06-Measurement of Temperature
06-Measurement of Temperature
06-Measurement of Temperature
Measuring of Temperature
Introduction
Not only are there several ways of measuring temperature, but there is more than
one scale on which it is measured. Two scales are in everyday use: the Celsius
(centigrade) scales ( oC) and the Fahrenheit scale (oF). There is also a third, the
thermodynamic scale or absolute scale, in which the unit of temperature is the Kelvin
(K).
A simple way to establish a temperature scale would be to take any device for
sensing temperature. Label the points 0 to 100 on the scale by measuring the
temperature of freezing and boiling water respectively (at a pressure of one standard
atmosphere). And then divide the scale into equal intervals, extrapolating for higher
and lower temperatures.
The trouble with this approach is that the readings given by differently constructed
instruments, measuring the same temperature would differ somewhat at practically
Olatunbosun, A Page 1 of 16 Univation
Shell Nigeria Graduate Training Programme Instrumentation
Measuring of Temperature
all points other than 0 and 100. Moreover, none of the readings on scales would be
in perfect agreement with the Celsius scale.
The reason for this is that none of the properties of materials used for temperature
sensing varies strictly linearly with temperature. This creates a calibration problem
for instrument manufacturers. To enable instruments to be calibrated in terms of the
Kelvin or Celsius scale, the International Practical Temperature Scale (IPTS) was
devised.
In the IPTS, values of temperature are assigned to the eleven reproducible 'fixed
points' shown in Table 1. Standard instruments, chosen for their repeatability and
stability, are used at temperature between the fixed points. Designated interpolation
equations are used to determine the temperature from the output of each standard
instrument. In this way the IPTS is made to agree with the Kelvin scale as closely as
the current state of measurement technology will allow.
Figure 1 shows the step response of the thermometer. Figure 1(a) shows how the
recorded temperature rises when the thermometer is placed in a hot liquid; Figure 1
(b) shows how the recorded temperature falls when the thermometer is placed in a
cold liquid. In both cases the recorded temperature takes a finite time to be as close
to the temperature of the liquid as the thermometer's minimum error will allow.
The speed of change can be characterised by the time constant τ, which is the time
interval in which the difference between the recorded and actual temperatures
changes by 63.2% of the initial temperature difference, ∆T.
All types of thermometer have a step response like the one shown in Figure 1,
though the time-scales along the axes can vary quite considerably. The smaller the
response time the faster the thermometer will respond to temperature changes.
• Bimetallic Strip
• Sealed-fluid thermometers
Bimetallic strip
A bimetallic strip is one of the simplest examples of the way in which the thermal
expansion of two materials can be used to measure temperature. It makes use of the
differential expansion of two materials. Strips of two different metals are bonded
together as in Figure 2(a). The metals are chosen so that if they were allowed to
expand freely, they would expand by different amounts for a given temperature
change. Since they are constrained to be the same length at their common surface,
a temperature change forces the combined strip to bend into a circular arc (Figure
2(b)). If one end of the strip is fixed, the temperature change can be calibrated in
terms of the deflection of the other end. Bimetallic strips made in the shape of a
spiral are used directly as thermometers, with a pointer attached to the free end of
the spiral (Figure 3).
The fractional increase in length per unit rise in temperature of a metal (when
allowed to expand freely) is called its coefficient of linear expansion. The greater the
difference between the coefficients of linear expansion of the two metals in a
bimetallic strip, the more the strip will bend for a given rise in temperature.
For this reason, one of the metals is usually invar (an alloy of iron and nickel), which
has a very low coefficient of linear expansion.
Liquid-in-glass thermometer
Sealed-fluid thermometers
Two other types of temperature sensing device using a sealed fluid are shown in
Figure 4. These can display a temperature reading at a distance from the point at
which the temperature is measured. In Figure 4(a) a liquid, such as mercury,
completely fills the bulb, capillary and pressure sensor, which are made of metal.
The temperature is sensed at the bulb. The capillary, which may be as much as 60m
long, connects the bulb to a pressure sensor such as a Bourdon tube. An increase in
the temperature of the bulb causes the fluid in it to expand. This increased volume
corresponds to an increased pressure in the Bourdon tube, which thus uncoils. The
resulting motion of the free end is used to drive a pointer.
If a voltmeter is inserted into the circuit as in Figure 6, it will indicate a voltage, which
depends upon the difference in temperature between the two junctions. If the
Across the junction of any two dissimilar metals there always appears a difference in
electric potential called the contact potential. In Figure 6 the contact potentials are
labelled V1 and V2. The contact potential between two metals varies with the
temperature at their junction, increasing in magnitude with temperature. The
voltmeter reading equals the difference between the two contact potentials V2 - V1.
Therefore, when the two contact potentials are equal the output will be zero.
Olatunbosun, A Page 7 of 16 Univation
Shell Nigeria Graduate Training Programme Instrumentation
Measuring of Temperature
This conclusion has some useful implications. Although Figure 6 shows only two
junctions between dissimilar metals, unless the voltmeter and the wires leading to it
are also made of metal B, there will be at least two more junctions of dissimilar
metals. Figure 7 shows the two additional contact potentials V3 and V4. If the two
connections between metal B and the copper wire are at the same temperature, V3
and V4 will be equal. The net voltage across the voltmeter will still be V2 - V1.
directly (as in Figure 7) because the third metal, copper, is common to both of the
low-temperature connections.
Reference junctions
One strategy for establishing the temperature of the reference junction is to place it
in some apparatus in which the temperature is accurately controlled. The most
accurate method, used primarily in laboratory situations where great accuracy is
required, is to use a triple-point apparatus, in which the temperature is maintained at
the triple point of water. Other methods are also available.
Extension wires
Measuring instruments must often be located far from the point at which the
measurements are to be made. With thermocouples this raises the problem of where
to pit the reference junction (which may well be located where the thermocouple
wires join the wires leading to the meter). If it cannot be located near where the
measurement is to be made, then long lengths of thermocouple wire is needed. Such
wire manufactured to very tight specifications is expensive. The usual alternative is
to use extension wires either made of the same material as the thermocouple wires,
but of a smaller gauge and to less exacting standards, or made of some other metal,
which closely matches the temperature characteristics of the thermocouple wires.
With the extension wires connected, the reference junction is effectively moved to
their ends.
The extension wires need only match the thermocouple wires over the range of
temperatures to which the extension wires are exposed, i.e. the temperatures at the
reference junction and at the junctions of the thermocouple and extension wires. The
temperature range to which the measuring junction of the thermocouple is exposed
may be much greater.
Since thermocouples have been widely used for temperature measurements for
many years, various standards for thermocouples have been developed and adopted
by international and national standards bureaux. The metals or alloys of metals used
for the seven thermocouples in this standard are listed in Table 2.
For each of the types listed, there are published tables giving their output voltages
for a wide range of temperatures at their measuring junctions, and giving allowable
deviations from these values. Most manufacturers of the listed types of thermocouple
manufacture them to meet the requirements of the standard. This not only ensures
that the table accurately describes the behaviour of the thermocouple but permits
interchangeability between thermocouples of different makes.
Table 2 shows that none of the seven BS4937 thermocouple types can be used
satisfactorily above a temperature of about 1500oC. However, as better refractory
materials become available, it becomes possible to use higher temperatures in
applications such as turbines, rocket engines and nuclear reactors. This generates a
need for reliable thermocouple sensors with a useful life at elevated temperatures.
A thermocouple probe at its simplest consists only of the two wires with their ends
soldered or welded to from a junction. In many applications that is all that is needed:
the junction of the bare wires is placed in contact with whatever is to have its
temperature measured and the measurements are taken.
In some situations however, something more elaborate is required. In such cases the
actual thermocouple wires and the measuring junction are encased in a metal
sheath, as shown in the X-ray photographs in Figure 11. Between the wires and the
sheath is a layer of insulation. This insulation is a mineral such as magnesium oxide.
For this reason, this type of probe is called a mineral-insulated thermocouple
probe. The function of the metal sheath is to protect the thermocouple from
environments in which it might be corroded, abraded or otherwise damaged.
Together with the seal at the end, the sheath also enables the thermocouple to be
used in a system, which must be sealed or must withstand high pressures.
Table 3 lists some commonly used sheath materials and their properties.
Magnesium oxide insulation provides a high insulation resistance (of the order of
mega-ohms) between the thermocouple wires and the metal sheath. At the same
time it has a high thermal conductivity to enable the probe to respond rapidly to
changes in the temperature to be measured.
The response time of the thermocouple depends upon the rate of heat transfer from
the material being measured and so varies not only from one thermocouple to
another but also from one application to another. However, in general, the response
time of a mineral-insulated probe increases as the probe diameter increases, so if a
rapid response is a requirement for a particular application then a small probe
diameter is essential. The best response times are obtained from bare wires.
A few terms need explanation. 'Grounded' means that one of the two-thermocouple
wires is in electrical contact with the casing. This reduces the time constant of the
thermocouple. A 'hypo' tips is an angled tip, as shown in the lower diagram; this
shape improves penetration. A 'patch' tip is, as its name suggests, a flattened tip for
use on a flat surface.
Figure 12 shows how the resistance of platinum, copper, tungsten, Balco (an alloy of
nickel) and nickel vary with temperature. The resistance RT at temperature T is
shown relative to the resistance Ro at 0oC. The metals are those most often used in
resistive temperature transducers.
Although the sensitivity of platinum is the smallest, it is the most stable and
reproducible type of temperature transducer. It is thus used to define the IPTS for
temperatures in the range from 13.81 K to 903.89 K (see Table 1). Platinum is
chemically inert and is resistant to contamination. This enables platinum resistance
thermometers to be used in a wide range of environments. Unfortunately, platinum is
relatively expensive.
Of the base metals used for resistance thermometers, nickel is the most widely used,
primarily in the temperature range from about 175 K to 600 K.