ELF magnetic field receiver: frequency
performance and natural signals detection
Receptor ELF de campo magn
´
etico: desempe
˜
no en frecuencia y detecci
´
on de se
˜
nales naturales
L. M. Carducci
∗
, R. Alonso
†
and W. G. Fano
†
∗
Laboratorio de Procesamiento de Se
˜
nales y Comunicaciones
†
Laboratorio de Radiaci
´
on Electromagn
´
etica
Facultad de Ingenier
´
ıa, Universidad de Buenos Aires, Argentina
Abstract—This article presents the analog stage analysis
of an extremely low frequency (ELF) magnetic field receiver.
Details are described about the characterization and modeling
of the antenna used, as well as the frequency behavior of the
system. To achieve higher sensitivity, a coil was manufactured
with a considerably high inductance (800 H), in addition
to a high-gain amplification stage (88 dB). An equivalent
circuit model was defined for the antenna, and its parasitic
elements were determined by laboratory measurements and
computational approximations. The acquisition system has a
digitization stage with a sampling frequency of 100 Hz. To
study its behavior against possible aliasing effects, an analysis
of the analog stage in the frequency domain was carried
out, determining the optimal operating conditions. Field
tests of the equipment were carried out in the mountains of
Villa Alpina, Cordoba, Argentina. Taking into account that
the system was designed for the measurement and study of
natural electromagnetic phenomena in the ELF band, the first
six Schumann resonances have been successfully captured and
detected.
Keywords: loop antenna; ELF; magnetic fields; acquisition
system; Schumann resonance.
Resumen— En este art
´
ıculo se presenta el an
´
alisis de
la etapa anal
´
ogica de un receptor de campo magn
´
etico
para frecuencias extremadamente bajas (ELF). Se describen
detalles sobre la caracterizaci
´
on y modelado de la antena
utilizada, as
´
ı como el comportamiento en frecuencia del
sistema. Para lograr una mayor sensibilidad, se fabric
´
o una
bobina con una inductancia considerablemente alta (800 H),
adem
´
as de una etapa de amplificaci
´
on de alta ganancia (88
dB). Se defini
´
o un modelo de circuito equivalente para la
antena y se determinaron sus elementos par
´
asitos mediante
mediciones de laboratorio y aproximaciones computacionales.
El sistema de adquisici
´
on tiene una etapa de digitalizaci
´
on
con una frecuencia de muestreo de 100 Hz. Para estudiar su
comportamiento frente a posibles efectos de aliasing, se realiz
´
o
un an
´
alisis de la etapa anal
´
ogica en el dominio de la frecuencia
determinando las condiciones
´
optimas de funcionamiento. Las
pruebas de campo del equipo se llevaron a cabo en las sierras
de Villa Alpina, C
´
ordoba, Argentina. Teniendo en cuenta
que el sistema fue dise
˜
nado para la medici
´
on y estudio de
fen
´
omenos electromagn
´
eticos naturales en la banda ELF, se
han logrado detectar con
´
exito las primeras seis resonancias
de Schumann.
Palabras clave: antena de lazo; ELF; campos magn
´
eticos;
sistema de adquisici
´
on; resonancias de Schumann.
I. INTRODUCTION
Electromagnetic Fields are present on the Earth surface,
due to natural and artificial emissions. The magnetic field
of Earth is not only static, but also includes very slow time
variations: secular, annual, 27 days, diurnal and substorm
magnetic bay type variations at very low frequency. The
geomagnetic pulsations phenomena, take place from 1 mHz
to 1 Hz. During the past few decades, a remarkable increase
in the quality and quantity of electromagnetic data recorded
before and during eruptions and earthquakes [1] evidence
that seismic movements are preceded by anomalous electro-
magnetic signals. The Electromagnetic signals as earthquake
precursors have been discussed in many publications and
are also studied in our research project [2], [3]. Aspects
of tectonomagnetism, volcanomagnetism and tectonoelec-
tricity, concentrate on various parts of the electromagnetic
spectrum from radio frequencies (RF) to submicrohertz
frequencies [1]. Other electromagnetic phenomena in the
ULF/ELF band are originated by particles impinging on
the magnetosphere causing electromagnetic emissions that
propagate inside the magnetosphere cavity [4]. Lightning
flashes are the main source of energy for the electromagnetic
background inside the ionospheric cavity. Starting from the
lower band ELF (few Hz) up to VHF (hundreds MHz) the
noise is originates mainly from the energy radiated by light-
ning strokes. The main relevant phenomena in the ELF lower
band are the Schumann Resonances (SR) [4]. The Earth and
the ionospheric layers appear as perfect conductors having
air in between, forming an Earth-ionosphere cavity, in
which electromagnetic radiation is trapped. Lightning strikes
within the troposphere radiate energy into this system and
the waves travel around the Earth. In the case of constructive
interference, Earth-ionosphere cavity resonances are excited
[4]. The SR oscillation detection is a complex procedure
which requires customized and high-quality measurement
systems. The detection of SR employs the limited energy
generated and dissipated by the global lighting activity. The
magnetic field amplitudes received are about few tenths
of pico Tesla. [5]. The most important naturally occurring
VLF signal is the whistler. A whistler is created from a
lightning stroke that passes first to the ionosphere and then
to the magnetosphere above. These particles are then guided
along the Earth’s magnetic field, returning to ground to the
opposite hemisphere [2].
The study of these very low frequencies electromagnetic
phenomena mentioned above, has motivated the construction
of a specific acquisition system for this electromagnetic
spectrum range. The main objective of this work is focused
Revista elektron, Vol. 5, No. 2, pp. 105-111 (2021)
ISSN 2525-0159
105
Recibido: 28/06/21; Aceptado: 09/09/21
lcarducci@fi.uba.ar
Creative Commons License - Attribution-NonCommercial-
NoDerivatives 4.0 International (CC BY-NC-ND 4.0)
https://doi.org/10.37537/rev.elektron.5.2.135.2021
Original Article
on the characterization of a high-sensitivity antenna used as
a sensor for magnetic waves in the ELF band and the system
performance as a whole. In the section II a circuit model
for the used antenna is proposed. The section III describes
the main system specifications, including the antenna, a
high-gain amplifier and the digitizing system. Then, the
section IV shows the results obtained in the antenna model
parameters characterization, the complete system frequency
response and the Schumann resonances measured in filed
test. Finally, in sections V and VI the general conclusions
and some goals that are expected to be carried out in the
near future are mentioned.
II. ANTENNA MODELING
A. Proposed design
For the applications of interest mentioned in this work, the
design, implementation and modeling of a high inductance
loop antenna with magnetic core is proposed. The chosen
design diagram can be seen in the Fig. 1. This antenna
consists of eight commercial copper windings with a large
number of turns each, connected in series around a high
permeability core. This configuration was chosen to have a
simpler and more practical assembly. On the other hand, a
plastic cover was also incorporated for protection. In later
sections more details and characteristics will be seen.
B. Equivalent Circuit Model
The proposed antenna is essentially an inductor, which
can be represented as an ideal inductance L. Thus, the
simple model for impedance of the antenna results Z(ω) =
jωL. However, a series resistance R
s
must be included due
to wire losses, obtaining Z(ω) = jωL + R
s
. On the other
hand, due to the high inductance for this antenna design,
a considerably low resonant frequency occurs, therefore
capacitive effect must also be taken into account. Then,
an improvement of the model contemplates an equivalent
parasitic capacity C which is caused by the copper winding.
In this case would be Z(ω) = (jωL + R
s
)//(jωC)
−1
.
Finally, others losses represented as an equivalent resistance
in parallel R
p
to the whole assembly can also be considered,
that is Z(ω) = (jωL + R
s
)//(jωC)
−1
//R
p
. In this way,
in Fig. 2 the definitive circuit model proposed for the loop
antenna with a magnetic core is shown. Likewise, the total
impedance can also be rewritten as a function of their
resistive and reactive parts: Z(ω) = R(ω) + jX(ω). In
Equations 1 and 2, both parts are expressed in terms of
all electrical parameters and the angular frequency.
Fig. 1. Proposed design for the loop antenna.
Fig. 2. Equivalent electric circuit of the receiver antenna: ideal inductance
L; series resistance R
s
; parasitic capacity C; parallel loss resistance R
p
.
R(ω) =
R
s
+
R
2
s
R
p
+ ω
2
L
2
R
p
1 − ω
2
LC +
R
s
R
p
2
+
ω
L
R
p
+ ωR
s
C
2
(1)
X(ω) =
ωL − ω
3
L
2
C − ωR
2
s
C
1 − ω
2
LC +
R
s
R
p
2
+
ω
L
R
p
+ ωR
s
C
2
(2)
III. MATERIALS AND SPECIFICATIONS
This section describes some characteristics and specifica-
tions for different components and main stages for a low
frequency magnetic field measurement system. In Fig. 3 the
sensing system simplified scheme can be seen. It includes
the following stages: loop antenna, high gain amplifier +
filter, and digital acquisition. These are briefly described
below.
A. Antenna Specifications
An iron core loop antenna with a square section and a
large number of wire turns has been built. The antenna is
made up of 8 reels of copper wire, as can be seen in Fig. 1,
with around 12000 turns each, obtaining a total 96000 turns.
In Table I its main design parameters are detailed. Likewise,
in Fig. 4 a picture with different parts of the built antenna
can be seen. A feature to highlight is the huge number
of turns and relatively high core permeability, since this
implies a very high inductance which improve its sensitivity
for detecting very weak magnetic field signals at very low
frequencies.
Fig. 3. Stages of the magnetic field measurement system.
TABLE I
LOOP ANTENNA SPECIFICATIONS
Parameter Specification
Core section rectangular
Side (m) 0, 0285
Core area (m
2
) 9 · 10
−4
Wire radius (mm) 0, 12
Wire resistance (kΩ) 18,6
Number of turns (total) 96000
µ
r
(approx.) 35
Resonant frequency (Hz) 200
Revista elektron, Vol. 5, No. 2, pp. 105-111 (2021)
ISSN 2525-0159
106
http://elektron.fi.uba.ar
Fig. 4. Loop antenna (high sensitivity). (a) eight reels on core; (b) a single
reel; (c) cover protection and antenna connector.
B. High Gain Amplifier and Anti-aliasing Filter
The voltage induced in the coil (close to micro-volt), is
amplified by a high gain amplifier (Minimum Open Loop
Amplifier, MOLA) [6] of up to G
0
[dB] ' 88 dB, with a very
low common mode rejection ratio. An anti-alising filter is
also included to mitigate the signals of 50 Hz interference
due to the public electric service and other higher frequency
components. It should be noted that in Villa Alpina the
closest public power grid is 8 km away, which reduces
the presence of this type of interference [6]. Even so, the
equipment grounding scheme was carefully planed, with all
ground references connected to the metal chassis of the
testing equipment. All the system was powered by a 12V/
75Ah battery and ±9 V voltage regulators. The amplifier
input resistance has been set at R
in
= 39 kΩ and the output
resistance at R
o
= 1, 2 kΩ. For the anti-aliasing filter two
capacitors were connected in parallel 4, 7 µF//470 nF to
the amplifier output. This result in a total capacity C
o
'
5, 17 µF, forming a simple first order low pass filter, whose
theoretical cut-off frequency is f
c
= (2πR
o
C
o
)
−1
' 25
Hz. Later, a more detailed analysis about this aspect will be
developed.
C. Data Acquisition System
At the amplifier output, the amplified signal is digitized by
an analog to digital converter (ADC) with 12-bit resolution
and 100 Hz sampling rate. The input resistance R
D
for this
ADC is considered much higher than amplifier resistance
output, that is R
D
>> R
o
. The acquired data are encoded
and stored in an 8 GB micro-SD memory, storing one minute
long data for each file, also including additional information
as date, time, and other variables detected. For this digital
part, a 32-bit microcontroller cortex-M3 LPC1769 was used.
All parts of the measurement system has been installed
inside a metal cabinet located inside a waterproof plastic
enclosure.
IV. METHODS AND EXPERIMENTAL RESULTS
A. Antenna model estimation
The electrical circuit of the loop antenna has been mea-
sured by mean of a LCR meter for frequencies between 12
Hz and 187 Hz. For measuring lower frequencies (< 12
Hz) a setup with an Oscilloscope and a function generator
was used. All results can be observed in the Table III. The
typical inductive behaviour of the loop antenna is depicted
in the measurement of the reactance X(ω) = =m{Z(ω)}
versus the frequency, as can be observed in the Figure
5. In this graph, the reactance can be approximated as a
constant slope line for low frequencies thus noticing its
inductive behaviour, but changing this trend when the fre-
quency approaches resonance. As for the real part, R(ω) =
<e{Z(ω)}, it increases with the frequency due to the Joule
effect in a conductor and the losses in the magnetic core, as
can be seen in Fig. 6.
The estimated curves for the antenna model (dotted lines)
are displayed using the best-fitting model to measured
data. Parameters values for this model were determined in
different ways. The series resistance R
s
= 18, 6 kΩ was
measured directly at DC. The inductance was deduced by
computing the slope of the reactance at low frequencies
adding a posterior numerical adjustment, resulting in a value
close to L ' 800 H. Likewise, there is a stray capacitance
due to the layers of the loop antenna windings. Therefore,
due to stray capacitance and huge inductance, a resonance at
unusually low frequencies was observed, resulting f
0
' 200
Hz. Then, for resonance condition (X(ω) = 0 in Eq. 2) the
stray capacitance is C = 0, 79 nF which can be determined
by C = L/(ω
2
0
L
2
− R
2
s
). Last, the parallel resistance
was estimated using numerical approximations to fit model,
obtaining R
p
= 1, 89 MΩ.
0 50 100 150 200
Frequency [Hz]
0
2
4
6
8
10
X [ ]
10
5
Estimated
Measured
Fig. 5. Imaginary part of the antenna impedance (estimated and measured).
0 50 100 150 200
Frequency [Hz]
0
0.5
1
1.5
2
R [ ]
10
6
Estimated
Measured
Fig. 6. Real part of the antenna impedance (estimated and measured).
Revista elektron, Vol. 5, No. 2, pp. 105-111 (2021)
ISSN 2525-0159
107
http://elektron.fi.uba.ar
B. Attenuation at the antenna-amplifier interface
In Fig. 7 the simplified electrical circuit which describes
the connection between antenna and amplifier is shown.
The induced voltage V
oc
, which occurs when a magnetic
field is present, is connected in series with the antenna
impedance itself Z(ω). The resulting effect is a frequency
dependent attenuation at the input interface represented
by R
in
. Considering this situation, the Eq. 3 defines the
attenuation A(ω) as the relationship between open circuit
voltage V
oc
(ω) and input voltage of the amplifier V
in
(ω).
Here, the antenna impedance and amplifier input resistance
produce a significant loading effect for the useful frequency
range (0-50 Hz) due to the antenna inductance which is
unusually high. This could lead to the conclusion that the
problem would be solved by increasing the amplifier input
resistance. However, as will be seen later, other conclusions
can be obtained by doing a more detailed analysis. In Fig.
8 multiple possible curves for the attenuation |A(ω)| are
shown. In solid blue lines attenuations parameterized with
different hypothetical values for the input resistance of the
amplifier R
in
can be seen. Looking at this attenuation in
isolation, can be seen that if the input resistance were high
(eg. 3.9 MΩ) the attenuation would be low, while for a low
resistance (eg. 680 Ω) the attenuation would be considerably
higher around the resonance. There, |A(ω)| for the true input
resistance R
in
= 39 kΩ (in black dotted line) can also be
seen. However, a whole analysis that includes the complete
analog stage is necessary to reach a better conclusion in
which the aliasing effects are also considered.
Fig. 7. Connection between the antenna equivalent circuit (induced voltage
V
oc
(ω) and antenna impedance Z(ω)) and amplifier input resistance R
in
.
Fig. 8. Input attenuation |A(f)| (f = 2π/ω) parameterized with R
in
,
for hypothetical values (solid blue lines) and true (black dotted lines).
A(ω) =
V
in
(ω)
V
oc
(ω)
=
R
in
R
in
+ Z(ω)
(3)
C. Amplifier frequency response
Considering the equivalent circuit and the voltages in-
dicated in Fig. 9, the transference V
out
(ω)/V
in
(ω) can be
defined as expressed in Eq. 4, where V
in
(ω) and V
out
(ω)
are the amplifier input and output voltages, respectively. In
Fig. 10 the theoretical and measured frequency response
magnitude can be seen, showing a cutoff frequency close
to f
c
= 25 Hz and a maximum gain around G
0
|
dB
= 88
dB.
G(ω) =
V
out
(ω)
V
in
(ω)
=
G
0
1 + jωR
o
C
o
(4)
D. Total frequency response and induced voltage
The total transference T (ω) relates V
out
(ω) with V
oc
(ω)
and is defined as the product between attenuation and
amplification stages. That is T (ω) = A(ω)G(ω), resulting
in Eq. 5.
T (ω) =
V
out
(ω)
V
oc
(ω)
=
R
in
R
in
+ Z(ω)
G
0
1 + jωR
o
C
o
(5)
In analogy to the curves shown for A(ω), in Fig. 11 the
frequency response modulus |T (ω)| parameterized with R
in
is plotted. There, it can be observed the frequency response
curve for R
in
= 39 kΩ (dotted black line). Also, it is shown
Fig. 9. Amplifier equivalent circuit, where R
in
is the input impedance,
G
0
is the maximum gain, R
o
the output resistance and C
o
the output
capacitance.
10
0
10
1
Frequency [Hz]
75
80
85
90
|G(f)| [dB]
Measured
Theoretical
Fig. 10. Amplifier and filter Frequency response |G(f )|.
Revista elektron, Vol. 5, No. 2, pp. 105-111 (2021)
ISSN 2525-0159
108
http://elektron.fi.uba.ar
10
1
10
2
10
3
10
4
10
5
Frequency [Hz]
-20
0
20
40
60
80
100
|T(f)| [dB]
Rin=3.9 M
f
N
= 50 Hz
Rin=618
Rin = 39 k
Nyquist Limit
Rin parameterized
Fig. 11. Total frequency response |T (f)| parameterized with R
in
, for
hypothetical values (solid blue lines) and true (black dotted lines).
that when R
in
is big, the frequency response is close to
G(ω), but when R
in
is small the behavior is notoriously
different.
There are two relevant aspects that can be defined how
optimal the measurement system is for the proposed antenna.
First, it is important to analyze the gain of the complete
system, since the attenuation A(ω) can significantly reduce
the gain provided by the amplifier. This is relevant because
the higher the gain, the better the sensitivity of the system.
Second, the digital acquisition device has a sampling rate of
f
s
= 100 Hz. This implies a limit on the available spectrum
given by the Nyquist frequency, which in this case is f
N
=
50 Hz. Therefore, to avoid a possible aliasing effect, it is
necessary that the frequency behavior of T (ω) maximizes
the energy in the useful band (0-50 Hz) and minimizes it in
the not useful band (50-∞ Hz). From Fig. 11, is important to
observe that for low R
in
the energy in the not useful range
seems to have a better rejection of aliasing but with a large
gain reduction, while for high R
in
, the gain is significantly
higher although with a poor behavior against aliasing. To
quantify this condition, the ratio between the filter energy
in the useful range (0 ≤ ω < ω
N
) and the total energy for
all range (0 ≤ ω < +∞) can be defined as a merit factor.
In this way, the Antialiasing Factor (AF) can be expressed
as indicated in the equation 6, where ω
N
= 2π50 [rad/s] is
the Nyquist angular frequency.
AF =
R
ω
N
0
|T (ω)|
2
dω
R
+∞
0
|T (ω)|
2
dω
(6)
From this definition, it is possible to analyze the per-
formance of the antialiasing filter by evaluating AF based
on the different hypothetical values of R
in
. In the Fig. 12
this function (expressed in percentage [%]) can be seen,
observing that for both low and high values of R
in
the filter
behavior is worse for aliasing filtering, while the optimal
value that maximizes AF is given for R
in
' 39 kΩ.
On the other hand, since it is also important to analyze the
total system gain, AF can be plotted in function of a refer-
10
3
10
4
10
5
10
6
Input Resistance Rin [ ]
60
70
80
90
100
AF [%]
Rin = 39 k
55 60 65 70 75 80 85 90
Gain T0 [dB]
60
70
80
90
100
110
AF [%]
Rin = 39 k
T0 = 84.3 dB
Rin = 3.9 M
T0 = 87.9 dB
Fig. 12. AF[%] as function of input resistance R
in
(above) and as function
of DC gain T
0
|
dB
(below).
ence gain corresponding for each value of R
in
. In this case,
the DC gain T
0
= |T (0)| has been chosen. Then, according
to Fig. 12, AF decreases significantly as T
0
approaches the
maximum possible value, that is 88 dB, which is precisely
when R
in
→ ∞. It is also noticeable that for the optimal
case (R
in
= 39 kΩ) the gain is T
0
= 84, 3 dB, maintaining
a gain relatively close to maximum. However, for a future
design, the gain could be enhanced by slightly increasing R
for a suboptimal but close to maximum AF value.
Knowing the total analog stage transference, it is possible
to obtain the induced voltage V
oc
(ω) computing the spec-
trum of signal v
out
(t) which is sampled by the acquisition
system. To do this, the spectrum V
out
(ω) can be equalized
by applying the inverse of T (ω) to obtain then V
oc
(ω).
In Fig. 13, the total transfer (blue solid and black dotted
lines) and the equalization function |T (ω)|
−1
(black solid
line) are plotted. It is important to clarify that for a more
general case, this type of equalization can have instability
problems if there are zeros located outside left half plane
in Laplace’s s space, or also noise amplification at high
frequencies. Regarding instability, in the presented system
there is no such problem since the T(s) only has one pole
in the left half plane, so its inverse doesn’t have any pole.
On the other hand, although there is noise amplification,
the useful spectrum for signal is relatively low for 100 Hz
sample rate and the noise amplification can be mitigated
by subsequent processing techniques. Then, according to
Eq. 7 the compensated spectrum magnitude |V
oc
(ω)| can
be determined.
|V
oc
(ω)| =
|Z(ω) + R
in
|
p
1 + (ωR
o
C
o
)
2
G
0
R
in
|V
out
(ω)| (7)
E. Application: Schumann resonances detection
Taking into account the objectives of this work, it is
important to verify the effectiveness of the proposed antenna
in combination with the rest of the system in order to
detect very small natural electromagnetic signals. In this
Revista elektron, Vol. 5, No. 2, pp. 105-111 (2021)
ISSN 2525-0159
109
http://elektron.fi.uba.ar
10
1
10
2
10
3
10
4
10
5
Frequency [Hz]
-100
-50
0
50
100
[dB]
f
N
= 50 Hz
not useful range
useful range
Transference |T(f)|
Useful transference
Ecualization |1/T(f)|
Nyquist Limit
Fig. 13. Total response |T (f )| and |T (f )|
−1
for R
in
= 39 kΩ.
case, the authors have measured the Schumann resonances
in the location of Villa Alpina, C
´
ordoba, Argentina. Fig.
14 shows the complete installation of the sensor station on
site, which has also been used in other works with similar
antennas [7], [6]. This location was chosen due to the very
low artificial noise of the power lines radiation and others
possible urban interference of higher frequencies [8]. In
order to see the spectrum of the amplifier output signal,
the power spectral density (PSD), S
out
(ω) [V
2
/Hz], can be
calculated from digitized and stored data. For the better
plotted and visibility of the Schumann Resonances peaks,
the average PSD can be used, according to some methods
seen in [7]. Then, the amplifier output voltage spectrum
can be computed as |V
out
(ω)| = S
1/2
out
(ω) [V/
√
Hz]. This
is shown in Fig. 15. There, up to the first six Schumann
resonances can be clearly seen. Table II summarizes the
mean and deviation of the frequencies estimated using 60
PSDs of one minute long each. Finally, the induced voltage
spectrum modulus |V
oc
(ω)| can be calculated by mean of
the equation 7 and using the estimation of |V
out
(ω)|, thus
obtaining the graph of Fig. 16. An increase in amplitude is
noted as the frequency rises, but this characteristic is normal,
TABLE II
SCHUMANN RESONANCES ESTIMATED FROM MEASUREMENTS
Parameter 1st 2nd 3rd 4th 5th 6th
Mean [Hz] 7.84 14.23 20.67 26.67 32.53 39.83
Deviation [Hz] 0.10 0.09 0.14 0.14 0.18 0.19
Fig. 14. Installation of the Sensing Station and the antenna in Villa Mar
´
ıa,
C
´
ordoba, Argentina.
0 10 20 30 40 50
Frequency [Hz]
0
0.5
1
1.5
2
2.5
3
3.5
4
|Vout(f)| [mV/ Hz]
Average
Smoothed
Fig. 15. Average PSD of the voltage measured (60 time windows of one
minute each) and smoothed curve. An additional digital filter has also been
applied to suppress 50 Hz components.
Fig. 16. Average PSD and smoothed curve of the induced voltage
estimated after of system equalisation.
as can be seen in similar works such as [9]. Likewise,
the last stage in the estimation, which is not part of this
work, would be to apply the antenna factor to obtain the
spectrum measured in units of magnetic field (tesla), that is
B(ω) = K(ω) V
oc
(ω). The antenna factor K tends to be an
inversely proportional to the frequency, so the noise floor
would be approximately flat after applying it.
V. CONCLUSIONS
In this work we have characterized a loop antenna at very
low frequencies to determinate its equivalent impedance
model. Also, numeric approximations have been used to
estimate other parameters of the model. With these result, a
huge inductance L ' 800 H has been verified, which implies
a very high sensibility for field sensing. An unusually low
frequency of resonance, near to 200 Hz, was observed due to
the huge inductance and the parallel stray capacitance. It was
found that the proposed antenna together with the rest of the
measurement system have a very good performance in the
Revista elektron, Vol. 5, No. 2, pp. 105-111 (2021)
ISSN 2525-0159
110
http://elektron.fi.uba.ar
TABLE III
MEASUREMENT OF THE LOOP ANTENNA WITH MAGNETIC CORE.
F [Hz] X [kΩ] R [kΩ] F [Hz] X [kΩ] R [kΩ]
3 30,29 26,34 70 387,04 175,93
5 47,25 25,4 75 414,22 211,34
7 57,44 35,68 80 452,39 226,19
9 75,55 33,73 85 471,05 239,11
12 97,26 51,46 90 501,02 267,93
14 107,4 57,44 94 525,65 287,24
15 113,47 59,72 100 560,46 325,85
16 116,62 60,42 104 587,45 341,54
18 128,48 69,07 110 627,56 389,79
20 140,12 73,74 115 666,21 432,6
22 144,31 78,86 120 694,42 472,39
25 161,01 82,57 125 730,42 529,29
26 164,67 83,59 130 762,01 581,68
28 174,7 89,59 136 785,3 665,51
30 190,38 90,66 140 819,83 738,59
35 213,31 101,58 145 845,47 782,84
40 231,72 108,79 150 901,01 883,34
45 257,01 122,39 160 864,57 1094,39
50 273,32 130,15 166 846,92 1366,01
55 307,22 143,56 176 723,22 1607,16
60 332,88 158,82 187 434,73 1890,15
65 359,4 171,47 - - -
range of frequencies of interest. This is verified by analyzing
the total transfer of the analog measurement stage, observing
an optimal behavior for aliasing filtering and guaranteeing
a high gain. The resulting portable equipment showed a
very high sensitivity, proving that it can detect very weak
natural signals at extremely low frequencies. As a particular
application case, the first six Schumann resonances have
been detected successfully with a great precision.
VI. FUTURE WORKS
It is expected to measure antenna factor to determinate
the magnetic filed B (tesla) from voltage measured. To do
this, a measurement bench with Helmholtz coils of a suitable
size will be used to generate a known magnetic field as a
calibration reference. Likewise, tests will be carried out with
other magnetic materials and also assembly three antennas
that allow monitoring the three spatial components of the
near field.
VII. ACKNOWLEDGMENT
The authors would like to thank to Edgardo Maffia
from Electronica Aplicada Company for the help during
experimental process. To Norman Trench for taking care
of the installation of equipment and the magnetic field
measurements. To Julio Zola and Enrique Zothner for pro-
viding a working prototype of the high gain amplifier. To
the Universidad de Buenos Aires for the project Grant
20020150100085.
REFERENCES
[1] M. J. S. JOHNSTON, “Review of electric and magnetic fields accom-
panying seismic and volcanic activity,” Surveys in Geophysics, vol. 18,
no. 5, p. 441–475, 1997.
[2] W. G. Fano, G. Quintana, and R. Alonso, “3D magnetic field measure-
ment at ELF and SLF frequencies,” IEEE 2018 GEMCCON, 2018.
[3] G. I. Quintana, R. Alonso, and W. G. Fano, “Loop antenna charac-
terization for ELF and SLF measurements,” Elektron, Universidad de
Buenos Aires, Argentina., vol. 2, no. 2, p. 95–100, 2018.
[4] C. Bianchi and A. Meloni, “Loop antenna characterization for ELF
and SLF measurements,” ANNALS OF GEOPHYSICS., vol. 50, no. 3,
p. 435–445, 2007.
[5] C. I. Votis, G. Tatsis, V. Christofilakis, S. K. Chronopoulos,
P. Kostarakis, V. Tritakis, and C. Repapis, “A new portable
ELF Schumann resonance receiver: design and detailed analysis
of the antenna and the analog front-end,” EURASIP J. Wirel.
Commun. Netw., vol. 2018, p. 155, 2018. [Online]. Available:
https://doi.org/10.1186/s13638-018-1157-7
[6] N. Trench, R. Alonso, L. M. Carducci, J. Kelly, E. Maffia,
E. Zothner, J. Zola, and W. G. Fano, “Sistema de medici
´
on remota
de campo magn
´
etico para se
˜
nales precursoras s
´
ısmicas de muy
bajas frecuencias,” in Latinmag 2019, Dec 2019. [Online]. Available:
https://www.elatinmag.com/aplimetodos-103-trench
[7] L. M. Carducci, R. Alonso, F. Luna, J. Zola, E. Zothner, and W. G.
Fano, “Aplicaciones y m
´
etodos de medici
´
on y procesamiento de se
˜
nales
para la detecci
´
on de anomal
´
ıas de campo magn
´
etico de muy bajas
frecuencias,” LATINMAG 2019, 2019.
[8] E. Maffia, V. Trainotti, W. G. Fano, and N. Trench, “Medici
´
on de la
resonancia de Schumann,” Latinmag Letter, vol. 1, no. A02, pp. 1–8,
2011.
[9] G. T. et al, “Design and implementation of a test fixture for ELF Schu-
mann resonance magnetic antenna receiver and magnetic permeability
measurements,” Electronics, 2020, 9, 171, 2020.
Revista elektron, Vol. 5, No. 2, pp. 105-111 (2021)
ISSN 2525-0159
111
http://elektron.fi.uba.ar