Tom, W8JI wrote:
"A traditional directional coupler works by comparing voltage across the
line at any one point to current in the line at that same point."

Almost. It compares a voltage sample to a current sample, both of which
have been converted into d-c voltages. These have been carefully crafted
to be exactly equal d-c voltages regardless of the power level in the
line.

I`m giving up on correcting line by line.

Important fact is that a reflection reverses the phase between the
voltage and current produced by a wave.

So when the samples from the forward wave are siummed, their total is
exactly 2x the value of either the voltage-derived sample or the
current-derived sample.

When the samples from the reflected wave are summed, being equal but
opposite in polarity, they add to ZERO. Calibration is so the total
produces the correct value on the power scale for the wave in the
forward direction.

To get the power in the reverse direction, the input and output are
effectively exchanged so that the forward power indication cancels and
the reverse power indication is produced by the sum of its voltage and
current d-c sample outputs.

Best regards, Richard Harrison, KB5WZI

Richard Harrison wrote:
Tom, W8JI wrote:
"A traditional directional coupler works by comparing voltage across the
line at any one point to current in the line at that same point."

Almost. It compares a voltage sample to a current sample, both of which
have been converted into d-c voltages. These have been carefully crafted
to be exactly equal d-c voltages regardless of the power level in the
line.

That's absolutely incorrect Richard.

If you get out the schematic of ANY directional coupler, you will see
the current sampling device is in series with a voltage sampling
device.

The radio frequency voltage ratios of sampling system are combined
BEFORE detection.

The dc voltage level does vary with both voltage and current (power),
and that is why the meter on the front of your watt meter goes up and
down with power levels.

Only a phase detector levels voltages.

73 Tom

On Tue, 14 Mar 2006 17:03:49 -0600, (Richard
Harrison) wrote:

Tom, W8JI wrote:
"A traditional directional coupler works by comparing voltage across the
line at any one point to current in the line at that same point."

Almost. It compares a voltage sample to a current sample, both of which
have been converted into d-c voltages. These have been carefully crafted
to be exactly equal d-c voltages regardless of the power level in the
line.

When you convert a sample of the RF voltage at a point to DC voltage
you have a DC voltage proportional to the magnitude of the RF voltage
at the point.

When you convert a sample of the RF current at the same point to DC
voltage you have a DC voltage proportional to the magnitude of the RF
current at the point.

You cannot use these two samples to determine generally the magnitude
of the reflection coefficient (and hence VSWR), much less independent
samples of the travelling waves.

No, neither a directional coupler nor a common reflectoter work as you
describe Richard.

Owen
--

Richard Harrison wrote:
Tom, W8JI wrote:
"A traditional directional coupler works by comparing voltage across the
line at any one point to current in the line at that same point."

Almost. It compares a voltage sample to a current sample, both of which
have been converted into d-c voltages. These have been carefully crafted
to be exactly equal d-c voltages regardless of the power level in the
line.

I`m giving up on correcting line by line.

Important fact is that a reflection reverses the phase between the
voltage and current produced by a wave.

So when the samples from the forward wave are siummed, their total is
exactly 2x the value of either the voltage-derived sample or the
current-derived sample.

When the samples from the reflected wave are summed, being equal but
opposite in polarity, they add to ZERO. Calibration is so the total
produces the correct value on the power scale for the wave in the
forward direction.

To get the power in the reverse direction, the input and output are
effectively exchanged so that the forward power indication cancels and
the reverse power indication is produced by the sum of its voltage and
current d-c sample outputs.

Thank you for this concise summary.

John Popelish wrote:
To get the power in the reverse direction, the input and output are
effectively exchanged so that the forward power indication cancels and
the reverse power indication is produced by the sum of its voltage and
current d-c sample outputs.

Thank you for this concise summary.

Except it is actually an incorrect concise summary.

The directional coupler adds RF voltage from a sampling across the line
directly to a sampling of RF current past that point.

It is only after the voltages, one proportional to current and one
proportional to voltage, are added that the resulting voltage is
rectified and used to drive a meter.

The directional effect can be analyzed using wave theory or simple
circuit theory. The results are the same.

73 Tom

Tom wrote, "The directional effect can be analyzed using wave theory or
simple
circuit theory. The results are the same."

Of course, "the directional effect" depends completely on having the
sampler calibrated to the impedance of the line into which it's
inserted. Otherwise, it's just resolving "forward" and "reverse"
_as_if_ the signal is in a line that has a characterisitc impedance
equal to the sampler's calibration impedance.

To the extent the samples are accurate for instantaneous currents and
voltages, the sampler does NOT depend on sinusoidal excitation. The
result is accurate for the current and voltage that exist at each
instant in time. Some directional couplers are very broadband; others
are not. We made the ones in the 8753 that Tom uses to be accurate
over a wide frequency range. And of course, if you don't just rectify
the output, you can extract phase information from it as well as
amplitude.

Cheers,
Tom

wrote:
John Popelish wrote:

To get the power in the reverse direction, the input and output are
effectively exchanged so that the forward power indication cancels and
the reverse power indication is produced by the sum of its voltage and
current d-c sample outputs.

Thank you for this concise summary.


Except it is actually an incorrect concise summary.

The directional coupler adds RF voltage from a sampling across the line
directly to a sampling of RF current past that point.

It is only after the voltages, one proportional to current and one
proportional to voltage, are added that the resulting voltage is
rectified and used to drive a meter.

The directional effect can be analyzed using wave theory or simple
circuit theory. The results are the same.

I appreciate the correction. I am weak in the area of RF
instrumentation, but am learning fast. It is an area I have somehow
avoided for a long time, but am getting interested in it, lately.

I would very much like to see a more complete report on the
measurements you have made, in relation to this thread. I am sure I
would learn from seeing that. I tried to find an operating manual or
application note on the network analyzer you used, but found little
that was helpful to teach me how it works, and how one applies it. It
seems to have 4 signal connectors (if I am interpreting what I have
found, correctly) and I can interpret your web account to mean several
possible things, so I am still under a bit of a cloud, here. Your
tutelage is much appreciated.

John Popelish wrote:

I appreciate the correction. I am weak in the area of RF
instrumentation, but am learning fast. It is an area I have somehow
avoided for a long time, but am getting interested in it, lately.

Good. It is an interesting area of electronics. If you are comfortable
with RF circuitry, RF instrumentation is only a small additional step.

I would very much like to see a more complete report on the
measurements you have made, in relation to this thread.

The problem is always time. I'm at the busiest time of the year for me,
so everything that isn't a fore is sitting. I really swore I wouldn't
get involved in an ungoing three year debate, but here I am anyway. I
guess I needed a break from a constant string of projects all with
tight deadlines.

I am sure I
would learn from seeing that. I tried to find an operating manual or
application note on the network analyzer you used, but found little
that was helpful to teach me how it works, and how one applies it. It
seems to have 4 signal connectors (if I am interpreting what I have
found, correctly) and I can interpret your web account to mean several
possible things, so I am still under a bit of a cloud, here. Your
tutelage is much appreciated.

This is the closest manual I could find.

http://www.home.agilent.com/cgi-bin/...OUNTRY_CODE=US

For this:

http://www.home.agilent.com/USeng/na...881282/pd.html

Agilent seems to obsolete things after seven years.

I have some useful equipment. Including an Impedance test set I paid
about 20K for in the 90's. It directly measures almost anything you
would every want to know.

The nice thing about having test gear is being able to build almost
anything. I have it because of work. You can do a lot with almost
nothing except a vector voltmeter and a test fixture, but the automated
measurements save me time.

73 Tom

wrote:
John Popelish wrote:
(snip)

I would very much like to see a more complete report on the
measurements you have made, in relation to this thread.

The problem is always time. I'm at the busiest time of the year for me,
so everything that isn't a fore is sitting. I really swore I wouldn't
get involved in an ungoing three year debate, but here I am anyway. I
guess I needed a break from a constant string of projects all with
tight deadlines.

I can appreciate that. I have recently gotten sucked into a wide
ranging study of ferrite rod antenna basics, and am having trouble
finding time to go to work or to bed. Almost every preconceived
notion I had about them I have been able to disprove by direct
measurement. Very educational.


I am sure I
would learn from seeing that. I tried to find an operating manual or
application note on the network analyzer you used, but found little
that was helpful to teach me how it works, and how one applies it.
(snip)

This is the closest manual I could find.

http://www.home.agilent.com/cgi-bin/...OUNTRY_CODE=US

For this:

http://www.home.agilent.com/USeng/na...881282/pd.html

I'll take a good look at these.

Agilent seems to obsolete things after seven years.

I have some useful equipment. Including an Impedance test set I paid
about 20K for in the 90's. It directly measures almost anything you
would every want to know.
(snip)

So far, I am working with an RF volt meter, a signal generator or two,
and an antique Boonton 160A Q meter. But I am finding lots of ways to
put them to use. I would love to have a vector volt meter or vector
impedance meter. A network analyzer is way beyond my budget.

John P. wrote, "A network analyzer is way beyond my budget."

Though I'd love for you to buy a nice new Agilent Vector Network
Analyzer, I have to say that for things up through low VHF at least,
the very economical project at
http://users.adelphia.net/~n2pk/index.html is well worth looking at.
The performance, when properly calibrated, needs no apologies, for
sure. Paul is one of the Good Guys in ham radio, and not just for
making this project available.

One of Tom's links probably won't work for you, but I'd highly
recommend some of the ap notes you can find after a somewhat diligent
search. I'm sorry to say that the search engine on the Agilent web
site is a poor relative of Google, but you should be able to find these
two ap notes the

Agilent AN 1287-1: Understanding the Fundamental Principles of Vector
Network Analysis

Agilent 1291-1B: 10 Hints for Making Better Network Analysis
Measurements.

They are in PDF files, and I just saved a copy of each...just in case
you can't find them.

A Google search on phrases like "network analysis application note" and
"VNA application note" should yield some interesting things. Here's
one thing I found, which has links to others:
http://na.tm.agilent.com/vnahelp/appnotes.html

Finally, there is an old HP ap note on S-parameters that you should try
to find. It _may_ be on the Agilent web site, but if not, a Google
search will probably turn it up. Though a Vector Network Analyzer does
not necessarily have to do S-parameter measurements, the 8753 is set up
to do them as its fundamental measurement, and they are generally
useful in making higher frequency measurements, since the standard
methodology in the industry is to use S parameters to characterize both
passive and active devices. If you look at network analyzers on eBay,
you may see ones offered without the S-parameter test set, and you can
find the S-parameter test sets offered separately; that's all fine so
long as you understand what you are looking at.

Cheers,
Tom