Calibration
AAplot has the capability of calibrating away errors that
are introduced by a two-port network that is placed between
the bridge of the RigExpert analyzer and the actual
impedance that is being measured.
Such a two-port network can be a length of coax
transmission line that is connected between the instrument
and the feed point of an antenna, or it can be a
measurement "fixture" that is built to measure unknown
impedances, and can even be errors that are introduced by
the SO-239 connector and internal wiring of the antenna
analyzer itself.
The correction process used by AAplot follows the
"One-port, 3-term Error Model" that is
described by D. Rytting of
Hewlett-Packard and used in HP and Agilent Network
Analyzers.
The calibration process consists of storing measurements of
three known impedances for a collection of measured
frequencies. AAplot saves the reflection coefficient
measurements as calibration profiles in the
Application
Support folder of your home directory's
Library folder. The reflection coefficients in a
profile are later interpolated for use at other frequencies
when you make measurements of unknown impedances.
As it is usually used, the Rytting calibration is most
often performed using three "known" impedances that are (1)
a short circuit, (2) an open circuit and (3) a known load
resistance that is close to the terminal impedance of the
analyzer (e.g., 50 ohms).
However, the bridge topology that is implemented in the Rig
Expert antenna analyzers produces very large errors when
measuring impedances in regions where the VSWR is large
(i.e., when the absolute value of the reflection
coefficient is near 1.0). Because of that, AAplot lets you
use a moderately low resistance (e.g., 15 to 25 ohms) in
place of a short circuit load, and a moderately large
resistance (e.g., between 150 to 300 ohms) in place of the
open circuit load.
The Calibration Manager item in the Calibration menu in
AAplot's main menu bar will open the Calibration Manager's
window. The Table View in this window displays the custom
profiles for your analyzer that you have measured and
saved.
As with all Table Views in Mac OS X, you select the
calibration profile to use by clicking on the table view's
row. The above figure shows a profile that was named
"AA-230 (22-270-47)."
You can unselect all calibrations by clicking on an
unoccupied row of the table view.
(The name of the profile that is selected is saved to your
plist file so that the next time you launch AAplot, the
previous calibration profile, if any, will remain in
effect.)
The + and - boxes of the segmented control below the table
view let you add new profiles or remove a selected profile.
When the + box is pressed to add a new profile, you will be
presented with the calibration window:
As shown in the figure above, there are three measurements
that has to be performed.
One measurement (topmost group) is made by using either a
short circuit or a low impedance load.
As mentioned earlier, it is not advisable to choose a true
short circuit with a RigExpert analyzer. A "good" value to
use is somewhere between 15 ohms and 25 ohms. If the
termination has an inductance in series with a resistance
and a parallel capacitance across the inductive resistance,
you can include those values in the
Series L and
Parallel C fields.
The second measurement is made with either an open circuit
or a high impedance load. Again, as mentioned earlier,
because of the circuit topology that is used in RigExpert
analyzers, avoid the use of a true open circuit with a
RigExpert analyzer. A value of between 150 ohms and 300
ohms works well.
Finally, the third measurement should be made with a
resistance that is close to 50 ohms. It does not have to be
precisely 50 ohms. However, you do want the three impedance
values to be as different from one another as you can
construct them, while keeping the VSWR of each load to be
no greater than 6.0:1. This will result in the most stable
matrix inverses, while also avoiding the places where the
RigExpert has the biggest problem with.
You can perform the three measurements in any order.
The progress bar near the bottom of the window shows the
progress of a measurement pass. You can abandon a
calibration attempt simply by closing the Calibration
window when it is not actively measuring (the progress bar
is not active).
Profile
Names
When all three measurements are completed, the
Profile text field and
Save button at the
bottom of the window will become enabled. You can use any
profile name that does not include a / character (the
profiles are saved as folders; a / in the folder name will
be converted by Mac OS X into something else).
If you decide later to change the name of a profile, you
can quit AAplot and go into the AAplot folder in the
Application Support folder to change the profile folder's
name there. The new name will appear the next time you
launch AAplot.
Similarly, you can temporarily remove any currently unused
profile from the Application Support folder to reduce
clutter of the Calibration Manager window, and move it back
in at a later date. The profiles are folders that contain
two text files (with measured impedances and target
reflection coefficients).
Profiles can be permanently removed by using the - box in
the Calibration Manager window.
Reference
Loads
For the HF region, the reference loads can be constructed
using ordinary non-inductive resistors. I find that garden
variety 1/8 watt 2% axial resistors work quite well. I
construct my load inside the body of a PL-259 coax plug.
One end of the resistor is threaded through and soldered to
the coax connector's center pin and the other end of the
resistor is bent back and fed out through one of the solder
holes of the plug and soldered to the body of the plug.
Keep the leads short and keep the body of the resistor away
from the inside wall of the plug.
Measure the actual resistance with a good DVM and use that
as the resistance of the load. This should be sufficient
for non-critical measurements in the HF region.
Example 1 --
Calibrated analyzer vs Uncalibrated
analyzer
The following shows the measurement done on an uncalibrated
AA-230PRO. A 150 ohm resistor that is mounted inside a
PL-259 plug is connected directly to the SO239 connector on
the analyzer:
The solid yellow line is the resistive component and the
dotted yellow line is the reactive component of the
measured impedance.
The following shows the corrected measurement of the same
load. The calibration profile is constructed by inserting
the three calibrating loads (22.0 ohms, 271.2 ohms and 46.8
ohms) that are inserted directly at the AA-230PRO's SO-239
connector.
Example 2 --
Calibrated coax cable with a known resistive
load
The following shows the measured impedance of a 3m length
of RG-58 cable that is terminated by the same 150 ohm
resistor above. The frequency is swept from 1 MHz to 29
MHz:
The typical constant SWR arc that is formed by typical
short length (thus low loss) of transmission line is easiiy
recognized. The SWR circle is centered at the
characteristic impedance of the coax cable, which can be
seen to be a tad off from 50.0 ohms for the Smith Chart
center.
A calibration profile is created for this 3m cable by
calibrating it with the same 22 ohm/270 ohm/47 ohm set of
load resistors that was used earlier.
The following shows the output when the 3m cable that is
terminated with the fixed 150 ohm load is measured with the
calibration profile for the cable:
Notice that the corrected measurements are so close to a
single value that they all fell inside (and under) the
green donut. The cable has been calibrated away and
practically disappeared. The Scalar plot looks like this:
Example 3 --
Calibrated coax cable with a complex load
The following shows the uncalibrated Smith chart for an
artificial complex impedance that is made up of a resistor,
an inductor and a capacitor all in series. (The load was
constructed to be close to resonant and near 50 ohms at the
low end of the 40m Amateur band.) The load is plugged
directly into the SO-239 connector of the AA-230PRO:
When this complex load is use to terminate the same 3m
RG-58 that is used in Example 2 above, the uncalibrated
output of the AA-230PRO shows this:
Notice that this bears very little resemblance to the
actual impedance sweep that is shown in the previous Smith
Chart.
We now apply the 3m coax cable profile that was used in
Example 2, and the corrected measurement shows:
There is some error at the higher frequencies. Overall
however, the result is quite close to the measurement of
the complex load directly at the analyzer's connector.