote:
On 07/10/2015 04:50, David Ryeburn wrote:
I recommend http://owenduffy.net/blog/?p=5442#more-5442, a new posting
by Owen Duffy (who in days of yore when rec.radio.amateur.antenna had a
number of people like him posting regularly here, was one of the best of
them).

Often the correctness of ideas can be tested by pushing their
applicability to extremes. He considers a two-wire air-insulated line
operating at 10 MHz, wonderfully short, just 30 mm long. Using it, he
shows that predictions made by some popular models for transmission line
loss cannot possibly be correct. To understand his argument you do not
need partial differential equations or Smith Charts or anything much
more complicated than Ohm's Law. This is what is nice about really short
lines, where for all intents and purposes the current and voltage do not
change along the line.

Caution: His argument clearly shows that loss is less in a mismatched
line with high load impedance than in a matched line, and more in a
mismatched line with low load impedance than in a matched line, for very
short lines. Do not apply his reasoning to longer lines. But his
argument does demolish the theory that additional loss depends only on
SWR.

He used to have a very nice line loss program on his old website, but
it's gone now. For a good one, one that does not just enter in the SWR
but instead uses the actual complex load impedance directly, see a Java
based program at http://fermi.la.asu.edu/w9cf/tran/ from Kevin
Schmidt, W9CF. Java can be dangerous to your computer's health, but his
program is OK. However opening it up on-line using Java will expose your
computer to evil things from other people while Java is running. You can
instead download his program, and then remove your computer from the
Internet while you run it.

David, VE7EZM and AF7BZ


I post is pretty much hogwash, it is implying that his results for a
very short line can be generalized and apply to any length of line. I
have not had time to go through his workings to see what other flaws
there are in them, BUT trying to draw conclusions from the analysis of a
short line is flawed anyway. A spice analysis may show the correct results.

Consider a very short discontinuity in the middle of a long 50 ohm line,
say a PL259/239 connector pair. The impedance of a PL259 is not 50 ohms
and usually somewhere nearer 100ohms. Do you see a 2:1 vswr when you use
it at 7MHz? No of course you don't, but you may well at 23cms. The
reason is the length of the discontinuity compared to the wavelength.

Jeff

The rule of thumb is that a line less than a tenth of wavelength long
shows negligible transmission line effects and can be viewed simply
as a shielded wire.

He is using a pair of wires, not shielded. However, I'm sure that it
doesn't matter because the line is so short.

On 10/7/2015 12:12 PM, Jeff wrote:
On 07/10/2015 14:33, John S wrote:
On 10/7/2015 1:45 AM, Jeff wrote:
On 07/10/2015 04:50, David Ryeburn wrote:
I recommend http://owenduffy.net/blog/?p=5442#more-5442, a new
posting
by Owen Duffy (who in days of yore when rec.radio.amateur.antenna had a
number of people like him posting regularly here, was one of the
best of
them).

Often the correctness of ideas can be tested by pushing their
applicability to extremes. He considers a two-wire air-insulated line
operating at 10 MHz, wonderfully short, just 30 mm long. Using it, he
shows that predictions made by some popular models for transmission
line
loss cannot possibly be correct. To understand his argument you do not
need partial differential equations or Smith Charts or anything much
more complicated than Ohm's Law. This is what is nice about really
short
lines, where for all intents and purposes the current and voltage do
not
change along the line.

Caution: His argument clearly shows that loss is less in a mismatched
line with high load impedance than in a matched line, and more in a
mismatched line with low load impedance than in a matched line, for
very
short lines. Do not apply his reasoning to longer lines. But his
argument does demolish the theory that additional loss depends only on
SWR.

He used to have a very nice line loss program on his old website, but
it's gone now. For a good one, one that does not just enter in the SWR
but instead uses the actual complex load impedance directly, see a Java
based program at http://fermi.la.asu.edu/w9cf/tran/ from Kevin
Schmidt, W9CF. Java can be dangerous to your computer's health, but his
program is OK. However opening it up on-line using Java will expose
your
computer to evil things from other people while Java is running. You
can
instead download his program, and then remove your computer from the
Internet while you run it.

David, VE7EZM and AF7BZ


I post is pretty much hogwash,


This conclusion seems a bit strong for you, Jeff.


it is implying that his results for a
very short line can be generalized and apply to any length of line.


I cannot find where he implies that. Can you point to the particular
section where he alludes to this?


I
have not had time to go through his workings to see what other flaws
there are in them, BUT trying to draw conclusions from the analysis of a
short line is flawed anyway. A spice analysis may show the correct
results.


Why is a short line analysis flawed? I kinda thought physics was the
same everywhere. Yes, please check it out and let us know your findings.


Consider a very short discontinuity in the middle of a long 50 ohm line,
say a PL259/239 connector pair. The impedance of a PL259 is not 50 ohms
and usually somewhere nearer 100ohms. Do you see a 2:1 vswr when you use
it at 7MHz? No of course you don't, but you may well at 23cms. The
reason is the length of the discontinuity compared to the wavelength.


But, that is not part of the analysis. Can you provide a disagreement
with his analysis under the same conditions?



The analysis relies on taking a line that is 1/1000th of a wavelength
long which has pretty much a constant voltage on the line.

In the case of a line that short the analysis is correct because the
load impedance dominates over the characteristic impedance of the line.
However, when a source is connected to a load via a “long” transmission
line, the line’s own characteristic impedance dominates over load
impedance in determining circuit behaviour. In other words, an
electrically “long” line acts as the principal component in the circuit,
its own characteristics overshadowing the load’s.

Because of this the claim in the Conclusions that "A very simple
transmission line scenario that could be solved accurately using basic
linear circuit analysis was designed as a basis for evaluation of some
published techniques for predicting TL loss under mismatch" is untrue as
the analysis is ONLY valid for a very short line.

Okay, I accept that. However, since you agree that his analysis is
correct for his scenario, then you must admit that the equations and the
statements in quotes are not correct as they do not give the same answer
in his scenario.

And of course the conclusion that "Reflections II did not reconcile with
the linear circuit analysis solution, and showed gross error" is correct
because of the flawed analysis.

See above.

Jeff

On 10/8/2015 4:09 AM, Jeff wrote:

His statement does not imply his results and analysis will apply to a
long line. He is talking about analyzing a short line and comparing to
the results obtained with the other analysis method for a short line.

You obviously have not fully understood the article. The whole thrust of
it is comparing the v short analysis with statements from Maxwell's book
that are specific to a long line, as are the references from the ARRL
handbook etc.. He even quotes Maxwell: "Of course, the attenuation is
greater when there is a load mismatch, because in addition to the
attenuation of the forward power, the reflected power is also attenuated
during its return to the transmatch"; which is obviously talking about
the 'normal' set-up of Tx, ATU , feeder, then antenna.

Yes, the article is probably more or less correct for a very short line,
BUT ONLY FOR A VERY SHORT LINE, taking the analysis and assumptions
further, as the author does from 'REFLECTIONS II' onwards, is NOT CORRECT.

Let me pose you a question, take an infinitely long lossy transmission
line and load it with Zo/3 as in the article; what is the impedance seem
by the Tx, and what is the loss 1000m along the line?

If the 'simple linear solution' is to be believed the impedance seen at
the Tx depends on the load, but wait the wave has not got to the load
yet, a real measurement would show the impedance to be that of the line
Zo, and the loss whatever the loss/unit length of the feeder X1000
happens to be.

Zo of a transmission line is also known as the Surge Impedance, and that
is the impedance seen before any reflections come into play due to
mismatches etc.

If the circuit in question handles low-frequencies, such short time
delays are introduced by a transmission line between when the AC source
outputs a voltage peak and when the source “sees” that peak loaded by
the terminating impedance (round-trip time for the incident wave to
reach the line’s end and reflect back to the source) are of little
consequence. The actual phase difference between start-of-line and
end-of-line signals is negligible, because line-length propagations
occur within a very small fraction of the AC waveform’s period. For all
practical purposes, we can say that voltage along all respective points
on a low-frequency, two-conductor line are equal and in-phase with each
other at any given point in time.

This of course lead us back to the articles simple analysis only being
correct when the line is short cf a wavelength.

By contrast, an electrically long line where the propagation time is a
significant fraction or even a multiple of the signal period the
reflected signals phase is different enough to be of concern.

When a source is connected to a load via a 'long' transmission line, the
line’s own characteristic impedance dominates over load impedance in
determining circuit behaviour. In other words a long line acts as the
principal component in the circuit, its own characteristics
overshadowing the load’s. With a source connected to one end of the
cable and a load to the other, current drawn from the source is a
function primarily of the line and not the load.

This is increasingly true the longer the transmission line is. Consider
the infinite length cable above, no matter what kind of load we connect
to one end of this line, the source (connected to the other end) will
only see Zo, because the line’s infinite length prevents the signal from
ever reaching the end where the load is connected. In this scenario,
line impedance exclusively defines circuit behaviour, rendering the load
completely irrelevant.

I would also question the articles use of some formulas, for example, I
think that in the limit the equation for the impedance of a parallel
line breaks down and is not accurate. Also as an aside I find the
constant use of the wording "Maxwell's Equations" annoying and
misleading as they have nothing to do with the 'real' Maxwell's Equations!!

Anyway all of the article can be blown out of the water by some practice
measurement of a real life situation, which will show that a 3:1
mismatch will produce the same loss regardless of whether it is Zo*3 or
Z0/3 when you are talking about feeding an antenna.

Jeff

Great, Jeff! I would support your suggestion to make measurements. All
you need to do is set up his scenario and collect some data. Please use
his identical set-up to confirm or refute his results.

On 10/8/2015 4:40 AM, Jeff wrote:
Anyway all of the article can be blown out of the water by some practice
measurement of a real life situation, which will show that a 3:1
mismatch will produce the same loss regardless of whether it is Zo*3 or
Z0/3 when you are talking about feeding an antenna.

Jeff

Great, Jeff! I would support your suggestion to make measurements. All
you need to do is set up his scenario and collect some data. Please use
his identical set-up to confirm or refute his results.


If you have understood what I have been saying at all then you would
understand that I am not refuting his results (at least to a first
approximation). What I am refuting is the extension of his results to
the 'normal' case of feeding an antenna, which he tries to do in the
later parts of the article.


Okay. He says in his section "Failure in thinking":

"The forward and reflected waves give rise to E and I that vary along a
real transmission line, and the loss is due to I^2R loss in conductors
and E^2G loss in dielectric, so the loss in any incremental length of
line depends on E and I at that point. The loss in any line then is the
sum of the incremental losses due to varying E and I along the line."

His use of "real transmission line" now turns my attention to the
practical situations. It seems to me that he means to point out that the
losses are not smoothed over the length, but the final averaged result
can be influenced by the incremental, summed losses rather than just the
reflection coefficient which the equations use.

In reality you do not need to do the experiments, they have been done
many times, and the loss in a feeder under mismatch conditions is well
know and documented. It is these results that the author is challenging
from the wrong standpoint of a short line which is not extensible to
other cases.

Please point our where he extended his analysis to other cases.

Jeff

Well, you suggested that the experiments be performed and I applaud you
for that. If you can provide data on experiments which refute his
analysis, please do so.

In message , Jeff writes
Anyway all of the article can be blown out of the water by some practice
measurement of a real life situation, which will show that a 3:1
mismatch will produce the same loss regardless of whether it is Zo*3 or
Z0/3 when you are talking about feeding an antenna.

Jeff

Great, Jeff! I would support your suggestion to make measurements. All
you need to do is set up his scenario and collect some data. Please use
his identical set-up to confirm or refute his results.


If you have understood what I have been saying at all then you would
understand that I am not refuting his results (at least to a first
approximation). What I am refuting is the extension of his results to
the 'normal' case of feeding an antenna, which he tries to do in the
later parts of the article.

In reality you do not need to do the experiments, they have been done
many times, and the loss in a feeder under mismatch conditions is well
know and documented. It is these results that the author is challenging
from the wrong standpoint of a short line which is not extensible to
other cases.

My understanding is that when there is a standing wave on a feeder, the
additional SWR loss occurs because the higher 'I-squared R' losses at
the current maxima outweigh the lower 'I-squared R' losses at the
current minima.

If the feeder is not long enough to have a current maximum (or anything
like it), and the load mismatch is higher than Zo, the typical current
on the feeder can be lower than it would be if it was correctly
terminated. If so, the feeder loss can actually be lower than would be
if it was correctly terminated.

As I confused?


--
Ian