Accelerometer Theory & Design

Download as pdf or txt
Download as pdf or txt
You are on page 1of 46

11

Chapter 2
Accelerometer Theory & Design

2.1

Introduction

An accelerometer is a sensor that measures the physical acceleration


experienced by an object due to inertial forces or due to mechanical
excitation. In aerospace applications accelerometers are used along with
gyroscopes for navigation guidance and flight control. Conceptually, an
accelerometer behaves as a damped mass on a spring. When the
accelerometer experiences acceleration, the mass is displaced and the
displacement is then measured to give the acceleration [17].
In

these

devices,

piezoelectric,

piezoresistive

and

capacitive

techniques are commonly used to convert the mechanical motion into an


electrical signal. Piezoelectric accelerometers rely on piezoceramics (e.g.
lead zirconate titanate) or single crystals (e.g. quartz, tourmaline). They
are unmatched in terms of their upper frequency range, low packaged
weight and high temperature range. Piezoresistive accelerometers are
preferred

in

high

shock

applications.

Capacitive

accelerometers

performance is superior in low frequency range and they can be operated


in servo mode to achieve high stability and linearity.

12

Modern

accelerometers

are

often

small

micro

electro-

mechanical systems (MEMS), consisting of little more than a


cantilever beam with a proof-mass (also known as seismic-mass)
realized in single crystal silicon using surface micromachining or
bulk micromachining processes.

2.2 Working principle of accelerometer

Casing

Fig. 2.1 Schematic of an accelerometer


The principle of working of an accelerometer can be explained by a
simple mass (m) attached to a spring of stiffness (k) that in turn is
attached to a casing, as illustrated in fig 2.1. The mass used in
accelerometers is often called the seismic-mass or proof-mass. In most
cases the system also includes a dashpot to provide a desirable damping
effect.

13

The dashpot with damping coefficient (c) is normally attached to


the mass in parallel with the spring. When the spring mass system is
subjected to linear acceleration, a force equal to mass times acceleration
acts on the proof-mass, causing it to deflect. This deflection is sensed by
a suitable means and converted into an equivalent electrical signal. Some
form of damping is required, otherwise the system would not stabilize
quickly under applied acceleration.
To derive the motion equation of the system Newtons second law is
used, where all real forces acting on the proof-mass are equal to the
inertia force on the proof-mass. Accordingly a dynamic problem can be
treated as a problem of static equilibrium and the equation of motion can
be obtained by direct formulation of the equations of equilibrium. This
damped mass-spring system with applied force constitutes a classical
second order mechanical system.
From the stationary observers point of view, the sum of all forces in the z
direction is,

=
+ + =

14

+ + =

(2.1)

Where
m = mass of the proof-mass
x = relative movement of the proof-mass with respect to frame
c = damping coefficient
k = spring constant
F = force applied
The equation of motion is a second order linear differential equation with
constant coefficients. The general solution x (t) is the sum of the
complementary function XC (t) and the particular integral Xp (t) [18].
(2.2)

= +
The complementary function satisfies the homogeneous equation

+ + = 0

(2.3)

The solution to is

(2.4)

Substituting (2.4) in (2.3)


(ms2 + cs + k) C e

st

=0

15

As cannot be zero for all values of t, then ( 2 + + ) = 0 called as


the auxiliary or characteristic equation of the system. The solution to
this equation for values of S is
1

1,2 = 2 ( 2 4 )

(2.5)

From the above equation 2.5, the following useful formulae are derived

n =

k
m

(2.6)

c/m = 2 n

(2.7)

= c/2

(2.8)

km

Where
n = undamped resonance frequency
k = spring constant
m = mass of proof-mass
c = damping coefficient

= damping factor
Steady state performance
In the steady state condition, that is, with excitation acceleration
amplitude a and frequency , the amplitude of the response is constant
and is a function of excitation amplitude and frequency . Thus for
static response =0, the deflection amplitude

16

X = 0 = F/k.
X = ma / k

(2.9)

Here the sensitivity S of an accelerometer is defined by,


S = X / a = m/k

(2.10)

Dynamic performance
For the dynamic performance it is easier to consider the Laplace
transform of eqn (2.1)

x( s )

a( s)

1
c
k
s2 s
m
m

(2.11)

It can be seen by comparing eqn (2.6) and (2.10) that the bandwidth of
an accelerometer sensing element has to be traded off with its sensitivity
since S 1/n2 (this trade off can be partly overcome by applying
feedback , i.e closed loop scheme).
The sensor response is determined by
system.

damping present in the

A damping factor ( ) between 0.6 to 1.2 results in high

response time, fast settling time, good bandwidth and linearity.

17

2.3 Specifications of the accelerometer


The accelerometer that is to be designed shall have a measuring
range of 30 g with a resolution of 50 milli g i.e a dynamic range of
600. The total non-linearity from the sensor element, electronics and
from other sources shall not be more than 1% of full scale output (FSO).
The sensor shall have a bandwidth ( 3dB) of >100 Hz and the cross-axis
sensitivity shall be limited to a maximum of 1% of FSO. The sensors bias
stability

and hysteresis values are specified as 0.15% of FSO each.

Finally the sensor shall have a response time of less than 1msec and it
shall perform over a temperature range of -20 to 80C.

2.4 Configuration of accelerometer


Various aspects are taken in to consideration before finalizing the
configuration of the accelerometer. Special attention is paid to the
available fabrication processes, signal conditioning electronic circuit,
material selection, electrical routing and packaging. The configuration of
the accelerometer is as shown in fig 2.2.
(i)

The sensor is configured to have a three wafer (glass-silicon-glass)


configuration.

(ii)

Differential capacitance transduction method is selected as it offers


the advantages of low temperature sensitivity, higher transduction

18

efficiency and the method can be readily adopted for closed loop
operation.
(iii)

Glass wafers are used for the top and bottom plates and a thin
film of aluminum material is deposited on the inner side of glass
wafers using E-beam metal evaporation process. The central wafer
consists of the active proof-mass which moves as a function of the
applied acceleration thereby causing change in capacitance.

(iv)

The proof-mass is supported on all four sides by L-shaped beams.


The proof-mass exhibits piston like movement and remain parallel
to electrodes at all accelerations. Also any geometrical change in
the beam length due to temperature variation, limits the proofmass to in-plane rotation only and it does not experience any out
of plane bending. This configuration reduces the overall sensor
chip size thereby improving the per wafer yield and also reduces
the

non-linearity

associated

with

cantilever

type

support

structures.
(v)

The mechanical support for the proof-mass is provided at the


central plane of the proof-mass. Positioning of the beams at
central plane of proof-mass will reduce the cross-axis sensitivity of
the sensor.

(vi)

The proof-mass to electrode gap is selected as 22 microns, this


eliminates the requirement of

complicated device level vacuum

19

sealing and also need for perforations on the proof-mass thus


simplifying the process.
(vii)

Bulk micro-machining process using KoH is considered for


realizing the micro structures.

(viii) Modular concept is used for realizing the final device. MEMS chip
and the signal conditioning electronics are realized separately and
packaged on a signal platform.

Top glass wafer

Top electrode

Si wafer

Support beams
Proof -mass

Bottom Glass wafer

Bottom Electrode

Fig 2.2 Accelerometer configuration

2.5

Material selection
Single crystal silicon (100) material is selected for accelerometer

structure. Silicon is almost an ideal structural material, it has about the


same youngs modulus as steel but is as light as aluminum. The melting
point of silicon is 14000 C and the thermal expansion coefficient is much

20

less than steel which makes it dimensionally stable even at high


temperatures. Silicon wafers are extremely flat, can accept coatings &
additional thin-film layers for building microstructures. Silicon exhibits
no mechanical hysteresis and precise geometrical features can be
realized using standard photolithography and etching techniques.
Electrically conductive silicon with resistivity 0.1 -cm is selected for the
proof-mass. Similarly Pyrex 7740 glass is chosen for top and bottom
wafers to reduce stray capacitance and to provide required sealing. The
glass wafers are bonded to silicon wafer using anodic bonding process.
Electrodes and electrical contact pads are realized by depositing submicron thickness Aluminum coating, using E-beam evaporation process.
The material properties of silicon and Pyrex glass are shown in Table 2.1
Material Property

Pyrex
glass
0.5-0.7

E (Youngs modulus)1011N/m2

1.69

400

0.28

0.17

(thermal expansion coefficient)10-6 mt/mt C

2.5

0.5

(density) g/cm3

2.3

2.225

y (yield strength) 109 N/m2

(Poissons ratio)

Silicon

Table 2.1: Material properties of silicon and Pyrex glass.

21

2.6 Analytical design


The following assumptions are made to begin the design work
(i)

The proof-mass size and spring stiffness are selected in such a


way that there shall be a capacitance change of around 1fF for 50
milli g (minimum resolution). This limitation comes from the
capacitance signal that can be handled comfortably by the
electronics scheme.

(ii)

For calculating the proof-mass to electrode gap and damping


aspects, the micro structure is considered as working under
ambient pressure conditions. This is due to fabrication facility
limitation in chip level sealing under vacuum.

The structural parameters of the proof-mass are optimized to


achieve the required sensitivity and bandwidth. The proof-mass length,
width, thickness are represented as l1, b1, h1. The L- Beam dimensions
are represented by l2, l3, b2, and h2 respectively, which are shown in
fig.2.3.
The optimized device dimension are given below
Proof-mass size (l1 X b1 X h1)

2500 X 2500X 300 m

Length of beam (l2)

3200 m

Length of beam (l3)

640 m

22

Beam Width (b2)

150 m

Beam thickness (h2)

55 m

Air gap

22 m

(i)

Area of the proof-mass (A) = l1 X b1 = 6.25 x10-6 m2

(ii) Mass of proof-mass (m) = V.


(Density of the silicon

= (A x h1) x = 4.313 x10-6 Kg

= 2300 kg/m3, V = Volume of proof-mass)

Mass of the beams (mb) = 4(l2+l3) x b2 x h2 x

= 2.91456 x10-7 Kg.

All dimensions are in microns

Fig 2.3 Accelerometer geometrical details

23

(iii)

Maximum force on the proof-mass

Let F be the force acting on the proof-mass, due to 1 g acceleration,


which is given by the relation.
F=mxa
Where m is the mass of the proof-mass and a is the acceleration
F = 4.313 e-6 X 9. 81 X 1 = 0.0423 mN
This force is shared by all the beams equally.
Force acting on each beam W = F/4 = 0.01057mN
(vi)

Moment of inertia of beams

b2 h23
I=(
) = 2.07 e-18 m4
12
(v)

Deflection of the beam


The L beam is rigidly fixed to the frame on one side and other side is

attached to the Proof-mass. This can be considered as two cantilever


springs in series, rigidly fixed on one side and guided on the other side.
The numerical equation for the above is given as

Deflection

Wl 3
12 EI

Where E = Modulus of elasticity for silicon E=1.69 x 1011 N/m2


I= moment of inertia of beam.

24
3

Wl 2
For beam having length l2 1
12 EI
3

For beam having length l3 2

= 1+ 2 =

W l13 l 23
12 EI

Wl3
12 EI

8.3166 10 8 m /g
At 30g the maximum deflection =2.49 x10-6 m
(vi)

Bending stress in the beam


The formula for determining the bending stress in a beam
under simple bending is given by

M
y
I

Where
b = Bending stress in the beam
M = Bending moment acting on the beam
y = the perpendicular distance to the neutral axis.
At 30 g

M W l 2 30
M 0.01057 10 3 3200 10 6 30
M 1.014 10 6 Nm

25

For maximum bending stress y h2

13.4MPa

(vii)

Factor of safety
Since silicon is a brittle material, the UTS value is taken for

calculating the factor of safety design margin over the theoretical design
capacity.
Calculating factor of safety at 30g
FOS = ultimate strength / maximum stress
= 7000 / 13.4 = 522

(viii) Natural frequency


The natural frequency fn of the spring mass system is given
by
fn =

1
2

k/m

Where k is the stiffness of the spring

From equation 2.12

= 508.62 N/m

fn = 1728Hz.

(2.12)

26

2.7 Electrical design


Capacitance is the ability of a body to hold an electric charge. The
capacitance between two parallel plate conductors can be calculated if
the geometry of the conductor and the dielectric properties of the
insulator between the conductors are known.
(i) Nominal Capacitance (Co)
The Nominal Capacitance of a parallel plate capacitor with overlapping
area `a separated by a distance `d is equal to

(2.13)

Where
= Relative permittivity of the dielectric medium (for Air =1)
0 = Permittivity of free space = 8.85 x 10-12 F/m
Co= 2.514 pF
(ii) Accelerometer Sensitivity
The initial gap between the proof-mass and the electrode is 22m. Let
C1 and C2 are capacitances between top electrode and proof-mass and
bottom electrode and proof-mass respectively under the application of
1 g. Since the system is a differential capacitor, under the influence of
gravitation force, as one side capacitance increases the other side it
decreases.

27

0 r a
d
C1 2.5047 pF
C1

0 r a
d
C 2 2.5237 pF
C2

Change in capacitance is therefore calculated as,


C = C2 C1

(2.14)

= 19 fF
Sensitivity = C / applied acceleration
= 19 fF/g

2.8 FEM modelling and simulation


The design and development of a MEMS device is highly challenging
task involving simulation of micro structure behavior under coupled
environmental load conditions. FEM is an essential tool for MEMS design
and it provides accurate stimulation of the static and dynamic behavior of
complex structures at micro scale. Several FEM tools are available in the
mrket for MEMS modelling and simulation. However in this case,
Coventorware software is used for accelerometer modelling and FEM
analysis, Intellisuite software is used for wet etching process simulation
and SABER for system level simulation.

In Coventorware solid models are built from 2-D layout tool with
process information and meshes are created on solid models in the

28

preprocessor. Coventorware analyzer module uses various numerical


approaches such as 3-D FEM of MEMMECH, 3-D BEM of MEMELECTRO
modules for solving the partial differential equations of mathematical
physics.
Since our accelerometer structure has regular shaped beams and plates,
an eight node manhattan brick element is used for meshing the model.
The element has an orthogonal geometry i.e all model faces are planar and
join at right angles. The model is meshed with uniform mesh density
through the model to reduce errors.

Fig 2.4 Meshed model of accelerometer structure

2.9 Mesh convergence study


The accuracy of a discrete solution of a partial differential equation
depends on the density of the mesh. Therefore, the accuracy of an
analysis can only be judged by the comparison of results on meshes of
increasing degrees of freedom i.e. studying mesh convergence. To study

29

mesh convergence multiple mesh models of different densities are


created. The mesh density on the beams is gradually increased by
reducing the element size while keeping the aspect ratio same. The proofmass is meshed with a mesh density of 600 throughout the study. The
results of the proof-mass displacement (microns) with 1 g acceleration
as a function of mesh density of the beams are plotted in fig 2.5.

Fig 2.5 Mesh convergence simulation result


From fig 2.5, it can be seen that for a mesh density of 800 elements and
more on the beams, the change in variation of proof-mass displacement
is less than 1%, hence convergence is achieved. All further analysis is
done considering the optimized mesh density. Elements on the L-beam

30

joining face to the frame are completely constrained and all other
elements have 6 degrees of freedom.

2.10 FEM simulation results


The model is subjected to acceleration load upto 30 g and the
response of the sensor for displacement, change in capacitance, bending
stresses and cross-axis sensitivity are studied.

2.10.1

Acceleration vs. displacement

The structure is subjected to 0 g to 30 g acceleration in +Z


direction, in steps of 3 g to analyze the proof-mass displacement. The
analysis results are shown in table 2.2. The maximum displacement of
the proof-mass in Z-direction is 3.48 microns.

Acceleration 0
3
6
9
12
15
18
21
24
27
30
(g)
Displacement 0 0.348 0.696 1.044 1.392 1.74 2.088 2.436 2.784 3.132 3.48
(m)

Table 2.2 Acceleration Vs displacement of proof-mass

As shown in table 2.2 & fig 2.6, the proof-mass displacement is linear
with respect to the applied acceleration.

31

Fig 2.6 Applied acceleration Vs Proof- mass displacement

2.10.2 Acceleration vs. capacitance


Coupled electro-mechanical simulation is done for the accelerometer
by applying a voltage of 5V on the top and bottom electrodes and
grounding the proof-mass. The accelerometer is subjected to acceleration
varying from 0 g to 30 g in both + Z and

direction. The

capacitances across top electrode and proof-mass, bottom electrode and


proof-mass are obtained. The nominal capacitance of the accelerometer
is 2.514 pF. The resultant change in capacitance value in steps of 3 g is
given in Table 2.3 and the sensitivity of the accelerometer is 27.4 fF/g.

32

ACCELERATION
(g)

CHANGE IN CAP. FOR


POSITIVE ACC. (pF)

CHANGE IN CAP. FOR


NEGATIVE ACC. (pF)

0.080249

-0.08011

0.16055

-0.16041

0.24119

-0.24095

12

0.321998

-0.32185

15

0.403377

-0.40324

18

0.485374

-0.48524

21

0.568108

-0.56797

24

0.651715

-0.65158

27

0.73633

-0.73619

30

0.822097

-0.82196

CHANGE IN CAPACITANCE

Table 2.3 Acceleration Vs change in capacitance

1
0.8
0.6
0.4
(pF) 0.2
0
-0.2 0
-0.4
-0.6
-0.8
-1

12

15

ACCELRATION (g)

18

21

24

27

30

Positive Acceleration
Negative Acceleration

Fig 2.7 Change in capacitance with acceleration

33

From the change in capacitance equation 2.14, it can be seen that


the change in capacitance is nonlinear with applied acceleration. The
above data is analyzed for non-linearity using least square method curve
fitting technique. The equation of the best fit line is y = 27.3684x +
(-3.5698) and the maximum non-linearity is 0.764% which is well within
the design requirement.

2.10.3 Bending stresses in the beams


Maximum bending stress occurs in L-beams due to the applied
acceleration in the Z direction. The maximum bending stress is occurring
at the point where the L-beam is anchored to the frame and at the sharp
corner of the L beam, fig 2.8 gives the stress distribution in the model.
The magnitude of maximum von Mises stress at 30 g acceleration is
11 MPa, which is much less than the UTS value of silicon which is
7000 MPa.

Fig 2.8 Stress distribution in the accelerometer at 30 g

34

2.10.4 Modal analysis


Modal analysis is done to obtain the first and second mode
frequencies of the accelerometer. The first mode resonant frequency of
the structure is, out of plane vibration along Z axis which is at 1470 Hz.
The thumb rule is that the first mode resonant frequency of the structure
shall be at least three times the required bandwidth. The second mode
frequency is, in X-Y plane at 3323 Hz. fig. 2.9 shows the modes of
vibration. Both the frequency modes of operation are far away from the
maximum operational bandwidth of 100Hz, hence no interference on
accelerometer functioning is expected.

Mode 1:1470Hz

Mode 2: 3323Hz

Fig 2.9 Modal analysis results of the accelerometer structure

35

2.11 Cross - axis sensitivity


The accelerometer shall sense acceleration along the Z-axis only
and it shall be insensitive to the applied acceleration along other two
axes. The cross-axis sensitivity specification of the accelerometer is < 1%
of FSO. As already presented earlier the proof-mass thickness is 300 m
and the supporting L-beam thickness is 55 m. From fabrication process
point of view, it is very easy to realize the beams aligned to the top or
bottom plane of the proof-mass by etching the beams to the required
thickness from one side. The accelerometer is symmetrical about x and y
axes. The meshed accelerometer model is subjected to cross acceleration
along X-axis for three different beam positions along the thickness of the
Proof-mass.
The results are shown in fig 2.10. The cross-axis sensitivity increases
as function of the distance from central plane of the proof-mass. For the
beam position, at the central plane of the proof-mass, the cross-axis
sensitivity obtained is 0.01% of FSO, hence the position of the support
beams is finalized at the center of the proof-mass.

36

Fig 2.10 Effect of beam position on the cross axis sensitivity

2.12

System level simulation of the sensor


Traditionally MEMS devices have been simulated using field

solvers, such as finite element method (FEM) and boundary-element


method (BEM) analysis tools. These tools solve complex partial
differential equations derived from the detailed description of the
physical design, but those equations are far from simple and take a lot of
time to solve.

37

However, in system level or high-level simulation method, simulations


are done based on the behavior of a device as expressed by reducedorder equations [19]. It simulates the overall behavior of complete model
instead of the interactive behavior of many finite elements that comprise
the model. The complex mathematical description used with high-level
models leads to a much smaller number of degrees of freedom that
reduces the number of computations performed by the solver, resulting
in much faster simulation runs. The higher level of abstraction and the
physical

analogy

between

translation,

rotation,

electronics,

thermodynamics, and other entities permits the interconnection of a


number of physical domains.
The system level simulation tool offered by Coventorware architect
module called SABER is used in the present design. SABER uses
behavioral model libraries.

The models include underlying code that

expresses the behavior of the individual components subjected to


electrical, mechanical or other domain stimuli.

2.12.1 System level modelling


The system level accelerometer model (fig 2.11) is built using
standard library components and consists of four support beam elements
and a rectangular proof-mass plate. Two electrode elements at the top
and bottom of the proof-mass allow electrostatic excitation and capacitive
detection due to vibration.

38

Fig.2.11 System level behavior model of the accelerometer

2.12.2

Small signal AC analysis

To determine the resonant behavior of the accelerometer structure a


small signal AC analysis is performed. The model is excited over a
frequency range of 1 Hz to 10 kHz. Fig 2.12 shows the maximum
response and phase angle details as a function of the applied
acceleration. It can be seen that the natural frequency of the system is
1432 Hz where the phase angle is 90 deg.

39

Fig 2.12 Small signal AC analysis results

2.12.3 Pull in analysis


During anodic bonding process of accelerometer, for bonding
silicon wafers with glass wafers, upto 1000 VDC is applied between the
wafers. At such high voltages due to electrostatic attraction the proofmass may come in complete area contact with electrodes on the glass
wafers and may not revert back to normal position due to stiction arising
out of large area of contact. Pull in analysis is done to know the safe
working voltage that can be applied between the proof-mass and
electrodes. The analysis is done by grounding the proof-mass and
varying bottom electrode voltage from 25 to 160 VDC.

40

Fig 2.13 Pull in analysis result


As shown in fig 2.13 pull in analysis result, the electrostatic force
overcomes spring force of the structure at pull in voltage of 147.97 V.
Hence during fabrication suitable bumps are provided on the proof-mass
to overcome stiction problem.

2.12.4 Transient analysis


Transient analysis is done to estimate the device sensitivity and
response time to the applied acceleration input. The aerospace sensors
need to have quick response and fast settling times. An acceleration of
1 g is applied on the structure in 0.1 micro-sec and withdrawn after
1m-sec in 1micro-sec. Fig 2.14 shows the result of transient analysis.
The result shows that the displacement and capacitance change are very
closely following the input 1 g signal. The system has response time of

41

less than 1 m-sec which is well within the design requirements of 1msec.

Fig 2.14 Transient analysis result with 1 g input signal

2.13

Dynamic analysis

In the current accelerometer design the sensing proof-mass is


capped on both sides using glass wafers at atmospheric pressure. As the
proof-mass moves towards the stationary electrodes, pressure between

42

the two layers increases developing damping forces. This pressure drives
out the entrapped air between the parallel plates. On the contrary, when
the proof-mass is moving away from the electrode the pressure in the gap
is reduced causing surrounding air to flow into the gap. In both cases the
force on the proof-mass caused by built-up pressure is always against
the movement of the plate. The work done by the plate is consumed by
the viscous flow of the air and transformed into heat. In other words, the
air film acts as a damper and this type of damping is called squeeze film
damping. The damping phenomenon is shown in fig.2.15. In the past,
considerable research was done in characterizing squeeze film damping
behavior in MEMS structures [20-29]. Veijola et al developed equivalent
circuit

model

of

squeeze

film

damping

applicable

to

MEMS

accelerometers using R-L elements [23]. Sadd et al considered the


incompressible effects of gas at low squeeze numbers [27].

Air Moving Inside

Movement of
Proof-mass

Proof-mass
Air Moving Outside

Fig.2.15 Damping phenomenon

43

2.13.1 Squeeze film damping

Starr proposed the following conditions which are to be satisfied to


obtain satisfactory behavior of damping in micro accelerometers, [20].

1. The behavior of squeeze film is governed by both viscous and


inertial effects. For small geometries inertial effects are neglected.
The specific condition of validity is:

d 2 / 1.0

(2.15)

Where
= frequency of oscillation of proof-mass
d = air gap
= density of air 1.16 e-18 kg /m3
= dynamic viscosity of the damping media i.e. air 1.86
x 10-11N-sec/m

1470 2 22 2 1.16 10 18
0.2
1.86 10 11

2. The air flow is assumed as continuum, slip flow condition at the


boundaries may reduce the effectiveness of damping.
To overcome this problem the film thickness shall be > (100 times
mean free path of air) [20] [29]. At room temperature, the mean free

44

path of air is 0.065 microns. Hence, the film thickness shall

be >

6.5 microns.

3. Squeeze no. which is dimensionless number, is a measure of


compression of fluid in the gap. If is low (close to zero implying
low speed), the gas film follows nearly incompressible viscous flow.
If
is large ( > 0.3), the film essentially acts as an incompressible
air spring and exhibits little energy dissipation.

12b 2
0.26
h02 Pa

(2.16)

Where = frequency of oscillation of proof-mass


b = half width of the plate
h0 = nominal film thickness
Pa = ambient gas pressure

4. For squeeze film damping condition, the damping coefficient (C)


is given by
w 3
2f ( ) n l
l
C
d 03

Where

w
f ( ) = Shape function
l

(2.17)

45

w, l = width and length of moving plate


n = Natural Frequency of the structure
do = Initial gap between moving and stationary plate
C = 0.05545 and
Cc (critical damping coefficient) = 2 m n = 0.1033

C 0.05545

0.54
Cc
0.1033

Where is damping factor.


As

the

damping

factor

is

less

than

one,

designed

accelerometer is under-damped system. For under-damped system


response time is given by
T(r) =

1
4

= 4.25 10-5 sec

2.13.2 Simulation results


To study the dynamic behavior of the accelerometer, a squeeze film
damper model is made with moving proof-mass, electrodes, air gap and
air column as shown in fig 2.16. The air is assumed to be moving in and
out of the gaps at the four side edges of the proof-mass.

46

Fig. 2.16 Squeeze film damping model

2.13.3 Squeeze film damping coefficient variation with


frequency
To realize a sensor with required linearity over the operational
frequency range, the damping forces shall be linear within the
operational frequency range and the spring force shall be low and
damping coefficient shall be constant. Fig 2.17 shows the variation of
damping force and spring force over large frequency range. At low
frequencies, air can escape with little resistance and the spring force is
small. At high frequencies, the air is held captive by its own inertia, there
is not enough time for the air to move out of the way as the structure
oscillates. The air gets compressed, resulting in an increased spring
force. The damping is caused by viscous forces and proportional to the
velocity of the oscillating structure. But, if the gas is compressed and
does not move much, then the damping force will be lower. This explains
why the damping gets smaller as the frequency increases above about 1
MHz.

47

Fig 2.17 Damping analysis results


over large frequency range

Fig 2.18 Damping analysis results


over small frequency range

Fig.2.19 Damping coefficient variation with frequency


However, as shown in fig 2.18, within the operational frequency
range of accelerometer damping force is proportional to the frequency of
operation and spring force is negligible. It can be seen from fig 2.19,
within the operational range of frequency damping co-efficient variation
is negligible.

48

2.13.4 Harmonic analysis


Harmonic analysis is performed using MEMMECH module of the
coventorware software to find the linearity of the accelerometer response
over the required bandwidth of operation under applied acceleration. The
accelerometer is subjected to a harmonic load of 30g at frequency
ranging from 1 Hz to 100Hz. The modal-damping coefficient for the

Displacement (m)

analysis is 0.055.

Fig.2.20. Harmonic analysis result

49

Fig 2.20 gives the displacement response of accelerometer at 30g input


up to required bandwidth frequency of 100 Hz. The frequency sensitivity
of displacement within the operation bandwidth is negligible.

2.14 Interface electronics


The MEMS Accelerometer is electrically equivalent to a differential
parallel plate capacitor structure with the capacitance change occurring
due to the change in the gap between the parallel plates. The switchedcapacitor charge integration method has been widely used in MEMS
capacitive sensor interface circuits [30-38]. The interface circuit shall
have the following features
Interface with the sensor with given nominal capacitance of 2.5pF
and also nullify any capacitance offset that is present.
The required output swing of the circuit is 0 to 5V with a DC bias
output voltage at 0g
The circuit should be able to handle a total capacitance change of
0.6pF
The circuit should be able to resolve a minimum capacitance
change of 1fF
The bandwidth should be more than 100Hz

50

2.14.1 Capacitance to voltage conversion scheme


The interface circuit for converting the variation in capacitance into
voltage is implemented using a standard capacitance to voltage
conversion ASIC, MS3110 from Irvine Sensors.
MS3110 [39] is a general purpose, ultra noise CMOS IC that
requires only a single +5V DC supply and some decoupling components.
The ASIC is capable of sensing capacitance changes down to 4aF/rtHz. It
can interface with either a differential capacitor pair or a single capacitive
sensor.
The salient features of the IC that enables its suitability for
integration with the accelerometer chips are listed below:
Capacitance resolution: 4aF/rtHz
Differential variable capacitance sensor interfacing
Gain and DC offset trim
Programmable bandwidth adjustment
On chip EEPROM for storage of program coefficients

51

The ASIC functional block diagram is shown in fig 2.21.

Fig 2.21 Block diagram of MS3110 ASIC

The fundamental operation of the conversion scheme is, that of


charge

amplification

by

the

trans-conductance

operational

amplifier interfacing the sensor capacitance bridge. The sense


nodes of the capacitance bridge are fed by a square wave signal
generated internally and whose amplitude oscillates between zero
g bias voltage and 0V.
The amplified signal is then low pass filtered in the next section.
The bandwidth of the LPF is programmable.

52

The last block is a gain amplifier with the provision of gain and
offset programmability. The various coefficients for capacitance
bridge balancing, bandwidth, sensitivity, gain and offset are stored
in an EEPROM.
The various timing signals for EEPROM read, write, and square
wave are generated internally.

The ASIC senses the change in capacitance between two capacitors


and provides a voltage output proportional to the change. The transfer
function of the ASIC is given as:
Vout = GAIN * V2P25 * 1.14 * (CS2T-CS1T)/CF + VREF
Where Vout is the output voltage
Gain = 2 or 4V/V
V2P25 = 2.25 VDC
CS2T = CS2IN + CS2
CS1T = CS1IN + CS1

CF is selected to obtain the required sensitivity


VREF can be set to 0.5V or 2.25V DC for CS = CS2T CS1T =0

53

The pin diagram of the ASIC and pin description are given in Fig 2.22
and Table 2.4 respectively.

Fig 2.22 MS3110 ASIC pin diagram


Name

Description

CHPRST

IC reset. Internally pulled up.

V2P25

2.25VDC reference

TESTSEL

Enables the user to bypass on-chip EEPROM

Pin
No

and program the IC directly


4

CS2IN

Capacitor sense input 2

CSCOM

Capacitor sense common

CS1IN

Capacitor sense input 1

SDATA

Serial Data input, used for serial data input


port for programming the EEPROM or the IC

54

registers directly.
8

SCLK

Serial clock input

NC

No Connection

10

HV16

16VDC input port, tied to 16V when writing to


EEPROM, grounded otherwise

11

WRT

Write select

12

NC

No connection

13

-V

Negative voltage rail, usually 0V

14

VO

IC Voltage output

15

+V

Positive voltage rail, usually 5V

16

NC

No connection
Table 2.4 MS 3110 pin diagram description

2.14.2 Features of the interface electronics


The circuit gives a DC bias of 2.25V at 0g, which is also equal to
the reference voltage of the ASIC
The sensitivity of the sensor can be programmed from 20mV/g to
66mV/g.
Thus the maximum output swing is 2V for 30g
The fabricated accelerometer chips show a deviation in nominal
capacitance values from the designed values.

55

The differential capacitance bridge is not balanced at zero g. As a


result there is an initial offset in the output voltage when at zero
g. The offset is nullified by using the internal capacitances in the
ASIC. In cases where the offset is much more than the limit of the
internal capacitances, provision is made to add an external
capacitor of suitable value in parallel with the lower capacitance in
the bridge.
Provision is made in the interface circuit board to make possible
the tuning of the ASIC coefficients after the final assembly of the
components on the PCB Board to cater for packaging effects also.

2.15 Results & discussion

An accelerometer with a range of 30 g and with linearity &


cross-axis sensitivity less than 1% of full scale output (FSO) is
configured.

Detailed mechanical and electrical design is done using FEA


simulation techniques and the results show that the design
meets required specifications of the sensor.

Comparison of analytical and simulation results are presented


in table 2.5.

56

Results

Analytical

FEM
simulation

System level
simulation

Proof-mass displacement

0.083

0.116

0.121

Stress at 30 g (MPa)

13.4

11

--

Natural frequency(Hz)

1728

1470

1432.5

Sensitivity (fF/g)

19.0

26.0

28.07

Response time (msec)

0.042

0.2

(m/g)

Table 2.5 Comparison of results

From the comparison of results presented above it can be seen that


the analytically estimated maximum displacement of the proofmass is almost 30% less than the simulated FEM results under the
applied 1 g acceleration.

This is due to the fact that during analytical calculations it is


estimated that the L-beams are fixed on one side and guided on
the other side. However, the smaller length (l3) of the L- beam
(640microns) which is 20% of larger length (l2) contributes to
twisting in addition to bending, hence the discrepancy.

You might also like