Comments on final(?) tests on 5-MHz antennas with mesh

Tests were performed on 22 April 2015, with mesh suspended
at about 280 feet.

These were 2-port VNA measurements taken on selected
combinations of antennas designed to fill in all the
points on a 6-port S-parameter matrix, taking heavy
advantage of reciprocity and of expected symmetries.
Some duplication was done, however, to serve in the
role of sanity and consistency checks.

The measurements were performed at the TX ends of the
various Heliax lines at the points where the lines
connect to the transmitter proper.  Note that these
lines have all been cut carefully to virtually exact
integer multiples of a halfwavelength minus a fixed
offset (of about 8 feet) so that the feedpoint currents
are accurately reflected at the midpoint of the various
directional couplers used to sample the line signals.
In the table to follow these notes, the complex
S-parameter values have been adjusted to reflect the
impedances at the actual dipole terminals, accounting
both for the 8-foot offset and for individual losses
in the heliax lines.

In addition, one file was taken as a duplicate of a
regular measurement but with all terminations on the
"unused" dipoles left open.  With this sole exception,
all the unused dipoles were terminated in 50 ohms for
all these measurements, as required for a legitimate
set of S-parameter measurements.  The purpose of this
test was to see how much the S-paraneters varied as a
function of source impedances seen by cross coupled
antennas.

The VNA setup was to measure on an exact harmonic
series based on 25 kHz, running through 100.025 MHz,
to allow some reasonable FDTDR capability as well as
providing good FD resolution on S-parameter plots such
as VSWR and cross coupling.  To wit:

Fstrt = 25 kHz;  Fstop = 100.025 MHz;  # points = 4001.

File names take the form of '3a5b.s2p' (for example),
where '3a' is the antenna designation connected to
port 1 of the VNA, and '5b' is that for port 2.  Thus,
in this example, dipole HF-3A is connected to VNA
port 1, and dipole HF-5B is connected to VNA port 2.

In general the VSWR values ranged in the 1.4 to 1.55
regime- not as good as I'd like to see, but hopefully
good enough.

Cross coupling coefficients ranged from about -40 dB
between a pair of cross dipoles, to about -22 dB for
certain combinations of dipoles on adjacent towers.

The reported complex values of S-paramters have been
adjusted from raw data to include the "8-foot offset"
in each transmitter- thus they should accurately
reflect the complex impedances at the dipole terminals.


RESULTS:

For starters, the simple stuff (all evaluated at 5.125 MHz):

1a1b.s2p:  VSWR(1A) = 1.523     VSWR(1B) = 1.494    |coupling| = 10.837e-3 = -39.3 dB
1a3a.s2p:  VSWR(1A) = 1.558     VSWR(3A) = 1.498    |coupling| = 44.684e-3 = -27.0 dB
1a3b.s2p:  VSWR(1A) = 1.525     VSWR(3B) = 1.430    |coupling| = 77.876e-3 = -22.2 dB
1a5a.s2p:  VSWR(1A) = 1.515     VSWR(5A) = 1.421    |coupling| = 40.566e-3 = -27.8 dB
1a5b.s2p:  VSWR(1A) = 1.537     VSWR(5B) = 1.524    |coupling| = 75.655e-3 = -22.4 dB
1b3a.s2p:  VSWR(1B) = 1.505     VSWR(3A) = 1.430    |coupling| = 67.933e-3 = -23.4 dB
1b3b.s2p:  VSWR(1B) = 1.480     VSWR(3B) = 1.463    |coupling| = 78.742e-3 = -22.1 dB
3a1a.s2p:  VSWR(3A) = 1.431     VSWR(1A) = 1.534    |coupling| = 46.017e-3 = -26.7 dB
3a5b.s2p:  VSWR(3A) = 1.434     VSWR(5B) = 1.521    |coupling| = 70.728e-3 = -23.0 dB
3b5b.s2p:  VSWR(3B) = 1.434     VSWR(5B) = 1.534    |coupling| = 44.669e-3 = -27.0 dB
5a1a.s2p:  VSWR(5A) = 1.420     VSWR(1A) = 1.541    |coupling| = 41.083e-3 = -27.7 dB
5a3a.s2p:  VSWR(5A) = 1.419     VSWR(3A) = 1.425    |coupling| = 39.021e-3 = -28.2 dB
5b5a.s2p:  VSWR(5B) = 1.531     VSWR(5A) = 1.421    |coupling| = 18.405e-3 = -34.7 dB

The following was taken with all unused dipole cables unterminated (open) at TX end
5b5anotm.s2p:  VSWR(5B) = 1.547     VSWR(5A) = 1.462    |coupling| = 17.614e-3 = -35.1 dB

VSWR ranges seen on each dipole (excluding the 'cables-unterminated' test):

HF-1A:  1.515 to 1.558
HF-1B:  1.480 to 1.505
HF-3A:  1.425 to 1.498
HF-3B:  1.430 to 1.463
HF-5A:  1.419 to 1.421
HF-5B:  1.521 to 1.534

I can't account for the wide range of variabilities, but wonder if wind blowing
the mesh around might have something to do with it.

It appears that the "curse" on HF-3A is not present in this data.  In the past
the curse has resulted in quite obvious abnormalities in the time-domain plot
of impulse reflections.  See  '3a_vs_1a_refl_impulse.pdf' (or '....fc7'). 


Still to come once I get to it:  
6-port S-parameter matrix for coupled antennas with mesh.


Dana Whitlow
Data taken 22 April 2015 in TX bldg
This file date 23 April 2015
final_dipole_meas.txt