Transducers

Definition
Transducers devices have been defined as components that may be used to interconnect like or unlike
systems and transmit energy between them. Although such a definition can be justified, a more generally
acceptable definition to electrical engineers is that a transducer is a device used to convert a physical
phenomenon into an electrical signal. In many cases the conversion may be via an intermediate stage, that
is, the measurand (for example, pressure) is first converted to a mechanical displacement which in turn is
converted to an electrical signal. The mechanical conversion is accomplished by one of two
fundamentally different methods, as follows.

Fixed reference devices

In these devices one part of the transducer is attached to a reference surface, and the other part connected,
either directly or via a linking mechanism, to the variable as in figure 5.1. Should the displacement be
small, some means of mechanical magnification or electrical amplification must be used to obtain a
satisfactory sensitivity.

Mass-spring or seismic device

In these transducers there is only one contact or anchor point, this being the attachment of the transducer
base to the point where the variation is to be measured. The motion of the measurand is inferred from the
relative motion (δ) of the mass (m) to that of the case (figure 5.2), and will depend on the size of the mass
and the stiffness of the spring (k) the magnitude of the damping being determined by the damper (C). The
seismic type of transducer is indispensable in the study of movements and vibrations in any form of
vehicle, it has a sensitivity-frequency characteristic similar to a recorder galvanometer (see figure 2.29),
and it may be shown that to avoid amplitude distortion when using a seismic displacement transducer ω
(the frequency of sinusoidal movement) must be greater than 2ω 0 (where ω0 is the internal resonant
frequency of the transducer), while for an accelerometer undergoing sinusoidal acceleration, the ratio
ω/ω0 must be less than 0.2.

In figure 5.3 the effects of damping magnitude on both of these characteristics are shown.
The relationships between sinusoidal displacements, velocities, and accelerations are shown in figure 5.4
from which

2 2
peak acceleration=ω S max m/s

2
ω S max
¿ ∈gs of acceleration
g

An accelerometer’s output performance is usually specified in terms of peak mV/g.

Transducer classification
This is an introductory text on instrumentation. However, because they form such an important part of any
instrumentation system it is desirable to examine the transduction methods used to convert physical
phenomena into electrical signals. Transducers may be classified either by the transduction method used,
or by the function they are capable of measuring, for example acceleration, displacement, etc.
In table 5.1 are tabulated the more common measurands and the possible methods of transduction. The
method of classification adopted in this

book is that of transduction method, it being considered that in a teaching/ introductory text this method
of classification is the more logical, whereas in an application-orientated book on transducers,
classification by use would be more appropriate. [1] It should be noted that in many cases the various
transduction methods have equal merit, and that each of the various transducer manufacturers, having
specialised on a particular transduction method, will tend to incorporate it in the majority of the
transducers that they make.
5.1 RESISTANCE CHANGE TRANSDUCERS

The methods of transduction involving a change of resistance are perhaps the easiest to understand and
will therefore be described first. The variable resistance method of transduction is not widely used but
serves to illustrate the underlying principles.

Consider figure 5.5 where the operation of this circuit as a position measuring system simply requires that
the sliding contact on the resistor be linked to the motion or displacement under observation. A change in
the measurand will cause a change in the resistance in the circuit and a consequent change in the current.
If the ammeter were calibrated in suitable units a continuous display of the measurand would thus have
been obtained. For such a transducer to operate satisfactorily the voltage must have a constant magnitude
irrespective of circuit resistance, so that the current in the circuit is only dependent on the variations in the
resistance R. Another drawback of this type of transducer is that if a plus and minus variation is to be
observed, current will be flowing in the circuit when the measurand is in the zero or reference position.

5.1.1 Potentiometric Transducers

The variable resistance transducer results in a current variation which is a function of the magnitude of
the measurand. In a potentiometric transducer the output is a voltage variation, that is, the resistive divider
is used as a potentiometer or potential divider (figure 5.6) and is in general a more satisfactory
arrangement. A potentiometric transducer should be supplied from a constant current source, the
transducer output being monitored by a high impedance instrument so that the loading effects across the
potentiometer are negligible. It will be seen that mechanically coupling the measurand to the sliding
contact of the potentiometer will result in an output voltage that will be proportional to the displacement
of the wiper from one end of the resistive path.
Since the resistance potentiometer may be used to convert a displacement into a proportional electrical
signal, the measurement of quantities such as flow, force, pressure, humidity, etc., can be performed by a
fixed reference potentiometric sensor when a suitable quantity to displacement converter is positioned
between the sliding contact and the measurand. Figure 5.7 illustrates some of the techniques used to
convert physical properties to displacements suitable for applying to the sliding contact.
To obtain measurements of velocity or acceleration from a fixed reference potentiometric transducer,
requires that the output voltage proportional to dis-placement is electrically differentiated once to obtain
velocity and differentiated twice to produce an electrical signal proportional to acceleration (see section
6.7.2). A signal proportional to acceleration can however be obtained directly from a seismic
potentiometric transducer (see figure 5.2) provided its mechanical resonant frequency ω 0 is much higher
than the frequency of the acceleration (ω), that is ω/ω0 < 0.2. Both the variable resistance and the
resistance potentiometer types of trans-ducer have a resistive element which may be wire wound,
deposited carbon film, platinum film, or a resistive compound such as conductive plastics. Since the
operation of an these transducers depends on the contact between a slider and the resistance element, their
life is related to the wear of this contact, and the frequency of operation limited to a few cycles of
displacement per second. The typical life expectancy of this type of transducer is around 3 × 106 cycles of
operation; they are comparatively inexpensive, and the circuitry associated with their use is simple
although instrument loading effects must be considered, it should also be remembered that if
displacement is the quantity being measured, a force has to be provided to overcome the friction of the
sliding contact and this may affect the magnitude of the observed displacement.

Example
A potentiometric displacement transducer with a 20 mm travel has a resistance of 10 kΩ. If it is connected
to a 10 V d.c. source of negligible output resistance, calculate the error in position if the output is
measured as 5 V (that is, ≡ 10 mm) on an instrument that has an input resistance of 20 kΩ. The circuit
appropriate to this example is figure 5.6. The wiper divides the potentiometer resistance into R and (10
kΩ — R), then
5.1.2 Resistance Strain Gauges
An important series of transducers is derived from the use of resistance strain gauges, and before
considering such transducers the principles and applications of resistance strain gauges should be
understood. If a length of electrical wire is subjected to a tensile force it will stretch, its length being
increased by an amount of δL (say), and provided that the elastic limit of the material is not exceeded, the
change in length is proportional to the load and the wire will revert to its original length when the load is
removed. Corresponding to this increase in length there will be a slight decrease in the cross-sectional
area of the wire (the increased length has to come from some-where), and since resistance of a conductor
= ρL/A, where p = resistivity of the material, L = length and A = cross-sectional area, the increase in
length and the decrease in area will both contribute to an increase in the resistance of the stretched wire.
In addition to the dimensional changes the resistivity of the material changes, this effect (termed
piezoresistivity) is small for metals but does make a contribution to the change in resistance. Obviously, it
is inconvenient to have long lengths of resistance wire attached to a test object so the general arrangement
of a strain gauge is one in which the resistance wire (which typically has a 0.025 mm diameter) is folded
into a grid and mounted on a backing of paper or Bakelite. A development from this is the foil gauge (see
figure 5.8), which is manufactured by techniques similar to those used in the production of printed
circuits. Such a process readily lends itself to the production of shaped, special purpose, strain gauges.
The size of a strain gauge will depend on the
intended application but resistance gauges are available in a variety of lengths from around 3 mm to 150
mm, there being a range of nominal resistance values, preferred magnitudes of which are 120 Ω and 600
Ω. It was stated above that the electrical resistance (R) of a straight wire will change when it is stretched;
let this change in resistance be δR. The definition of mechanical strain is the ratio ‘change in length/the
original length’ or δL/L and is denoted by the symbol ε. To relate these two quantities, a factor known as
the strain sensitivity is used such that:
δR per R δR / R
strain sensitivity= = (5.2)
δL per L ε

and has a characteristic value dependent on the type of resistance wire. Note: ε is often quoted as
microstrain (με), for example, micrometres per metre. When the wire is bent into a grid to form a strain
gauge there are small lengths of the wire at each bend that are no longer acted on by the strain parallel to
the axis of the gauge. This part of the gauge resistance is termed ‘dead resistance’ and the ratio of dead
resistance to total gauge resistance will depend on the type of fold and length of the gauge. When a gauge
is attached to an object that is stressed parallel to the axis of the strain gauge it may sense two strains: (a)
the principal or longitudinal strain which is in line with the gauge axis, and (b) the transverse or Poisson
strain which is at right angles to the gauge axis. If the gauge has appreciable dead resistance it will exhibit
a transverse strain sensitivity, which is typically 1-2 per cent of the longitudinal sensitivity; this effect
may be reduced in foil gauges by increasing the width of material at the bend (see figure 5.8). The output
resulting from this Poisson strain acts to oppose that from the principal strain, thus reducing the overall
gauge sensitivity.
The advantages of foil gauges over the wire gauges may be summarised as follows.
(a) The foil is thinner and therefore more flexible.
(b) The cross-sensitivity of the foil gauge is much lower due to the large cross- sectional area of the ends.
(c) The surface area of the conductor is much greater; it is therefore easier to dissipate the power loss in
the strain gauge.
(d) A better adhesion of the conductor to the backing is obtained.
(e) Large area terminal connections are easily provided.
(f) The photographic production techniques provide consistency in manufacture.
(g) The dimensions of the gauge may be made very small, enabling the measure-ment of localised strains
and strains on curved surfaces.

Gauge factor
The strain sensitivity of a manufactured strain gauge is termed the gauge factor, and is less than that of
the resistance wire used in its construction, as it will include the effects of the dead resistance, the Poisson
ratio of the surface to which the gauge is attached, and the transverse sensitivity. The majority of
resistance gauges have a gauge factor of around 2, copper-nickel wires giving values between 1.9 and 2.1;
however, iron-chromium-aluminium and iron- nickel-chromium alloys may be used to give a gauge factor
of 2.8 to 3.5.
Vibratory movement
When a strain gauge is used in a dynamic situation, for example, for the deter-mination of strains caused
by vibrations or movement, the frequency response of the gauge must be considered. Since a gauge will
average the strain over its active length, if the wavelength of the strain variations are of the same order of
magni-tude as the gauge length, considerable errors will be introduced. In the worst case, if the gauge
length is equal to the vibratory wavelength, the effective average strain and hence the output, will be zero
(see figure 5.9). If the frequency were increased above this point, output would again be obtained but its
ampli-tude could not be easily related to the measurand.
To determine the frequency of vibration at which zero output would occur requires the application of the
expression relating frequency and wavelength to
the velocity of sound, that is fL = u, where / is the frequency of vibration in the test object, L is the active
length of the strain gauge and u is the velocity of sound in the medium to which the strain gauge is
bonded. If, as is commonly the case, the gauge is attached to a steel member, in which u is 5 × 103 m/s
(approximately), the frequency for zero output is /= (5 × 103)/L Hz and if L = 5 mm, /= 1 MHz.
Alternatively if L = 150 mm, /= 33 kHz. These values of frequency are theoretical values, and in practice
(due to the effects of bonding) the frequency at which zero output occurs will be consider-ably less. To
obtain an error of 1 per cent or less in the output, the theoretical values should be divided by a factor of
20, giving practical upper frequency limits of 50 kHz and 1.6 kHz for the 5 mm and 1 50 mm gauges
respectively. Since instrument specifications normally quote frequency limits as 3 dB points (where
response is 30 per cent down) (see p. 303), these may be determined for the strain gauge by f = u/4L,
giving bandwidths B for the above gauges of 250 kHz and 8 kHz respectively. To obtain the bandwidth of
the measurement system—when the gauge is joined to a display instrument—the empirical expres-sion
for the system bandwidth (as shown in equation 1.42) is used, remembering that the resulting system
bandwidth should be divided by 5 to obtain a system frequency range over which the measurement errors
should be of the order of 1 per cent.
Temperature effects
Another factor which may affect the performance of the resistance strain gauge is that of temperature, and
its effects are threefold. First, the gauge filament will have a temperature coefficient of resistance, and
since this may be as large as 50 p.p.m. per °C for some of the copper-nickel alloys (constantan ± 20
p.p.m. per °C), it cannot be lightly ignored. The second temperature effect results from
the presence of dissimilar metals in the construction of the gauge, it being com-mon practice to form the
leads from nickel clad copper, and the thermal e.m.f. generated by various resistance wire materials at
their junctions with copper varies from about 2 μV/°C for manganin to about 47 μV/°C for constantan.
The third temperature effect results from the difference in the temperature coefficients of expansion of the
materials of the test object and the resistance wire of the gauge. For example, should the test object
expand (thermally) at a greater rate than the gauge wire, the latter will be stretched and to the measuring
instrument it will appear as if the test object is being stressed, this phenomenon being known as apparent
strain.
To compensate for these temperature effects a ‘dummy’ or nonactive gauge is often incorporated in the
measuring circuit. A dummy gauge is one which is nominally identical to the active gauge, but is attached
to a nonstressed piece of the test material in the same environment as the active gauge. The resistance
changes due to the temperature effects can then be made to counterbalance each other in the measuring
circuit.

5.L3 Measuring Circuits

The circuit commonly used in strain gauge work is some form of d.c. Wheatstone bridge circuit. Since the
variations in strain are dynamic the bridge circuit is operated in the ‘unbalanced condition’ (see section
3.3.4), that is, an output proportional to the variation in the resistance of the active gauge is obtained. To
increase the sensitivity, two or four active gauges may be used, in which case the need to use dummy
gauges is removed.
Example
An unbalanced Wheatstone bridge arrangement incorporating two active gauges is to be used to monitor
the strain in a cantilever. If the unstrained resistance of each gauge is 120 Ω, the bridge is completed with
two further 120 Ω resistors, and the gauges (which have a gauge factor of 2) are subject to a strain of 100
microstrain; calculate the output voltage from the bridge when it is supplied from a 10 V d.c. source.
Now the gauge factor is (δR/Rε) or δR = 2 × 100 × 120/106. Hence δR = 0.024 Ω. For an equal arm
bridge with two active arms (increments of opposite sign)

This output signal of 1 mV is fairly typical of the signal levels encountered in operating strain gauges, and
shows a possible amplification requirement.
Bridge Balancing
When strain gauges are used to form a bridge circuit it is unlikely that the bridge will be balanced in the
zero or no-load condition. The unstrained gauges are un-likely to have exactly the same resistance;
additionally, unequal strains are likely to occur when the gauges are mounted. It thus becomes necessary
to add a

variable component to the bridge circuit so as to provide for this initial balance. Figure 5.10 illustrates
some of the methods used for this purpose. The series resistance arrangement is only satisfactory if one or
two strain gauges are being used, whereas the apex and parallel balancing arrangements are both suitable
for use with two or four active gauge circuits. For the apex circuit Rv should be small, perhaps 10 to 25
per cent of the strain gauge resistance. However, for the parallel circuit if 2R + AV is made too small a
reduction in the sensitivity of the bridge results. Combining the parallel balancing network with an
operational amplifier as shown in figure 5.11 produces a practical solution to the small signal and initial
unbalance problems.

Lead resistance
In many applications the measuring/display instruments will be remote from the
active gauges, and the effect of the wires connecting the active gauges to the remainder of the circuit is
equivalent to inserting a small resistance in series with it. This resistance will vary due to temperature
effects and to reduce its influence additional connections as given in figure 5.12a and b are required and
then, the lead resistances are approximately self-cancelling.
The above discussion of strain gauge measuring techniques has assumed that all methods use a d.c.
supply. While the tendency is towards such methods due to the development of sensitive, drift free d.c.
amplifiers, digital voltmeters and data logging system, the a.c. carrier system has not completely
disappeared from use. The a.c. system had a considerable advantage in its stability of amplification, but
stray capacitance and the need to balance its effect from the measuring circuits are a considerable
disadvantage. Figure 5.13 illustrates the general principles involved in an a.c. carrier system of
measurement.
Strain Gauge Attachment
In the majority of strain gauge applications it is necessary to attach the gauge to the test object by means
of an adhesive. The thickness of the adhesive layer should be made small to avoid errors in transmitting
the size changes in the test object to the strain gauge. This is best performed by using a clamp, curved
metal plate, rubber pad and nonadhesive foil film as indicated in figure 5.14, the excess adhesive being
squeezed out from under the centre of the gauge. An alternative

method is to stick the gauge upside down to a piece of adhesive tape, then having aligned the gauge with
its desired position, start at one end and press it firmly on to the test object. While satisfactory results are
obtainable by using this method it is not as reliable as using a clamp. Having attached the gauge to the test
object it may be desirable that the guage is covered with a waterproofing agent to prevent it from the
effects of a moist environment. Waterproofing compounds are available for this purpose but many do not
adhere to the PVC insulation of leads commonly used to connect the gauge to the measuring instrument.
To overcome this difficulty clamps or crimped-on sleeves may be used over the connection leads.

Applications
The uses of resistance strain gauges are almost limitless but their direct uses may be summarised as
applications involving the measurement of stress and strain in existing structures such as aircraft, railway
wagons, bridges, cranes, reinforced concrete, automobiles, buildings, etc. It is usually necessary to
investigate a large
number of points and ease of attachment and connection is often a very import-ant factor. This type of
application of the resistance strain gauge is usually in connection with research or development
investigations. They are also extensively used in sensors for monitoring, and in control systems, where
they form the active part of a transducer.
5.1.4 Resistance Strain Gauge Transducers
To operate, a strain gauge transducer requires that the phenomenon under investigation shall first be
converted into mechanical strain, that is, the gauge is attached to an elastic member within the transducer,
which is subjected to a force proportional to the measurand. If the force is small, for example, as in the
measurement of small pressure differences, an unbonded strain gauge system may be adopted where the
strain gauge wire itself acts as the elastic member.

Figure 5.15 shows such a sensing element, which may be incorporated in a wide range of pressure
transducers simply by varying the area or thickness of the diaphragm that operates the force rod. This
type of assembly contains four resistance strain gauges, two being relaxed and two stretched when a force
is applied to the spring element. The four gauges are connected to form a Wheat-stone bridge, unbalance
in the bridge arms in the unloaded condition being compensated for during construction by the addition of
small resistances in the appropriate arms. In a good quality unbonded strain gauge transducer, consistency
of operation is ensured by temperature and pressure cycling during manufacture,, the sensing unit being
sealed in a container of dry helium, quality checks being
conducted at each stage of manufacture of the transducer. The use of unbonded strain gauge transducers is
not limited to pressure gauges—figure 5.16 illustrates its application in the construction of a seismic
accelerometer. Since many physical
phenomena may be converted to a variation of force, a strain gauge element of the type shown in figure
5.15 could be incorporated into transducers to measure such properties as weight, temperature, humidity,
flow and viscosity.
The use of strain gauges in the construction of transducers is by no means limited to unbonded gauges;[4]
common examples of the use of bonded gauges in transducer construction are in the measurement of
force (load cells) figure 5.17 and torsion, figure 5.18. A more recent development has been the applica-
tion of thin film techniques to the manufacture of transducers. [5]

5.2 REACTANCE CHANGE TRANSDUCERS

This is a group of transducers in which a displacement is used to modulate either a capacitive or inductive
reactance. Since variations in reactance may only be measured when the supply is an alternating voltage
(or current) the excitation source for reactance change transducers must be alternating. This may,
however, make the transducer incompatible with the remainder of the measuring system, and to overcome
this a number of reactance change transducers are manufactured with an internal d.c. to a.c. converter and
an output demodulator so that both the transducer input and output are d.c. while the internal excitation is
a.c. [1]
5.2.1 Capacitance Variation
It has been indicated that a change in almost any physical phenomenon may be converted to a
displacement and as the capacitance between two parallel con-ducting plates is approximately
proportional to εA/d, (where ε is the dielectric constant of the material between the plates of area A,
separated by a distance cI) varying any of these quantities will result in a change in the capacitance
between the plates.
Variable dielectric
Figure 5.22 shows diagrammatically two types of dielectric variation transducer, one in which a dielectric
sleeve is slid between coaxial electrodes to vary capaci-tance, the other being a device for measuring the
height of a liquid in a container.
In both cases it is important that the dielectric constant of the material between the electrodes is
substantially different from that of air so that an appreciable change in capacitance is obtained.

Variable plate area

Transduction by variation of plate area may be performed by arrangements similar to those shown in
figure 5.23, where linear or angular movement of the actuator causes one plate, or a set of plates, to vary
the capacitance of the transducer.

Variable plate separation

The third form of variation, namely that involving a change in the electrode separation (see figure 5.24) is
perhaps the most widely used for the construction of such devices and is comparatively simple, as the
forces required to move one plate in relation to another may be made exceedingly small, and the
alignment of

the fixed and moving parts may not be as critical as it is when some other forms of transducer are used.
Figure 5.25 illustrates the use of this technique to produce a capacitance strain gauge, a major advantage
of this being its potential use at elevated temperatures. [8,9]

Measuring techniques
The measuring circuitry associated with capacitance change transducers usually involves some form of
bridge. The simplest form would have two resistive and two capacitive arms. A slightly more
sophisticated form of bridge is shown in figure 3.28, where a ratio transformer is incorporated together
with a resistance circuit to compensate for any leakage resistance effects associated with the capacitance
of the transducer. Either circuit may be used in the balanced or un-balanced condition provided that for
the latter mode of operation the unbalance is less than 10 per cent.
Some capacitance change transducers are produced with internal signal con-ditioning equipment which
uses a d.c. input to supply an oscillator, for operation of the capacitance bridge, and conversion of the
bridge output to d.c. Alternat-ively an oscillator circuit may be used so that a change in capacitance
causes a change in output frequency.

5.2.2 Inductance Variation

The inductance of a coil depends on the manner in which magnetic flux links its turns, thus by using
suitable ‘measurand to displacement’ converters to cause a change in the magnetic flux linkages in a coil,
a series of transducers exhibiting a change in inductance proportional to the measurand may be obtained.
This change in inductance can be measured either as an amplitude change in a bridge balance, or as a
change of resonant frequency in an oscillatory circuit. The methods by which this variation of inductance
(flux linkages) is obtained are as follows.

(a) By changing the reluctance (or magnetic resistance) of the flux path by the movement of an
armature, so that the displacement may be either linear or angular (see figure 5.26) this being
incorporated in transducers to measure such quantities as pressure, acceleration, force, and
displacement or position.
 (b) By the displacement of a slug of material having a high magnetic permeance that is constrained to
move inside a centre-tapped coil which is wound on a ferromagnetic core (see figure 5.27). When
the slug is positioned centrally the inductance of the two halves will be equal; on movement to
one side the inductance of one half will increase while that of the other decreases.

The inductor may form the arms of a bridge circuit, or be incorporated in an oscillator circuit (see figure
5.28a and b). The phase sensitive detector (see section 2.1.1) is included so that the d.c. output has a sign
corresponding to the direction of the displacement. The phase shift circuit between the oscillator and the
p.s.d. is necessary because at zero displacement there will be a phase angle between the reference voltage
and the centre tap on the inductor (see also p. 316).
5.2.3 Linear Variable Differential Transformer (L.V.D.T.)
In this type of transducer, the change of inductance is a variation of mutal coupling between windings
rather than a variation of self inductance. This varia-tion of mutual inductance is obtained by the
movement of a ferromagnetic core within a coil arrangement of the type shown in figure 5.29. There is
usually one primary and two secondary coils. For a linear device the coils are wound on a nonmagnetic
and insulating former while in angular devices of this type a ferro-magnetic core for the coils may be
used. The operational circuits used for the differential transformer range from those such as that in figure
5.30a which has an a.c. output voltage that increases in magnitude for displacement either side of the
centre or zero position. This circuit can easily be modified so that a d.c. output is obtained having a
polarity dependent on the direction of displacement (see figure 5.30b). More complex arrangements are
available commercially, in which a d.c. input is converted to an a.c. voltage, that is, varied by the
transducer operation and then restored to d.c. so that the output voltage has a magnitude and direction
proportional to displacement.

The differential transformer transducer is a durable, low impedance device, having little or no friction,
and capable of measuring total, steady state, and dynamic displacements from 100 μm to 25 mm. It has
become extensively used for measuring a wide range of quantities by positioning a suitable ‘measurand to
displacement’ converter between the quantity and the differential transformer.

'Synchro'

A type of transducer used for the measurement and display of angular position or movement is the
‘Synchro’, in which two similar units are used, one termed the transmitter and the other the receiver. Each
is constructed with a two-pole rotor and a stator having three windings distributed 120° apart (see figure
5.31).

In operation the rotors of both units are fed, via slip rings, with an alternating current (usually 50 or 400
Hz) and if the rotors of the two units are in the same relative position to the stator windings no currents
will flow between them. However, should the rotor of the transmitter be moved relative to its stator
winding, the voltages induced in the two stators will be different and currents will flow in the connecting
leads. These currents will produce a torque on the rotor of the receiver and cause it to move in sympathy
with the rotor of the transmitter. The positioning accuracy of such a system is limited by the friction of
the bearings in the receiver, and the calibration of the dial face attached to it.

5.4.1 Electromagnetic Transducers

To provide self-generation the magnetic flux in electromagnetic transducers is usually derived from a
permanent magnet, while the coil system may be air cored or wound on a core of ferrous material.

Linear velocity transducer

The simplest form of electromagnetic transducer is one in which a permanent magnet attached to an
actuating shaft is free to move in a cylindrical coil, move-ment of the object/actuating shaft/magnet
causing a voltage to be generated at the coil terminals; the magnitude of the voltage will be proportional
to the velocity of the movement. A variation on this arrangement is shown in figure 5.36 where the
permanent magnet is supported between springs, and fitted with low friction bearing rings. If such a
transducer is attached to an object that is vibrating at a frequency in excess of the low natural resonant
frequency of the transducer, the magnet would appear to an observer to be remaining stationary in space
while the case and coil oscillate around it, resulting in an output volt-

age whose magnitude would be proportional to the amplitude of the vibration, with a frequency equal to
that of the vibration. The relationships between sinusoidal displacement, vibration, and acceleration are
shown in figure 5.4.
Instead of the magnet being the moving member it is possible for the coil to be moved, common examples
being (a) some loudspeakers in which the coil has a linear movement, and (b) a moving coil instrument
where the motion is rotary.
Angular devices
The use of transducers to determine angular motion has been widespread for many years, for it includes
devices such as the d.c. tachometer or tachogenerator, in which the stator’s magnetic field is derived
either from a permanent magnet or from a separately excited field coil, while the rotor winding (armature)
is of the form normally used in a d.c. generator fitted with a commutator. The output voltage of a
tachogenerator is proportional to rotary speed (or angular velocity) it being about 5 V per 1000 rev/min
for the permanent magnet tachometer, and around three times this value for the separately excited
transducer. In both forms the polarity of the voltage is dependent on the direction of rotation.
If the rotor contains the permanent magnet field, then the magnetic inter-action between it and the stator
coils produces an a.c. voltage at the stator terminals having a magnitude and frequency which are both
proportional to rotor speed. In many applications it is advantageous to operate with variations of fre-
quency rather than voltage level, since frequency is unaffected by circuit impedance, loading, and
temperature effects.
Toothed rotor tachometer
Tachometers using a toothed rotor made of a ferromagnetic material, and a trans-duction coil wound
around a permanent magnet are probably the commonest form of frequency output angular velocity
transducers. Figure 5.37 illustrates the principle of operation of such a transducer, in which the magnetic
field sur-rounding the coil is distorted by the passing of a tooth causing a pulse of output voltage from the
coil. If a multi-toothed rotor is used, a pulse is generated for each tooth that passes the magnet, and if six
teeth are used the frequency of the output in Hz will be 6 × rev/min/60 or rev/min/10. The r.m.s. value of
the out-put voltage will increase with a reduction in the clearance between the rotor and the pickup, with
an increase in the rotor speed, and with an increase in tooth size.

5.4.2 Piezoelectric Transducers

These devices utilise the piezoelectric characteristics of certain crystalline and ceramic materials to
generate an electrical signal. The piezoelectric effect was discovered by Pierre and Jacques Curie in 1880
when they found that by placing
weights on to a quartz crystal an electrical charge could be generated. Subsequent researchers have shown
that there are about 40 crystalline materials which when subjected to a ‘squeeze’ (the Greek word
‘piezein’ means to squeeze) generate an electrical charge. The Curie brothers also discovered the reverse
effect, that is, a dimension change is obtained when an electric stress is applied to a quartz crystal.
The crystal materials are of two basic types, natural crystals and synthetic crystals, the latter type being
mainly ceramic ‘crystals’ of which barium titanate was the first to be commercially used. The addition of
controlled impurities such as calcium titanate was found to improve some of the ‘crystal’s’
characteristics. Research on this and other synthetic ‘crystals’ has led to them being used more frequently
than natural crystals in the production of piezoelectric transducers.
The force applied to the crystal may be that of shear, compression, or bend-ing, depending on the
particular application for which the transducer has been designed. The merits of operating the crystal with
the various force modes will not be discussed here—they are described in detail in reference 1. It must,
how-ever, be remembered that since piezoelectric transducers belong to the self-generating category their
use is predominantly in dynamic applications, a common form being as seismic accelerometers, although
the piezoelectric principle may be effectively used in any situation where the measurand can be converted
to a force that results in the stressing of a suitable crystal.
All piezoelectric crystals have a high output impedance, and for this reason the instrumentation to which
the piezoelectric transducer is connected must have a high input impedance, or for most applications a
charge amplifier (see section 6.7.1) must be connected between the transducer and the display instru-
ment. A disadvantage is the additional expense involved. The equivalent circuit of a piezoelectric
transducer is shown in figure 5.38, and this in practice may for simplicity be considered as a voltage
source whose series impedance is a capaci-tor having a value of a few picofarads. The internal resistance
(Ri) will normally
exceed 20 GΩ and can thus be ignored when considering the overall performance of the transducer;
similarly the effects due to the lateral inductance are beyond the upper frequency limit of the transducer
and can also be ignored.
5.4.3 Thermoelectric Transducers
In 1822 John Seebeck reported that if a magnetic needle was held near a circuit made from conductors of
two different materials, the magnetic needle was deflected when heat was applied to part of the circuit,
thus implying the presence of an electric current. In 1834, a French watchmaker, Jean Charles Peltier,
found
that when a current was passed through a junction of two different conductors it changed temperature. In
1854 William Thompson (Lord Kelvin) discovered that if a current carrying conductor is heated at one
point along its length, points at equal distances on either side of the heat source along the length of the
con-ductor will be maintained at different temperatures, points on the negative potential side of the heat
source being lower in temperature than those on the positive potential side.
Thermocouple
The Seebeck effect, which is a combination of the Peltier and Thompson effects, is utilised in what is
probably the most widely used method of thermometry, namely the thermocouple. These transducers
consist of a pair of bars or wires of dissimilar metal joined at both ends: one end is used as the hot
(sensing) junction, while the other end is used as a cold or reference junction, as shown in figure 5.39.
The term thermo

couple has through usage come to mean a single thermojunction, consisting of two lengths of dissimilar
wire, insulated from each other but joined at one end. In conforming with this usage it becomes necessary
to describe the combination of a ‘hot’ and ‘cold’ junction as a thermocouple circuit.
A suitable cold junction may be obtained by the use of melting ice, but such a reference is not always
convenient or for that matter necessary; for example if a thermocouple is being used to measure a
temperature that is of the order of 1000 K, variations of 5 °C or so in the temperature of the reference
junction will have only a small effect on the accuracy of the measurement. Thus ambient air temperature
may satisfactorily be used as a reference for some applications.

Output measurement
A thermocouple may be used as a voltage source, when its output should be measured by a high
impedance instrument, for example, a digital voltmeter or a d.c. potentiometer, the displayed millivolts
being dependent on the unknown temperature. Alternatively a low impedance instrument may be used to
com-plete the circuit as in figure 5.40 when the magnitude of the current in the circuit will be dependent
on the magnitude of the unknown temperature. When near ambient temperatures are to be measured with
a thermocouple and it is in-convenient to use a fixed temperature reference junction, compensating
circuits must be incorporated in the measuring system. One method of arranging auto-

matic compensation is shown in figure 5.41 where a temperature-sensitive bridge is included in the
thermocouple circuit, such that variations in the ambient tem-perature level are compensated for by the
changes in the resistance Rc and the compensating junction. Multi-channel thermocouple systems
sometimes use a heated, thermostatically controlled reference junction (see figure 5.42) where by the use
of regulated temperature control of the reference junction, a system accuracy of ± 0.3 °C may be
maintained. When the reference junction is not held at 0 °C, the observed value must be corrected by
adding to it a voltage that would have resulted from a temperature difference equal to the amount by
which the reference junction is above 0 °C, that is ET = (Et + Er) where ET is the total e.m.f. at
temperature Tf Et is the e.m.f. due to the temperature differ-ence between the sensing and reference
junctions, and Er is the e.m.f. due to the temperature of the reference junction above 0 °C. Since the
temperature-voltage characteristic of a thermocouple is nonlinear, it is important that temperatures
are determined by the above process (rather than by converting an e.m.f. to a temperature and then adding
this to the ambient temperature), also the reference junction should be maintained at, or near, the
temperature used during calibra-tion of the thermocouple or errors will result.

Thermocouple materials
The materials commonly used to form thermocouples are
Type T: copper-constantan 0-370 °C (270-640 K)
Type J: iron-constantan 0-760 °C (270-1030 K)
Type K: nickel/chromium-nickel/aluminium, 0-1260°C (270-1500K)
Type—: tungsten-rhenium, up to 2760 °C (3000 K) and
Type R: platinum-rhodium alloy 13% which may be used up to 1750 K, but is
also used to define the International Temperature scale from 800-1340 K (see also reference 10)
The conductors of a thermocouple must be insulated from each other, and except at low temperatures, it
has become the practice to use mineral insulation
and sheath the conductors and insulator with stainless steel. It is possible to purchase thermocouples in a
variety of sizes, from 0.25 mm to 3.0 mm dia., and shape of sensing junction (see figure 5.43). In addition
to this the locating of the measuring junction is important, considerable care being necessary to ensure
that the temperature recorded is the unaffected temperature of the measuring point.
When several thermocouples made from the same materials are connected in series the hot junctions all
being at one temperature and the cold junctions at another, they are said to form a thermopile. The output
voltage from a thermo-pile is equal to the output voltage from a single thermocouple multiplied by the
number of thermocouples in the thermopile assembly, provided that the materials are connected in the
correct sequence and the reference junctions are all kept at the same temperature. If the thermopile is
arranged in the configuration shown in figure 5.44 so that the hot junctions are at the focal point of an
optical system.

then the device so produced forms the sensor for a thermal radiation pyrometer, and may be used to
determine the temperatures of heated surfaces without physical contact. [11] The advantages of
noncontacting methods of temperature measurement have led to the development of infrared
thermometers which use vacuum thermo-couples, metal film bolometers, or photodiodes as the detector
of radiant energy. Such instruments vary in complexity from portable, hand held thermometers,which
may be used to determine the temperature at a point on the surface of an object, to the sophisticated
thermal imaging systems in which the focal point of measurement is scanned across the surface of the
object under investigation. The output from the pyrometer head in this type of instrument is used to
modulate the brightness of a spot that is scanning the screen of a c.r.t. enabling a ‘television type’ picture
of temperature variations to be built up. [12] This type of equip-ment can be used to detect temperature
differences as small as 0.1 °C and has many applications both in the industrial and medical fields.
5.5 ULTRASONIC TRANSDUCERS
These devices were originally developed using the magnetostrictive properties of ferromagnetic materials.
This characteristic relates to the changes in shape of a ferromagnetic material caused by the variation in
alignment of the domains when such a material is subjected to an alternating magnetic field. The
phenomenon is reversible, that is, if a ferromagnetic material is extended and compressed in a regular
manner an alternating voltage will appear at the terminals of a coil wound around it. This property is used
in the construction of ultrasonic transducers [13] in, for example, underwater (sonar) detection.
The development of piezoelectric materials has resulted in a considerable extension of the use of the
ultrasonic principle for many applications such as flow, thickness, fault detection and intruder detection
systems. In general an ultrasonic transducer consists of two piezoelectric crystals, both in contact with the
medium to be measured, but separated from each other by a short distance. One crystal acts as the
transmitter of an ultrasonic pulse and the other receives it after a delay dependent on the geometry of the
conditions in which the trans-ducer is operating. By detecting the phase displacement between the
transmitted and received signals the magnitude of measurand can be established.
5.6 DIGITAL TRANSDUCERS
True digital transducers are those whose output is represented by a number of discrete increments. The
term is generally used, however, to include those devices which have a pulse output that can be applied to
a digital counter. Examples of this latter type are the toothed rotor tachometer, and a similar effect
obtained by using photoelectric transduction, for example, figure 5.45 where light shines through a
segmented disc or plate which, when in motion, modulates the light falling on to the light sensor. The
devices that generate a single pulse train can-

not, in general, give an indication of direction, but this difficulty may be over-come for the toothed rotor
tachometer by using two sensing heads positioned relative to each other so that their electrical outputs
have a phase displacement that will be lagging or leading depending on the direction of movement of the
rotor. In the case of the photoelectric transducer, determination of direction
may be performed by the addition of a second track of opaque segments with its own illuminating sensor.
A higher degree of resolution in photoelectric trans-ducers may be obtained by using interference pattern
techniques derived from the use of optical gratings. [3, 14] A number of different pattern types exist, but
the most widely known is the ‘moiré’ pattern. To obtain this a strip with altern-ate parallel opaque and
transparent segments is moved past a similarly striped plate that is set at an angle to the moving plate.
Positioning a line of light sensors on one side of the stripes and shining light through the stripes on to
them results in a variation in the particular sensors illuminated, or obtains considerable varia-tions in the
sensor’s output level with the movement of the striped strip. The outputs from such arrangements will
approximate to a sinewave, the frequency of which will be dependent on the number of segments on the
strips.

Digital encoder
The above transducers are all pulse producing devices, capable of detecting a change in the measurand
rather than describing its absolute or unique state. In addition to this latter requirement a true digital
transducer should have an out-put that is in a form suitable for direct entry into a digital computer or data
handling system. This latter type of transducer is known as a digital encoder and may read continuously
or be momentarily stopped to obtain a readout (depending on the design and code used). Encoders are
generally an integral part of a more complex transducer, for example, a wind direction indicator.
While digital encoders may be made for linear applications, one of the com-monest is the shaft position
encoder. It consists of a disc or drum with a digitally coded scale which may be formed either from a
combination of translucent and opaque segments, or a combination of conducting and insulating surfaces,
which are coded so that there exists a unique form for each discrete position on the disc as shown in
figure 5.46. Such discs are manufactured with diameters from 50 to 250 mm and give unique codes for
between 100 and 50 000 positions per
360°. These coded positions are read either by a senes of brushes or by a light source and sensor
arrangement, and fed into an appropriate number of channels. The sequence and order of indications
represent the shaft in the coded form. A special code known as the Gray code is commonly used in the
operation of these devices.

WORKING PRINCIPLE OF A THERMOCOUPLE

A typical circuit diagram of a thermocouple is shown in Figure. In the Figure, two dissimilar
metals ‘A’ and ‘B’ are joined at the two junctions ‘P’ and ‘Q’. Here the ‘P’ junction is measuring
junction or hot junction whereas the junction ‘Q’ is the reference junction or cold junction. And a
PMMC instrument is connected in this arrangement as shown in Figure 1
When these junctions are kept at different temperatures, generally cold junction is kept at 0oC
and measuring junction is kept at an unknown temperature which is to be measured (i.e. the
temperature of the junction is raised by heating it). An e.m.f. will be generated in this circuit due
to the temperature difference of the junctions. This e.m.f. is in the order of millivolts. And the
e.m.f. can be measured with the help of a PMMC instrument by connecting it in the circuit as
shown in Figure. When both the junctions are at the same temperature, e.m.f. generated at both
junctions will be the same. No current will flow through the circuit. And there will be no
deflection in the meter. When both the junctions are at different temperatures, a current will flow
through the meter. And the meter will show the deflection. As the generated e.m.f. is
proportional to the temperature difference, the amount of current flow will also be proportional
to the temperature difference. And therefore, the meter can be calibrated directly in terms of
temperature.

WORKING PRINCIPLE OF A THERMISTOR

The thermistor works on the simple principle of change in resistance due to a change in
temperature. When the ambient temperature changes the thermistor starts self-heating its
elements. its resistance value is changed with respect to this change in temperature. This change
depends on the type of thermistor used. The resistance temperature characteristics of different
types of thermistors are given in the following section.