Temperature Measurement
Temperature Measurement
Example 8.8 Calculate the length and volume for a 200 cm on a side copper cube at
20°C, if the temperature is increased to 150°C.
−6
New length = 200(1 + 9.4 × 10 × [150 − 20] × 9/5)
= 200(1 + .0022) = 200.44 cm
3 −6
New volume = 200 (1 + 29 × 10 × [150 − 20] × 9/5)
3 3
= 200 (1 + .0068) = 8054400 cm
In a gas, the relation between the pressure, volume, and temperature of the
gas is given by
P1V1 P2V 2
= (8.13)
T1 T2
8.4.1 Thermometers
Mercury in glass was by far the most common direct visual reading thermome-
ter (if not the only one). The device consisted of a small bore graduated glass tube
with a small bulb containing a reservoir of mercury. The coefficient of expansion
128 Chapter Eight
Figure 8.3 Shows (a) the effect of temperature change on a bimetallic strip and (b) bimetal-
lic strip thermometer.
Temperature and Heat 129
Liquid filled thermometer works on the same principle as the liquid in glass
thermometer, but is used to drive a Bourdon tube. The device has good linear-
ity and accuracy and can be used up to 550°C.
Vapor-pressure thermometer system is partially filled with liquid and vapor
such as methyl chloride, ethyl alcohol, ether, toluene, and so on. In this system
the lowest operating temperature must be above the boiling point of the liquid
3000
Bourdon 2500
Vapor Pressure kPa
Tube
2000
1500
1000
Bulb 500
0
0 10 20 30 40 50 60 70 80 90 100
Temperature °C
(a) (b)
Figure 8.4 Illustrates (a) pressure filled thermometer and (b) vapor pressure curve for
methyl chloride.
130 Chapter Eight
Example 8.9 What is the resistance of a platinum resistor at 250°C, if its resistance
at 20°C is 1050 Ω?
Resistance at 250°C = 1050 (1 + 0.00385 [250 − 20])
= 1050 (1 + 0.8855)
= 1979.775 Ω
Material Coeff. per degree Celsius Material Coeff. per degree Celsius
8.4.4 Thermistors
Thermistors are a class of metal oxide (semiconductor material) which typically
have a high negative temperature coefficient of resistance, but can also be pos-
itive. Thermistors have high sensitivity which can be up to 10 percent change per
degree Celsius, making them the most sensitive temperature elements available,
but with very nonlinear characteristics. The typical response times is 0.5 to 5 s
with an operating range from −50 to typically 300°C. Devices are available with
the temperature range extended to 500°C. Thermistors are low cost and manu-
factured in a wide range of shapes, sizes, and values. When in use care has to be
taken to minimize the effects of internal heating. Thermistor materials have a
temperature coefficient of resistance (a) given by
∆R 1
α= (8.15)
RS ∆T
where ∆R is the change in resistance due to a temperature change ∆T and RS
the material resistance at the reference temperature.
The nonlinear characteristics are as shown in Fig. 8.5 and make the device dif-
ficult to use as an accurate measuring device without compensation, but its sen-
sitivity and low cost makes it useful in many applications. The device is normally
used in a bridge circuit and padded with a resistor to reduce its nonlinearity.
8.4.5 Thermocouples
Thermocouples are formed when two dissimilar metals are joined together to
form a junction. An electrical circuit is completed by joining the other ends of the
dissimilar metals together to form a second junction. A current will flow in the
circuit if the two junctions are at different temperatures as shown in Fig. 8.6a.
10000
1000
Resistance Ω
100
10
0
0 50 100 150 200 250 300 350
Temperature °C
Figure 8.5 Thermistor resistance temperature curve.
132 Chapter Eight
Figure 8.6 (a) A thermocouple circuit, (b) thermocouples connected to form a thermopile,
and (c) focusing EM rays onto a thermopile.
The current flowing is the result of the difference in electromotive force developed
at the two junctions due to their temperature difference. In practice, the voltage
difference between the two junctions is measured; the difference in the voltage is
proportional to the temperature difference between the two junctions. Note that
the thermocouple can only be used to measure temperature differences. However,
if one junction is held at a reference temperature the voltage between the ther-
mocouples gives a measurement of the temperature of the second junction.
Three effects are associated with thermocouples. They are as follows:
In practice, the Seebeck voltage is the sum of the electromotive forces gener-
ated by the Peltier and Thompson effects. There are a number of laws to be
observed in thermocouple circuits. Firstly, the law of intermediate tempera-
tures states that the thermoelectric effect depends only on the temperatures
of the junctions and is not affected by the temperatures along the leads. Secondly,
the law of intermediate metals states that metals other than those making up
the thermocouples can be used in the circuit as long as their junctions are at the
same temperature, i.e., other types of metals can be used for interconnections
Temperature and Heat 133
and tag strips can be used without adversely affecting the output voltage from
the thermocouple. The various types of thermocouples are designated by letters.
Tables of the differential output voltages for different types of thermocouples
are available from manufacturer’s thermocouple data sheets. Table 8.6 lists
some thermocouple materials and their Seebeck coefficient. The operating range
of the thermocouple is reduced to the figures in brackets if the given accuracy
is required. For operation over the full temperature range the accuracy would
be reduced to about ±10 percent without linearization.
Thermopile is a number of thermocouples connected in series, to increase the
sensitivity and accuracy by increasing the output voltage when measuring low
temperature differences. Each of the reference junctions in the thermopile is
returned to a common reference temperature as shown in Fig. 8.6b.
Radiation can be used to sense temperature. The devices used are pyrome-
ters using thermocouples or color comparison devices.
Pyrometers are devices that measure temperature by sensing the heat radiated
from a hot body through a fixed lens that focuses the heat energy on to a thermopile;
this is a noncontact device. Furnace temperatures, for instance, are normally meas-
ured through a small hole in the furnace wall. The distance from the source to the
pyrometer can be fixed and the radiation should fill the field of view of the sensor.
Figure 8.6c shows the focusing lens and thermocouple set up in a thermopile.
Figure 8.7 shows plots of the electromotive force (emf) versus temperature of
some of the types of thermocouples available.
8.4.6 Semiconductors
Semiconductors have a number of parameters that vary linearly with temperature.
Normally the reference voltage of a zener diode or the junction voltage variations
are used for temperature sensing. Semiconductor temperature sensors have a lim-
ited operating range from –50 to 150°C but are very linear with accuracies of ±1°C
or better. Other advantages are that electronics can be integrated onto the same
die as the sensor giving high sensitivity, easy interfacing to control systems, and
making different digital output configurations possible. The thermal time con-
stant varies from 1 to 5 s, internal dissipation can also cause up to 0.5°C offset.
Semiconductor devices are also rugged with good longevity and are inexpensive.
For the above reasons the semiconductor sensor is used extensively in many appli-
cations including the replacement of the mercury in glass thermometer.
134 Chapter Eight
70 E Type
60
J Type K Type
EMF Output (mV)
50
40
30
R Type
20
T Type
S Type
10
0
0 400 800 1200 1600
Temperature °C
Figure 8.7 Thermocouple emf versus temperature for various types.
8.5.1 Selection
In process control a wide selection of temperature sensors are available.
However, the required range, linearity, and accuracy can limit the selection. In
the final selection of a sensor, other factors may have to be taken into consid-
eration, such as remote indication, error correction, calibration, vibration sen-
sitivity, size, response time, longevity, maintenance requirements, and cost.
The choice of sensor devices in instrumentation should not be degraded from a
cost standpoint. Process control is only as good as the monitoring elements.