--- Page 1 ---
Implementation, Characterization, and Evaluation of an Inexpensive
Low-Power Low-Noise Infrasound Sensor Based on a Micromachined
Differential Pressure Transducer and a Mechanical Filter
OMAR MARCILLO AND JEFFREY B. JOHNSON
Department of Earth and Environmental Science, New Mexico Institute of Mining and Technology,
Socorro, New Mexico
DARREN HART
Sandia National Laboratories, Albuquerque, New Mexico
(Manuscript received 14 June 2011, in ﬁnal form 31 January 2012)
ABSTRACT
The implementation, characterization, and evaluation of a low-cost infrasound sensor developed at the
Infrasound Laboratory at the New Mexico Institute of Mining and Technology (Infra-NMT) are described.
This sensor is based on a commercial micromachined piezoresistive differential pressure transducer that uses
a mechanical high-pass ﬁlter to reject low-frequency outband energy. The sensor features a low-noise,
2.02-mPa rms (0.5–2 Hz), 5.47-mPa rms (0.1–20 Hz), or 5.62-mPa rms (0.05–20 Hz), ﬂat response between
0.01 and at least 40 Hz; inband sensitivity of 45.13 6 0.23 mV Pa21; and a nominal linear range from 2124.5
to 1124.5 Pa. Intended for outdoor applications, the inﬂuence of thermal changes in the sensor’s response
has been minimized by using a thermal compensated pressure transducer powered by an ultralow drift
(,5 ppm 8C21) and noise (,4mV from peak to peak) voltage reference. The sensor consumes a minimum of
24 mW and operates with voltages above 8 V while drawing 3 mA of current. The Infra-NMT speciﬁcations
described above were independently veriﬁed using the infrasound test chamber at the Sandia National
Laboratories’ (SNL’s) Facility for Acceptance, Calibration, and Testing (FACT), and the following pro-
cedures are for comparison calibration against traceable reference stands in voltage and pressure. Because
of the intended broad frequency response of this sensor, the testing chamber was conﬁgured in a double-
reference sensor scheme. A well-characterized microbarometer (with a ﬂat-amplitude response between 0.01
and 8 Hz) and a microphone (with a ﬂat-amplitude response above 8 Hz) were used simultaneously in this
double-reference test conﬁguration.
1. Introduction
Infrasound refers to atmospheric elastic waves prop-
agating with frequencies between 0.02 and 20 Hz.
Infrasound can be produced by natural (e.g., volcanic
eruptions, earthquakes, and meteorites) and anthro-
pogenic events (e.g., nuclear or chemical explosions).
Because it experiences low atmospheric attenuation,
infrasound can travel long distances and provide an ideal
remote sensing tool to study source phenomenology and
atmospheric wave propagation. For example, in volcanic
environments, mass ﬂux, gas velocity, and precise source
location have been estimated from infrasonic records
(Johnson et al. 2003; Vergniolle et al. 2004). In atmo-
spheric studies, infrasound is one method applied to
probe the atmosphere by estimating thermal gradients
and wind ﬁelds (Antier et al. 2007; Le Pichon et al. 2005;
Marcillo and Johnson 2010).
The characteristics of infrasound recorded in a speciﬁc
location depend upon the type and location of the source
and the conditions of the intervening atmosphere. The
magnitude of infrasonic signals from geophysical events,
measured in units of pressure, typically range from
millipascals (mPa) to several thousands of pascals (Pa;
Arrowsmith et al. 2010). Infrasound sensors require high
resolution and a wide dynamic range (DR) to detect
within the infrasonic ﬁeld. Infrasound sensors are also
Corresponding author address: Omar Marcillo, Department of
Earth and Environmental Science, MSEC 208 801 Leroy Place,
Socorro, NM 87801.
E-mail: omarcill@ees.nmt.edu
SEPTEMBER 2012
M A R C I L L O E T A L .
1275
DOI: 10.1175/JTECH-D-11-00101.1
 2012 American Meteorological Society
Unauthenticated | Downloaded 09/17/26 05:50 PM UTC
--- Page 2 ---
required to reject outband energy, such as meteorolog-
ical pressure ﬂuctuations (very low frequency, ,0.01 Hz)
and wind and anthropogenic noise (high frequency,
.20 Hz), which can saturate the sensor’s response.
Thus, the design and operation of infrasound sensors
require special considerations to maximize inband sen-
sitivity and dynamic range, and to reduce the inﬂuence
of noise.
Microphones sensitive to subaudible frequencies, mi-
crobarometers, and differential pressure sensors have
been adapted to measure the infrasonic range. These
sensors use different mechanisms to convert pressure
ﬂuctuations into analog or digital signals. Proper cali-
bration and characterization of infrasound sensors are
required to ensure that this conversion provides an ac-
curate representation of the infrasonic ﬁeld (Kromer
et al. 2007). This paper shows the implementation of an
inexpensive low-power infrasonic sensor developed
at the Infrasound Laboratory at the New Mexico In-
stitute of Mining and Technology (Infra-NMT), which
is based on a commercial micromachined differen-
tial pressure transducer. The mechanisms used for
signal transduction and noise ﬁltering are explained,
and the characterization and evaluation processes are
described.
2. Infrasound sensors
High-sensitivity barometers, used ﬁrst to detect in-
frasonic waves in the atmosphere (Bedard and Georges
2000), have been replaced by very specialized sensors
intended to enhance the response in the infrasonic
range. For example, the MB2000, MB2005 (CEA/
DAM 2009), and Chaparral 50 (Chaparral Physics
2011) [used in the Comprehensive Nuclear-Test-Ban
Treaty (CTBT) International Monitoring System (IMS)]
have noise ﬂoors that are 8 times lower than the mean
atmospheric acoustic noise (Kromer and McDonald
2000) and are normally deployed along with atmo-
spheric noise-reduction systems (Hedlin and Raspet
2003) to reduce the inﬂuence of the wind. This section
provides an overview on the operation of various com-
mercial infrasound sensors and compares their opera-
tion to that of the Infra-NMT sensor introduced here.
Operation of infrasound sensors
Infrasonic sensors, or microbarometers, operate by
measuring perturbations in the background atmospheric
pressure (;100 kPa at sea level). Absolute and differ-
ential pressure sensors can be used to detect infrasonic
waves. An absolute pressure sensor would require
high sensitivity and dynamic range to resolve infrasonic
signals superimposed on atmospheric pressure. For
example, an absolute pressure sensor that complies with
IMS standards would cover the full extent of atmo-
spheric pressure (;100 kPa at sea level) with a resolu-
tion of at least 0.63-mPa rms (IMS sensor ﬂoor noise).
This sensor would have a DR of 164 dB0.63 mPa rms,
equivalent to 27 bits of resolution. This high resolu-
tion in sensor output and digital conversion imply
special
requirements
for
datalogging
capabilities
and sensor operation, such as high-power consump-
tion and low-noise electronic environments. Absolute
microbarometers, such as the MB2000 or MB2005,
provide barometric (absolute) and differential (high-
pass-ﬁltered version of the absolute channel) outputs with
different sensitivities to facilitate digital conversion.
Differential pressure sensors, on the other hand,
measure the infrasound signal compared to a reference
pressure (e.g., average atmospheric pressure). Again,
using the IMS standards, a differential pressure sensor
should resolve signals as small as 0.63-mPa rms and as
large as 158-Pa rms. This sensor would require only an
18-bit digitizer to properly resolve the full range. Dif-
ferential pressure sensors have two ports interconnected
by a sensitive membrane or diaphragm that deforms
with a differential pressure applied to the ports. Bellows
(Jones and Forbes 1962), with silicon or metallic mem-
branes, or a more complex system, such as ﬁber optics
(Zumberge et al. 2003), have been used as sensing di-
aphragms. Displacement sensors (Jiang et al. 2010),
capacitors, or piezoresistive elements can be used to
measure the diaphragm deﬂection (Zuckerwar et al.
2003) and convert that to an electrical signal.
Condenser microphones (CMs), like those in phones
and personal computers, use one of the plates of a ca-
pacitor like a diaphragm; a pressure change over the
plate deforms the diaphragm and changes the capaci-
tance that can be detected with electronic circuits. CMs
are mass produced to operate in audible frequencies
(20–20 000 Hz) and can be used for infrasound de-
tection if electronic conditioning is applied such that
subaudible frequencies are ampliﬁed. However, the
frequency response of these elements at low frequen-
cies can vary from one element to the next, and re-
peatability is not ensured. Nonetheless, the low cost of
these elements made them popular as sensing elements
for infrasonic sensors (Johnson et al. 2003: Ripepe et al.
2007).
Piezoresistivity pressure transducers (PPTs), manu-
factured using microelectromechanical systems (MEMS)
technology, are extensively used in industrial and med-
ical applications for low-pressure measurements (Barlian
et al. 2009). Lower cost and higher performance in
MEMS sensors have been driven by sustained research
and development in MEMS technology (Ko 2007). In
1276
J O U R N A L O F A T M O S P H E R I C A N D O C E A N I C T E C H N O L O G Y
VOLUME 29
Unauthenticated | Downloaded 09/17/26 05:50 PM UTC
--- Page 3 ---
MEMS PPTs, materials that change resistance under
applied stress are connected to micromachined silicon-
based diaphragms. The deﬂection of the diaphragm and
the corresponding change in resistance is measured us-
ing a Wheatstone bridge (Bae et al. 2004). MEMS PPTs
can sense static and dynamic pressures with high pre-
cision. Because of the response of MEMS PPTs at very
low frequencies, their use as a differential sensing ele-
ment for an infrasound sensor depends on the ability to
reject static pressure that can saturate the diaphragm
and make them insensitive to pressures in the targeted
frequency band. Pneumatic ﬁlters can provide a way to
ﬁlter out the undesirable effect. The Infra-NMT sensor
is based on a MEMS PPT transducer and uses a pneu-
matic ﬁlter to reject saturation from very low frequency
changes in the barometric pressure (,0.01 Hz).
3. Conﬁguration of the Infra-NMT sensor
The implementation of the infrasound sensor de-
scribed here is based on the following three main com-
ponents (see Fig. 1): 1) a micromachined pressure
transducer (MEMS PPT), 2) an ultrastable voltage
source, and 3) a mechanical ﬁlter based on a volume el-
ement and a capillary tube. We used a 0.5 INCH-D-MV
pressure transducer from All Sensors (All Sensor, Inc.
2011), which is linearly sensitive to static and dynamic
pressures between 2124.5 and 124.5 Pa (differential
0.0-in. H2O pressure). This transducer has two ports
(A and B) that are interconnected by a ﬂexible silicon
diaphragm (see Fig. 1b). A Wheatstone bridge is used
to sense the deﬂection by measuring the changes in the
piezoresistive elements attached to the diaphragm
FIG. 1. Conﬁguration of the Infra-NMT sensor. (a) Electronic scheme; (b) diagram of the mechanical ﬁlter and its
elements and geometry; and (c) Infra-NMT enclosure, consisting of a PVC electric box with two outlets (one used for
the analog output and power, and the other used to connect the pressure transducer to the atmosphere).
SEPTEMBER 2012
M A R C I L L O E T A L .
1277
Unauthenticated | Downloaded 09/17/26 05:50 PM UTC
--- Page 4 ---
(Fig. 1a). When powered by a voltage source, the sen-
sor’s output is proportional to supply voltage (ratio-
metric). In this conﬁguration, the sensitivity of the
bridge is proportional to the voltage of the source.
Therefore, the voltage input to the sensor’s transducer
is important to keep stabilized and well regulated. To
deliver a very stable and well-regulated voltage we used
the voltage reference LT1021-7 (Linear Technology
2011). The LT1021-7 delivers the required voltage (7 V)
with low noise (4 mV from peak to peak between 0.1 and
10 Hz), and consumes less than 2 mA; this reference
voltage features also ultralow thermal drift (5 ppm 8C21),
long-term stability [7 ppm (10000 h)21], and low line
(1 ppm V21) and load regulation (12 ppm mA21). The
LT1021-7 was chosen to reduce the inﬂuence of the
Infra-NMT sensor to thermal changes and to ensure
voltage stability in the long term. In addition, the voltage
required for the operation of the regulator can be as
low as 8 V, which allows the use of common 9- or 12-V
batteries while maximizing the sensor’s sensitivity.
In the Infra-NMT sensor, port A of the pressure
transducer is open to the atmosphere and port B is
connected to a closed volume chamber [also referred
as backing volume, see Mutschlecner and Whitaker
(1997)], with a capillary tube acting as a leak (Fig. 1b).
The volume–capillary pair, connected to port B, acts like
a low-pass ﬁlter for the atmospheric pressures outside
our frequency band of interest. The difference in pres-
sure between ports is a high-pass-ﬁltered version of the
atmosphere. The characteristics of the pneumatic ﬁlter
are controlled by the geometry of the elements used in
the ﬁlter. The next section provides an overview of the
physics of this type of mechanical ﬁlter and a full de-
scription of the characteristics of the ﬁlter used in the
Infra-NMT sensor.
Using a capillary tube and backing volume as
a mechanical ﬁlter
Mechanical, or pneumatic, ﬁlters based on passive
acoustic elements (capillary tubes and chambers or
volume elements) have been used along with differential
pressure transducers to sense speciﬁc portions of the
acoustic spectrum. For example, Richiardone (1993)
used, and Mutschlecner and Whitaker (1997) described,
mechanical ﬁlters to attenuate very low frequency at-
mospheric events
and high-frequency wind noise.
Pneumatic ﬁlters are modeled by analogy with electric
circuits, such that, depending on the geometry and size
of acoustic elements, electronic counterparts can be
identiﬁed. For example, sound moving through a cap-
illary tube can be modeled as electrical current in
a resistor. Thus, the capillary tube (assuming a small
diameter compared to the length) can be characterized
by its acoustic resistance, deﬁned as (Keefe 1984)
R 5 8h
pr4l,
(1)
where h is the shear viscosity of the ambient air, and r and
l the radius and length of the capillary, respectively. Ac-
cording to Eq. (1), the longer and narrower the capillary
tube, the higher its resistivity to airﬂow. Chambers or
volume elements are able to temporarily ‘‘store’’ differ-
ential pressure and behave like an electronic capacitor;
these elements can be modeled by deﬁning its acoustic
capacitance as (Mutschlecner and Whitaker 1997)
C 5 V
gP 5 V
rc2,
(2)
where V is the backing volume, c is the speed of sound, r
is the air density, P is the average background pressure,
and g 5 1.403 is the adiabatic gas constant, assuming
adiabatic regime.
The processes in the backing volume are not always in
the adiabatic regime but can be in isothermal or transi-
tional (from adiabatic to isothermal) regimes depending
on the frequency range, air conditions, and geometry of
the chamber. The transition frequency for the sensor’s
backing volume (Mentink and Evers 2011), from more
adiabatic to more isothermal behavior, is f 5 3.45 Hz,
where f 5 h/pL2, h 5 k/rcp is the thermal diffusivity, k 5
0.025 W m21 K21 is the thermal conductivity, rcp 5
1.1 3 103 J m23 K21 is the heat capacity per unit volume,
L 5 V/S is the characteristic size, V 5 1 cm3 is the internal
volume, and S 5 7 cm2 is the internal surface. The value of
g tends to 1.403 and 1 for the adiabatic (high frequencies)
and isothermal limits (low frequencies). In the transition
area, the value of g is replaced by (Richiardone 1993)
g 5 Lg,
(3)
where jLj 5 f[A(g 2 1) 2 g]2 1[B(g 2 1)]2g1/2, arg(L) 5
(p/2) 1 tan21[A(g 2 1) 2 g=B(g 2 1)], A 1 iB 5 1 2
[2J1(z)=zJ0(z)] (assuming the chamber is a long cylinder),
z 5 (22i) 1/2[rcpv/(2k)]1/2(V=S), v is angular frequency,
and J0(z) and J1(z) are ﬁrst-kind order 0 and 1 Bessel
functions. Thus, g depends of the frequency and the
ﬁlter geometry and has values between 1 and 1.403.
Figure 2b shows how g changes in the frequency range
of interest for the Infra-NMT sensor.
The parameters of acoustic resistance (R) and ca-
pacitance (C; with the correction for the frequency de-
pendent coefﬁcient g) are used to model the response
to this acoustic low-pass ﬁlter using the electronic anal-
ogy of an RC ﬁlter connected in series,
1278
J O U R N A L O F A T M O S P H E R I C A N D O C E A N I C T E C H N O L O G Y
VOLUME 29
Unauthenticated | Downloaded 09/17/26 05:50 PM UTC
--- Page 5 ---
fcutoff 5
1
2pRC 5 rc2
16h
r4
Vl


5 gP
16h
r4
Vl


5 gLP
16h
r4
Vl


.
(4)
This acoustical low-pass ﬁlter is connected to one of
the ports of the differential pressure sensor and the
other is open to the atmosphere. The differential pres-
sure across the diaphragm [Dp(t) in time or DP(v) in
frequency] is the difference in pressure between un-
ﬁltered and ﬁltered ports,
Dp(t) 5 p(t) 2 hlp(t)*p(t),
(5)
where the asterisk (*) is the convolution operator, p(t)
and P(v) are the atmospheric pressure in time and fre-
quency, and hlp(t) and Hlp(v) 5 1/(1 1 ivRC) is the
impulse and frequency response of an acoustic single-pole
low-pass ﬁlter. Using the convolution theorem, the
frequency counterpart of Eq. (5) can be written as
DP(v) 5 P(v)[1 2 Hlp(v)] or
DP(v) 5 P(v)

1 2
1
1 1 ivRC

.
(6)
Equation (6) can be reorganized as
DP(v) 5 P(v)

ivRC
1 1 ivRC

,
where the term Hhp(v) 5 ivRC/(1 1 ivRC) is the fre-
quency response of a single-pole high-pass ﬁlter with corner
frequency controlled by the product RC; the correspond-
ing impulse response for this ﬁlter is hhp(t). Equations (5)
and (6) can be rewritten as Dp(t) 5 hhp(t)*p(t) and
FIG. 2. (a) Theoretical and experimental frequency response for the Infra-NMT sensor with standard and
test ﬁlters. The theoretical frequency responses for a standard Infra-NMT (blue) and test ﬁlters (magenta)
assuming adiabatic (dashed lines) and transitional (solid lines) regimes, respectively. The frequency re-
sponses of two Infra-NMT sensors (green and cyan lines) are shown, as is the frequency response of the test
ﬁlter sensor (black line with circles). The models assuming high (light gray; 15% in capillary tube length and
backing volume and 5% capillary tube radius) and low (dark gray; 10% in capillary tube length and backing
volume and 5% capillary tube radius) tolerances are represented, respectively. These tolerance areas are
centered at the transitional and adiabatic responses for the standard Infra-NMT and test sensor, re-
spectively. The transition frequency from adiabatic to isothermal regimes (plain black line) is shown. (b) g
and the correction factor L. In the frequency range of the Infra-NMT, g is replaced by g 5 gL.
SEPTEMBER 2012
M A R C I L L O E T A L .
1279
Unauthenticated | Downloaded 09/17/26 05:50 PM UTC
--- Page 6 ---
DP(v) 5 P(v)Hhp(v), respectively. Thus, the entire sys-
tem behaves as a single-pole high-pass ﬁlter. Table 1 shows
the parameters used and the corresponding corner fre-
quencies for mechanical ﬁlters in different infrasound
sensors that are used for sensor evaluation of the Infra-
NMT. In these calculations, the molecular shear viscosity is
1.8013 kg m21 s21 and the atmospheric pressure is equal
to 81 kPa; both are considered constants.
Figure 2a shows the theoretical and experimental
frequency response of the ﬁlters described in Table 1
for the Infra-NMT sensor implementation. The solid
and dashed blue lines are the theoretical frequency re-
sponses for a standard Infra-NMT ﬁlter, assuming adi-
abatic and transitional regimes, respectively. The green
and cyan lines show the frequency responses of two
Infra-NMT sensors. The light and dark gray areas are
centered at the transitional response (for the standard
Infra-NMT) and represent the model assuming high and
low tolerances, respectively, for the geometrical para-
meters of the mechanical ﬁlter. These tolerances are
introduced to account for variations resulting from the
manual construction of the mechanical ﬁlter. Sensors 1
and 2 agree better with the transitional than adiabatic
model. However, sensor 2 still lies outside of the higher-
tolerance region; this deviation can be related to a par-
ticularly long capillary tube and/or a big backing volume.
Also, a nonstandard Infra-NMT mechanical ﬁlter (with
a wider capillary tube) was tested (black line with cir-
cles). The response for this test ﬁlter is closer to the
adiabatic (tolerance areas for this sensor are centered at
the adiabatic model) than transitional response. How-
ever, this response is also outside of the high-tolerance
area, and we suspect a particularly short capillary tube
and/or backing volume. The diameter of the capillary
tube is the most stable of the geometrical parameters of
the ﬁlter and the tolerance was kept at 5%.
4. Sensor characterization and evaluation: Double-
reference technique to determine the Infra-NMT
frequency response
The Infra-NMT infrasound sensor is conﬁgured to re-
spond to a wide frequency range (i.e., 0.01–40 Hz) with
a ﬂat response and sensitivity of 46.85 mV Pa21 (using
manufacturer’s speciﬁcations). This sensitivity is de-
termined by scaling the nominal span, that is, 10 mV for
124.5 Pa (0.5 in. of H2O) at 12 V, to the new voltage
reference (7 V). The response time of the pressure
sensor to get 90% of any applied pressure is less than
100 mS; this ensures a high-frequency corner of several
hundreds to a few thousands of hertz, which is well
above the infrasound band. The low-frequency corner of
0.01 is controlled by geometry of the elements of the
mechanical ﬁlter, where the capillary tube has a length
of 1.1 cm and a diameter of 25 mm, and the volume of
the chamber is 1 cm3. The total current consumption
(3 mA) was determined by adding the consumption of
the voltage reference (1.8 mA) and the consumption of
the transducer (1.2 mA).
To evaluate the performance of this sensor, the in-
frasonic testing chamber at Sandia National Laboratory
(SNL) Facility for Acceptance, Calibration, and Testing
(FACT) was used. Figure 3 shows a diagram of the
calibration system, where a Smart24 digitizer (Geotech
Instruments 2011) provides up to six channels for analog-
to-digital conversion. The active element in this testing
setup is based on a mechanical piston (APS300; APS
TABLE 1. Parameters used in different sensors in acoustic ﬁlters. Equation (3) was used for the determination of the cutoff frequency using
h 5 1.8013 and atmospheric pressure of 81 kPa.
Sensor
Capillary tube
Backing volume
Cutoff frequency
Length (m)
Radius (mm)
(1026 m3)
Model (Hz)
Expt (Hz)
Mutschlecner and Whitaker (1997)
0.5
200
200
0.0062
Richiardone (1993)
1.012
254
10 690
0.0001
Infra-NMT (sensor 1)
0.01
25
1
0.01
0.004
Infra-NMT (sensor 2)
0.01
25
1
0.01
0.01
Infra-NMT (test ﬁlter)
0.01
75
1
1.22
2.5
FIG. 3. Infrasound sensor testing chamber at the SNL FACT site.
The chamber is located 5 m below ground to prevent that diurnal
temperature variations and wind noise affect the evaluation.
1280
J O U R N A L O F A T M O S P H E R I C A N D O C E A N I C T E C H N O L O G Y
VOLUME 29
Unauthenticated | Downloaded 09/17/26 05:50 PM UTC
--- Page 7 ---
Dynamics 2011) that is connected to one of the digital-
to-analog channels of the Smart24 digitizer. The cham-
ber has a nominal volume of 50 L, which was reduced to
30 L to allow the generation of high-pressure signals
(e.g., signals with full-scale amplitudes in the 10–150-Pa
range). The infrasound test chamber is installed in
an underground enclosure, approximately 5-m under-
ground, which reduces the inﬂuence of diurnal tem-
perature variation and wind noise on the component
evaluation.
Standard infrasound sensor evaluation, following
comparison calibration procedures, uses one reference
(e.g., MB2000) to evaluate the new device under test.
However, because of the broad frequency response of the
Infra-NMT sensor design (i.e., 0.01–40 Hz), which in-
cludes relatively high frequencies, a ‘‘double reference’’
sensor scheme was setup using an MB2000 micro-
barometer and a calibrated GRAS 42AE microphone
(GRAS Sound and Vibration 2008). The MB2000 mi-
crobarometer features ﬂat response (within 3dB) for
frequencies between 0.01 and 8 Hz and noise levels less
than 2 mPa between 0.02 and 4 Hz. The GRAS 42AE
microphone features a ﬂat response (within 3 dB) from
1 to 100 Hz. By overlapping the ﬂat responses of these
two well-calibrated references we are able to comment
of the full range of the response of the Infra-NMT
sensor.
Two types of tests were performed on a group of
seven Infra-NMT sensors, using the standard mechan-
ical ﬁlter. The testing tools at the SNL FACT site are
fully described by Kromer et al. (2007) and include
1) Infrasound Sensor Piston-Phone Linearity Veriﬁca-
tion (IS-PLV), which uses sinusoid pressure waves of
increasing amplitude at a single frequency to comment
on the sensors linearity of voltage output versus pres-
sure; and 2) Infrasound Sensor Frequency/Amplitude/
Phase Response Veriﬁcation (IS-FAPR), which uses a
broadband white noise pressure signal to verify the
amplitude and phase response of the sensor design
across the frequencies of interest (i.e., 0.01–40 Hz).
The IS-PLV, for this evaluation, used a single tone of
1 Hz and amplitudes of 0.05, 0.2, 0.4, 2, 4, 23, 50, 100, and
121 Pa on the seven Infra-NMT sensors. The sensors
showed linear responses in the tested amplitudes (see
Fig. 4a), and we expect the sensors to be linear in the
manufacturer’s full-scale linear range (up to 124 Pa).
Results from the IS-PLV test show an average sen-
sitivity of 45.13 mV P21, with a standard deviation of
0.23. Figure 4b shows the calculated sensitivity at the
different tested amplitudes by using the estimated noise
FIG. 4. IS-PLV results and analysis: (a) voltage response of the Infra-NMT sensor vs applied
pressure, and the linear regression of sensor output voltage (Y) vs input pressure (X) and the
coefﬁcient of determination; (b) sensitivity of the Infra-NMT. The sensitivity of the Infra-NMT
is determined by averaging the sensitivity of all sensors at the different amplitudes. The con-
ﬁdence intervals are determined by using the noise’s rms, determined for each input pressure.
SEPTEMBER 2012
M A R C I L L O E T A L .
1281
Unauthenticated | Downloaded 09/17/26 05:50 PM UTC
--- Page 8 ---
during the IS-PLV test for individual sensors; uncertainty
intervals are estimated for sensitivity. For low-pressure
amplitudes, these intervals are big because the sensor
noise and input signal have comparable levels (low signal-
to-noise ratio); for high amplitudes, the intervals increase
due to distortions in the pressure generator and/or re-
sponse of the chamber at higher pressures.
For the IS-FAPR test, we used a normally distributed
white noise signal with a zero-peak amplitude of 23 Pa
and 1000-Hz bandwidth. A two–channel coherence tech-
nique (Merchant and Hart 2011) was applied to two
Infra-NMT sensors to determine the noise spectrum
and conﬁdence intervals (Fig. 5). Using this frequency
spectrum, noises in speciﬁc bands were determined
as follows: 1) 2.02-mPa rms between 0.5 and 2 Hz, 2)
5.47-mPa rms between 0.1 and 20 Hz, and 3) 5.62-mPa
rms between 0.05 and 20 Hz. These intervals were se-
lected to facilitate comparison with other sensors. Table 2
summarizes the complete set of characteristics of the
Infra-NMT sensor determined by the evaluation process.
The cost of the elements for the implementation of an
Infra-NMT single unit is several hundred dollars.
5. Conclusions
We have packaged and calibrated a low-cost infra-
sound sensor with low noise (2.2-mPa rms at 0.5–2 Hz);
high dynamic range (up to 95 dB for the 0.5–2-Hz band);
and sensitivity of 45.13 mV Pa21, which is appropriate
for infrasonic studies that may encounter high levels
of background infrasound noisy and/or near-source en-
vironments. Its low-cost, low-power consumption, and
small size (cf. to other infrasound sensors with higher
sensitivity and lower noise) favor its use in network and
array conﬁguration requiring large quantities of sensors,
in near-source locations (where pressure signals are in-
tense), and/or in higher noise environments. The Infra-
NMT has been used successfully in many environments
and applications [e.g., volcano (Goto and Johnson
2011), atmospheric (Marcillo and Johnson 2010), and
thunder (Arechiga et al. 2010) studies]. A simple model
FIG. 5. Self-noise spectrum of the Infra-NMT sensor. This self-noise response is shown of the
Infra-NMT sensor, the acoustic low-noise model (ALNM) by Bowman et al. (2005), and an
example signal from a 500-lb ammonium nitrate/fuel oil (ANFO) explosion recorded at 5.5 km.
TABLE 2. Summary of the characteristics of the Infra-NMT sensors.
Parameter
Value
Units
Sensitivity
45.13 6 0.23
mV Pa21
Nominal linear range
6124.5
Pa
Noise
mPa rms
0.5–2 Hz
2.02
0.1–20 Hz
5.47
0.05–20 Hz
5.62
Power consumption
(using a 9-V source)
27
mW
Sensor voltage range (SVR)
8 # SVR # 40
V
Current (using a 9-V source)
3
mA
1282
J O U R N A L O F A T M O S P H E R I C A N D O C E A N I C T E C H N O L O G Y
VOLUME 29
Unauthenticated | Downloaded 09/17/26 05:50 PM UTC
--- Page 9 ---
describing the mechanical ﬁlter was tested by comparing
theoretical and experimental responses of different ﬁlter
implementations. We suspect that deviations from the
ideal response are related to the construction of the
mechanical ﬁlter. However, assuming a tolerance up to
15% in the parameters controlled by the construction
(capillary tube length and backing volume), the Infra-
NMT provides a ﬂat response (within 3 dB) throughout
the infrasonic range above the speciﬁed corner fre-
quency. The implementation of the double-reference
scheme in the testing chamber at the SNL FACT site
provided an effective tool to evaluate the wide fre-
quency response of the Infra-NMT sensors.
Acknowledgments. This work was partially supported
through funding from NSF EAR 0838562 and NSF EAR
0738802.
REFERENCES
All Sensor, Inc., 2011: Millivolt output pressure sensors. Datasheet
Rev. C, 4 pp. [Available online at http://www.allsensors.com/
datasheets/DS-0091-Rev-C1.pdf.]
Antier, K., A. Le Pichon, S. Vergniolle, C. Zielinski, and M. Lardy,
2007: Multiyear validation of the NRL-G2S wind ﬁelds using
infrasound from Yasur. J. Geophys. Res., 112, D23110,
doi:10.1029/2007JD008462.
APS
Dynamics,
2011:
APS
300
portable
shaker-ampliﬁer.
Datasheet, Rev. April 2011, 2 pp. [Available online at http://
www.apsdynamics.com/images/stories/Prospekte/APS_Shaker/
APS_300/APS_300_Data_Sheet_en.pdf.]
Arechiga, R. O., J. B. Johnson, H. E. Edens, R. J. Thomas, and
W. Rison, 2010: Acoustic localization of triggered lightning.
J. Geophys. Res., 116, D09103, doi:10.1029/2010JD015248.
Arrowsmith, S. J., J. B. Johnson, D. P. Drob, and M. A. H. Hedlin,
2010: The seismoacoustic waveﬁeld: A new paradigm in study-
ing geophysical phenomena. Rev. Geophys., 48, RG4003,
doi:10.1029/2010RG000335.
Bae, B., B. R. Flachsbart, K. Park, and M. A. Shannon, 2004:
Design optimization of a piezoresistive pressure sensor con-
sidering the output signal-to-noise ratio. J. Micromech. Mi-
croeng., 14, 1597, doi:10.1088/0960-1317/14/12/001.
Barlian, A., W. Park, J. Mallon Jr., A. Rastegar, and B. Pruitt, 2009:
Review: Semiconductor piezoresistance for microsystems.
Proc. IEEE, 97, 513–552.
Bedard, A. J., and T. M. Georges, 2000: Atmospheric infrasound.
Phys. Today, 53, 32.
Bowman, J. R., G. E. Baker, and M. Bahavar, 2005: Ambient in-
frasound noise. Geophys. Res. Lett., 32, L09803, doi:10.1029/
2005GL022486.
CEA/DAM, 2009: MB2000 and MB2005 microbarometers.
[Available online at http://www-dase.cea.fr/public/dossiers_
thematiques/microbarometres/description_en.html.]
Chaparral Physics, 2011: Operation manual for the Model 50A
infrasound sensor. Operation manual, Rev. May 2010, 17 pp.
[Available
online
at
http://www.gi.alaska.edu/ﬁles/Model
%2050A%20Manual%20beta.pdf.]
Geotech Instruments, 2011: The Smart Series. Datasheet Rev.
August 2010. [Available online at http://www.geoinstr.com/
ds-smart24.pdf.]
Goto, A., and J. B. Johnson, 2011: Monotonic infrasound and
Helmholtz resonance at Volcan Villarrica (Chile). Geophys.
Res. Lett., 38, L06301, doi:10.1029/2011GL046858.
GRAS Sound and Vibration, 2008: Low-frequency calibrator
type 42AE. Datasheet, Rev. March 2008. [Available online
at
http://www.grasinfo.dk/documents/pd_42AE_ver_19_03_
2008.PDF/.]
Hedlin, M. A. H., and R. Raspet, 2003: Infrasonic wind-noise re-
duction by barriers and spatial ﬁlters. J. Acoust. Soc. Amer.,
114, 1379–1386.
Jiang, B., K. Yang, and J. Wang, 2010: Design of the opti-
cal ﬁber MEMS infrasound sensor. Proc. 2010 Int. Conf.
Measuring
Technology
and
Mechatronics
Automation
(ICMTMA), Changsha, China, IEEE Computer Society,
1049–1051.
Johnson, J. B., R. C. Aster, M. C. Ruiz, S. D. Malone,
P. J. McChesney, J. M. Lees, and P. R. Kyle, 2003: In-
terpretation and utility of infrasonic records from erupting
volcanoes. J. Volcanol. Geotherm. Res., 121, 15–63.
Jones, R., and S. Forbes, 1962: A microbarograph. J. Sci. Instrum.,
39, 420–427.
Keefe, D. H., 1984: Acoustical wave propagation in cylindrical
ducts: Transmission line parameter approximations for iso-
thermal and nonisothermal boundary conditions. J. Acoust.
Soc. Amer., 75, 58–62.
Ko, W. H., 2007: Trends and frontiers of MEMS. Sens. Actuators A,
136, 62–67.
Kromer, R. P., and T. S. McDonald, 2000: Infrasound sensor
models and evaluation. Proc. 22nd Annual DoD/DOE Seismic
Research Symp.: Planning for Veriﬁcation of and Compliance
with the Comprehensive Nuclear-Test-Ban Treaty (CTBT),
New Orleans, LA, Department of Defense and Department
of Energy. [Available online at http://handle.dtic.mil/100.2/
ADA523399.]
——, D. M. Hart, and J. M. Harris, 2007: Test deﬁnitions for the
evaluation of infrasound sensors. Sandia National Laborato-
ries Tech. Rep. SAND2007-5038, 11 pp. [Available online at
http://prod.sandia.gov/techlib/access-control.cgi/2007/075038.
pdf.]
Le Pichon, A., E. Blanc, and D. Drob, 2005: Probing high-altitude
winds using infrasound. J. Geophys. Res., 110, D20104,
doi:10.1029/2005JD006020.
Linear Technology, 2011: LT1021—Precision Reference. Da-
tasheet, Rev. C. [Available online at http://cds.linear.com/
docs/Datasheet/1021fc.pdf.]
Marcillo, O., and J. B. Johnson, 2010: Tracking near-surface at-
mospheric conditions using an infrasound network. J. Acoust.
Soc. Amer., 128, EL14–EL19.
Mentink, J. H., and L. G. Evers, 2011: Frequency response and
design parameters for differential microbarometers. J. Acoust.
Soc. Amer., 130, 33–41.
Merchant, J., and D. M. Hart, 2011: Component evalua-
tion testing and analysis algorithms. Sandia National Lab-
oratory SAND2011-8265, 136 pp. [Available online at
http://prod.sandia.gov/techlib/access-control.cgi/2011/118265.
pdf.]
Mutschlecner, J. P., and R. W. Whitaker, 1997: The design and
operation of infrasonic microphones. Los Alamos National
Laboratory Tech. Rep. LA-13257, 27 pp. [Available online
SEPTEMBER 2012
M A R C I L L O E T A L .
1283
Unauthenticated | Downloaded 09/17/26 05:50 PM UTC
--- Page 10 ---
at http://www.osti.gov/bridge/servlets/purl/481856-QT6CGV/
webviewable/481856.pdf.]
Richiardone, R., 1993: The transfer function of a differential mi-
crobarometer. J. Atmos. Oceanic Technol., 10, 624–628.
Ripepe, M., E. Marchetti, and G. Ulivieri, 2007: Infrasonic moni-
toring at Stromboli Volcano during the 2003 effusive eruption:
Insights on the explosive and degassing process of an open
conduit system. J. Geophys. Res., 112, B09207, doi:10.1029/
2006JB004613.
Vergniolle, S., M. Boichu, and J. Caplan-Auerbach, 2004: Acoustic
measurements of the 1999 basaltic eruption of Shishaldin
Volcano, Alaska: 1. Origin of Strombolian activity. J. Volca-
nol. Geotherm. Res., 137, 109–134.
Zuckerwar, A. J., T. R. Kuhn, and R. M. Serbyn, 2003: Back-
ground noise in piezoresistive, electret condenser, and ce-
ramic microphones. J. Acoust. Soc. Amer., 113, 3179–
3187.
Zumberge, M. A., J. Berger, M. A. H. Hedlin, E. Husmann,
S. Nooner, R. Hilt, and R. Widmer-Schnidrig, 2003: An optical
ﬁber infrasound sensor: A new lower limit on atmospheric
pressure noise between 1 and 10 Hz. J. Acoust. Soc. Amer.,
113, 2474.
1284
J O U R N A L O F A T M O S P H E R I C A N D O C E A N I C T E C H N O L O G Y
VOLUME 29
Unauthenticated | Downloaded 09/17/26 05:50 PM UTC