1
KEVIN T. CABANTE AUGUST 24, 2015
EE5-1
MECHANICAL SENSORS
Definition
Mechanical sensors are transducers that measure mechanical phenomena and
convert it into a signal which can be read by an observer or by an instrument. They
include pressure sensor, force and torque sensor, inertial sensor and flow sensor. In terms
of its sensing techniques, mechanical sensors use either of the following scientific
principles: piezoresistivity, piezoelectricity, capacitive techniques, inductive techniques
and resonant techniques.
1. STRAIN GAUGES
Strain gauges are sensors for measuring various mechanical quantities indirectly
by measuring the strain the quantity produces on an object. Strain can be related to
stress, force, torque and a host of other quantities including displacement, acceleration
or position.
At the heart of all strain gauges is the change in resistance of materials due to
change in their length due to strain. This change in resistance can be mathematically
expressed as
dRR
= d + d(L/A)
L/A
where R is the object’s electrical resistance, σ is the resistivity of the material, L is the length
of the material and A is the cross-sectional area of the material. As shown, the change in
resistance of a material is dependent on (1) change in conductivity and (2) change in
deformation. For small deformations (linear deformation), both terms on the right hand
side are linear functions of strain . Bundling both effects together (that is, the change in
conductivity and deformation) we can write,
2
dRR
= Ss
where is Ss is the sensitivity of the strain gauge, also known as the gauge factor.
For any given strain gauge the gauge factor is a constant. It ranges between 2 to
6 for most metallic strain gauges and from 40 to 200 for semiconductor strain gauges. The
strain gauge relation gives a simple linear relation between the change in resistance of
the sensor and the strain applied to it.
Strain Gauge as Force Sensor
Strain gauge can be used as a force sensor. In the diagram below, it can be shown
that deformation is proportional to stress applied to the material.
Given the conductor shown above and applying a force along its axis, the stress
is given by
= FA
= EdLL
= E
where is the stress [N/m2], E is Young’s modulus of the material (modulus of elasticity)
[N/m2 ] and is the strain (dL/L). Herein, strain is a normalized linear deformation of the
material.
Strain gauges come in many forms and types. Any material, combination of
materials or physical configuration that changes its resistance due to strain can be used
as a strain gauge. However, there are two types that account for most of the strain
gauges in use today: (1) wire (or metal) strain gauges and (2) semiconductor strain
gauges.
Metallic strain gauges in its simplest form is a length of wire held between two
posts. When a force is applied to them, the force will deform the wire causing a change
in the wire’s resistance. Sometimes, multiple lengths of wire were used. This method was
used in the past and is valid although it is not very practical.
3
A more practical strain gauge is built out of a thin layer of conducting material
deposited on an insulating substrate (plastic, ceramic, etc.) and etched to form a long,
meandering wire. Constantan (60% copper, 40% nickel) is the most common material
used in this method due to its negligible temperature coefficient of resistance. Other
materials commonly used are shown in the table.
4
Table 1. Materials for resistive strain gauges and their properties.
Material Gage
factor
Resistivity
[mm2/m]
Thermal
coeff. of
expansion
[10/K]
Expansion
coeff.
[10/K]
Maximum
temperatrure
[C]
Constantan (Cu60Ni40) 2.0 0.48 5 12.5 400
Nichrome (Ni80Cr20) 2.0 1.3 100 18 1000
Manganine
(Cu84Mn12Ni4)
2.2 0.43 10 17
Nickel -12 0.11 6000 12
Chromel (Ni65Fe25Cr10) 2.5 0.9 300 15 800
Platinum 5.1 0.1 2450 8.9 1300
Elinvar
(Fe55Ni36Cr8Mn0.5)
3.8 0.84 300 9
Platinum-Iridium
(Pt80Ir20)
6.0 0.36 1700 8.9 1300
Platinum Rhodium
(Pt90Rh10)
4.8 0.23 1500 8.9
Bismuth 22 1.19 300 13.4
Strain gauges may also be used to measure multiple axis strains by simply using
more than one gauge or by producing them in standard configurations. Some of these
are shown next.
Two-axis Three-axis rosette
5
Semiconductor strain gauges operate like resistive strain gauges but their
construction and properties are different. The gauge factor for semiconductors is much
higher than for metals. The change in conductivity due to strain is also much larger than
in metals. This type is typically smaller than metal types and is often more sensitive to
temperature variations that it requires temperature compensation.
Like metallic materials, all semiconductor materials exhibit changes in resistance
due to strain. The most common material used is silicon because of its inert properties and
ease of production. The base material is doped, by diffusion of doping materials (usually
boron or arsenide for p or n type) to obtain a base resistance as needed. The substrate
provides the means of straining the silicon chip and connections are provided by
deposition of metal at the ends of the device.
45-degrees rosette 45-degrees stacked rosette
Membrane rosette
6
One of the important differences between conductor and semiconductor strain
gauges is that semiconductor strain gauges are essentially nonlinear devices with
typically a quadratic transfer function, expressed as
dRR
= S1 + S12
They can also be either positive or negative temperature coefficient.
Applications
Strain gauge must be made to react to a force. The strain gauge is attached to
the member in which strain is sensed, usually by bonding. Special bonding agents exist
for different applications and types of materials and are usually supplied by the
manufacturers of strain gauges or specialized producers.
Strain gauges are often used for bending strain, twisting (torsional and shear strain)
and longitudinal tensioning/deformation (axial strain) of structures (engine shafts, bridge
loading, truck weighing and many others).
The properties of strain gauges vary by application. Most metal gauges have a
nominal resistance between 100 and 1000 ohms, gauge factor between 2 to 5 and
dimensions from less than 3x3 mm to lengths in excess of 150 mm. Rosettes (multiple axis
strain gauges) are also available with 45, 90 and 120-degree axes as well as diaphragm
and other specialized configurations.
Semiconductor strain gauges are usually smaller than most resistive strain gauges
and can be made with higher resistances. However, their use is limited to low
temperatures. They can be much less expensive than metal strain gauges and are often
part of another device.
Strain gauges are subject to a variety of errors. Resistance, especially in
semiconductors, is affected by temperature in the same way as by strain. In metal
7
gauges, this is usually small through the use of materials with low temperature coefficients
of resistance. In semiconductors, temperature compensation is sometimes provided on
board or a separate sensor may be used for this purpose. A third source of error is due to
the strain itself, which, over time, tends, to permanently deform the gauge. This error can
be eliminated by periodic re-calibration and can be reduced by ensuring that the
maximum deformation allowed is small and below the recommended for the device.
Specialized Strain Gauges
There are other types of strain gauges for very specialized applications. One
example is the optical fiber strain gauge. It operates on the principle that the change in
length of the fiber due to strain changes the phase of the light through the fiber. By
measuring the light phase, either directly or in an interferrometric method, it can produce
readings of minute strain that cannot be obtained in other strain gauges. The device and
the electronics necessary is far more complicated than standard gauges.
There are also liquid strain gauges which rely in the resistance of an electrolytic
liquid in a flexible container which can be deformed. Another type of strain gauge that
is used on a limited basis is the plastic strain gauge. These are made as ribbons or threads
based on graphite or carbon in a resin as a substrate and used in a way similar to other
strain gauges. They have very high gauge factors (up to about 300) but they are
otherwise difficult to use, inaccurate and mechanically unstable, severely limiting their
practical use.
2. TACTILE SENSORS
Tactile sensors are force sensors but as the definition of “tactile” action is broader,
the sensors are also more diverse. One view is that tactile action as simply sensing the
presence of force. Then a simple switch is a tactile sensor. This approach is commonly
used in keyboards where membrane or resistive pads are used.
The simplest tactile sensors are made of conductive polymers or elastomers or with
semi-conductive polymers, called piezoresistive sensors or force sensitive resistive (FSR)
sensors. In these devices, the resistance of the material is pressure dependent.
In many tactile sensing applications it is often important to sense a force
distribution over a specified area (such as the “hand” of a robot). Either an array of force
sensors or a distributed sensor may be used. These are usually made from piezoelectric
films which respond with an electrical signal in response to deformation (passive sensors),
such as shown in the figure below.
8
In the figure, the polyvinylidene fluoride (PVDF) film is sensitive to deformation. The
lower film is driven with an ac signal. It contracts and expands mechanically and
periodically. When the upper film is deformed, its signal changes from normal and the
amplitude and or phase of the output signal is now a measure of deformation.
Another example of tactile sensors is shown below. It is used to detect breathing.
3. ACCELEROMETERS
Accelerometers measures the acceleration of an object indirectly by measureing
other quantities related to acceleration. It can accomplish the task using three methods:
by force, magnetic or electrostatic means.
In resistive method, a force sensor, such as a strain gauge, is used. By virtue of
Newton’s second law (F = ma) a sensor may be made to sense acceleration by simply
9
measuring the force on a mass. At rest, acceleration is zero and the force on the mass is
zero. At any acceleration a, the force on the mass is directly proportional given a fixed
mass. This force may be sensed in any method of sensing force but, again, the strain
gauge will be representative of direct force measurement.
Magnetic methods and electrostatic (capacitive) methods are quite commonly
used. A magnetic sensor can be used by measuring the field of a magnetic mass. The
higher the acceleration, the closer (or farther) the magnet from a fixed surface and
hence the larger or lower the magnetic field. On the other hand, the distance between
the mass and a fixed surface, which depends on acceleration can be made into a
capacitor. Capacitance increases (or decreases) with acceleration.
Methods of acceleration sensing start with the mechanical model of a mass shown
below.
The mass, moves under the influence of forces, has a restoring force (spring) and a
damping force (which essentially prevents it from oscillating). Under these conditions,
and assuming the mass can only move in one direction (along the horizontal axis),
Newton’s second law may be written as
ma = kx bdxdt
which essentially tells us that the mass has moved a distance x under the influence of
acceleration a and where k is the restoring (spring) constant and b is the damping
coefficient. Given the mass m and the constants k and b, a measurement of x gives an
indication of the acceleration a. Therefore, for a useful acceleration sensor (often called
accelerometer), it is sufficient to provide a component which can move relative to the
sensor’s housing and a means of sensing this movement. A displacement sensor (position,
proximity, etc.) can be used to provide an appropriate output proportional to
acceleration.
In strain gauge accelerometers, the mass is suspended on a cantilever beam and
a strain gauge senses the bending of the beam, as shown in the figure below.
10
Magnetic accelerometers uses variable inductance set-ups. A rod is connected
such as it is moving with the mass which in turn is linked to a coil. The inductance of the
coil is proportional to the position of the mass, as shown below.
In electrostatic accelerometers, one plate of a small capacitor is fixed and
connected physically to the body of the sensor. A second plate serves as the inertial
mass of the sensor is free to move and connected to a restoring spring. The restoring
force is then provided by springs or by a cantilever’s fixed end. Several configurations are
given below.
11
In the first figure, the distance between the plates changes with acceleration. In
the second, the effective area of the plates changes while the distance between the
plates stays constant. In either case, acceleration either increases the capacitance or
decreases it, depending on the direction of motion. In a practical accelerator, the plates
must be prevented from touching by stoppers and some kind of damping mechanism
must be added to prevent the springs or the beam from oscillating. A more practical set-
up is shown below.
Other type of accelerometers also employ a moving mass in one form or another.
One particular example is the heated gas accelerometer. A gas in cavity is heated to an
equilibrium temperature. Two (or more) thermocouples are provided and are equidistant
from the heater. Under rest conditions, the two thermocouples are at the same
temperature. Their reading (one thermocouples is the sense thermocouple, the second
the reference thermocouple) is zero. Under acceleration, the gas shifts to the direction
12
opposite the motion (the gas is the inertial mass) causing a temperature gradient which
is calibrated in terms of acceleration. The set-up is shown below.
Other accelerometers use optical means, for example, activating a variable
shutter by means of the moving mass. Optical fiber accelerometers use an optical fiber
position sensor. Other uses vibrating reeds whose vibration rate changes with
acceleration.
4. VELOCITY SENSORS
Velocity sensing is more complicated than acceleration sensing. However, one
can always measure something proportional to velocity. For example, we may infer the
velocity of a car from the rotation of the wheels, the transmission shaft ( a common
method of velocity measurement in cars) or count the number of rotation of a shaft per
unit time in an electric motor.
A free-standing sensor that measures velocity directly is much more difficult to
produce. One approach that may be used is the induction of emf in a coil due to a
magnet. This requires that the coil be stationary. If the velocity is constant (no
acceleration) the magnet cannot move relative to the coil. For changing velocity
(when acceleration is not zero), the set-up below maybe useful.
13
The emf induced in the coils is governed by Faraday’s law, written mathematically
as
emf = Nddt
where N is the number of coils, Ф is the flux produced by the magnet and t is the time.
The time derivative indicates that the magnet must be moving to produce a nonzero
change in flux. The most common approach to velocity sensing is to use an
accelerometer and integrate its output using an integrating amplifier. Practically,
constant velocity (zero acceleration) cannot be sensed by the set-up but relative
velocity of objects is easily measured.
Aside from the one illustrated, there are also fluid velocity sensors, doppler effect
sensors, time of flight devices and others to measure velocity.
5. PRESSURE SENSORS
Sensing of pressure is only second in importance to sensing of strain in mechanical
systems. These sensors are used either in their own right, (to measure pressure), or to sense
secondary quantities such as force, power, temperature and the like. One of the reasons
for their prominence is that in sensing gases and fluids, force is not an option – only
pressure can be measured and related to properties of these substances. Another reason
for their widespread use and of exposure of most people to them is their use in cars,
atmospheric weather prediction, heating and other consumer oriented devices. The
“barometer” hanging on many a wall and the use of atmospheric pressure as indication
of weather conditions has helped popularize the concept of pressure and pressure
sensing.
Pressure is force per unit area. Sensing it follows the same principle as the sensing
of force – measuring the displacement of an appropriate member of the sensor in
14
response to pressure. The range of methods is quite large and includes thermal, optical
as well as magnetic and electrical principles.
Mechanical Pressure Sensors
Earliest sensors were purely mechanical and involves direct transduction from
pressure to mechanical displacement. These devices are actuators that react to pressure
and are still as common today as ever. Some mechanical devices have been combined
with other sensors to provide electrical output. Others are still being used in their original
form. The most common of these is the Bourdon tube, shown below.
The most common devices used for pressure sensing are the thin plate and the
diaphragm or membrane. Their behavior and response to pressure is different. In relation
to the figure below, the deflection of the center of a membrane (maximum deflection)
under radial tension S and the stress in the diaphragm are given as
dm = r2P4S
, m = St
15
where P is the applied pressure difference between the top and bottom of the
membrane, r its radius and t its thickness. If the thickness is not negligible, the behavior
or response of the membrane is given as
dm = 3 1 v2 r4P
16Et2, m = 3r2P
4t2
where E is the Young’s modulus and V is the Poisson’s ratio. The displacement is linear
with pressure hence their widespread use for pressure sensing.
Pressure sensors come in four basic types. Absolute pressure sensors senses pressure
relative to absolute vacuum. Differential pressure sensors measures the difference
between two pressures on two ports of the sensor. Gage pressure sensors measures the
pressure relative to ambient pressure and is the most common. Sealed gage pressure
sensor measures the pressure relative to a sealed pressure chamber (usually 1 atm at sea
level or 14.7 psi).
Piezoresistive Pressure Sensors
Piezoresistor is a semiconductor strain gauge. Most modern pressure sensors use it
rather than the conductor type strain gauge. However, resistive (metal) strain gauges are
still used at higher temperature or for specialized applications. Piezoresistors may be
fabricated of silicon, easy to construct and allows on board temperature compensation,
amplifiers and conditioning circuitry. Its basic construction is shown below.
The change in resistance of the two piezoresistos is given as
R1
R1 =
R2
R2 = 1
2 y x
where π is an average sensitivity (gauge) coefficient and σx and σy are the stresses in
the transverse directions. Piezoresistors and the diaphragm are fabricated of silicon. A
vent is provided, making this a gage sensor. If the cavity under the diaphragm is
hermetically closed and the pressure in it is P0, the sensor becomes a sealed gage
16
pressure sensor sensing the pressure P-P0. A differential sensor is produced by placing the
diaphragm between two chambers, each vented through a port.
Another approach is to use a single strain gauge and passing a current through it.
Pressure is applied perpendicular to the current. The voltage across the element is
measured as an indication of the stress and therefore pressure.
Various pieroresistive pressure sensors are shown below.
Miniature Type
Pitran Absolute Pressure Sensors 150psi Differential Pressure Sensor
17
Capacitive Pressure Sensors
The deflection of the diaphragm constitutes a capacitor in which the distance
between the plates is pressure sensitive. These sensors are very simple and are particularly
useful for sensing of very low pressure. At low pressure, the deflection of the diaphragm
may be insufficient to cause large strain but can be relatively large in terms of
capacitance. The capacitance may be part of an oscillator, where the change in its
frequency may be quite large making for a very sensitive sensor. It has advantages of
being less temperature dependent and not sensitive to overpressures. Overpressures of
2-3 orders of magnitude larger than rated pressure may be easily tolerated without ill
effects. The sensors are linear for small displacement but at larger pressures the
diaphragm tends to bow causing nonlinear output.
Magnetic Pressure Sensors
In large deflection sensors, an inductive position sensor may be used or an LVDT
attached to the diaphragm. For low pressures, variable reluctance pressure sensor is
more practical. In magnetic pressure sensors, the diaphragm is made of a ferromagnetic
material and is part of a magnetic circuit, as shown below.
100psi Pressure Sensor in TO5 Can 15 and 30 psi Differential Pressure Sensors
18
The reluctance is directly proportional to the length of the air gap between the
diaphragm and the E-core. Gap changes with pressure and the inductance of the two
coils changes and sensed directly. A very small deflection can cause a very large
change in inductance of the circuit making this a very sensitive device. Magnetic sensors
are almost devoid of temperature sensitivity allowing these sensors to operate at
elevated temperatures.
Optoelectronic and Thermal Pressure Sensors
Fabri-Perot optical resonator is used to measure small displacements, and
consequently, pressure. Light reflected from a resonant optical cavity is measured by a
photodiode to produce a measure of pressure sensed.
A very old method of sensing low pressures (often called vacuum sensors) is the
Pirani gauge. It is based on measuring the heat loss from gases which is dependent on
pressure. The temperature is sensed and correlated to pressure, usually in an absolute
pressure sensor arrangement.
Semiconductor-based Pressure Sensors
There are also semiconductor-based pressure sensors available. However, they
can only operate at low temperatures (-50 to +150˚C). Temperature dependent errors
can also be high unless properly compensated either externally or internally. The range
of these sensors can exceed 50,000 psi and can be as small as a fraction of psi.
Impedance is anywhere between a few hundred ohms to about 100 kiloohms,
depending on device. Linearity is typically between 0.1 to 2%.
6. GYROSCOPES
Gyroscopes come to mind usually as stabilizing devices in aircraft and spacecraft
in such applications as automatic pilots. However, they are much more than that and
much more common than one can imagine. The gyroscope is a navigational tool. Its
purpose is to keep the direction of a device or vehicle. It is used in all satellites, in smart
weapons and in all other applications that require attitude and position stabilization.
Eventually will find their ways into consumer products such as cars although they have
already found their ways into toys.
The basic principle involved is the principle of conservation of angular momentum:
“In any system of bodies or particles, the total angular momentum relative to any point
in space is constant, provided no external forces act on the system”.
Mechanical Gyroscopes
Mechanical gyroscopes are the best known of the existing gyros and the easiest
to understand. It consists of a rotating mass (heavy wheel) on an axis in a frame which
19
provides the angular momentum. If one tries to change the direction of the axis, by
applying a torque to it, a torque is developed in directions perpendicular to the axis of
rotation. This forces a precession motion. This precession is the output of the gyroscope
and is proportional to the torque applied to its frame.
The relation between applied torque and the angular velocity of precession is
T = I
where T is the applied torque, the angular velocity, I the inertia of the rotating mass,
is the angular velocity of precession and I is the angular momentum.
Two or three axes gyroscope are built by duplicating this structure with rotation
axes perpendicular to each other. This type of gyroscope has been used for many
decades in aircraft but it is a fairly large, heavy and complex which could not be easily
adapted to small systems. It also has other problems, associated with the spinning mass
(bearings, friction, balancing, etc.)
Coriolis Force Gyroscopes
Coriolis acceleration has been used to devise much smaller and more cost
effective gyroscopic sensors. It is built in silicon by standard etching methods. The
rotating mass of mechanical gyroscopes is replaced by a vibrating body where the
coriolis acceleration is used for sensing. The idea is based on the fact that if a body
moves linearly in a rotating frame of reference, an acceleration appears at right angles
to both motions as shown in the figure below.
20
Linear motion is supplied by the vibration of a mass, usually a harmonic motion.
Under normal conditions, the coriolis acceleration is zero and the force associated with
it is zero. If the sensor is rotated in the plane perpendicular to the linear vibration, an
acceleration is obtained, proportional to the angular velocity Ω.
An example of a coriolis based gyroscope is shown below.
Optical Gyroscopes
One of the more exciting developments in gyroscopes is the optical gyroscope.
It has no moving members and can be used extensively for guidance and control. It is
based on the Segnac effect. The Segnac effect is based on propagation of light in
optical fibers and can be explained using the figure below.
21
The ring is at rest and two laser beams travel the length of the ring, in opposite
directions. The time it takes either beam to travel the length of the ring is Δt=2πR/nc
where n is the index of refraction of the optical fiber and c is the speed of light in the
fiber. Suppose that the ring rotates clockwise at an angular velocity Ω. The CW beam
will travel a distance 2πR +ΩRΔt and the CCW beam a distance 2πR - ΩRΔt. The
difference between the two paths is then given by
l = 4R2
nc
Thus, there is linear relation between Ω (the stimulus in this case) and the change in
length traveled. The challenge is to measure this change in length. This can be done in
a number of ways. One method is to build an optical resonator.
A resonator is any device which has a dimension equal to multiple half
wavelengths of the wave. A ring is built as shown in the following figure.
Light is coupled through the light coupler (beam splitter). At resonance, which
depends on the circumference of the ring, maximum power is coupled into the ring and
minimum power is available at the detector. The incoming beam frequency is tuned to
do just that. If the ring rotates at an angular velocity Ω, the light beams in the ring change
in frequency (wavelength) to compensate for the change in apparent length of the ring.
The relation between frequency, wavelength and length can then be calculated as
dff
= d
= dll
22
The light wavelength increases in one direction and decreases in the other. The net effect
is that the two beams generate a frequency difference which is given as
f' = 4Anl
where f’ is the output linearly proportional to Ω.
In coil optical fiber gyroscopes, the resonator is replaced by a coiled optical fiber
fed from a polarized light source through a beam splitter to ensure equal intensity and
phase (the phase modulator adjusts for any variations in phase between the two beams).
The beams propagate in opposite directions. When returning to the detector, they
are at the same phase in the absence of rotation. If rotation exists, the beams will induce
a phase difference at the detector which is dependent of the angular frequency Ω.
Although very practical, optical gyroscopes are not cheap. But they are orders of
magnitude cheaper than the spinning mass gyroscope and much smaller and lighter.
They also have a very large dynamic range (as high as 10000) and so they can be used
for sensing angular frequency over a large span. Optical fiber gyroscopes are also
immune to electromagnetic fields as well as to radiation and hence can be used in very
hostile environments including space.
23
LEVEL SENSORS
Definition
Level sensors detect the level of liquids and other fluids and fluidized solids,
including slurries, granular materials, and powders that exhibit an upper free surface. The
level measurement can either be continuous or point values. Continuous level sensors
measure level within a specified range and determine the exact amount of substance in
a certain place, while point-level sensors only indicate whether the substance is above
or below the sensing point.
Introduction
The simplest and oldest industrial level measuring device is, of course, the sight
glass. A manual approach to measurement, sight glasses have always had a number of
limitations. The material used for its transparency can suffer catastrophic failure, with
ensuing environmental insult, hazardous conditions for personnel, and/or fire and
explosion. Seals are prone to leak, and buildup, if present, obscures the visible level. It
can be stated without reservation that conventional sight glasses are the weakest link of
any installation. They are therefore being rapidly replaced by more advanced
technologies.
Other level-detection devices include those based on specific gravity, the
physical property most commonly used to sense the level surface. A simple float having
a specific gravity between those of the process fluid and the headspace vapor will float
at the surface, accurately following its rises and falls. Hydrostatic head measurements
have also been widely used to infer level.
When more complex physical principles are involved, emerging technologies
often use computers to perform the calculations. This requires sending data in a machine-
readable format from the sensor to the control or monitoring system. Useful transducer
output signal formats for computer automation are current loops, analog voltages, and
digital signals. Analog voltages are simple to set up and deal with, but may have serious
noise and interference issues. 4-20 mA current loops (where the loop current varies with
the level measurement) are the most common output mechanism today. Current loops
can carry signals over longer distances with less degradation. Digital signals coded in any
of a number of protocols (e.g., Foundation Fieldbus, Hart, Honeywell DE, Profibus, and RS-
232) are the most robust, but the older technologies such as RS-232 can handle only
limited distances. New wireless capabilities can be found in the latest transmitters' signals,
allowing them to be sent over tremendous distances with virtually no degradation.
As for the more advanced measurement technologies (e.g., ultrasonic, radar, and
laser), the more sophisticated digital encoding formats require digital computer
intelligence to format the codes. Combining this requirement with the need for
advanced communication capabilities and digital calibration schemes explains the
trend toward embedding microprocessor-based computers in virtually all level
measurement products.
24
Level measurement determines the position of the
level relative to the top or bottom of the process
fluid storage vessel. A variety of technologies can
be used, determined by the characteristics of the
fluid and its process conditions.
Floats
Floats work on the simple principle of placing a buoyant object with a specific
gravity intermediate between those of the process fluid and the headspace vapor into
the tank, then attaching a mechanical device to read out its position. The float sinks to
the bottom of the headspace vapor and floats on top of the process fluid. While the float
itself is a basic solution to the problem of locating a liquid's surface, reading a float's
position (i.e., making an actual level measurement) is still problematic. Early float systems
used mechanical components such as cables, tapes, pulleys, and gears to
communicate level. Magnet-equipped floats are popular today.
Early float level transmitters provided a simulated analog or discrete level
measurement using a network of resistors and multiple reed switches, meaning that the
transmitter's output changes in discrete steps. Unlike continuous level-measuring devices,
they cannot discriminate level values between steps.
Hydrostatic Devices
Displacers, bubblers, and differential-pressure transmitters are all hydrostatic
measurement devices. Any change in temperature will therefore cause a shift in the
liquid's specific gravity, as will changes in pressure that affect the specific gravity of the
vapor over the liquid. Both result in reduced measurement accuracy.
25
Displacers work on Archimedes' principle. As shown in the figure, a column of solid
material (the displacer) is suspended in the vessel. The displacer's density is always greater
than that of the process fluid (it will sink in the process fluid), and it must extend from the
lowest level required to at least the highest level to be measured. As the process fluid
level rises, the column displaces a volume of fluid equal to the column's cross-sectional
area multiplied by the process fluid level on the displacer. A buoyant force equal to this
displaced volume multiplied by the process fluid density pushes upward on the displacer,
reducing the force needed to support it against the pull of gravity. The transducer, which
is linked to the transmitter, monitors and relates this change in force to level.
A bubbler-type level sensor is shown in the figure below. This technology is used in
vessels that operate under atmospheric pressure. A dip tube having its open end near
the vessel bottom carries a purge gas (typically air, although an inert gas such as dry
nitrogen may be used when there is danger of contamination of or an oxidative reaction
with the process fluid) into the tank. As gas flows down to the dip tube's outlet, the
pressure in the tube rises until it overcomes the hydrostatic pressure produced by the
liquid level at the outlet. That pressure equals the process fluid's density multiplied by its
depth from the end of the dip tube to the surface and is monitored by a pressure
transducer connected to the tube.
26
A differential pressure (DP) level sensor is shown in the figure below. The essential
measurement is the difference between total pressure at the bottom of the tank
(hydrostatic head pressure of the fluid plus static pressure in the vessel) and the static or
head pressure in the vessel. As with the bubbler, the hydrostatic pressure difference
equals the process fluid density multiplied by the height of fluid in the vessel. The unit in
the figure uses atmospheric pressure as a reference. A vent at the top keeps the
headspace pressure equal to atmospheric pressure.
In contrast to bubblers, DP sensors can be used in unvented (pressurized) vessels.
All that is required is to connect the reference port (the low-pressure side) to a port in the
vessel above the maximum fill level. Liquid purges or bubblers may still be required,
depending on the process's physical conditions and/or the transmitter's location relative
to the process connections.
Load Cells
A load cell or strain gauge device is essentially a mechanical support member or
bracket equipped with one or more sensors that detect small distortions in the support
member. As the force on the load cell changes, the bracket flexes slightly, causing output
signal changes. Calibrated load cells have been made with force capacities ranging
from fractional ounces to tons.
To measure level, the load cell must be incorporated into the vessel's support
structure. As process fluid fills the vessel, the force on the load cell increases. Knowing the
vessel's geometry (specifically, its cross-sectional area) and the fluid's specific gravity, it is
a simple matter to convert the load cell's known output into the fluid level.
While load cells are advantageous in many applications because of their
noncontact nature, they are expensive and the vessel support structure and connecting
piping must be designed around the load cell's requirements of a floating substructure.
The total weight of the vessel, piping, and connecting structure supported by the vessel
will be weighed by the load cell system in addition to the desired net or product weight.
This total weight often creates a very poor turndown to the net weight, meaning that the
net weight is a very small percentage of the total weight. Finally, the supporting structure's
growth, caused by uneven heating (e.g., morning to evening sunshine) may be reflected
as level, as can side load, wind load, rigid piping, and binding from overturn-prevention
27
hardware (for bottom-mounted load cells). In short, load cell weighing system
requirements must be a paramount consideration throughout initial vessel support and
piping design, or performance is quickly degraded.
Magnetic Level Gauges
These gauges are the preferred replacement for sight glasses. They are similar to
float devices, but they communicate the liquid surface location magnetically. The float,
carrying a set of strong permanent magnets, rides in an auxiliary column (float chamber)
attached to the vessel by means of two process connections. This column confines the
float laterally so that it is always close to the chamber's side wall. As the float rides up and
down with the fluid level, a magnetized shuttle or bar graph indication moves with it,
showing the position of the float and thereby providing the level indication. The system
can work only if the auxiliary column and chamber walls are made of nonmagnetic
material.
Many manufacturers provide float designs optimized for the specific gravity of the
fluid being measured, whether butane, propane, oil, acid, water, or interfaces between
two fluids, as well as a large selection of float materials. This means the gauges can
handle high temperatures, high pressures, and corrosive fluids. Oversized float chambers
and high-buoyancy floats are available for applications where buildup is anticipated.
Chambers, flanges, and process connections can be made from engineered
plastics such as Kynar or exotic alloys such as Hastelloy C-276. Special chamber
configurations can handle extreme conditions such as steam jacketing for liquid asphalt,
oversized chambers for flashing applications, and cryogenic temperature designs for
liquid nitrogen and refrigerants. Numerous metals and alloys such as titanium, Incoloy,
and Monel are available for varying combinations of high-temperature, high-pressure,
low-specific-gravity, and corrosive-fluid applications. Today's magnetic level gauges can
also be outfitted with magnetostrictive and guided-wave radar transmitters to allow the
gauge's local indication to be converted into 4-20 mA outputs that can be sent to a
controller or control system.
28
Capacitance Transmitters
These devices operate on the fact that process fluids generally have dielectric
constants, , significantly different from that of air, which is very close to 1.0.
Oils have dielectric constants from 1.8 to 5. Pure glycol is 37; aqueous solutions are
between 50 and 80. This technology requires a change in capacitance that varies with
the liquid level, created by either an insulated rod attached to the transmitter and the
process fluid, or an uninsulated rod attached to the transmitter and either the vessel wall
or a reference probe. As the fluid level rises and fills more of the space between the
plates, the overall capacitance rises proportionately. An electronic circuit called a
capacitance bridge measures the overall capacitance and provides a continuous level
measurement.
Magnetostrictive Level Transmitters
The advantages of using a magnet containing a float to determine liquid level
have already been established, and magnetostriction is a proven technology for very
precisely reading the float's location. Instead of mechanical links, magnetostrictive
transmitters use the speed of a torsional wave along a wire to find the float and report its
position.
In a magnetostrictive system, the float carries a series of permanent magnets.
29
A sensor wire is connected to a piezoceramic sensor at the transmitter and a
tension fixture is attached to the opposite end of the sensor tube. The tube either runs
through a hole in the center of the float or is adjacent to the float outside of a
nonmagnetic float chamber.
To locate the float, the transmitter sends a short current pulse down the sensor
wire, setting up a magnetic field along its entire length. Simultaneously, a timing circuit is
triggered ON. The field interacts immediately with the field generated by the magnets in
the float. The overall effect is that during the brief time the current flows, a torsional force
is produced in the wire, much like an ultrasonic vibration or wave. This force travels back
to the piezoceramic sensor at a characteristic speed. When the sensor detects the
torsional wave, it produces an electrical signal that notifies the timing circuit that the
wave has arrived and stops the timing circuit. The timing circuit measures the time interval
(TOF) between the start of the current pulse and the wave's arrival. From this information,
the float's location is very precisely determined and presented as a level signal by the
transmitter. Key advantages of this technology are that the signal speed is known and
constant with process variables such as temperature and pressure, and the signal is not
affected by foam, beam divergence, or false echoes. Another benefit is that the only
moving part is the float that rides up and down with the fluid's surface.
30
Ultrasonic Level Transmitters
Ultrasonic level sensors measure the distance between the transducer and the
surface using the time required for an ultrasound pulse to travel from a transducer to the
fluid surface and back (TOF). These sensors use frequencies in the tens of kilohertz range;
transit times are ~6 ms/m. The speed of sound (340 m/s in air at 15°C (1115 fps at 60°F)
depends on the mixture of gases in the headspace and their temperature. While the
sensor temperature is compensated for (assuming that the sensor is at the same
temperature as the air in the headspace), this technology is limited to atmospheric
pressure measurements in air or nitrogen.
Laser Level Transmitters
Designed for bulk solids, slurries, and opaque liquids such as dirty sumps, milk, and
liquid styrene, lasers operate on a principle very similar to that of ultrasonic level sensors.
Instead of using the speed of sound to find the level, however, they use the speed of
light.
31
A laser transmitter at the top of a vessel fires a short pulse of light down to the
process liquid surface, which reflects it back to the detector. A timing circuit measures
the elapsed time (TOF) and calculates the distance. The key is that lasers have virtually
no beam spread (0.2° beam divergence) and no false echoes, and can be directed
through spaces as small as 2 in.2 Lasers are precise, even in vapor and foam. They are
ideal for use in vessels with numerous obstructions and can measure distances up to 1500
ft. For high-temperature or high-pressure applications, such as in reactor vessels, lasers
must be used in conjunction with specialized sight windows to isolate the transmitter from
the process. These glass windows must pass the laser beam with minimal diffusion and
attenuation and must contain the process conditions.
Radar Level Transmitters
Through-air radar systems beam microwaves downward from either a horn or a
rod antenna at the top of a vessel. The signal reflects off the fluid surface back to the
antenna, and a timing circuit calculates the distance to the fluid level by measuring the
round-trip time (TOF). The fluid's dielectric constant, if low, can present measurement
problems. The reason is that the amount of reflected energy at microwave (radar)
frequencies is dependent on the dielectric constant of the fluid, and if r is low, most of
the radar's energy enters or passes through. Water ( r = 80) produces an excellent
reflection at the change or discontinuity in r.
In through-air radar systems, the radar waves suffer from the same beam divergence that
afflicts ultrasonic transmitters. Internal piping, deposits on the antenna, and multiple
reflections from tank buildup and obstructions can cause erroneous readings. To
overcome these problems, complex algorithms using fuzzy logic must be incorporated
into the transmitter. Transmitter setup can be tedious and changes in the process
environment (buildup, etc.) can be problematic.
Guided wave radar (GWR) systems can be the answer. A rigid probe or flexible
cable antenna system guides the microwave down from the top of the tank to the liquid
level and back to the transmitter. As with through-air radar, a change from a lower to a
higher r causes the reflection. Guided wave radar is 20 × more efficient than through-air
radar because the guide provides a more focused energy path. Different antenna
32
configurations allow measurement down to r = 1.4 and lower. Moreover, these sytems
can be installed either vertically, or in some cases horizontally with the guide being bent
up to 90° or angled, and provide a clear measurement signal.
GWR exhibits most of the advantages and few of the liabilities of ultrasound, laser,
and open-air radar systems. Radar's wave speed is largely unaffected by vapor space
gas composition, temperature, or pressure. It works in a vacuum with no recalibration
needed, and can measure through most foam layers. Confining the wave to follow a
probe or cable eliminates beam-spread problems and false echoes from tank walls and
structures.
References:
www.sensorsmag.com/sensors/leak-level/a-dozen-ways-measure-fluid-level-and-how-
they-work-1067. Accessed August 25, 2015.
www. ee.ascs3.uakron.edu/ida/sensors/chapter6.ppt. Accessed August 24, 2015.