Brought to you by
Radiocommunications Agency Netherlands
This material is posted here with permission of het IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of Radiocommunications Agency Netherlands’
products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new
collective works for resale or redistribution must be obtained from the IEEE by writing to pubns-permissions@ieee.org. By choosing to view this document, you agree to all provisions of the
copyright laws protecting it.
Measuring the Isolation of the Circularly Polarized
Characteristic Waves in NVIS Propagation
Ben A. Witvliet1,2, Erik van Maanen1, George J. Petersen1, Albert J. Westenberg1,
Mark J. Bentum2, Cornelis H. Slump2, Roel Schiphorst2
1
2
Radiocommunications Agency Netherlands, Spectrum Management Department
P. O. Box 450, Groningen, The Netherlands
Tel: +31 6512 48341; Fax: +31 50 5877 400
E-mail: ben.witvliet@agentschaptelecom.nl
University of Twente, Center for Telecommunications and Information Technology
P. O. Box 217, Enschede, The Netherlands
Tel: +31 6512 48341; Fax: +31 53 489 1060
E-mail: b.a.witvliet@utwente.nl
Cite:
Witvliet, B. A.; van Maanen, E.; Petersen, G. J.; Westenberg, A. J.; Bentum, M. J.; Slump, C. H.; Schiphorst, R.,
"Measuring the Isolation of the Circularly Polarized Characteristic Waves in NVIS Propagation [Measurements
Corner]," in Antennas and Propagation Magazine, IEEE , vol.57, no.3, pp.120-145, June 2015.
doi: 10.1109/MAP.2015.2445633
URL: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7214390&isnumber=7214355
measurements Corner
Brian E. Fischer
Ivan J. LaHaie
Measuring the Isolation of the Circularly Polarized
Characteristic Waves in NVIS Propagation
Ben A. Witvliet, Erik van Maanen, George J. Petersen, Albert J. Westenberg,
Mark J. Bentum, Cornelis H. Slump, and Roel Schiphorst
Ionospheric Radio
Wave Propagation
Ionospheric radio wave propagation can
be used to bridge hundreds of kilometers
with a direct radio link [1]. This makes
ionospheric radio communication valuable when the independence of satellite
or terrestrial networks is required, e.g., in
regions without telecommunication infrastructure [2], for disaster relief operations
in areas where the telecommunication
infrastructure is destroyed [3]–[5], and for
defense operations [6]. When the frequency is properly chosen, typically 3–10
MHz, radio waves sent upward are
Digital Object Identifier 10.1109/MAP.2015.2445633
Date of publication: 21 August 2015
120
1045-9243/15©2015IEEE
EDITORS’ NOTE
This issue’s “Measurements Corner” article describes a very accurate measurement of
the propagation characteristics of near vertical incidence ionospheric skywaves. In it,
the authors describe and characterize the unique morning and evening phenomena
associated with the ordinary and extraordinary waves, which the authors have
dubbed “Happy Hours.” Of course, this may raise the question for many of our reader
as to what exactly goes on during a morning Happy Hour!
reflected by the ionosphere to create a
large continuous coverage area (400 km #
400 km) around the transmitter [7]. The
antenna system has to concentrate the
transmit power at high elevation angles
[7], typically 70–90º, hence the name of
Ionosphere
S
eparate excitation of the characteristic waves in the ionosphere results
in two orthogonal propagation
channels on the same frequency, which
may be used in diversity and multipleinput, multiple-output (MIMO) systems.
In this article, a method to measure the
isolation between these paths is proposed
and demonstrated in a near vertical incidence skywave (NVIS) experiment at a
frequency of 7 MHz over a 105-km distance. Characteristic wave isolation
exceeding 25 dB is measured during
“Happy Hour”: the interval when the
propagation path just opens or closes and
only the extraordinary wave propagates.
NVIS
Antenna
Coverage Area
Figure 1. An example of NVIS.
Ionospheric radio wave propagation
can be used to cover a continuous
area with a radius of several hundred
kilometers using a single transmitter.
Frequencies between 3 and 10 MHz
are used, and radio waves must be
radiated at steep angles.
june 2015
the propagation mechanism: near vertical
incidence skywave. A simplified illustration of the NVIS is given in Figure 1, and
the propagation mechanism is described
in detail in [6].
Ionospheric radio wave propagation
adds fading to the received signal, decreasing the link reliability and throughput, but
this may be countered with diversity reception or MIMO [8]. Improved diversity
reception can be obtained by adapting the
polarization of the antenna to the circular
polarization of the characteristic waves
propagating in the ionosphere, thereby
creating two independent propagation
paths from the transmitter to the receiver
[9]. In this article, a method to measure
the isolation between these paths is proposed and demonstrated in an NVIS experiment at a frequency of 7 MHz over
a 105-km distance. The Happy Hour
propagation phenomenon that facilitates
this measurement as well as the equipment needed and its calibration and accuracy is described.
IEEE Antennas & Propagation Magazine
Characteristic Waves
in NVIS Propagation
Figure 3 shows the simulated paths of the
ordinary and extraordinary waves, in red
and green, respectively, through the ionosphere, over a horizontal distance of
approximately 90–120 km. The reflection
takes place in the F2 layer at a 180–
280-km height. The reflection height and
path geometry vary over the day. PropLab
Pro 3 ionospheric ray-tracing software
[13] was used and simulations were made
at a frequency of 7 MHz on 9 March
2014, with a smoothed sunspot number
(SSN) of 79 and an effective sunspot
number (IGN) of 164.
In 24 h, the propagation varies as follows. At night, the electron density of the
ionosphere is too low to support NVIS
propagation at the selected operating
IEEE Antennas & Propagation Magazine
F1' F1" F '
2
G
F2"
F1' F "
1
F2' F2"
(a)
(b)
Figure 2. Pulse delay measurements of Appleton in 1932. (a) The transmitted pulses
are first received via ground wave G, then twice via the F1-layer (F1' and F1"), and twice
via the F2-layer (F2' and F2"). (b) A 1,115-Hz sine wave, serving as time reference. (Figure
adapted from [10].)
path height is lowered due to a further
frequency and both waves pass through
increase in electron density (see Figures
the ionosphere (Figure 3). In the morn6 and 7). In the evening, the electron
ing, the radiation of the sun causes a
density slowly decreases again and the
steep rise in the electron density of the
path altitude increases again (Figure 8).
ionosphere, until it is sufficiently high
to reflect the extraordinary
wave at the operating frequency. However, the ordiIonospheric radio wave
nary wave still passes through
the ionosphere (Figure 4). The
propagation can be used
received signal consists of the
to bridge hundreds of
extraordinary wave only and
kilometers with a direct
is circularly polarized. This
radio link.
situation remains until the
electron density has increased
enough for the ordinary wave
At a certain instant, the electron density
to reflect as well (Figure 5). The received
becomes too low to reflect the ordinary
signal is then a summation of the ordiwave and it passes through the ionosphere
nary and the extraordinary wave comwhile the extraordinary wave is still
ponents. As each varies in strength and
reflected (Figure 9). Again, the received
delay, the received signal shows rapidly
signal consists of only the extraordinary
changing polarization.
wave and is circularly polarized. This situThroughout the day, both waves propation remains until the electron density
agate from transmitter to receiver. Their
300
250
200
150
100
50
0
0
06:00 UTC
Height (km) "
The Happy Hour
Propagation Interval
G
Height (km) "
Appleton and Builder [10] discovered
that electromagnetic pulses sent toward
the ionosphere were received as pulse
pairs at a distance of 5 km from the transmitter. One of their registrations is reproduced in Figure 2. The transmit pulse is
first received via the ground wave (G),
then twice via the F1-layer reflection (F1'
and F1"), and twice via the F2-layer reflection (F2' and F2"). Appleton explained this
phenomenon with his magneto-ionic theory [11], showing that, under the influence of the Earth’s magnetic field, an
electromagnetic wave of arbitrary polarization is split into two circularly polarized
waves of opposite direction of rotation
when entering the ionosphere. Rathcliffe
[12] showed that that only these waves
propagate in the ionosphere and named
them “characteristic waves.” Each characteristic wave follows its own path and
experiences a particular (variable) attenuation and delay. The “ordinary wave” follows a path similar to the path that it
would have followed in the absence of a
magnetic field. The other characteristic
wave is named the “extraordinary wave.”
In the Northern Hemisphere, the downward ordinary wave has left-hand circular
polarization (LHCP) and the greater
delay while the extraordinary wave has
right-hand circular polarization (RHCP)
and the lesser delay.
20 40 60
80 100 120
Distance (km) "
Figure 3. The ionospheric paths
of the ordinary wave (red) and the
extraordinary wave (green) at 06:00
Coordinated Universal Time (UTC).
Daylight is from 06:02 to 16:28 UTC.
The ionization of the ionosphere is
not sufficient to reflect either of the
characteristic waves.
june 2015
300
250
200
150
100
50
0
0
06:30 UTC
Happy
Hour
20 40 60
80 100 120
Distance (km) "
Figure 4. At 06:30 UTC, the ionization
of the ionosphere has sufficiently
increased to reflect the extraordinary
wave (green), and the ordinary
wave (red) is not reflected yet.
The downward wave has circular
polarization. This is the morning
Happy Hour.
121
20 40 60
80 100 120
Distance (km) "
Height (km) "
Figure 5. At 07:00 UTC, both the
ordinary and extraordinary waves
are reflected. Received polarization is
highly variable.
300
250
200
150
100
50
0
11:00 UTC
0
20
40 60 80 100 120
Distance (km) "
Height (km) "
Figure 6. At 11:00 UTC, the reflection
height is lowered due to the increased
electron density of the ionosphere.
300
250
200
150
100
50
0
15:00 UTC
0
20
40 60 80 100 120
Distance (km) "
Figure 7. The propagation of both
waves continues at 15:00 UTC.
has decreased so much that the extraordinary wave also passes through the ionosphere (Figure 10). This situation remains
until the solar radiation builds up ionization again the next morning.
We identified two exceptional intervals (Figures 4 and 9) nicknamed
“Happy Hours,” in which only the
extraordinary wave propagates and
RHCP is received. This phenomenon
was predicted (but not observed) in
[14] and experimentally confirmed in
[15]. At sunrise, the ionization shows a
steep gradient, and consequently, the
122
Measuring NVIS
Characteristic Wave Isolation
This propagation phenomenon can be
used to measure the isolation between
the ordinary and extraordinary waves. If
substantial isolation between the RHCP
and LHCP waves can be demonstrated
in the Happy Hour intervals, this will
provide strong support for the assumption that the characteristic waves travel
independent paths through the ionosphere with little crosstalk, and that they
can be used effectively in NVIS diversity
and MIMO. Therefore, we propose the
following experiment. A beacon transmitter is connected to a linearly polarized NVIS antenna and the transmit frequency is chosen so that stable NVIS
layer propagation is present during a
major part of the day. The transmit frequency is chosen so that stable NVIS
layer propagation is present during a
major part of the day. A receive station
located approximately 100 km away continuously measures the signal strength of
the RHCP and the LHCP components of the incoming wave. The ratio
between the LHCP and RHCP signal
strength is calculated and plotted against
time. For this experiment, a measurement system is realized using commercially off-the-shelf equipment, completed with a few components specially
designed for this experiment. An overview of the system components and their
interconnections is shown as a block diagram in Figure 11.
Beacon Transmitter
A software-defined radio transmittertype Flex-Radio FLEX-6500 is used, followed by a Trans World Electronics
T1000 linear amplifier with a radio fre-
Height (km) "
100
50
0
0
300
250
19:30 UTC
200
150
100
50
0
0
20
40 60 80 100 120
Distance (km) "
Figure 8. After 16:38 UTC (sunset), the
ionization decreases. At 19:30 UTC,
waves already penetrate much further
into the ionosphere and the reflection
height increases.
Height (km) "
07:00 UTC
morning Happy Hour is short, typically
30 min at midlatitudes in winter. The
evening Happy Hour often lasts more
than an hour because of the slower recombination processes.
300
250
200
150
100
50
0
20:30 UTC
Happy Hour
0
20
40 60 80 100 120
Distance (km) "
Figure 9. At 20:30 UTC, the electron
density of the ionosphere has decreased
so much that the ordinary wave is no
longer reflected, but the extraordinary
wave is still supported and the
downward wave has circular polarization.
This is the evening Happy Hour.
Height (km) "
Height (km) "
300
250
200
150
300
250
200
150
100
50
0
21:30 UTC
0
20
40 60 80 100 120
Distance (km) "
Figure 10. At 21:30 UTC, the ionization
of the ionosphere has decreased so
much that none of the characteristic
waves are reflected.
quency (RF) output power of 300 W.
The measured transmitter frequency
stability is better than
0.1 Hz/24 h, and the measured output power stability is
better than 0.1 dB/24 h. The
The high frequency and time
Weak Signal Propagation
accuracy allow for precise
Reporter (WSPR) protocol
[16] is used to transmit the
filtering at the receiver.
june 2015
IEEE Antennas & Propagation Magazine
station identification and geographical
coordinates using 1.5-baud four-frequency shift keying, with a necessary
bandwidth of 6 Hz. The WSPR protocol
has a 2-min periodicity, consisting of
110.6-s transmissions followed by a 9.4-s
silence, synchronized to a standard time
server accessed over the Internet using
Dimension 4 software [17]. The high
frequency and time accuracy allow for
precise filtering at the receiver. A halfwave dipole antenna is used as a transmit
antenna. It is suspended horizontally at a
height of approximately 4 m (0.09 m)
above farmland soil. The antenna produces linear polarization with a broad
main lobe toward the zenith. For high
angles, the radiation pattern is omnidirectional. For low angles, the antenna
radiates lengthwise with vertical polarization. Therefore, to minimize ground
wave coupling, it is oriented perpendicular to the direction of the receiver.
NVIS
Propagation
LHCP
RHCP
Turnstile
Antenna
Phasing
+/-90°
PA
SDR TX
PC
WSPR
BPF
Lab
View
Dimension 4
Internet
Time
Server
Dimension 4
RX
PC
Turnstile Antenna
A turnstile antenna [18] was selected to
measure the field strength of both characteristic waves. This antenna consists of
two quadrature-fed perpendicular halfwave dipole antennas and exhibits circular polarization for the steep elevation
angles used in NVIS propagation. The
polarization sense can be reversed by
changing the phase difference of the
dipoles from +90º to -90º. On frequencies between 3 and 10 MHz, this turnstile antenna can be realized as wire
dipole elements suspended in an “inverted vee” configuration from a single
extendable mast, as shown in Figures 12
and 13. The copolar and cross-polar
antenna diagrams shown in Figures 14
and 15 are calculated using NEC-4.2
method-of-moments antenna simulation
software [19]. The Sommerfeld ground
model [20] is used to obtain realistic
results near real ground. The model was
created and analyzed with 4Nec2 [21].
Farmland soil was used in the calculations. Practical realization of the antenna
can be observed in Figure 16.
Balance Transformers
and Feed Lines
Both dipole antennas are fed through 1:1
balance-unbalance transformers (baluns),
IEEE Antennas & Propagation Magazine
Figure 11. A block diagram of the experimental measurement system.
A (left) beacon transmitter is connected to a linearly polarized NVIS antenna. A
(right) receive station—located approximately 100 km from the transmitter—
continuously measures the signal strength of the RHCP and the LHCP components
of the incoming wave. SDR: software-defined radio; Tx: transmitter; PA: power
amplifier; BPF: bandpass filter; Rx: receiver; PC: personal computer.
type Diamond BU-50. As any difference
in phase delay or attenuation of these baluns would degrade quadrature, five baluns are measured pairwise. Individual
phase delays vary from 5.9º to 7.0º. A
matched pair is selected with an attenuation difference smaller than 0.05 dB and a
phase difference smaller than 0.1º. The
baluns are connected through identical
lengths (50 m) of EcoFlex 10 doubly
shielded coaxial cable. Both cables are
taped on opposite sides of the antenna
support (mast), and ferrite clamps are
added every meter to suppress commonmode current that would otherwise influence the antenna radiation diagram. The
horizontal part of the feed lines is buried
approximately 70 cm below the ground to
avoid coupling with the antenna. The
electrical length of both feed lines is measured and the difference in phase delay is
less 0.16º. From these measurements, the
overall difference of cable and baluns is
expected to be lower than 0.1 dB and 0.5º.
june 2015
Figure 12. The turnstile antenna
made of two quadrature-fed halfwave dipole antennas suspended in
an “inverted vee” configuration from a
single extendable mast.
10.05 m
7.5 m
2.5 m
Figure 13. The dimensions of the
dipole elements of the turnstile
antenna (wire radius: 1 mm).
123
75°
60°
Copolar
45°
75°
0 dBr
60°
45°
-10 dBr
30°
15°
90°
Cross Polar
0°
30°
-20 dBr
15°
-30 dBr
0 dBr = 5.5 dBi (CP)
0°
Figure 14. The vertical antenna diagram showing copolar
(red) and cross-polar (blue) circular polarization antenna
gain of the turnstile antenna, simulated using NEC-4.2
method-of-moments software. Farmland soil was used in
the calculations.
Phasing Network
Figure 16. The turnstile antenna installed at the measurement
location. The farmhouse shown in the picture is approximately
50 and 80 m from the antenna. Other buildings are >800 m from
the antenna.
phasing network is shown in Figure 17.
The practical realization mounted in a
transportation box is shown in Figure 18.
The phase and amplitude difference of
the completed phasing unit was measured using an Agilent E5062A network
analyzer and was carefully aligned. The
total quadrature error of the phasing
unit is <0.1º and <0.05 dB. This makes
the overall quadrature error in the turnstile antenna <0.6º and <0.15 dB.
surement system depends on the cross
polarization of the measurement antenThe quadrature feed for the turnstile
na. To evaluate the influence of the
antenna is realized using switched coaxiquadrature error on the cross polarizaal delay lines. The feed lines coming
tion of the turnstile antenna, the NECfrom the perpendicular dipoles are con4.2 antenna model from the “Turnstile
nected to a phasing box, in which one
Antenna” section is used. In the model,
cable is lengthened with a quarter wave
the dipole elements are kept perfectly
phasing line to provide 90º phase shift.
identical and perpendicular on perfectly
The other feed line is lengthened by a
flat ground, but quadrature errors are
half-wave phasing line that can be
introduced to produce cross-polarization
bypassed using coaxial relays. Dependlevels of -20, -25, and -30 dB at an eleing on the position of these relays, the
vation angle of 80º. These quadrature
phase difference between the dipoles
Antenna Cross Polarization
errors are plotted in the graph of Figantennas is now either 90º or -90º. The
The maximum characteristic wave isolaure 19 and ellipses are drawn through
phasing lines are connected to an RF
tion that can be measured with our meathese points to delimit the
combiner (Merrimac PDNLareas in which a certain cross
20 -100). The RF combiner has
0°
polarization is achieved.
a measured phase error of
330°
30°
0 dBr
A quadrature error of 0.6º
<0.03º and an attenuation
Copolar
and
0.15 dB will result in a
error of <0.04 dB. The coaxial
-10 dBr
cross
polarization of approxirelays are Tohtsu CX-600M,
300°
60°
mately -32 dB. However, the
with a measured insertion loss
-20 dBr
real cross-polarization value
of 0.01 dB and an isolation
Cross Polar
will also depend on the physigreater than 80 dB. The phas-30
dBr
cal symmetry of the antenna
ing lines are made of Belden
270°
90°
wires and the homogeneity of
H-155 doubly shielded coaxial
the ground below them. For
cable. Their insertion losses
example, adding a random
are 0.25 and 0.5 dB, respec240°
120°
height error between 0 and
tively. To compensate for this,
30 cm to the end height of the
attenuators of 0.25 and 0.5 dB
four dipole legs, while keeping
are inserted as indicated in the
210°
150°
the leg length constant, will
diagram. Two 6-dB attenuators
180° 0 dBr = 5.0 dBi (CP)
degrade the cross polarization
are inserted between the
to values between -25 and -29
antenna feeder line and the
dB. Some specific combinaphasing unit to make the phase Figure 15. The horizontal antenna diagram at 70º
elevation showing copolar (red) and cross-polar (blue)
tions that cause worse degrashift less dependent on the circular polarization antenna gain of the turnstile antenna,
dation can also be found. This
source impedance of the simulated using NEC-4.2 method-of-moments software.
must be kept in mind when
dipoles. A block diagram of the Farmland soil was used in the calculations.
124
june 2015
IEEE Antennas & Propagation Magazine
Measurement Receiver
The HF radio environment puts high
demands on measurement receiver
performance. A 24-h registration
shows that the maximum total power
at the antenna terminals of a dipole
antenna is -40 dBm, due to the accumulated power of high-power shortwave broadcast stations. At the same
time, the minimum discernible signal
power is -135 dBm. Therefore, the
intermodulation free dynamic range
of the receiver must exceed 95 dB.
For our experiment, a Rohde and
Schwarz FSMR26 measurement
receiver was selected. This receiver
provides a combined measurement
uncertainty of 0.3 dB for 95% confi-
{0 + 90°
-6 dB
-0.25 dB
-0.5 dB
{0
R
-6 dB
{0 + 180°
Figure 17. A block diagram of the
phasing network for the turnstile
antenna using coaxial phasing lines
to produce either -90º or +90º phase
difference. The attenuation of the
phasing lines is compensated with small
attenuators (0.25 and 0.5 dB).
dence on 7 MHz, with a 100-Hz
This leaves only 7 dB of headroom;
receiver bandwidth and a root-meantherefore, a 7-MHz bandpass filter is
square detector. The specified -1 dB
added at the receiver input. This reducinput compression point of the receives the maximum total input power to
er is +13 dBm, the third-order input
-20 dBm and increases the headroom
intercept point (IIP3) is +17 dBm,
to 13 dB. Filter passband attenuation is
and the second-order input intercept
only 0.23 dB. A 100-Hz receiver bandpoint is +35 dBm.
We measured an IIP3 of
+18 dBm at 7 MHz for a 200Both data acquisition
kHz spacing and a displayed
and polarization sense
average noise level (DANL)
of -135 dBm. The maximum
are controlled by a laptop
allowed input power Pmax for
computer using LabView
which the third-order intersoftware.
modulation products remain
beneath the receiver noise
floor can be calculated as
width is chosen for the measurements.
Pmax =IIP3 - ^IIP3 -DANL h /3
This is sufficiently large to ensure fast
^
h
settling of the detector and selective
=18 dBm - 18 dBm +135 dBm /3
(1)
enough to reduce the probability of
=-33 dBm.
cochannel interference. Measurements
using LHCP and RHCP are alternated
every 5 s. Both data acquisition and
1 dB
-20 dB
polarization sense are controlled by a
-25 dB
0.5 dB
-30 dB
laptop computer using LabView soft0 dB
ware. Time synchronization between
transmitter and receiver is achieved by
-0.5 dB
synchronizing both to the same stan-1 dB
dard time server, which is accessed over
-10° -5° 0° 5° 10°
the internet with Dimension 4 [17] soft! Phase Error "
ware. The synchronization error is
Figure 19. The simulated cross
lower than 0.05 s.
polarization of the turnstile antenna as
a function of the amplitude and phase
Measurement Results
error of the phasing unit, assuming
Using the system described in the preperfectly identical dipoles. The
vious section, dual circular polarization
elevation angle is 80º.
! Amplitude Error "
installing the measurement antenna system and when interpreting the measurement results.
7.042
60
Frequency (MHz) "
7.041
IEEE Antennas & Propagation Magazine
40
7.039
30
7.038
20
7.037
7.036
7.035
16:00
Figure 18. The practical realization of
the phasing network for the turnstile
antenna.
50
7.04
10
Beacon
16:10
0
16:30
16:20
Time (UTC) "
16:40
16:50
-10
Figure 20. The spectrogram showing the strong and cyclic signal of the beacon
transmitter and other radio signals on adjacent frequencies. The color scale is received
signal strength in decibel microvolts. The measurement was taken on 9 March 2014.
june 2015
125
Signal Strength
(dBuV) "
60
40
Saturday,
8 March 2014
Extraordinary (R)
Ordinary (L)
20
0
RH/LHCP
(dB) "
40
30 25 dB
20
10
0
-10
00:00 03:00
Isolation (L/R)
06:00
09:00 12:00 15:00
Time (UTC) "
18:00
21:00
00:00
Signal Strength
(dBuV) "
Figure 21. The start of the measurements on Saturday, 8 March 2014 at 14:51 UTC.
The signal strength of the extraordinary wave is shown in green, and the ordinary
wave is shown in red. The ratio of the two is shown in blue. Daylight ends at 16:28
UTC. NVIS propagation ends around 20:35 UTC. The evening Happy Hour shows
approximately 25-dB wave isolation.
60
40
Sunday,
9 March 2014
Extraordinary (R)
Ordinary (L)
20
0
RH/LHCP
(dB) "
40
30 25 dB
20
10
0
-10
00:00 03:00
Isolation (L/R)
06:00
09:00 12:00 15:00
Time (UTC) "
18:00
21:00
00:00
Signal Strength
(dBuV) "
Figure 22. The continuation of the measurements on Sunday, 9 March 2014.
Daylight is from 06:02 to 16:28 UTC. NVIS propagation starts around 06:03 UTC
and ends around 20:25 UTC. Both Happy Hours show approximately 25-dB wave
isolation. The short signal loss at 02:30 UTC due to beacon failure shows that the
signal-to-noise ratio (SNR) is >20 dB at night.
60
40
Monday,
10 March 2014
Extraordinary (R)
Ordinary (L)
20
0
RH/LHCP
(dB) "
40
30 25 dB
20
10
0
-10
00:00 03:00
Isolation (L/R)
06:00
09:00 12:00 15:00
Time (UTC) "
18:00
21:00
00:00
Figure 23. The continuation of the measurements on Monday, 10 March 2014. Daylight
is from 06:02 to 16:28 UTC. NVIS propagation starts around 06:10 UTC and ends around
21:40 UTC. Both Happy Hours show approximately 25-dB wave isolation.
126
june 2015
measurements were performed from
Saturday 8 March 2014 14:51 UTC to
Tuesday 11 March 2014 at 00:00 UTC.
The beacon transmitter was located
53.18058º north and 6.29503º east.
The measurement system was located
52.26153º north and 6.62175º east.
Both locations are in rural areas in The
Netherlands. The path length was
104.5 km, and the azimuthal direction
was 188º. As shown in [7], this distance
is sufficient for the NVIS signal to
dominate the ground wave. At the
time of the measurements, the SSN
was 65. Ionization was sufficiently high
to use a frequency near 7 MHz to
obtain stable NVIS propagation for a
large part of the day. This frequency
was also high enough to ensure E-layer
transparency. The expected elevation
angle, obtained by simulations, varies
between 75º and 80º during daytime
propagation and between 79º and 86º
when the propagation path opens or
closes. The ionosondes located at
Dourbes (50.1º north, 4.6º east),
Juliusruh (54.6º north, 13.4º east), and
Chilton (51.5º north, -1.3º east) were
monitored for sporadic E-layer patches that could disturb the measurements, however, none were observed.
Signal Identification
and SNR
During the 57-h measurement interval,
a spectrogram (frequency–time graph or
waterfall diagram) was recorded every
2.5 s, alternating on both antenna polarizations. In this spectrogram, the beacon
signal is easily identified, first by its
transmit frequency and second by its
precisely defined on–off pattern, as can
be observed in Figure 20. The “off” periods are used to verify the absence of onchannel interference. Fortunately, no
data had to be discarded because of
interference and the beacon signal was
sufficiently strong and the frequency was
clear. The “off” periods in the beacon
signal were also used to estimate the
instantaneous SNR, which was greater
than 65 dB during daylight hours. At
night, with no apparent NVIS propagation, the SNR was still greater than
20 dB. The latter can be observed on 9
March 2014 at 02:30h UTC (Figure 22),
IEEE Antennas & Propagation Magazine
Analysis and Discussion
Figures 21–23 prove the existence of a
morning and evening Happy Hour interval in which only RHCP waves are
received, consistent with preliminary
measurements in 2009 [15]. Since the
previous measurements were made near
the sunspot cycle minimum and these
extended measurements were made
near the sunspot cycle maximum, independence of the position in the sunspot
cycle is demonstrated. In all five Happy
Hour intervals, instantaneous RHCP/
LHCP ratios of up to 35 dB are
observed, as well as a 2.5-min average of
approximately 25 dB.
IEEE Antennas & Propagation Magazine
Signal (dBuV) "
RH/LHCP (dB) "
The beacon frequency was filtered
from the spectrogram data and the
remaining data were time gated to
retain only the samples in which the
beacon transmitter was switched on.
The signal strength of the beacon for
both LHCP and RHCP reception is
plotted in Figures 21–23. For ease of
interpretation, the 57-h continuous
measurement is presented in 24-h
intervals. Signal strength of the extraordinary wave (RHCP) is shown in green
and that of the ordinary wave (LHCP)
is shown in red. The lighter colored pixels are the individual measurement
samples recorded every 5 s; the solid
lines show a 2.5-min floating average.
Daylight on 9 March 2014 was from
06:02 to 16:28 UTC at path midpoint
(52.72º north, 6.46º east). NVIS propagation started every morning around
sunrise at approximately 06:10 UTC
and ended between 20:00 and 22:00
UTC, several hours after sunset, with a
large day-to-day variation. A blue trace
is added below the measured signal
strength curves, showing the ratio of
the signal strength of the ordinary and
extraordinary waves. The morning and
evening Happy Hour can be clearly
seen. The morning Happy Hour is
shorter than the evening Happy Hour,
as predicted in the section “Characteristic Waves in NVIS Propagation.”
The measurements during the morning
Happy Hour of 9 and 10 March 2014
are shown in Figures 24 and 25. Again,
the lighter colored pixels represent the
individual measurement samples
recorded every 5 s; the solid lines show
a 2.5-min floating average. The morning intervals consistently started a few
minutes after sunrise with a sudden
increase of the received signal strength
of the extraordinary wave. The signal
strength is increased by 35 dB in a
60
10-min interval, just as the propagation
channel “switches on.” After that
moment, the received polarization is
nearly perfectly RHCP. The rise of the
signal strength caused by the ordinary
wave started later and was more gradual than that of the extraordinary wave.
In the interval where the extraordinary
wave propagated and the ordinary wave
did not, a characteristic wave isolation
of 25 dB was measured. The onset of
the propagation before the Happy
Hour interval was gradual; a slight
Extraordinary (R)
40
Ordinary (L)
20
30 25 dB
20
Isolation (L/R)
10
0
-10
Sunday, 9 March 2014
05:00
05:15
05:30
05:45 06:00 06:15
Time (UTC) "
06:30
06:45
Figure 24. The signal strength of the ordinary (red) and extraordinary waves
(green) and their ratio (blue), measured during the morning Happy Hour of
Sunday, 9 March 2014. The peaks in the blue trace show the characteristic wave
isolation during Happy Hour.
60
Signal (dBuV) "
Dual Circular Polarization
Measurements
Morning Happy Hour Observations
RH/LHCP (dB) "
when the beacon transmitter has a short
failure due to human error.
Extraordinary (R)
40
Ordinary (L)
20
30 25 dB
20
Isolation (L/R)
10
0
-10
Monday, 10 March 2014
05:30
05:45
06:00
06:15
Time (UTC) "
06:30
06:45
Figure 25. The signal strength of the ordinary (red) and extraordinary waves
(green) and their ratio (blue), measured during the morning Happy Hour of
Monday, 10 March 2014.
june 2015
127
RH/LHCP (dB) "
Signal (dBuV) "
60
Extraordinary (R)
50
40
Ordinary (L)
30
20
30 25 dB
20
Isolation (L/R)
10
0
-10
Saturday, 8 March 2014
19:00
19:15
19:30
19:45 20:00 20:15 20:30 02:45 21:00
Time (UTC) "
Figure 26. The signal strength of the ordinary (red) and extraordinary waves
(green) and their ratio (blue), measured during the evening Happy Hour of
Saturday, 8 March 2014.
increase in the signal strength of the
extraordinary wave began 1–4 h before
the Happy Hour interval started.
Therefore, the baseline value for these
ratio values was calculated over a longer
time interval.
Evening Happy Hour Observations
The measurements during the evening
Happy Hour intervals of 8–10 March
2014, which occurred several hours after
sunset, are shown in Figures 26–28. The
NVIS propagation of the ordinary wave
started to decay between 1 and 2 h earli-
RH/LHCP (dB) "
Signal (dBuV) "
60
er (8 March and 10 March, respectively)
than the NVIS propagation of the
extraordinary wave. During the Happy
Hour interval, RHCP waves were
received at the measurement site. The
extraordinary wave exhibited a stable signal level up to the end of the NVIS
propagation period, after which the signal strength dropped abruptly. The evening Happy Hour had a longer duration
than the morning Happy Hour and its
onset and duration showed a larger dayby-day variation. The measured characteristic wave isolation was 25 dB.
Extraordinary (R)
50
40
Ordinary (L)
30
20
30 25 dB
20
Isolation (L/R)
10
0
-10
Sunday, 9 March 2014
19:00
19:15
19:30
19:45 20:00 20:15 20:30 02:45 21:00
Time (UTC) "
Figure 27. The signal strength of the ordinary (red) and extraordinary waves
(green) and their ratio (blue), measured during the evening Happy Hour of
Sunday, 9 March 2014.
128
june 2015
Interpretation of Measured
Characteristic Wave Isolation
The measured characteristic wave isolation values were consistently around
25 dB. This is 12 dB greater than the isolation measured during earlier experiments [15], which is attributed to the
significantly improved quadrature feeding network and antenna symmetry. The
true characteristic wave isolation is possibly still higher. In Figure 29, showing the
morning Happy Hour of 10 March 2014,
we see the steep rise of the extraordinary
wave signal (green) and the smoother ascend of the ordinary wave signal (red).
The expected slope of the red trace is
added as a dashed black line. We see an
abrupt step in the red trace when the
green trace rises. This indicates a leakage from the RHCP channel to the
LHCP channel in the measurement system, rather than an ionospheric phenomenon. There are three possible
interpretations.
1) The cross polarization of the measurement antenna is not limiting the
measurement and the characteristic
wave isolation is exactly 25 dB.
2) The cross polarization of the measurement antenna limits the measurement range to 25 dB, and the
characteristic wave isolation is greater than 25 dB.
3) The characteristic waves are slightly
elliptically polarized, with a characteristic wave isolation greater than
25 dB. The cross polarization of the
measurement antenna is greater than
25 dB for perfectly circular polarization, however, the cross polarization
is lower for the incoming waves that
are elliptical.
The measurements performed here are
not conclusive on this issue. Exclusion of
option 1) is only possible by in situ measurement of the cross polarization of the
measurement antenna. However, due to
the abrupt step observed in the ordinary
wave signal, authors favor interpretation
2) or 3). To differentiate between these
two, absolute polarization measurements
are necessary. This could be achieved by
simultaneously measuring both amplitudes and the phase difference on both
ports of the turnstile antenna using a
synchronous dual channel measurement
IEEE Antennas & Propagation Magazine
receiver. If the incoming waves are
slightly elliptical, orthogonality can be
restored by adapting phase and amplitude in the receiver [22].
Signal (dBuV) "
60
40
Ordinary (L)
30
20
30 25 dB
20
Isolation (L/R)
10
0
-10
Monday, 10 March 2014
20:00
20:30
21:00
21:30
Time (UTC) "
22:00
22:30
Figure 28. The signal strength of the ordinary (red) and extraordinary waves
(green) and their ratio (blue), measured during the evening Happy Hour of
Monday, 10 March 2014.
60
Signal Strength (dBuV) "
After the NVIS propagation path
closed, a clearly discernable beacon signal remained. This beacon signal had
the typical flutter fading normally associated with ionospheric reflection or
scattering. The character of this signal
was unlike the ground wave, which is
more stable. Although the beacon signal was 45 dB lower than during daytime, the SNR was still more than 20
dB. This can be observed in Figure 22;
at 02:30 UTC, the beacon transmitter
had a short failure and the recorded signal strength decreased approximately
20 dB. Previous researchers assumed
that the residual propagation at night is
either due to scattering on irregular
patches of higher ionization in the ionosphere [23] or due to side scatter on the
ground at a large distance [24]. No
means were available during this experiment to measure azimuth and elevation angle or absolute polarization,
therefore, no further analysis could be
made of this phenomenon. If the night
time propagation is due to a scattering
mechanism, dual circular polarization
diversity will probably not be effective
at night, as the polarization will probably be lost in the process.
RH/LHCP (dB) "
Nighttime Propagation
Observations
Extraordinary (R)
50
Extraordinary (R)
50
Ordinary (L)
Antenna
Characteristic
Cross
Polarization? Wave
Isolation?
40
30
20
10
05:30
06:00
06:30
Time (UTC) "
07:00
07:30
Figure 29. The measurement during the morning Happy Hour of 10 March 2014
shows an abrupt step in the ordinary wave signal, possibly indicating insufficient
antenna cross polarization.
Conclusions
The measured isolation between the
ordinary and extraordinary waves in
NVIS propagation exceeds 25 dB. The
measurements were performed using a
dual circularly polarized measurement
antenna. Observations suggest that higher isolation can be achieved by further
adapting the antenna to the polarization
of the incoming waves that may be
slightly elliptical. Two highly isolated
paths can be created on the same frequency using dual circularly polarized
antennas on both transmit and receive
sides of the link, effectively doubling the
data transfer capacity of the link. Alternatively, without modification on the
transmit side, a dual circular receive
IEEE Antennas & Propagation Magazine
antenna can be used to implement an
effective receive diversity system to fight
fading in NVIS links [9]. The measurements were done at midlatitudes in the
Northern Hemisphere (53º north, 6º
east), and the results may depend on the
latitude chosen. The measurements
made use of Happy Hour phenomenon:
when the NVIS propagation path first
opened up or nearly closes, only the
extraordinary wave propagates and the
ordinary wave passes through the ionosphere or is absorbed.
in Ambt Delden for the measurements, J. Mielich of the Leibniz Institute of Atmospheric Physics Kühlungsborn for providing verified ionosonde
data, and G. Visser of the Radiocommunications Agency Netherlands for
assistance with accurate phase delay
measurements and technical discussions. We also thank the Radiocommunications Agency Netherlands for the
use of their Rohde and Schwarz
FSMR26 measurement receiver.
Acknowledgments
Ben A. Witvliet (b.a.witvliet@utwente.
nl) received his B.Sc. degree in electronics and telecommunications in 1988
We would like to thank Mr. and Mrs.
Overbeek for the use of their property
june 2015
Author Information
129
from the Hogeschool voor Techniek en
Gezondheidszorg in Enschede, The
Netherlands. He has working experience
in electrical and electronic maintenance
in Israel, in international telecommunication network management in The
Netherlands, as a chief engineer of the
high-power shortwave radio station of
Radio Netherlands World Service in
Madagascar, and as a manager of a team
of technical specialists for TV, FM, and
The measured isolation
between the ordinary and
extraordinary waves in NVIS
propagation exceeds 25 dB.
AM broadcast transmitter operator
NOZEMA in The Netherlands. Since
1997, he has been working for Radiocommunications Agency Netherlands,
currently as a technical advisor. Since
2011, he has been combining his work
with part-time Ph.D. research in the
Telecommunication Engineering group
of the University of Twente, The Netherlands. He is a Senior Member of the
IEEE, a member of the IEEE Antennas
and Propagation Society, and a member
of the European Association on Antennas and Propagation.
Erik van Maanen (erik.vmaanen@
agentschaptelecom.nl) worked for Delft
University of Technology, The Netherlands, for five years and has been with
the Radiocommunications Agency
Netherlands since 1993, currently as a
technical advisor. His areas of expertise
are short-range devices, antenna technology, digital signal processing, measurements, instrument control, and
simulation and scenario tools. He was a
contributor and chapter coordinator of
the Spectrum Monitoring Handbook
from 1995 to 2005 of the International
Telecommunication Union (ITU) and
authored ITU-R recommendations on
helicopter antenna measurements and
radio noise measurements. He participates in several international working
parties on radio equipment standardization and frequency management.
130
George J. Petersen (george.petersen
@agentschaptelecom.nl) studied telecommunications at the Royal Military Academy (1989). He received his M.Sc. degree
in business management from Radboud
University in 2004. He worked at several
positions in the Ministry of Defense. Since
1998, he has been working for the Radiocommunications Agency Netherlands, currently as a public safety specialist.
Albert J. Westenberg (westberg@
xs4all.nl) served in the Dutch
Navy for 21 months. He
worked as development engineer in a laboratory of the
Ministry of Defense for almost
eight years. In 1978, he started
working for the Radiocommunications Agency Netherlands.
Until his retirement in 2005,
he was involved in maritime
regulations, including equipment type
approval, frequency planning, and international coordination.
Mark J. Bentum (m.j.bentum@utwente.nl) received his M.Sc. degree in
electrical engineering (with honors) in
1991 and his Ph.D. degree in 1995, both
from the University of Twente, Enschede, The Netherlands. From December 1995 to June 1996, he was a
research assistant at the University of
Twente. In June 1996, he joined The
Netherlands Foundation for Research
in Astronomy (ASTRON). In 2008, he
became an associate professor in the
Telecommunication Engineering Group
at the University of Twente. He is now
involved with research and education in
mobile radio communications. His current research interests are short-range
radio communications, novel receiver
technologies in the field of radio astronomy, channel modeling, interference
mitigation, sensor networks, and aerospace. He is a Senior Member of the
IEEE, chair of the Dutch URSI committee, initiator and chair of the IEEE
Benelux AES/GRSS Chapter, board
member of the Dutch Electronics and
Radio Society NERG, board member of
the Dutch Royal Institute of Engineers
KIVI NIRIA, member of the Dutch
Pattern Recognition Society, and has
acted as a reviewer for various conferences and journals. Since December
2013, he has also been the program director of electrical engineering at the
University of Twente.
Cornelis H. Slump (c.h.slump@
utwente.nl) received his M.Sc. degree in
electrical engineering from Delft University of Technology, The Netherlands,
in 1979 and his Ph.D. degree in physics
from the University of Groningen, The
Netherlands, in 1984. From 1983 to
1989, he was with Philips Medical Systems in Best, The Netherlands, as the
head of a predevelopment group on
X-ray image quality and cardiovascular
image processing. In 1989, he joined the
Department of Electrical Engineering
from the University of Twente,
Enschede, The Netherlands. In June
1999, he was appointed as a full professor in signal processing. His main
research interest is in detection and estimation, interference reduction, pattern
analysis, and image analysis as a part of
medical imaging. He is a Member of the
IEEE and of SPIE.
Roel Schiphorst (schiphorst@
gmail.com) received his M.Sc. degree
(with honors) in electrical engineering
in 2000 and his Ph.D. degree in 2004
from the University of Twente, The
Netherlands. From 2004 to 2014, he
was a senior researcher of the chair Signals and Systems. He is the author or
coauthor of over 70 papers, published
in technical journals or presented at
international symposia. His research
interests include coexistence studies in
wireless applications and digital signal
processing in wireless communication
(physical layer). He is a Member of the
IEEE, COST-TERRA, Network of
Excellence ICT ACROPOLIS, and CRplatform NL. Since 2013, he has been
with BlueMark Innovations, a technology firm that specializes in detecting and
locating smartphones.
june 2015
IEEE Antennas & Propagation Magazine
References
[1] K. Davies, Ionospheric Radio. London: Peter
Peregrinus Ltd., 1990.
[2] M. Hervás, J. L. Pijoan, R. Alsina-Pagès, M.
Salvador, and D. Altadill, “Channel sounding and
polarization diversity for the NVIS Channel,”
presented at the Nordic HF, Faro, Sweden, Aug.
2013.
(continued on page 145)
that he had thought of himself as a tinbasher until coming to Berkeley, but
he found experimental work so much
more difficult there that he focused on
theoretical works, developing a body of
theory around the reciprocity relation
and the reaction integral. Another one
of the authors, Jack Welch, was in his
second year as a graduate student in the
Electrical Engineering Department at
UCB and had just done some measurements of diffraction by a slit with Samuel
Silver. When Vic arrived, Jack, intrigued
by Vic’s reaction concept, joined him as
a Ph.D. student. With Vic, he extended
this reaction concept to fields of arbitrary
time dependence and applied it, e.g., to
the scattering of an impulse by a cylinder
of finite length.
Vic continued his work on frequencyindependent antennas and obtained a
solution to the planar equiangular spiral
problem with another Ph.D. student,
Bernie Chow. More recently, this concept of frequency-independent antennas
has made possible extremely wideband
Measurements Corner
feeds for radio telescopes, such as the
Allen Telescope Array at the Hat Creek
Radio Observatory.
In 1965, Henry Booker was asked to
form a department of applied electrophysics at the new University of California at San Diego (UCSD). So, of course,
he called Vic. With the help of Ken
Bowles and Marshall Cohen, Vic was
persuaded to move, once again, to San
Diego in 1966. One of the authors (Bill
Coles) was just starting a Ph.D. project
with Vic at UCB and went to UCSD
with him. The plan was to stay a few
years to help kick off the new department. Of course, UCSD proved to be an
extraordinary place, starting, as it did,
from the top down, so both Vic and Bill
stayed at UCSD until their respective
retirements. Over the next two decades,
the Applied Electrophysics Department changed its name three times as it
evolved into a full-service electrical and
computer engineering department. Vic
was a major factor in broadening out the
department wisely.
At UCSD, Vic turned his attention to
the problem of forward scattering of radio
waves propagating through a turbulent
medium, such as the atmosphere, ionosphere, solar wind, or interstellar plasma.
This form of scattering is dominated by
diffractive effects, first analyzed by Jack
Ratcliffe and his students in the Cambridge ionosphere group. Vic was one of
the first to bring attention to the importance of refractive effects and to show
how they modulate the diffractive scattering. Vic brought another of us to UCSD
(Barney Rickett) to help apply this work to
astrophysics. The work has continued at
UCSD since Vic retired, with a focus on
the remote-sensing aspects of scattering.
Vic lived through a turbulent period.
He made a very positive impact on his
field and on the lives of many individuals. He was fortunate to be able to enjoy
the full command of his marvelous brain
throughout, and we were fortunate to
enjoy him as instructor, research advisor,
mentor, colleague, and friend. We will not
forget him.
(continued from page 130)
[3] A. Kwasinski, W. W. Weaver, P. L. Chapman,
and P. T. Krein, “Telecommunications power
plant damage assessment caused by Hurricane
Katrina—Site survey and follow-up results,” presented at the Annu. Int. Telecommunications
Energy Conf., Providence, RI, Sept. 2006.
[4] A. Tang, D. C. Ray, D. Ames, C. V. R. Murty,
S. K. Jain, S. R. Dash, H. B. Kaushik, G. Mondal,
G. Murugesh, G. Plant, J. McLaughlin, M. Yashinsky, M. Eskijian, and R. Surrampallih, “Lifeline
systems in the Andaman and Nicobar Islands
(India) after the December 2004 great Sumatra
earthquake and Indian Ocean tsunami,” Earthq.
Spectra, vol. 22, no. S3, pp. S581–S606, June
2006.
[5] M. Kobayashi, “Experience of infrastructure
damage caused by the great East Japan Earthquake and countermeasures against future disasters,” IEEE Commun. Mag., vol. 52, no. 3, pp.
23–29, Mar. 2014.
[6] D. M. Fiedler and E. J. Farmer, Near Vertical Incidence Skywave Communication: Theory,
Techniques and Validation. Sacramento, CA:
Worldradio Books, 1996.
[7] B. A. Witvliet, E. van Maanen, G. J. Petersen,
A. J. Westenberg, M. J. Bentum, C. H. Slump,
and R. Schiphorst, “Near vertical incidence skywave propagation: Elevation angles and optimum
antenna height for horizontal dipole antennas,”
IEEE Antennas Propagat. Mag., vol. 57, no. 1, pp.
129–146, Feb. 2015.
IEEE Antennas & Propagation Magazine
[8] H. J. Strangeways, “Estimation of signal correlation at spaced antennas for multimoded ionospherically reflected signal and its effect on the
capacity of SIMO and MIMO HF links,” presented at the IRST, Beijing, China, July 2006.
[9] B. A. Witvliet, E. van Maanen, G. J. Petersen,
A. J. Westenberg, M. J. Bentum, C. H. Slump,
and R. Schiphorst, “Characteristic wave diversity
in near vertical incidence skywave propagation,”
presented at the European Conf. Antennas Propagation, Lisbon, Portugal, Apr. 2015.
[10] E. V. Appleton and G. Builder, “Wireless
echoes of short delay,” Proc. Phys. Soc., vol. 44,
no. 1, pp. 76–87, 1932.
[11] E. V. Appleton and G. Builder, “The ionosphere as a doubly-refracting medium,” Proc.
Phys. Soc., vol. 45, no. 2, pp. 208–220, 1933.
[12] J. A. Rathcliffe, The Magneto-Ionic Theory
and its Application to the Ionosphere. Cambridge,
U.K.: Cambridge Univ. Press, 1962.
[13] Proplab-Pro Version 3, Solar Terrestrial Dispatch, Stirling, AB, Canada, 2010.
[14] L. F. McNamara, T. W. Bullett, E. Mishin,
and Y. M. Yampolski, “Nighttime above-the-MUF
HF propagation on a midlatitude circuit,” Radio
Sci., vol. 43, no. 2, pp. 1–8, Apr. 2008.
[15] B. A. Witvliet, E. van Maanen, G. J. Petersen, A. J. Westenberg, M. J. Bentum, C. H. Slump,
and R. Schiphorst, “The importance of circular
polarization for diversity reception and MIMO in
NVIS propagation,” presented at the European
june 2015
Conf. Antennas Propagation, The Hague, The
Netherlands, Apr. 2014.
[16] J. H. Taylor. (2014, Feb. 1). Weak signal
propagation reporter. [Online]. Available: http://
physics.princeton.edu/pulsar/ K1JT/wspr.html
[17] (2014, Feb. 1). Dimension 4, Thinkman Software. [Online]. Available: http://www.thinkman.
com/dimension4/default.htm
[18] J. D. Kraus, Antennas, 2nd ed. New York:
McGraw-Hill, 1988, pp. 726–729.
[19] G. J. Burke, E. K. Miller, and A. I. Poggio,
“The numerical electromagnetics code (NEC)—A
brief history,” presented at the IEEE Int. Symp.
Antennas Propagation, Monterey, CA, June 2004.
[20] G. J. Burke, E. K. Miller, J. N. Brittingham, D.
L. Lager, R. J. Lytle, and J. T. Okada, “Computer
modeling of antennas near the ground,” Electromagnetics, vol. 1, no. 1, pp. 29–49, 1981.
[21] A. Voors, “NEC based antenna modeler and
optimizer,” 4Nec2 version 5.8.12, Dec. 2012.
[22] T. S. Chu, “Restoring the orthogonality of two
polarizations in radio communication systems—I,”
Bell Syst. Tech. J., vol. 50, no. 9, pp. 3063–3069, Nov.
1971.
[23] J. L. Wheeler, “Transmission loss for ionospheric propagation above the standard MUF,” Radio Sci.,
vol. 1, no. 11, pp. 1303–1308, Nov. 1966.
[24] A. J. Gibson and P. A. Bradley, “A new formulation for above-the-MUF loss,” presented at the
IRST, Edinburgh, U.K., July 1991.
145