sensors

Article
Development of a High-Sensitivity Optical
Accelerometer for Low-Frequency
Vibration Measurement
Rui-Jun Li 1 ID , Ying-Jun Lei 1 , Zhen-Xin Chang 1 , Lian-Sheng Zhang 1, * and
Kuang-Chao Fan 1,2 ID
1 School of Instrument Science and Opto–Electronics Engineering, Hefei University of Technology,
Hefei 230009, China; rj-li@hfut.edu.cn (R.-J.L.); 2016170008@mail.hfut.edu.cn (Y.-J.L.);
2017110054@mail.hfut.edu.cn (Z.-X.C.); fan@dlut.edu.cn (K.-C.F.)
2 School of Mechanical Engineering, Dalian University of Technology, Dalian 116024, China
* Correspondence: lszhang@hfut.edu.cn; Tel.: +86-136-9654-0049




Received: 5 July 2018; Accepted: 30 August 2018; Published: 1 September 2018


Abstract: Low-frequency vibration is a harmful factor that affects the accuracy of

micro/nano-measuring machines. Low-frequency vibration cannot be completely eliminated by
passive control methods, such as the use of air-floating platforms. Therefore, low-frequency vibrations
must be measured before being actively suppressed. In this study, the design of a low-cost
high-sensitivity optical accelerometer is proposed. This optical accelerometer mainly comprises
three components: a seismic mass, a leaf spring, and a sensing component based on a four-quadrant
photodetector (QPD). When a vibration is detected, the seismic mass moves up and down due
to the effect of inertia, and the leaf spring exhibits a corresponding elastic deformation, which is
amplified by using an optical lever and measured by the QPD. Then, the acceleration can be calculated.
The resonant frequencies and elastic coefficients of various seismic structures are simulated to attain
the optimal detection of low-frequency, low-amplitude vibration. The accelerometer is calibrated
using a homemade vibration calibration system, and the calibration experimental results demonstrate
that the sensitivity of the optical accelerometer is 1.74 V (m·s−2 )−1 , the measurement range of the
accelerometer is 0.003–7.29 m·s−2 , and the operating frequencies range of 0.4–12 Hz. The standard
deviation from ten measurements is under 7.9 × 10−4 m·s−2 . The efficacy of the optical accelerometer
in measuring low-frequency, low-amplitude dynamic responses is verified.

Keywords: optical accelerometer; low-frequency vibration; leaf spring; four-quadrant photodetector

1. Introduction
In ultra-precision measurement, ultra-precision machining, metrological verification and similar
applications, the low-frequency micro-vibration caused by human activity and natural factors has
great influence. Its amplitude is extremely small (in the micron range) and it cannot be eliminated
completely by passive control methods, such as the use of air-floating platforms. Thus, a vibration
control technique is a good solution. The main problems of active vibration isolation techniques involve
the design of appropriate sensor techniques and control schemes. In terms of vibration frequency,
the vibration generated by certain human activities, transportation, and mechanical devices ranges
of 1–10 Hz [1,2]. Accordingly, the accelerometer used in measuring environmental vibration must be
designed specifically for low frequency and high sensitivity.
Table 1 shows a summary of the existing accelerometers for low-frequency, low-amplitude
vibration, which include piezoelectric (PZT) accelerometers, strain accelerometers, fiber Bragg grating
(FBG) accelerometers, micro-electro-mechanical system (MEMS) accelerometer, optical accelerometers
and other types.
Sensors 2018, 18, 2910; doi:10.3390/s18092910 www.mdpi.com/journal/sensors
Sensors 2018, 18, 2910 2 of 15

Tian et al. [3] designed a high sensitivity and low transverse effect PZT accelerometer.
The piezoelectric accelerometer fabricated on an n-type single crystal silicon wafer and the sensor
chips were wire-bonged to printed circuit boards (PCBs). The sensitivity of the accelerometer is
9 mV/g, the linearity is 0.0205, and the hysteresis error is 0.0033. Zou et al. [4], Nishshanka et al. [5],
and Tims et al. [6] have also made progress in this field and the detailed data is shown in
Table 1. From Table 1, we can see that the PZT accelerometers have a wide bandwidth. However,
their sensitivities are low, and they cannot be used to detect low frequency micro-vibrations.
Kamentse et al. [7] presented a strain gauge accelerometer. A nonconductive ceramic beam was
used as elastic element. Four strain gauges were attached on the surface of the beam using screen
printing process. This accelerometer has a wide detection range. Santana et al. [8] have also made
progress in this field and the detailed data is shown in Table 1. Strain gauge accelerometers have
widely used in industry because of high stability and low price. However, the accuracy of the strain
gauge accelerometer is low, and the detection of low frequency micro-vibration cannot be achieved.
Liu et al. [9] applied FBG to an accelerometer. In their design, two symmetrical bended spring
plates were used as elastic elements so as to double the wavelength shift of the FBG. This accelerometer
has a low-frequency response range of 0.7–20 Hz. However, it cannot describe the output when the
outside acceleration is small (less than 1 m·s−2 ). Gao et al. [10], Zhang et al. [11], Zhang et al. [12],
Zeng et al. [13], and Li et al. [14] have also made progress in this field and their results is shown in
Table 1. It can be seen that the FBG accelerometers can achieve the detection of low-frequency vibration.
However, it is difficult for them to detect the vibration of low-frequency and low-amplitude since their
resolution is limited.
Sabato et al. [15–17] developed a wireless MEMS accelerometer for micro-vibrations.
This accelerometer used voltage to frequency conversion (V/F) other than analog to digital
conversion (ADC) to promote the accelerometer’s performance. However, the validations for
the wireless accelerometer in the low-frequency, ultra-low amplitude range have not been
reported. Swartz et al. [18], Cho et al. [19], Whelan et al. [20] Meyer et al. [21], Rice et al. [22],
and Kohler et al. [23,24] have also made progress in this field and their results are shown in
Table 1. We can find that the MEMS accelerometer has a wide frequency response range. However
low-frequency and low-amplitude vibration detection cannot be achieved because their low sensitivity.
Several accelerometers were developed based on the high precision displacement or angle sensor
modified from DVD pick up head [25–27]. These optical accelerometers have high sensitivity and
enable low amplitude vibration detection. The cantilever beam produces angular motion when sensing
vibration, thereby causing measurement errors.
In addition to the above technologies, Zheng et al. [28] proposed a new maglev sensor to
measure ultra-low frequency (ULF) vibration, which uses hybrid-magnet levitation structure with
electromagnets and permanent magnets as the supporting component. Jiang et al. [29] proposed an
all-metal double metal diaphragm-based optical fiber accelerometer with low transverse sensitivity.
Lin et al. [30] designed a compact-size fiber optic accelerometer to achieve both high resolution and
wide dynamic range.
In summary, for the cost-effective and accurate measurement of low-frequency, low-amplitude
vibration, a new optical accelerometer system is developed in this study. A centrally symmetric leaf
spring fixed on a seismic mass was used as a vibration sensitive unit, four-quadrant photodetector
(QPD) based optical sensor was used to sensor the seismic mass’s displacement. Section 2 introduces
the structure and principles. Section 3 explains the design, analysis, and fabrication of the optical
accelerometer. The performances of the optical accelerometer are discussed in Section 4. The summary
and prospect of the optical accelerometer are presented in Section 5.
Sensors 2018, 18, 2910 3 of 15

Table 1. Summary of accelerometers of low-frequency micro-vibration. BW: bandwidth; PZT:

piezoelectric; FBG: fiber Bragg grating; MEMS: micro-electro-mechanical system.

Principle Sensitivity (V/g) Range (g) BW (Hz) Noise-Density (µg Hz−1/2 ) Study
9 mV/g – – – [3]
X-0.93 mV/g
Y-1.13 mV/g >0.04 <100 – [4]
PZT
Z-0.88 mV/g
15.6 mV/g – 60–1.5 k 1.7 [5]
2.82 – 2–500 – [6]
– 0–5 <100 – [7]
Strain
Variable ±20,000 µε <100 70 [8]
* – 0.7–20 – [9]
0.135 0.1–2 80–800 – [10]
0.362 < 0.5 1–10 – [11]
FBG
* – 0–25 – [12]
* 0.5–1.5 20–70 – [13]
* 0.1–0.4 5–15 – [14]
1.2 ±3 0.2–1500 0.3 [15,16]
2 ±1 0–50 70 [18]
1 ±2 0–50 140 [19]
MEMS 0.66 ±2 0–100 30 [20]
0.66/0.22 ±6/±2 0–100 30 [21]
0.66 ±2 0–50 50 [22]
1.2 ±3 0–1500 0.3 [23,24]
12.28 <0.017 F 3–24 20 [25]
Optical 24.36 <0.0023 F 3–6 – [26]
22.9 <0.08 F 0.5–50 30 [27]
– – 0.2–0.4 – [28]
Others * – 5–400 0.09 [29]
– – 20–140 0.048 [30]
“*” indicates that the parameter is not comparable; “–” indicates that the parameter is not mentioned; “F ” indicates
that the parameter is not mentioned, and read from the picture of the paper.

2. Structure and Principle

As shown in Figure 1, a common contact accelerometer comprises a spring, a seismic mass, and a
displacement sensor arranged within a housing attached to a base. In operation, the base is mounted
on the vibrating structure to be measured, and the relative displacement between the seismic mass
and the base is recorded by the displacement sensor. Following Newton’s second law, the force acting
on the seismic mass m can be expressed as:

ma = k( xm − xb ), (1)

where xm is the displacement of the seismic mass and xb is the displacement of the base; thus,
the relative displacement between the seismic mass and the base can be expressed as:
m
( xm − xb ) = a. (2)
k
If the acceleration a is constant, then the larger the seismic mass m and the smaller the elastic
coefficient of spring k would result in the larger relative displacement between the seismic mass and the
base. Therefore, the stronger the output signal of the displacement sensor is, the higher the sensitivity
of the accelerometer will be. Nevertheless, if the seismic mass m is too heavy and the elastic coefficient
k of the spring is too low, then the seismic mass is liable to generating a lateral yaw and increasing the
measurement error. Thus, the elastic coefficient k of the elastic support and the mass m of the seismic
mass directly affect the accelerometer performance. The appropriate spring and seismic mass must
Sensors 2018, 18, 2910 4 of 15

be selected to realize the high-sensitivity detection of low-frequency vibration. The third part of this
paper emphasizes
Sensors 2018, 18, xthe design of spring and seismic mass. 4 of 15

Housing
Sensors 2018, 18, x 4 of 15
Spring
(t) k Displacement
Housing sensor
Seismic
massSpring
(t) k m Displacement
sensor
(t) Seismic
mass
m Base

(t)
Figure
Figure 1. 1. Typicalaccelerometer
Typical accelerometer structure
structure diagram.
diagram.
Base
On the basis of the typical accelerometer structure shown in Figure 1, a leaf spring is used as the
On the basis
elastic support ofinthethistypical
study. accelerometer
As shown structure
in accelerometer
Figure shown
2, thestructure
leaf spring inisFigure
mounted 1, on
a leaf springofisthe
the frame used as
Figure 1. Typical diagram.
the elastic support in
accelerometer. A this
seismicstudy.
massAs fixed shown
to thein Figure
lower surface2, theof leaf
the leafspring
spring is is
mounted
used as aon the frame of
vibration
the accelerometer.
system, Onandthe basisA QPD
the seismic
of the massaccelerometer
intypical
associationfixed to the
with lower
a laser diode
structure surface ofFigure
is regarded
shown in theasleaf aspring
a1,pick-up is used
sensor.
leaf spring asasa the
Mirror
is used 2vibration
is
attached
system,elastic
and the to
supportthe
QPD bottom
in in
this of the seismic
study. As shown
association mass.
with in When vibration
Figurediode
a laser 2, the leaf occurs, the leaf
spring is mounted
is regarded spring
as a pick-up undergoes
on thesensor. elastic
frame ofMirrorthe 2 is
deformation,
attachedaccelerometer.
to the bottom andA the of seismic
seismic mass
the seismicmass fixed moves
to the
mass. uplower
When and vibration
down.
surfaceThe position
of occurs,
the leaf the of the
springleafprojected
used aslight
is spring spot on elastic
a vibration
undergoes
the four-quadrant
system, and the QPD photodetector
in association thenwithshifts. To add
a laser a fine-tuning device
as a for optical path adjustment,
deformation, and the seismic mass moves up anddiodedown. is regarded
The position pick-up
of the sensor.
projected Mirror 2 is
light spot on
Mirror
attached 1 istoadded on theofleft
the bottom theside
seismicand mass.
installedWhenon the fine-tuning
vibration occurs, device.
the leaf spring undergoes elastic
the four-quadrant
Figure 2 shows
deformation,
photodetector
and thethe seismic
then
opticalmass shifts.
pathmovesof the up To
sensoradd a
and system.
fine-tuning
down. The Theposition device
laser beam of the
for
from optical
the laser
projected
path
diode
light
adjustment,
spotwithon
Mirroran1 is added
theincident on the left
angle isphotodetector
four-quadrant side
reflected by the and installed
mirrors
then shifts.and on
To thenthe fine-tuning
add aprojected
fine-tuning onto device.
the QPD.
device A vertical
for optical pathdisplacement
adjustment,
Figure
∆h
Mirror 2 shows
of Mirror1 is added theon
2 causes optical leftpath
a lateral
the shift
side of∆lthe
and sensor
of the
installed onsystem.
light spot TheQPD.
on the
the fine-tuning laser Thebeam
device. from thebetween
relationship laser diode∆h and with an
∆l angle
incident is expressed
Figure by: the
2 shows
is reflected by optical
the mirrorspath ofand the sensor system. The
then projected ontolaserthe beam
QPD. from the laser displacement
A vertical diode with ∆h
of Mirror 2 causes a lateral shift ∆l of the light spot on
an incident angle is reflected by the mirrors and then
hthe QPD. The relationship between ∆h and ∆l is
projected onto the QPD. A vertical displacement
∆h of Mirror 2 causes a lateral shift ∆l of the l  spot on, the QPD. The relationship between ∆h and
light (3)
expressed by: sin 
∆l is expressed by: ∆h
where α is the angle between the laser and ∆l = 2. Equation
Mirror , (3) shows that the vertical motion of (3)
−1hα
sin
 l  ,
the seismic mass is magnified by the factor of (sinα) . Provided that the value of angle α is smaller, (3)
where α is the angle between the laser and Mirrorsin 2.  Equation (3) shows that the vertical motion of
the sensitivity of the sensor is higher, but this situation also increases the lateral size of the
the seismic −1 . Provided that the value of angle α is smaller,
wheremass α is the
accelerometer. is magnified
angle between
Considering bythe theelastic
the factor
laser and ofMirror
(sinα)2.optical
structure, Equation path,(3) and
shows that the
circuit vertical
design, whichmotion of
affect
the sensitivity
the seismic
accelerometer of themasssensor is higher,
is magnified
sensitivity, by but
the existing this situation
the factor
amplifierof (sinα)
design −1also
canincreases
. Provided
meet the the
the lateral
value ofsize
thatrequirements angle
well.ofTherefore,
αthe accelerometer.
is smaller, in
Considering
the design
the the elastic
sensitivity structure,
of theofelastic
the sensor optical
structure and path,
is higher, butand
optical thiscircuit
path, situation
although design, which
alsobasic
the increases affecttheaccelerometer
performance lateral size of sensitivity,
requirements the
of
accelerometer.
the
the existing amplifierConsidering
accelerometer are met,can
design wethemust
meet elastic
thestructure,
consider opticalwell.
the difficulty
requirements path,
of and circuit
processing
Therefore, andin design,
reduce the
the design which
cost.ofaffect
Thus,
the elastic
structure and optical path, although the basic performance requirements of the accelerometerin
accelerometer
45° is selected sensitivity,
as the value the
of existing
the amplifier
included angle design
α. Mirrorcan meet
1, the requirements
which is introduced well.
into Therefore,
the design to
are met,
the design
adjust of the path,
the optical elasticisstructure
mountedand on aoptical path, although
2D fine-tuning device, the
andbasictheperformance
angle can ◦ be requirements
fine-tuned. of
we must consider the difficulty of processing and reduce the cost. Thus, 45 is selected as the value
the accelerometer are met, we must consider the difficulty of processing and reduce the cost. Thus,
of the included
45° is selected angle α. Mirror
as the value of1, thewhich
included is introduced
Floating
angle α. plate
Mirror into the design
1, which to adjust
is introduced intothethe optical
design topath, is
mounted on a 2D fine-tuning device, and the angle can be fine-tuned.
adjust the optical path, is mounted on a 2D fine-tuning device, and the angle can be fine-tuned.

Floating plate Seismic

mass
Frame
Mirror 2
Seismic
mass
Fine- Frame
tuning
Mirror 2
device QPD
Z Fine-
Mirror 1
tuning QPD
device QPD
X
Z
Mirror 1 Laser diode QPD
Figure 2. The schematic of an
X optical accelerometer. QPD: four-quadrant photodetector.
Laser diode
Figure 2. The schematic of an optical accelerometer. QPD: four-quadrant photodetector.
Figure 2. The schematic of an optical accelerometer. QPD: four-quadrant photodetector.
Sensors 2018, 18, 2910 5 of 15

3. Design, Analysis and Fabrication

In this study, the accelerometer is intended to measure the movement of the environmental
vibration (1–10 Hz) [1,2]. The frequency of the acceleration cannot be correctly measured until it
is lower than 20% of the resonance frequency of the accelerometer when the acceleration acts on
the base. This limits the frequency response design of the accelerometer. Therefore, the required
resonance frequency value of the accelerometer should be approximately 50 Hz. In addition to the
resonance frequency of the accelerometer, the elastic coefficient of the accelerometer’s elastic part must
be considered in the design.
As shown in Figure 2, the elastic part of the optical accelerometer contains a seismic mass, a leaf
spring, a floating plate, and a mirror. The symmetrical structure has the same deformation; thus,
the horizontal position of Mirror 1 is maintained to avoid the lateral yaw motion. The design of the leaf
spring must consider its structural symmetry and elasticity coefficient. Given these two factors and
previous experience in leaf spring design, we propose four types of leaf springs (as shown in Figure 3).
The leaf spring of shape (a) has a simple structure and a large elastic coefficient. The leaf spring of
shape (b) has a complex structure but a small elastic coefficient. The leaf springs of shapes (c) and (d)
are simpler than that of shape (b), and the elasticity coefficient is more moderate than those of (a)
and (b). The outer rings of the four types of leaf springs are the same size (33 mm). In accordance with
the laboratory’s existing leaf spring manufacturing process technology, leaf spring of two thicknesses
(0.1 and 0.15 mm) are produced. To reduce the processing difficulty of the accelerometer, a 10 mm
diameter cylindrical structural steel block is selected as the seismic mass. Its mass is 6.25 g. Simulation
software ANSYS 14.0 (ANSYS Co., Ltd., Pittsburgh, PA, USA) is used to simulate the eight seismic
structures. The simulation results are shown in Table 1, in which the elastic coefficient k is calculated by:
mg
k= , (4)
∆x
where m is the mass of the seismic mass, g is the acceleration of gravity, and ∆x is the static deformation
of the seismic structure. Following the results of the analysis in Table 2, the two seismic structures
formed by the leaf spring of shape (a) have resonance frequencies higher than the design requirements;
however, their elasticity coefficient is large, which may reduce the sensitivity of the accelerometer.
The resonant frequencies of the two types of seismic structures that compose the leaf spring of shape (b)
are much lower than the design requirements, and the first three resonant frequencies have a small
difference, which can easily generate higher-order resonance and cannot be used in accelerometers.
The problem of the two types of seismic structures formed by the leaf spring of shape (d) is similar to
that of the structure of the leaf spring of shape (a). The most appropriate seismic structure consists
of the leaf spring with (c) and a thickness of 0.1 mm, in which the resonant frequency of the seismic
structure is 48 Hz, the static deformation is 107.9 µm, and the elastic coefficient is 0.56 mN/µm,
as shown in Figure 4. This seismic structure has a modest rigidity and can satisfy the requirements
of the seismic structure in this study. Thus, the leaf spring with shape (c) and 0.1 mm thickness is
selected for fabrication. The leaf spring is fabricated from beryllium bronze, and the outer and inner
diameters are specified as 33 and 16 mm, respectively. The seismic mass is fabricated from 4041 alloy
steel. The dimensions of the completed optical accelerometer are as follows: length, 70 mm; width,
50 mm; and height, 70 mm. Figure 5 presents a photograph of the completed optical accelerometer.
The laser diode and the QPD are fixed to the accelerometer frame by a fixing seat. Mirror 1 is attached
to the fine-tuning device, which is fixed to the accelerometer frame by screws. As shown in Figure 5,
in order to reduce optical system noise, the optical part of the proposed accelerometer is sealed in a
housing shell, which can avoid the disturbance from an external light source.
The model of the QPD we used in the accelerometer is SPOT-4D (OSI Optoelectronics Co.,
Ltd., Hawthorne, CA, USA), which has an active area of 1.3 mm × 1.3 mm, a responsivity of about
0.4 A W−1 corresponding with a wave length of 630 nm, a maximum dark current of 1 nA, and a noise
equivalent power (NEP) of 8.7 × 10−15 W (Hz1/2 )−1 . The four-channel photocurrent signal output
(µA) by the QPD is converted into a voltage signal (using a resistor of 33 kΩ) output by a low-power
Sensors 2018, 18, 2910 6 of 15
Sensors 2018, 18, x 6 of 15
Sensors 2018, 18, x 6 of 15
by the QPD is converted
quad-operational intoLM124
amplifier a voltage signal
(Texas (using a resistor
Instruments of 33
Co., Ltd., kΩ) output
Dallas, by a low-power
TX, USA). A low-passquad-
filter
by the QPD
operational is converted into a voltage signal (using a resistor of 33 kΩ) output by a low-power quad-
circuit with amplifier LM124 (Texas
a cut-off frequency of 40Instruments
Hz was usedCo.,toLtd., Dallas,
reduce the TX, USA). A from
interference low-pass
highfilter circuit
frequency
operational
with a cut-offamplifier LM124of 40(Texas Instruments Co., Ltd., Dallas, TX, USA). A low-pass filter circuit
electrical noisefrequency
and the power Hz was
supply. used to reduce the interference from high frequency electrical
with
noiseaand
cut-off
the frequency of 40 Hz was used to reduce the interference from high frequency electrical
power supply.
noise and the power supply. Table 2. Modal analysis and static analysis.
Table 2. Modal analysis and static analysis.
Table 2. Modal
Resonant analysis
Frequency and static analysis.
(Hz)
ThicknessLeafLeaf
Thickness Spring
Spring Resonant Frequency (Hz) Statics
Statics Elastic
Elastic Coefficient
Coefficient
(mm) Shape 1st Order 2nd
Resonant Order
Frequency 3rd Order Deformation (µm) (mN/µm)
Thickness
(mm) Leaf Spring
Shape 1st Order 2nd Order (Hz)
3rd Order Statics (μm) Elastic
Deformation Coefficient
(mN/μm)
(mm) a Shape
a 1st70.33
70.33Order 2nd Order
2156.2
2156.2 3rd Order
2158.9
2158.9 Deformation
50.527 (μm)
50.527 (mN/μm)
1.211.21
b ab 70.33
20.82
20.82 2156.2
24.01
24.01 2158.9
24.11
24.11 50.527
577.53
577.53 1.21
0.110.11
0.1 0.1
c bc 4820.82
48 1245.6
24.01
1245.6 1246.5
24.11
1246.5 81.61
577.53
81.61 0.560.56
0.11
0.1 d cd 76.02 2411.9 2414.2 43.03
48
76.02 1245.6
2411.9 1246.5
2414.2 81.61
43.03 1.421.42
0.56
a da 76.02
121.87
121.87 2411.9
135.85
135.85 2414.2
136.77
136.77 43.03
16.96
16.96 1.42
3.613.61
b ab 31.26
121.87
31.26 35.66
135.85
35.66 35.73
136.77
35.73 256.54
16.96
256.54 0.240.24
3.61
0.15 0.15
c bc 89.35
31.26
89.35 108.73
35.66
108.73 109.57
35.73
109.57 24.92
256.54
24.92 1.961.96
0.24
0.15 d c 135.64 3299.2 3306.5 13.53 4.52
d 89.35
135.64 108.73
3299.2 109.57
3306.5 24.92
13.53 1.96
4.52
d 135.64 3299.2 3306.5 13.53 4.52

Figure 3. Leaf spring shapes. (a) cross; (b) serpentine; (c) double circular; (d) V shape.
Figure 3. Leaf spring shapes. (a) cross; (b) serpentine; (c) double circular; (d) V shape.
Figure 3. Leaf spring shapes. (a) cross; (b) serpentine; (c) double circular; (d) V shape.

48 Hz 107.9 μm
48 Hz 107.9 μm

Modal Static
Modal Static
Figure 4. The modal and static structural analysis result of simulation software ANSYS (shape c).
Figure 4. The modal and static structural analysis result of simulation software ANSYS (shape c).
Figure 4. The modal and static structural analysis result of simulation software ANSYS (shape c).
Sensors 2018, 18, 2910 7 of 15
Sensors 2018, 18, x 7 of 15
Sensors 2018, 18, x 7 of 15
Floating plate
Floating plate
Leaf spring
Leaf spring
Seismic mass
Seismic mass

QPD
QPD
Fine-tuning
device
Fine-tuning
Fine-tuning
device device
Fine-tuning
Laser diode
device
Laser diode

Base
Base

(a) (b)
(a) (b)
Figure 5. Optical accelerometer (a) internal schematic, (b) photo.
Figure5.5.Optical
Figure Optical accelerometer (a)internal
accelerometer (a) internalschematic,
schematic,
(b)(b) photo.
photo.
4. Experimental Setup and Measurement Results
4. Experimental Setup
4. Experimental Setupand
andMeasurement
Measurement Results
Results
In general, the accelerometer is calibrated by the recording of its output under a particular
In In
general,
general, thetheaccelerometer
accelerometer is
is calibrated
calibrated bythe
by the recordingofofitsits output under a particular
excitation force and then a comparison of this output withrecording
that generated by output under
a reference a particular
accelerometer
excitation
excitation force
force andand thenaacomparison
then comparison of this this output
output with
with that
thatgenerated
generated bybya reference
a accelerometer
reference accelerometer
under the same excitation conditions [26]. This type of calibration is known as the comparative
under
under thethesamesameexcitation
excitationconditions
conditions [26].
[26]. This type
typeof(OptoNCDT
This(LDS) ofcalibration
calibration is is
known
known as as
thethe
comparative
comparative
method. In this study, a laser displacement sensor 1402-10, MICRO-EPSILON Co.,
method.
method. In In this
this study,
study, a
a laser
laser displacement
displacement sensor (LDS)
sensor (OptoNCDT
(LDS) (OptoNCDT 1402-10, MICRO-EPSILON
1402-10, MICRO-EPSILON Co.,
Ltd., Ortenburg, Germany) is applied as the reference sensor. The measurement range is 10 mm, the
Co.,Ltd., Ortenburg,
Ltd., Ortenburg, Germany) is applied as the reference sensor. The measurement range is 10 mm, themm,
resolution is 1 μm, Germany) is applied
and the bandwidth isas1.5the reference
kHz. sensor.
The displacement The values
measurement
from therange LDS iscan10 be
resolution
theconverted
resolution is 1 μm, and the bandwidth is 1.5 kHz. The displacement values from the LDS can be be
to is 1 µm, andby:
acceleration the bandwidth is 1.5 kHz. The displacement values from the LDS can
converted to acceleration by:
converted to acceleration by:
a  (2 f )222 D, (5)
aa 
=(2  ff)) D
(2π D,, (5) (5)
where D is the measured amplitude of vibration by LDS and f is the frequency of the sinusoidal
where
where D the
D is is the measuredamplitude
measured amplitude of of vibration
vibration by by LDS
LDSand andf fis isthe
thefrequency
frequency of the
of thesinusoidal
sinusoidal
excitation signals applied to the calibration vibration generator.
excitation
excitation signals
signals applied
applied toto the
the calibration
calibration vibration
vibration generator.
generator.
Figure 6 illustrates the experimental setup used for the comparison calibration test. In this
Figure
Figure 6 illustratesthe
6 illustrates theexperimental
experimental setup setup used for the comparison calibration test. In this
arrangement, a vibration generator (Modalshop used 2075E, forThe
theModalshop
comparison MTS calibration
Systems test. In this
Co., Ltd.,
arrangement,
arrangement, a vibration
a vibration generator
generator (Modalshop 2075E, The Modalshop MTS Systems Co., Ltd.,
Cincinnati, OH, USA) was used to (Modalshop
provide the 2075E,excitationTheacceleration.
ModalshopAMTS
Cincinnati, OH, USA) was used to provide the excitation acceleration. A high-precision wave
Systems Co.,
high-precision wave Ltd.,
Cincinnati,
generator OH, USA) was used to provide the excitation acceleration. A high-precision wave generator
generator (Keysight
(Keysight 33519B,
33519B, Keysight Technologies Co.,
Keysight Technologies Co., Ltd.,
Ltd., Santa
Santa Rosa,
Rosa,CA,CA,USA)
USA)was wasused
usedtoto
(Keysight
generate 33519B,
the Keysight
sinusoidal Technologies
excitation signal, Co., Ltd.,
which was Santa
then Rosa, CA,
amplified by USA)
a power was used to
amplifier generate the
(Modalshop
generate the sinusoidal excitation signal, which was then amplified by a power amplifier (Modalshop
sinusoidal
2100E21, excitation signal, which was then Co.,amplified by a power
OH, amplifier (Modalshop 2100E21,
2100E21, TheThe Modalshop
Modalshop MTS MTS Systems
Systems Co., Ltd., Cincinnati,
Ltd., Cincinnati, OH, USA) toto drive
USA) drive thevibration
the vibration
The Modalshop
generator. The MTS
output Systems
signals ofCo.,
the Ltd.,
proposed Cincinnati, OH,
accelerometer USA)
and LDS to
generator. The output signals of the proposed accelerometer and LDS were acquired to the computer drive
were the
acquired vibration
to the generator.
computer
The byoutput
by signals acquisition
aasynchronous
synchronous of the proposed
acquisition card accelerometer
card (Keysight
(Keysight U2542A,
U2542A, and LDS were
Keysight
Keysight acquiredCo.,
Technologies
Technologies to Ltd.,
Co., the
Ltd.,computer
SantaRosa,
Santa by a
Rosa,
CA, USA).
synchronous
CA, USA). acquisition card (Keysight U2542A, Keysight Technologies Co., Ltd., Santa Rosa, CA, USA).

Acquisition Card
Acquisition Card
(Keysight U2542A)
(Keysight U2542A)

PC
PC
vibration
Force vibration

Optical
Optical LDS
LDS
direction
direction

Accelerometer
Accelerometer Power
Power Wave
Wave
Amplifier
Amplifier Generator
Generator
Vibration
Vibration Generator (Modalshop
(Modalshop (Keysight
(Keysight
Force

(Modalshop
(Modalshop 2075E) 2100E21)
2100E21) 33519B)
33519B)

(a)
(a) (a)
(a)
Figure 6. Cont.
Sensors 2018, 18, x 8 of 15
Sensors 2018, 18, 2910 8 of 15
Sensors 2018, 18, x Optical 8 of 15
accelerometer
Optical
accelerometer
LDS

LDS
Left view

Left view

Vibration
Vibration
generator
generator
(b)
(b)
Figure 6. 6.
Figure Experimental
Experimentalsetup
setup for
for the
the designed opticalaccelerometer
designed optical accelerometer calibration.
calibration. (a) (a) Diagram,
Diagram, (b) (b)
Figure 6. Experimental setup for the designed optical accelerometer calibration. (a) Diagram, (b) photo.
photo. LDS: laser displacement sensor.
photo. LDS: laser displacement sensor.
LDS: laser displacement sensor.
4.1.4.1.
Resonance
Resonance Frequency
FrequencyMeasurement
Measurement
4.1. Resonance Frequency Measurement
TheThe resonancefrequency
resonance frequencyof ofthe
the optical
optical accelerometer
accelerometerwas wasobtained
obtained using thethe
using experimental set- set-
experimental
up The resonance
shown in frequency
Figure 6. The of
wave the optical
generator accelerometer
was used to was
output obtained
a swept using
sine the
waveform
up shown in Figure 6. The wave generator was used to output a swept sine waveform signal, which experimental
signal, set-up
which
shown
waswas in Figure
input
input to tothe6. The
thepower wave generator
poweramplifier, was
amplifier, amplified,used to
amplified, and output
and then a swept
thensupplied sine waveform
suppliedto tothethe
vibration
vibrationsignal, which
generator. The was
generator. The
input to the
outputting power
voltageamplifier,
from amplified,
the optical and then supplied
accelerometer was to the
recorded vibration
in the generator.
range
outputting voltage from the optical accelerometer was recorded in the range of 2–26 Hz with an of 2–26 The
Hz outputting
with an
interval of the
voltage 2 Hz, and inaccelerometer
the range of 26–57 Hz with anininterval of 1ofHz. TheHzresults are presentedofin2 Hz,
intervalfrom
of 2 Hz, optical
and in the range of 26–57 was recorded
Hz with an the range
interval of 1 2–26
Hz. The with an are
results interval
presented in
and Figure
in the 7.range of 26–57 Hz with an interval of 1 Hz. The results are presented in Figure 7.
Figure 7.

12
12
Voltage (V)

8
Voltage (V)

0 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56
Frequency (Hz)
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56
Figure 7. Frequency responseFrequency
test of the designed
(Hz) optical accelerator.

Figure 7. Frequency response test of the designed optical accelerator.

Figure 7. Frequency response test of the designed optical accelerator.
Sensors 2018, 18, 2910 9 of 15
Sensors 2018, 18, x 9 of 15

The first
The first resonance
resonance frequency
frequency of
of the
the optical
optical accelerometer
accelerometer isis approximately
approximately 4545 Hz,
Hz, which
which is
is in
in
agreement with
agreement with the
the simulation
simulation results
results obtained
obtained byby ANSYS
ANSYS 14.0
14.0 (Figure
(Figure 5).
5). The
The difference
difference is
is mainly
mainly
due to the machining and the overall assembly errors of the leaf spring and the accelerometer.
due to the machining and the overall assembly errors of the leaf spring and the accelerometer.

4.2. Acceleration
4.2. Acceleration Sensitivity
Sensitivity Measurement
Measurement at
at Low-Frequency
Low-Frequency
Theacceleration
The accelerationsensitivity
sensitivityofof
thethe optical
optical accelerometer
accelerometer waswas measured
measured at frequency
at frequency far lower
far lower than
than
the the resonance
resonance frequency,
frequency, whichwhich
can makecan sure
makethatsure
thethat the proposed
proposed accelerometer
accelerometer works steadily.
works steadily. A sine
A sine waveform
waveform with a constant
with a constant frequency frequency of 5 generated
of 5 Hz was Hz was generated
to drive theto drive the vibration
vibration generatorgenerator
with the
with the amplitude
amplitude changingchanging
from 1 Vppfrom 1 Vpp
to 10 Vpp.toThe
10 Vpp. The measured
measured displacementdisplacement
amplitude amplitude by LDS
by LDS changes
changes
from from
24 µm to 24
282μm
µm. to The
282 converted
μm. The converted output acceleration
output acceleration signal ofsignal
the LDS of the
andLDS and thevoltage
the output output
voltage
of of theaccelerometer
the optical optical accelerometer are presented
are presented in Figurein8.Figure 8. Thefrom
The results results from
data data
fitting fitting indicated
indicated that the
that the sensitivity
sensitivity of the
of the optical optical accelerometer
accelerometer is 1.739 V (m s−2 )−1V. The
is ·1.739 (m·sR-square
−2)−1. The is
R-square is 0.9998,
0.9998, thereby thereby
indicating
indicating
that that accelerometer
the optical the optical accelerometer
has excellent has excellent
linearity linearity characteristics.
characteristics.

0.5

0.4
Voltage (V)

0.3

0.2 y=1.7397x-0.0012
R2=0.9998
0.1

0.0
0.00 0.05 0.10 0.15 0.20 0.25 0.30
Acceleration (m s-2)
Figure 8.
Figure 8. The
The sensitivity
sensitivity of
of the
the optical
optical accelerometer.
accelerometer.

To test the stability of the accelerometer’s output, repetitive tests were performed. A homemade
To test the stability of the accelerometer’s output, repetitive tests were performed. A homemade
high-performance vibration generator [31] is used to provide an acceleration excitation signal. The
high-performance vibration generator [31] is used to provide an acceleration excitation signal.
vibration generator can achieve steady output displacement with frequency range of 0.6–50 Hz, an
The vibration generator can achieve steady output displacement with frequency range−2of 0.6–50 Hz,
analytical displacement resolution of 3.1 nm and an acceleration range of 0–1.93 m·s . In order to
an analytical displacement resolution of 3.1 nm and an acceleration range of 0–1.93 m·s−2 . In order
monitor the displacement of the homemade vibrator, an eddy current sensor [32,33] is used as a
to monitor the displacement of the homemade vibrator, an eddy current sensor [32,33] is used as
reference sensor, the output sensitivity of the sensor is 0.2312 V·μm−1, the − 1
measurement range is 50
a reference sensor, the output sensitivity of the sensor is 0.2312 V·µm , the measurement range
μm, the resolution is 0.72 nm, and the maximum nonlinearity error is less than 1%. We used a sine
is 50 µm, the resolution is 0.72 nm, and the maximum nonlinearity error is less than 1%. We used
waveform with a constant frequency of 8 Hz as the excitation signal and varied the input amplitude
a sine waveform with a constant frequency of 8 Hz as the excitation signal and varied the input
of the vibration generator within the range of 12–120 Vpp (corresponding amplitude ranging of 2.2–
amplitude of the vibration generator within the range of 12–120 Vpp (corresponding amplitude
19.7 μm). The reference sensor and optical accelerometer output signals were recorded. The
ranging of 2.2–19.7 µm). The reference sensor and optical accelerometer output signals were recorded.
experiment was repeated ten times. The output voltage of the optical accelerometer was converted
The experiment was repeated ten times. The output voltage of the optical accelerometer was converted
into the corresponding acceleration by using the sensitivity value obtained from Figure 8. As shown
into the corresponding acceleration by using the sensitivity value obtained from Figure 8. As shown
in Figure 9, the deviation between the measured acceleration and the reference acceleration served
in Figure 9, the deviation between the measured acceleration and the reference acceleration served
as the ordinate, and the excitation acceleration served as the abscissa. The standard deviation of ten
as the ordinate, and the excitation acceleration served as the abscissa. The standard deviation of ten
measurements ranges of 1.4 × 10−4–7.9 × 10−4 m·s−2.
measurements ranges of 1.4 × 10−4 –7.9 × 10−4 m·s−2 .
Sensors 2018, 18, 2910 10 of 15
Sensors 2018, 18, x 10 of 15
Sensors 2018, 18, x 10 of 15
0.004
0.004

Standard deviation
Standard deviation
0.002
0.002

Deviation s-2s)-2)
(m(m
0.000
Deviation
0.000

-0.002
-0.002

-0.004
-0.0040.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07
0.00 0.03 0.01 0.02
0.04 -2 0.05 0.06 0.07
Acceleration (m s )
Acceleration (m s-2)
Figure 9. The
The stability of
of the optical
optical accelerometer. (The points with same shape
shape and
and color
color are the
the
Figure 9. The stability
Figure 9. stability of the
the optical accelerometer. (The points
accelerometer. (The points with same
with same shape and color are
are the
results
results of the same measurement).
results of
of the
the same
same measurement).
measurement).
4.3. Frequency Response Range, Acceleration Detection Range and Phase-Frequency Response
4.3. Frequency Response
4.3. Frequency Response Range,
Range, Acceleration
Acceleration Detection
Detection Range
Range and
and Phase-Frequency
Phase-Frequency Response
Response
The frequency response of the proposed optical accelerometer was tested using the setup in
The
The frequency
frequency response
response ofof the
the proposed
proposed optical
optical accelerometer
accelerometer waswas tested
tested using
using the
the setup
setup in
in
Figure 6. The wave generator was used to output sine waveform signals with frequencies in the range
Figure 6. The wave generator was used to output sine waveform signals with frequencies
Figure 6. The wave generator was used to output sine waveform signals with frequencies in the range in the
of 0.4–1 Hz with an interval of 0.1 Hz, and in the range of 1–12 Hz with an interval of 1 Hz. Then, the
range
of 0.4–1ofHz0.4–1
withHzanwith an interval
interval of and
of 0.1 Hz, 0.1 Hz, and
in the in the
range of range
1–12 Hz of with
1–12 an
Hzinterval
with anofinterval of 1 Hz.
1 Hz. Then, the
outputting voltage from the optical accelerometer and the LDS was recorded at the same time. And
Then, the outputting voltage from the optical accelerometer and the LDS was recorded
outputting voltage from the optical accelerometer and the LDS was recorded at the same time. And at the same
the converted output acceleration signal of the LDS and the output voltage of the optical
time. And the converted
the converted output acceleration
output acceleration signal of signal
the ofLDS
the LDS
and and
the the output
output voltageofofthe
voltage the optical
optical
accelerometer are presented in Figure 10. The sensitivity of the accelerometer in Figure 10 measured
accelerometer
accelerometer are arepresented
presentedin inFigure
Figure10.
10.The
Thesensitivity
sensitivityofofthe
theaccelerometer
accelerometer inin
Figure 1010
Figure measured
measured at
at different frequencies is consistent with the sensitivity measured at different amplitudes in Figure
different frequencies is consistent with the sensitivity measured at different amplitudes
at different frequencies is consistent with the sensitivity measured at different amplitudes in Figure in Figure 8.
8. That is to say, the operation frequency range is 0.4–12 Hz.
That is to
8. That say,
is to thethe
say, operation frequency
operation frequencyrange is 0.4–12
range Hz.Hz.
is 0.4–12
1.2
1.2

1.0
1.0
12 Hz
0.8 12 Hz
0.8
(V)
(V)

0.6
Voltage

0.6
Voltage

0.4
0.4 y=1.7417x+0.0057
y=1.7417x+0.0057
0.4 Hz R2=0.9995
0.2
0.4 Hz R2=0.9995
0.2

0.0
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
0.0 0.1 0.2 0.3
Acceleration 0.4 s-2) 0.5
(m 0.6 0.7
-2
Acceleration (m s )
Figure 10. Frequency response range of the optical accelerometer.
Figure
Figure 10.
10. Frequency
Frequency response
response range
range of
of the
the optical
optical accelerometer.
accelerometer.
The acceleration detection resolution and range experiments used a sine waveform with a
The acceleration detection resolution and range experiments used a sine waveform with a
constant frequency ofdetection
The acceleration 1 Hz as the excitation
resolution andsignal, and the amplitude
range experiments of the
used a sine excitation
waveform wasa adjusted
with constant
constant frequency of 1 Hz as the excitation signal, and the amplitude of the excitation was adjusted
from smallofto1large
frequency Hz asuntil the accelerometer
the excitation produced
signal, and a clearly
the amplitude ofsinusoidal waveform.
the excitation Then, from
was adjusted the output
small
from small to large until the accelerometer produced a clearly sinusoidal waveform. Then, the output
voltage of the optical accelerometer was converted into the corresponding acceleration
to large until the accelerometer produced a clearly sinusoidal waveform. Then, the output voltage by dividing
voltage of the optical accelerometer was converted into the corresponding acceleration by dividing
by the sensitivity. The experimental results are shown in Figure 11. It can be seen that the resolution
by the sensitivity. The experimental results are shown in Figure 11. It can be seen that the resolution
Sensors 2018, 18, 2910 11 of 15

Sensors 2018, 18, x 11 of 15

of the optical accelerometer was converted into the corresponding acceleration by dividing by the
sensitivity. The experimental
of the accelerometer is 0.003 m·results are shown in The
s−2 approximately. Figure 11. It can
maximum be seen that
acceleration thethe
that resolution of the
accelerometer
accelerometer is 0.003 − 2
·s approximately.
can measure amax can bemcalculated by: The maximum acceleration that the accelerometer can
measure amax can be calculated by:
Vmax
amax= Vmax
amax , (6)
s , (6)

where ss is
where is the
the sensitivity
sensitivityof of the
the accelerometer
accelerometerand andVVmax is the
max is the maximum
maximum voltage
voltage that
that the
the accelerometer
accelerometer
can output. The maximum voltage that the accelerometer can output
can output. The maximum voltage that the accelerometer can output is 12.7 V, and its sensitivity is 12.7 V, and its sensitivityis
− 2 − 1
is 1.74
1.74 V (m
V (m· s−2)·−1
s . Therefore,
) . Therefore, the maximum
the maximum acceleration
acceleration that the accelerometer
that the accelerometer can measure canismeasure is,
7.29 m·s−2
7.29
and m the·s− 2 , and the measurement range of the accelerometer is 0.003–7.29
measurement range of the accelerometer is 0.003–7.29 m·s−2. m · s−2 .
Experiments
Experiments were were also
also conducted
conducted to to investigate
investigate the the phase-frequency
phase-frequency performance
performance of of the
the
accelerometer
accelerometer with with frequencies
frequenciesin inthe
therange
rangeofof0.4–1
0.4–1Hz Hzwith
withananinterval
intervalofof
0.10.1 Hz,
Hz, andandin in
thethe range
range of
of 1–
1–12 Hz with an interval of 1 Hz. The LDS was configured to output analog
12 Hz with an interval of 1 Hz. The LDS was configured to output analog signal of 4–20 mA, which was signal of 4–20 mA, which
was converted
converted to voltage
to voltage signal signal and collected
and collected togethertogether with the
with the output output
signal signal
of the of the accelerometer
accelerometer are collected
are collected into a computer for further processing. Origin 2017 software
into a computer for further processing. Origin 2017 software (OriginLab Co., Ltd., Northampton, (OriginLab Co., Ltd.,
MA,
Northampton, MA, USA) was used to fit the two signals, and then the
USA) was used to fit the two signals, and then the phase difference between the two fitted waveforms phase difference between
the
wastwo fitted waveforms
calculated, which is shown wasin calculated,
Figure 12. which is shown in
The experimental Figure
results 12. that
show Thetheexperimental
accelerometer results
has a
show
nonlinear phase error of less than 0.011 rad, which can cause the corresponding relative errorthe
that the accelerometer has a nonlinear phase error of less than 0.011 rad, which can cause of
corresponding relative error
acceleration measurement of of acceleration
about measurement
0.006% (calculated of about
by: 1−cos 0.006%
0.011). This(calculated
is relativelyby: 1−cos
quite 0.011).
small and
This
can beis neglected.
relatively quite small and can be neglected.

0.004
Original signal
Sine fitting
0.002
Acceleration (m s-2)

0.000

-0.002

-0.004
-3 -2 -1 0 1 2 3
Time (s)
Figure 11. The resolution of the accelerometer.
Figure 11. The resolution of the accelerometer.
Sensors 2018, 18, 2910 12 of 15
Sensors 2018, 18, x 12 of 15
Sensors 2018, 18, x 12 of 15

0.000
0.000

-0.002
-0.002

error (rad)
(rad) -0.004
-0.004
Phaseerror

-0.006
Phase

-0.006

-0.008
-0.008

-0.010
-0.010
1 10
1 10
Frequency (Hz)
Frequency (Hz)
Figure 12.
Figure 12. The
The phase
phase error
error of
of the
the optical
optical accelerometer.
accelerometer.
Figure 12. The phase error of the optical accelerometer.
4.4. Noise Equivalent Acceleration Measurement.
4.4.
4.4. Noise
Noise Equivalent
Equivalent Acceleration
Acceleration Measurement.
Measurement.
To measure the noise equivalent acceleration (NEA) of the designed optical accelerometer, the
To
To measure the noise
measure the noiseequivalent
equivalentacceleration
acceleration(NEA)
(NEA) of of
thethe designed
designed optical
optical accelerometer,
accelerometer, the
output voltage noise was recorded by a DAQ card when the vibration generator was turned off. The
the output voltage noise was recorded by a DAQ card when the vibration generator
output voltage noise was recorded by a DAQ card when the vibration generator was turned off. The was turned off.
spectral density of the NEA was calculated using the sensitivity value shown Figure 8. Figure 13 plots
The spectral density of the NEA was calculated using the sensitivity value shown Figure
spectral density of the NEA was calculated using the sensitivity value shown Figure 8. Figure 13 plots 8. Figure 13
the NEA spectral density in frequency domain. The electronic circuit of the current optical
plots
the NEAthe NEA spectral
spectral density
density in frequency
in frequency domain.
domain. TheThe electroniccircuit
electronic circuitofofthe
the current
current optical
optical
accelerometer generated an electrical noise of less than 160 (μm·s−2) −(Hz −1/2) over the frequency
2 ) (Hz range
−1/2 ) over the frequency
accelerometer generated an electrical noise of less than 160 (µm ·
−2 s
accelerometer generated an electrical noise of less than 160 (μm·s ) (Hz ) over the frequency range
−1/2
of 0.5–50 Hz. Considering that the NEA in this study is lower than the ones from related literature
range of 0.5–50
of 0.5–50 Hz. Considering
Hz. Considering that the that
NEAthe in NEA in this
this study is study is lower
lower than than from
the ones the ones from
related related
literature
[12–16,19,21], the NEA in this study has been promoted visibly.
literature [12–16,19,21],
[12–16,19,21], the NEA in thethis
NEA in this
study hasstudy has been promoted
been promoted visibly. visibly.
160
160
140
140
120
density

120
density
)] )]
1/21/2

100
)/(Hz

100
s-2s)/(Hz
spectral

80
spectral
-2

80
[(μm

60
[(μm
NEA

60
NEA

40
40
20
20
0
0 0 10 20 30 40 50
0 10 20
Frequency 30
(Hz) 40 50
Frequency (Hz)
Figure 13. Noise equivalent acceleration (NEA) spectral density.
Figure 13. Noise equivalent acceleration (NEA) spectral density.
Figure 13. Noise equivalent acceleration (NEA) spectral density.
4.5. Signal Drift of the Optical Accelerometer
4.5. Signal Drift of the Optical Accelerometer
During the operation of the accelerometer, its accuracy is usually affected by a Gaussian laser
During the operation of the accelerometer, its accuracy is usually affected by a Gaussian laser
beam, angular drift and other influencing factors. To determine the effect of these factors on the
beam, angular drift and other influencing factors. To determine the effect of these factors on the
accelerometer, the accelerometer’s drift was tested for more than 6 h in a resting state. The
accelerometer, the accelerometer’s drift was tested for more than 6 h in a resting state. The
Sensors 2018, 18, 2910 13 of 15

4.5. Signal Drift of the Optical Accelerometer

During the operation of the accelerometer, its accuracy is usually affected by a Gaussian laser beam,
angular drift and other influencing factors. To determine the effect of these factors on the accelerometer,
Sensors 2018, 18, x 13 of 15
the accelerometer’s drift was tested for more than 6 h in a resting state. The experimental results are
shown in Figure
experimental 14. From
results Figure in
are shown 14,Figure
we can14. seeFrom
that the 6-h drift
Figure of the
14, we canaccelerometer is about
see that the 6-h drift 37
of mV,
the
whose equivalent acceleration is 0.021 m · s −2 .
accelerometer is about 37 mV, whose equivalent acceleration is 0.021 m·s . −2

0.045

0.040

0.035
Voltage (V)

0.030

0.025 0.021 m s−2

0.020

0.015

0.010

0.005

0 1 2 3 4 5 6 7
Time (h)

Figure
Figure 14.
14. The
The drift
drift of
of the
the optical
optical accelerometer.
accelerometer.

5. Discussions and Conclusions

5. Discussions and Conclusions
This work presents an optical accelerometer that consists of a QPD-based sensor, a leaf spring,
This work presents an optical accelerometer that consists of a QPD-based sensor, a leaf spring,
and a seismic mass for low-frequency, low-amplitude vibration. The principle and structure of the
and a seismic mass for low-frequency, low-amplitude vibration. The principle and structure of the
optical accelerometer are introduced. The elastic structure is analyzed, designed, simulated, and
optical accelerometer are introduced. The elastic structure is analyzed, designed, simulated, and
fabricated. The experimental results show that the accelerometer has a sensitivity of 1.74 V (m· s−2)−1,
fabricated. The experimental results show that the accelerometer has a sensitivity of 1.74 V (m·s−2 )−1 ,
the measurement range of the accelerometer is 0.003–7.29 m·s−2, the operating frequencies range of
−2
the measurement range of the accelerometer is 0.003–7.29 m·s , the operating frequencies range of
0.4–12 Hz, and electrical noise of less than 160 (μm· s2−2) (Hz−−1/2
− ) is found over the frequency range of
0.4–12 Hz, and electrical noise of less than 160 (µm·s ) (Hz 1/2 ) is found over the frequency range of
0.5–50 Hz.
0.5–50 Hz.
The accelerometer is useful in the detection of low-frequency, low-amplitude vibration and can
The accelerometer is useful in the detection of low-frequency, low-amplitude vibration and can be
be used as a sensor in an active vibration isolation system. Furthermore, different elastic structures
used as a sensor in an active vibration isolation system. Furthermore, different elastic structures can be
can be designed (such as by changing the shape or thickness of the leaf spring and increasing or
designed (such as by changing the shape or thickness of the leaf spring and increasing or decreasing
decreasing the mass of the seismic mass), and the amplification of the signal processing circuit can be
the mass of the seismic mass), and the amplification of the signal processing circuit can be modified to
modified to widen the frequency response range of the accelerometer. The proposed system can be
widen the frequency response range of the accelerometer. The proposed system can be applied to the
applied to the acceleration measurement of a broad range of vibrational fields, such as the monitoring
acceleration measurement of a broad range of vibrational fields, such as the monitoring and diagnosis
and diagnosis of machine tools and structural health monitoring.
of machine tools and structural health monitoring.
Author Contributions:
Author Contributions: R.-J.L.
R.-J.L. conceived
conceived and
and designed
designed the
the accelerometer.
accelerometer. Y.-J.L. performed the
Y.-J.L. performed the experiments
experiments and
and
wrote the manuscript with L.-S.Z. Z.-X.C. analyzed and interpreted the data. K.-C.F. edited the manuscript.
wrote the manuscript with L.-S.Z. Z.-X.C. analyzed and interpreted the data. K.-C.F. edited the manuscript.
Funding: This
Thisresearch
researchwas
wasfunded by the
funded by National Natural
the National ScienceScience
Natural Foundation of China,ofgrant
Foundation number
China, grant[51675157,
number
51705123], and
[51675157, the Anhui
51705123], and Provincial
the Anhui Natural Science
Provincial Foundation,
Natural grant number
Science Foundation, [1708085QE105].
grant number [1708085QE105].
Acknowledgments:
Acknowledgments: We We would
would like
like to
to thank
thank Professor
Professor Yong-bin
Yong-bin Liu
Liu and
and Associate
Associate Professor Bin Ju
Professor Bin Ju of
of Anhui
Anhui
University (China) for the support of some experimental equipment.
University (China) for the support of some experimental equipment.
Conflicts of Interest: The authors declare no conflict of interest.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Bard, P.Y. Microtremor measurements: A tool for site effect estimation. Eff. Surf. Geol. Seism. Motion 1999,
3, 1251–1279.
2. Seht, I.V.; Wohlenberg, J. Microtremor measurements used to map thickness of soft sediments. Bull. Seismol.
Soc. Am. 1999, 89, 250–259.
Sensors 2018, 18, 2910 14 of 15

References
1. Bard, P.Y. Microtremor measurements: A tool for site effect estimation. Eff. Surf. Geol. Seism. Motion 1999, 3,
1251–1279.
2. Seht, I.V.; Wohlenberg, J. Microtremor measurements used to map thickness of soft sediments. Bull. Seismol.
Soc. Am. 1999, 89, 250–259.
3. Tian, B.; Liu, H.; Yang, N.; Zhao, Y.; Jiang, Z. Design of a Piezoelectric Accelerometer with High Sensitivity
and Low Transverse Effect. Sensors 2016, 16, 1587. [CrossRef] [PubMed]
4. Zou, Q.; Tan, W.; Kim, E.S.; Loeb, G.E. Highly symmetric tri-axis piezoelectric bimorph accelerometer.
In Proceedings of the IEEE International Conference on MICRO Electro Mechanical Systems, Maastricht,
The Netherlands, 25–29 January 2004; pp. 197–200.
5. Hewa-Kasakarage, N.N.; Kim, D.; Kuntzman, M.L.; Hall, N.A. Micromachined Piezoelectric Accelerometers
via Epitaxial Silicon Cantilevers and Bulk Silicon Proof Masses. J. Microelectromech. Syst. 2013, 22, 1438–1446.
[CrossRef]
6. Tims, A.C.; Davidson, R.L.; Timme, R.W. High sensitivity piezoelectric accelerometer. Rev. Sci. Instrum. 1975,
46, 554–558. [CrossRef]
7. Kamentser, B. New Generation of Strain Gauge Accelerometers. SAE Tech. Paper 1994. [CrossRef]
8. Santana, J.; Hoven, R.V.D.; van Liempd, C.; Colin, M.; Saillen, N.; Zonta, D.; Trapani, D.; Torfs, T.; Van Hoof, C.
A 3-axis accelerometer and strain sensor system for building integrity monitoring. Sens. Actuators A Phys.
2012, 188, 141–147. [CrossRef]
9. Liu, F.; Dai, Y.; Karanja, J.M.; Yang, M. A Low Frequency FBG Accelerometer with Symmetrical Bended
Spring Plates. Sensors 2017, 17, 206. [CrossRef] [PubMed]
10. Gao, X.; Wang, Y.; Yuan, B.; Yuan, Y.; Dai, Y.; Xu, G. Low-cost vibration sensor based on dual fiber Bragg
gratings and light intensity measurement. Appl. Opt. 2013, 52, 6782–6787. [CrossRef] [PubMed]
11. Zhang, J.H.; Qiao, X.G.; Hu, M.L; Feng, Z.; Gao, H.; Yang, Y.; Zhou, R. Flextensional fiber Bragg grating-based
accelerometer for low frequency vibration measurement. Chin. Opt. Lett. 2011, 9, 25–28.
12. Zhang, X.; Rong, Q.; Sun, H.; Yang, S.; Yuan, L.; Hu, M. Low-frequency fiber Bragg grating accelerometer
based on a double-semicircle cantilever. Opt. Fiber Technol. 2014, 20, 190–193. [CrossRef]
13. Zeng, Y.J.; Wang, J.; Yang, H.Y. Fiber Bragg Grating Accelerometer Based on L-Shaped Rigid Beam and
Elastic Diaphragm for Low-Frequency Vibration Measurement. Chin. Opt. Lett. 2015, 35, 90–98. [CrossRef]
14. Li, K.; Chan, T.H.; Yau, M.H.; Nguyen, T.; Thambiratnam, D.P.; Tam, H.Y. Very sensitive fiber Bragg grating
accelerometer using transverse forces with an easy over-range protection and low cross axial sensitivity.
Appl. Opt. 2013, 52, 6401–6410. [CrossRef] [PubMed]
15. Sabato, A.; Feng, M.Q. Feasibility of Frequency-Modulated Wireless Transmission for a Multi-Purpose
MEMS-Based Accelerometer. Sensors 2014, 14, 16563–16585. [CrossRef] [PubMed]
16. Sabato, A.; Feng, M.Q.; Fukuda, Y.; Carní, D.L.; Fortino, G. A Novel Wireless Accelerometer Board for
Measuring Low-Frequency and Low-Amplitude Structural Vibration. IEEE Sens. J. 2016, 16, 2942–2949.
[CrossRef]
17. Zhu, L.; Fu, Y.; Chow, R.; Spencer, B.F.; Park, J.W.; Mechitov, K. Development of a High-Sensitivity Wireless
Accelerometer for Structural Health Monitoring. Sensors 2018, 18, 262. [CrossRef] [PubMed]
18. Swartz, R.A.; Lynch, J.P.; Zerbst, S.; Sweetman, B.; Rolfes, R. Structural monitoring of wind turbines using
wireless sensor networks. Smart Struct. Syst. 2010, 6, 183–196. [CrossRef]
19. Cho, S.; Yun, C.-B.; Lynch, J.P.; Zimmerman, A.T.; Spencer, B.F., Jr.; Nagayama, T. Smart wireless sensor
technology for structural health monitoring of civil structures. Int. J. Steel Struct. 2008, 8, 267–275.
20. Whelan, M.J.; Janoyan, K.D. Design of a robust, high-rate wireless sensor network for static and dynamic
structural monitoring. J. Intell. Mater. Syst. Struct. 2009, 20, 849–863. [CrossRef]
21. Meyer, J.; Bischoff, R.; Feltrin, G. Microelectromechanical systems (MEMS). In Encyclopedia of Structural
Health Monitoring; Boller, C., Chang, F.K., Fujino, Y., Eds.; Wiley: New York, NY, USA, 2009.
22. Rice, J.A.; Spencer, B.F., Jr. Structural health monitoring sensor development for the Imote2 platform.
Proc. SPIE 2008, 6932, 693234–693244.
23. Kohler, M.D.; Hao, S.; Mishra, N.; Govindan, R.; Nigbor, R. ShakeNet: A Portable Wireless Sensor Network
for Instrumenting Large Civil Structures; Technical Report 2015-1134; United States Geological Survey:
Los Angeles, CA, USA, 2015; p. 31.
Sensors 2018, 18, 2910 15 of 15

24. Kohler, M.D.; Heaton, T.H.; Cheng, M.-H. The Community Seismic Network and Quake-Catcher Network:
Enabling Structural Health Monitoring through Instrumentation by Community Participants. In Proceedings
of the SPIE International Society for Optical Engineering, Bellingham, WA, USA, 19 April 2013; Volume 8692,
p. 86923.
25. Chu, C.L.; Lin, C.H. Development of an optical accelerometer with a DVD pick-up head. Meas. Sci. Technol.
2005, 16, 2498. [CrossRef]
26. Liu, Y.C.; Fan, K.C.; Chu, C.L.; Werner, C.A.; Jäger, G. Development of an optical accelerometer for
low-frequency vibration using the voice coil on a DVD pickup head. Meas. Sci. Technol. 2008, 19, 817–822.
[CrossRef]
27. Chu, C.L.; Lin, C.H.; Fan, K.C. Two-dimensional optical accelerometer based on commercial DVD pick-up
head. Meas. Sci. Technol. 2007, 18, 265. [CrossRef]
28. Zheng, D.; Liu, Y.; Guo, Z.; Zhao, X.; Fan, S. Theory and experiment research for ultra-low frequency maglev
vibration sensor. Rev. Sci. Instrum. 2015, 86, 105001. [CrossRef] [PubMed]
29. Jiang, D.; Zhang, W.; Li, F. All-metal optical fiber accelerometer with low transverse sensitivity for seismic
monitoring. IEEE Sens. J. 2013, 13, 4556–4560. [CrossRef]
30. Lin, Q.; Chen, L.; Li, S.; Wu, X. A high-resolution fiber optic accelerometer based on intracavity
phase-generated carrier (PGC) modulation. Meas. Sci. Technol. 2011, 22, 15303–15308. [CrossRef]
31. Li, R.J.; Lei, Y.J.; Zhang, L.S.; Chang, Z.X.; Fan, K.C.; Cheng, Z.Y.; Hu, P.H. High-precision and low-cost
vibration generator for low-frequency calibration system. Meas. Sci. Technol. 2018, 29, 034008. [CrossRef]
32. Wang, H.; Liu, Y.; Li, W.; Feng, Z. Design of ultrastable and high resolution eddy-current displacement sensor
system. In Proceedings of the IECON 2014—40th Annual Conference of the IEEE Industrial Electronics
Society, Dallas, TX, USA, 29 October–1 November 2014; Volume 2015, pp. 2333–2339.
33. Wang, H.; Feng, Z. Ultrastable and highly sensitive eddy current displacement sensor using self-temperature
compensation. Sens. Actuators A Phys. 2013, 203, 362–368. [CrossRef]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).