Well, After many trials and much learning I believe the 8405a is currently
working and actually making measurements. The breakthrough was a web article
showing Tau ranging from -1 to +1. This finally made sense of the comments
posted here.

Now to the setup and the measurements. I have built a loaded vertical for 2
meters and placed it over a 3 foot by 5 foot ground screen. The exact frequency
of the antenna is unknown. This is assembly moved into another room and
connected via an unknown length of 50 Ohm cable. The technique is labor
intensive, but in the end produces consistent results.

The 8405a is connected to a dual directional coupler. The connector lengths are
tuned to produce zero phase. The coupler is driven by a signal generator and
connected to the antenna assembly.

Now the tuning process: It involves tuning the antenna for a zero phase return
and then temporally replacing the antenna with a 25 Ohm terminator. The 8405a is
set to zero phase adjusted to he terminator. The antenna is then re-substituted
for the terminator and the frequency is adjusted to produce the next zero phase.
The cycle is the frequency is adjusted to the antenna and the meter zeroed phase
adjusted to the terminator. This is repeated.

The first attempt did not resolve. That is it the values did not trend toward a
common frequency and phase. However the second attempt, which was just a few MHz
away did. The result was a measurement that was consistent and repeatable. The
meter was set to the 6 degree scale and the measurements were sensitive to 5 kc
differences in frequency. The results were consistent to 1/5 a degree.

The Smith Chart shows for a purely resistive load that the return phase from a
fixed cable and frequency is independent of R. That is the return phase is
constant for a frequency, fixed cable, and pure resistive load. That would imply
that when the phase return is the same for a resistive load and an unknown then
the unknown is a pure resistive load. This is a way to normalize the effects of
an unknown cable to determine resonance.

I think this actually works. Is there a better way to do this?

Thanks - Dan

"dansawyeror" wrote in message
...
Well, After many trials and much learning I believe the 8405a is currently
working and actually making measurements. The breakthrough was a web
article showing Tau ranging from -1 to +1. This finally made sense of the
comments posted here.

Now to the setup and the measurements. I have built a loaded vertical for
2 meters and placed it over a 3 foot by 5 foot ground screen. The exact
frequency of the antenna is unknown. This is assembly moved into another
room and connected via an unknown length of 50 Ohm cable. The technique is
labor intensive, but in the end produces consistent results.

The 8405a is connected to a dual directional coupler. The connector
lengths are tuned to produce zero phase. The coupler is driven by a signal
generator and connected to the antenna assembly.

Now the tuning process: It involves tuning the antenna for a zero phase
return and then temporally replacing the antenna with a 25 Ohm terminator.
The 8405a is set to zero phase adjusted to he terminator. The antenna is
then re-substituted for the terminator and the frequency is adjusted to
produce the next zero phase. The cycle is the frequency is adjusted to the
antenna and the meter zeroed phase adjusted to the terminator. This is
repeated.

The first attempt did not resolve. That is it the values did not trend
toward a common frequency and phase. However the second attempt, which was
just a few MHz away did. The result was a measurement that was consistent
and repeatable. The meter was set to the 6 degree scale and the
measurements were sensitive to 5 kc differences in frequency. The results
were consistent to 1/5 a degree.

The Smith Chart shows for a purely resistive load that the return phase
from a fixed cable and frequency is independent of R. That is the return
phase is constant for a frequency, fixed cable, and pure resistive load.
That would imply that when the phase return is the same for a resistive
load and an unknown then the unknown is a pure resistive load. This is a
way to normalize the effects of an unknown cable to determine resonance.

I think this actually works. Is there a better way to do this?

Thanks - Dan

Why are you using a 25 ohm termination? The procedure for calibration
requires a short, open, and 50 ohm load. It is also important that you know
the coupler directivity at the test frequency. There is nothing wrong with
calibrating the test fixture at the end of an unknown length of coax;
although you will experience some degradation in the dynamic range of the
return loss.

Regards,

Frank

Frank,

The Smith Chart shows that a termination value other then 50 Ohms is required to
calibrate the 8405a to zero phase for the frequency and cable. That said there
are two zero points, I will have to think about the effect of these on the
measurements.

Frank wrote:
"dansawyeror" wrote in message
...

Well, After many trials and much learning I believe the 8405a is currently
working and actually making measurements. The breakthrough was a web
article showing Tau ranging from -1 to +1. This finally made sense of the
comments posted here.

Now to the setup and the measurements. I have built a loaded vertical for
2 meters and placed it over a 3 foot by 5 foot ground screen. The exact
frequency of the antenna is unknown. This is assembly moved into another
room and connected via an unknown length of 50 Ohm cable. The technique is
labor intensive, but in the end produces consistent results.

The 8405a is connected to a dual directional coupler. The connector
lengths are tuned to produce zero phase. The coupler is driven by a signal
generator and connected to the antenna assembly.

Now the tuning process: It involves tuning the antenna for a zero phase
return and then temporally replacing the antenna with a 25 Ohm terminator.
The 8405a is set to zero phase adjusted to he terminator. The antenna is
then re-substituted for the terminator and the frequency is adjusted to
produce the next zero phase. The cycle is the frequency is adjusted to the
antenna and the meter zeroed phase adjusted to the terminator. This is
repeated.

The first attempt did not resolve. That is it the values did not trend
toward a common frequency and phase. However the second attempt, which was
just a few MHz away did. The result was a measurement that was consistent
and repeatable. The meter was set to the 6 degree scale and the
measurements were sensitive to 5 kc differences in frequency. The results
were consistent to 1/5 a degree.

The Smith Chart shows for a purely resistive load that the return phase
from a fixed cable and frequency is independent of R. That is the return
phase is constant for a frequency, fixed cable, and pure resistive load.
That would imply that when the phase return is the same for a resistive
load and an unknown then the unknown is a pure resistive load. This is a
way to normalize the effects of an unknown cable to determine resonance.

I think this actually works. Is there a better way to do this?

Thanks - Dan


Why are you using a 25 ohm termination? The procedure for calibration
requires a short, open, and 50 ohm load. It is also important that you know
the coupler directivity at the test frequency. There is nothing wrong with
calibrating the test fixture at the end of an unknown length of coax;
although you will experience some degradation in the dynamic range of the
return loss.

Regards,

Frank

"dansawyeror" wrote in message
...
Frank,

The Smith Chart shows that a termination value other then 50 Ohms is
required to calibrate the 8405a to zero phase for the frequency and cable.
That said there are two zero points, I will have to think about the effect
of these on the measurements.

Dan,

I am not sure how you arrived at the above conclusion. The procedure should
be as follows: Short the end of the coax at the antenna location, then
adjust the line stretcher for a 180 degrees phase shift. This is your
reference reflection coefficient of -1 (i.e. 1 180). Remove the short and
verify the reflection coefficient is 1 0. Connect a known 50 ohm load,
which should provide a return loss of 30 dB. The angle of the reflection
coefficient is irrelevant. If the return loss of the 50 ohm load is
significantly less than 30 dB, then either your load is not very good, or
you have poor coupler directivity. Attach the antenna, measure, and record,
the magnitude and angle of the reflection coefficient. Repeat the procedure
for all frequencies of interest. You will then be able to plot the result
on the Smith Chart. At the frequencies you are working, which I believe is
in the 100 to 200 MHz range, it is doubtful that the non-ideal short/load
standards parasitics will effect the results significantly.

I have done such measurements countless times, although with HP VNAs, which
makes the procedure far less tedious.

Hope this helps,

Frank

Frank,

Your comment about my post is correct upon analysis it is not relevant.

I am trying to identify the resonance frequency of an antenna. When that point
is found I am trying to measure the input impedance (should be R + 0j).

Assuming 50 Ohm source and cable, the smith chart shows for lengths of cable
terminated in values other the 50 Ohms, say 25R 0j, those points will plot of a
constant SWR = 2 circle. I am using that circle below:

I am assuming that for a frequency where the antenna is resonant that the phase
read from the antenna and the phase read from non 50 Ohm 0j load will be the
same. If that is true then 'zeroing' the meter by adjusting the phase offset
will not effect the frequency of resonance. It is simply a convenience.

This method should yield a direct resonant frequency reading.

Dan

Frank wrote:
"dansawyeror" wrote in message
...

Frank,

The Smith Chart shows that a termination value other then 50 Ohms is
required to calibrate the 8405a to zero phase for the frequency and cable.
That said there are two zero points, I will have to think about the effect
of these on the measurements.


Dan,

I am not sure how you arrived at the above conclusion. The procedure should
be as follows: Short the end of the coax at the antenna location, then
adjust the line stretcher for a 180 degrees phase shift. This is your
reference reflection coefficient of -1 (i.e. 1 180). Remove the short and
verify the reflection coefficient is 1 0. Connect a known 50 ohm load,
which should provide a return loss of 30 dB. The angle of the reflection
coefficient is irrelevant. If the return loss of the 50 ohm load is
significantly less than 30 dB, then either your load is not very good, or
you have poor coupler directivity. Attach the antenna, measure, and record,
the magnitude and angle of the reflection coefficient. Repeat the procedure
for all frequencies of interest. You will then be able to plot the result
on the Smith Chart. At the frequencies you are working, which I believe is
in the 100 to 200 MHz range, it is doubtful that the non-ideal short/load
standards parasitics will effect the results significantly.

I have done such measurements countless times, although with HP VNAs, which
makes the procedure far less tedious.

Hope this helps,

Frank

On Sun, 22 Jan 2006 09:29:38 -0800, dansawyeror
wrote:

Frank,

Your comment about my post is correct upon analysis it is not relevant.

I am trying to identify the resonance frequency of an antenna. When that point
is found I am trying to measure the input impedance (should be R + 0j).

Assuming 50 Ohm source and cable, the smith chart shows for lengths of cable
terminated in values other the 50 Ohms, say 25R 0j, those points will plot of a
constant SWR = 2 circle.

Only when the cable Zo = 50 +j0.

"dansawyeror" wrote in message
...
Frank,

Your comment about my post is correct upon analysis it is not relevant.

I am trying to identify the resonance frequency of an antenna. When that
point is found I am trying to measure the input impedance (should be R +
0j).

Assuming 50 Ohm source and cable, the smith chart shows for lengths of
cable terminated in values other the 50 Ohms, say 25R 0j, those points
will plot of a constant SWR = 2 circle. I am using that circle below:

I am assuming that for a frequency where the antenna is resonant that the
phase read from the antenna and the phase read from non 50 Ohm 0j load
will be the same. If that is true then 'zeroing' the meter by adjusting
the phase offset will not effect the frequency of resonance. It is simply
a convenience.

This method should yield a direct resonant frequency reading.

Dan

Dan, The reflection coefficient of 25 ohms is 0.333 180, so if you do trim
the line stretcher for 180 degrees, then attaching the antenna will
determine how close you are to nominal R + j0. You need to reset your line
stretcher at each frequency until you obtain a 180 degree phase shift on the
return loss from the antenna. It is reasonable to assume that your antenna
has an impedance in the region of 37 ohms, so the 180 phase shift is
probably correct. For the reflection coefficient at 0.333 0 the input
impedance is 100 ohms. For such an antenna it is probably easer to just
tune the antenna for minimum reflection coefficient and forget the phase
angle.

Frank

I am assuming the reflection angle of 0j terminations would all be the same (you
said 180), this includes both resistive loads and an antenna at resonance. The
method I am using relies on that.

I am confused about one thing in the smith chart program. Cables identified as
stubs create a circle that goes through infinity and do not exhibit a constant
SWR, while simple cables create a constant SWR circle with the center and 50 Ohm
0j.

I believe the setup I am testing exhibits the constant SWR characteristic.

Dan

Frank wrote:
"dansawyeror" wrote in message
...

Frank,

Your comment about my post is correct upon analysis it is not relevant.

I am trying to identify the resonance frequency of an antenna. When that
point is found I am trying to measure the input impedance (should be R +
0j).

Assuming 50 Ohm source and cable, the smith chart shows for lengths of
cable terminated in values other the 50 Ohms, say 25R 0j, those points
will plot of a constant SWR = 2 circle. I am using that circle below:

I am assuming that for a frequency where the antenna is resonant that the
phase read from the antenna and the phase read from non 50 Ohm 0j load
will be the same. If that is true then 'zeroing' the meter by adjusting
the phase offset will not effect the frequency of resonance. It is simply
a convenience.

This method should yield a direct resonant frequency reading.

Dan


Dan, The reflection coefficient of 25 ohms is 0.333 180, so if you do trim
the line stretcher for 180 degrees, then attaching the antenna will
determine how close you are to nominal R + j0. You need to reset your line
stretcher at each frequency until you obtain a 180 degree phase shift on the
return loss from the antenna. It is reasonable to assume that your antenna
has an impedance in the region of 37 ohms, so the 180 phase shift is
probably correct. For the reflection coefficient at 0.333 0 the input
impedance is 100 ohms. For such an antenna it is probably easer to just
tune the antenna for minimum reflection coefficient and forget the phase
angle.

Frank

dansawyeror wrote:
I am assuming the reflection angle of 0j terminations would all be the
same (you said 180), this includes both resistive loads and an antenna
at resonance. The method I am using relies on that.

I am confused about one thing in the smith chart program. Cables
identified as stubs create a circle that goes through infinity and do
not exhibit a constant SWR, while simple cables create a constant SWR
circle with the center and 50 Ohm 0j.

I believe the setup I am testing exhibits the constant SWR characteristic.

No real-world transmission line exhibits a constant SWR since
no real-world transmission line is lossless. Constant SWR
circles are approximations. The actual SWR curve is a spiral
from a point on that constant SWR circle that, in the limit,
spirals down to the Z0 of the line.

For instance, using Owen's feedline calculator, the SWR at a
500 ohm antenna fed with 100 feet of RG-58A is 10:1 and the
SWR at the other (source) end is 1.42:1 for 146 MHz.
--
73, Cecil http://www.qsl.net/w5dxp

I am assuming the reflection angle of 0j terminations would all be the same
(you said 180), this includes both resistive loads and an antenna at
resonance. The method I am using relies on that.

I am confused about one thing in the smith chart program. Cables
identified as stubs create a circle that goes through infinity and do not
exhibit a constant SWR, while simple cables create a constant SWR circle
with the center and 50 Ohm 0j.

I believe the setup I am testing exhibits the constant SWR characteristic.

Dan

Dan, if the load is resistive, and less than 50 ohms (assuming the center of
the Smith Chart is considered to be 50 ohms), then the angle of the
reflection coefficient is 180 degrees. If the load is resistive, and
greater than 50 ohms, then the reflection coefficient angle is 0 degrees.
As you approach the center of the Smith Chart the angle becomes less and
less relevant.

Shunt stubs - either open or shorted - always pass through "zero" (And
infinity - which you will not see since it is in parallel with the load
impedance) and the selected impedance on the Smith Chart. This is based on
the effect that an open stub, at quarter wave multiples, will exhibit
successive open and short circuits; as will a shorted stub.

Your method will exhibit a constant reflection coefficient circle (and
VSWR), with the angle varying from 180 degrees through zero degrees and then
back, through negative angles, to 180 degrees.

Frank