Fundamentals of Temperature Measurement
Fundamentals of Temperature Measurement
Fundamentals of Temperature Measurement
This course covers the basic methods used to measure temperature in commercial,
municipal and industrial applications. It is intended to be an introductory course and
somewhat broad in discussing early-to-current methods of measurement. It will also
discuss applications and conditions that can influence the accuracy and repeatability of
various types of electronic temperature sensors; including thermocouples, thermistors,
RTDs, and infrared technology.
In the early 1600s an Italian physicist named Santorio Santorio developed a crude
thermometer-like device which he referred to as a “thermoscope”. In some ways
resembling the liquid-filled thermometer of today, it lacked any sort of calibrated scale;
and the tube was open to the atmosphere, which adversely affected its accuracy due to
variations in barometric pressure.
Also in the early 1600s, among his other scientific credits, Galileo Galilei reportedly
invented another style of thermoscope, sometimes referred to as a “Galileo
Thermometer”. This type of instrument, which is still available as more of a decorative
item than a functioning thermoscope, operates on the principal of buoyancy. As the
temperature of the clear fluid contained within its vertical glass cylinder increases, the
fluid expands and its density decreases, causing small glass spheres that are each filled
with a uniquely-colored fluid that has a slightly different specific density corresponding
to a specific temperature, to sink within the cylinder as their density becomes greater than
that of the clear fluid in which they are submerged. The temperature is read from a small
engraved metal disc on each sphere. The current approximate temperature is indicated by
the lowest floating glass sphere within the cylinder, such as the yellow liquid filled glass
sphere appearing in the thermoscope shown in Figure No. 1 below.
Further development of temperature measurement, still using a fluid as the indicator and
a calibrated scale to measure the fluids displacement in relation to temperature is credited
to such scientists as Ole Christensen Røemer, who in 1701 reportedly made one of the
first practical thermometers using red wine as the indicator; Daniel Gabriel Fahrenheit,
who soon after developed a more accurate mercury-filled thermometer and thermometer
scale; and Anders Celsius (also in the 1700s), who developed a metric scale of
temperature measure. William Thompson (Lord Kelvin) and William Rankin were
credited in the mid-1800s with having broadened the calibrated temperature measurement
scale range to include “absolute zero”, thus making it more useful for research purposes.
As technology in the areas of physics, metallurgy and electronics have progressed since
the 1800s, so has temperature measurement technology and measurement instrumentation
in general.
Figure No. 1 – Galileo Thermometer (Image Source: Wikipedia)
Liquid filled temperature indicators were followed by the development of bimetallic and
gas-filled instruments, which were later followed by electric and more recently solid state
electronic devices. Please note that I said “followed by” and not “replaced by” the next
generation of instrumentation. Virtually all of the mentioned technologies are still in use
and still commercially available; and that is because each has inherent qualities that make
it more suitable than another type for a particular application. For that reason there are
many hundreds of variations of the various types and styles of temperature measurement
instruments.
Measurement Scales
There are several commonly used scales of temperature measurement that were created in
the 16th and 17th centuries that are still in use today. They include Fahrenheit, Centigrade
(also referred to as Celsius), Kelvin, and Rankin. Each is based on a different scale, and
must be mathematically converted from one to another in order to determine the
equivalent value in another scale. For the most part, outside of scientific applications, the
United States uses the Fahrenheit scale, while most of Europe and other parts of the
world have standardized on the centigrade scale. The Fahrenheit scale uses 32 degrees as
the freezing point of water and 120 degrees as the boiling point of water. By comparison,
the centigrade scale uses zero as the freezing point of water and 100 degrees as the
boiling point of water. Therefore, in order to convert a temperature from Fahrenheit to
centigrade, one would have to use the equation
The centigrade scale conforms to the metric system, which is based upon units of ten.
Kelvin and Rankin scales are primarily used in scientific research and reporting, where
temperature conditions outside of the normal day to day environment are investigated and
explored. It is not necessary that we know each day how close the outdoor or indoor
temperature, for example, are to absolute zero. Absolute zero is considered to be minus
273.15 centigrade, or zero degrees Kelvin - the theoretical temperature at which
molecular movement, which is the fundamental source of heat energy, ceases. In other
words, there is a total absence of heat at absolute zero. On a Kelvin scale the freezing
point of water is 273.15 and the boiling point of water is 373.15 degrees.
We will not go further in depth in our discussion of the various temperature scales, as
they are easily converted from one to another and the rest of this course topic will use
either Fahrenheit or centigrade as the reference scale.
Early Methodology
Another, non-fluid method of measuring temperature that was developed much later uses
a helically-wound bimetallic strip to operate a dial type indicator, such as that shown in
Figure No. 3. A bimetallic strip consists of two thin layers of different metals that have
different coefficients of expansion, such as brass and steel that have been bonded
together. With a given change in temperature, the two bonded metals expand at different
rates, causing the strip to distort (bow) in the direction of the metal with the lesser
coefficient of expansion. By arranging the bimetallic strip in a helical pattern, the
differences in the expansion rates of the bonded metals cause the helix to partially coil or
uncoil as their temperature changes. That movement, when mechanically attached to the
dial of an indicator, causes the dial to move somewhat proportionally in response to a
change in sensed temperature.
Older style, non-digital, wall-mounted room thermostats that control central residential
and commercial heating and air conditioning systems have an integral thermometer
mounted in the thermostat cover. By carefully removing the cover and examining the
back of it, you can typically see the helical bimetallic element of the thermometer. U.S.
patents are held on the helical bimetallic strip designs that are currently on the market.
The advent of the remote bulb thermometer made it possible to monitor temperature
without being very close to the point of measurement. When the accessibility of the
measurement location is difficult under normal circumstances, such as reading the
temperature of water flowing through a pipe that is fifty feet in the air, having a remote
bulb that allows the indicator dial to be down at eye level is very desirable. An added
advantage of this type of expanding fluid device is that it operates without electrical
current. Therefore, it cannot produce an electrical spark during normal operation, making
it intrinsically safe for use in hazardous areas where a flammable or explosive vapor may
be present and ignited by the occurrence of a spark.
Development of light emitting diodes (LEDs), liquid crystal displays (LCDs), and large
scale integration of electronics to a size and cost that are attractive has driven changes in
the technology of temperature measurement, as it has many other things, in recent years.
Although still popular and somewhat cost-effective, small portable thermometers such as
the one appearing on the left in Figure No. 5 are being gradually replaced with a digital
version resembling that of the digital pocket thermometer appearing on the right in Figure
No. 5.
Digital thermometers have the advantage of being more easily read and to a more precise
value, while dial thermometers have the advantage of not requiring a battery.
As mentioned previously, each device has its own qualities that make it better suited for a
particular application. When determining what type of temperature measurement device
is most suitable for your specific application, consider accuracy, repeatability, reliability,
durability, suitability and cost. In terms of suitability, the environment in which it will be
used is a very important consideration. Other things, such as the degree of accuracy and
reliability needed, will tend to drive the average cost of the device.
Many applications that require remote monitoring and/or remote control of temperature
rely on a solid state device for electric and or electronic control, such as a thermocouple,
thermistor, or a resistance temperature detector (RTD). So what are these devices, how
do they differ from one another, and when and where do you use one type rather than
another?
Thermocouples
Let’s start with the one that has been around the longest, and has historically seen the
most use – the thermocouple. A thermocouple is a device that consists simply of a
bonded juncture of two dissimilar metal wires that produces a very small amount of
dielectric current that varies with the temperature to which the primary junction,
sometimes referred to as the “bead” is exposed. This phenomenon is referred to as the
“Seebeck Effect”, which was named after Thomas Johann Seebeck, a physicist who is
credited with its discovery in 1821.
Voltmeter
(millivolts)
Another advantage of thermocouple is the ability to use them to measure relatively high
and relatively low temperatures. Therefore, they are commonly used in industrial and
combustion applications.
A disadvantage of these devices is that, in order to function properly, the same dissimilar
metals used at the junction must be used for the signal wiring all of the way back to the
indicator or recorder as shown in Figure No. 6. Any termination devices, such as
terminal blocks, must also have the same dissimilar metals at each junction. If dissimilar
metals were used as splices or conductors, we would be creating another junction of
dissimilar metals which would act as another thermocouple. Thermocouple wire tends to
be more expensive than standard signal wiring consisting of stranded copper conductors
that is used for the other types of temperature sensors we will discuss. It also must be
ordered from suppliers that stock such wire, which is less common than standard copper
wire.
The thermocouple wire considered to be industry standard are given letter descriptions,
such as the J or K-type. Each uses different dissimilar metals and has different
temperature ranges over which they are accurate, as illustrated in Chart No. 1.
Type J and K thermocouples are the most common types that are readily available and in
use today. Type J thermocouples utilize iron and constantan wire; and Type K
thermocouples utilize chromel and alumel wire and can be used in higher temperature
applications than Type J thermocouples, as can be seen on Chart No.1.
Thermistors are a type of resistor that varies their resistance in inverse proportion to a
change in their temperature. In other words a thermistor’s resistance decreases as its
temperature increases. According to historians the thermistor was invented by Samuel
Rubin in 1830.
Modern thermistors are typically made from either a ceramic disc or a cast chip of a
semiconductor material such as metal oxide, although they are also available with
stainless steel jackets to protect them from harsh environments. They function under the
principal that raising the temperature of a semiconductor device increases the number of
electrons that are able to move about and carry a charge.
An encapsulated thermistor, such as the one shown in Figure No. 8 is a typical style and
resembles a small ceramic capacitor. The encapsulation protects the thermistor junction
from physical damage from the environment.
with a, b and c being constants; ln = natural log; R = thermistor resistance in ohms; and T
= the absolute temperature expressed in degrees Kelvin
The upper and lower temperature limits in which thermistors may be used in applications
is generally considered to be in the range of -100 degrees to 600 degrees F.
Their very compact size makes them somewhat ideal for use in monitoring temperature in
very confined spaces, such as at motor windings or bearings
The third type of electronic temperature sensor that we will discuss is the resistance
temperature detector, which is typically referred to by the acronym “RTD”. The
application of the tendency of electrical conductors to increase their electrical resistance
with a rise in temperature was first identified by Sir William Siemens in 1871. In 1885,
Calender-Van Duesen invented the platinum resistance temperature detector.
Most RTD sensing elements consist of a length of very fine coiled wire that is wrapped
around a ceramic or glass insulated core. The fine wire of the element tends to be very
fragile, so it is often enclosed within a protective metal probe referred to as a “sheath”.
The wire or thin film sensing portion of the RTD is most commonly made of either
platinum, nickel, or copper.
The resistance of the RTD sensor is dependent upon its temperature and the type of wire
wound element or thin film sensing material. Note that in Chart No. 2 the platinum RTD
has a resistance of 100 ohms at zero degrees centigrade. Platinum RTDs have an
accuracy of approximately ±0.1 °C. By comparison, a nickel RTD would typically have
a resistance of 120 ohms at zero degrees centigrade and an accuracy of approximately
±0.5 °C.
Depending upon the specific type of application, RTDs are commercially available in
various configurations that are largely dependent upon the environment in which they
will be used. In harsh or hazardous environments the sensor may be enclosed in stainless
steel sheaths or wells and the electrical terminations may be enclosed within an
explosion-rated “head” such as those illustrated in Figure No. 9 below.
Figure No. 9 – Various RTD Configurations
Similar protective sheath and termination head configurations shown in Figure No. 8 and
described above for RTDs are also available for thermistors and/or thermocouples.
The simplest resistance thermometer configuration uses two wires (refer to Figure No.
10) and is only used when high accuracy is not required. In a two-wire configuration the
resistance of the signal wires is added to the resistance of the sensor, adding some degree
of error to the measured resistance value (signal).
A Wheatstone Bridge, named after its creator, is essentially an electronic “bridge” circuit
that acts as a signal transducer by producing an electrical output signal (voltage) that
varies in response to an imbalanced condition of the resistor bridge. The RTD acts as one
leg of the resistor bridge circuit. As the temperature of the RTD varies, its resistance
varies, causing an increase or decrease in the balance of the resistor bridge. More
information regarding RTD bridge circuits, and how they are configured for 2, 3 and 4-
wire RTDS, can be found by going to http://www.tempco.com/Sensors/rtd_circuits.htm
There are a number of conditions that may adversely affect the accuracy of an electronic
temperature sensor. They include:
When a sensor is installed in a protective metal well, such as the one shown in Figure No.
12, it is important that the sensor itself be directly in contact with the well, so that good
thermal conductance occurs from the measured fluid (steam, water, oil, etc…) outside the
well through the metallic wall of the well and into the sensor itself. This can be more
readily accomplished by applying a thermally conductive grease between the sensor and
interior surface of the well (thermal grease is a far better conductor of heat than an air
gap) or by utilizing spring tension to press the sensor against the wall or end of the well.
Both of those items are commercially available. Failure to do so can substantially slow
down the response time of the sensor to changes in the temperature of the measured
variable (sometimes referred to as “lag time”), or result in measurement error. Response
time is often critical in control system applications. Thermocouples are inherently more
responsive than RTDs.
Figure No. 12 – Thermowell (left) and Sensor with Termination Head (right)
Exposure of the sensor to other environmental conditions not only influences the
measurement accuracy; but also plays a significant role in the reliability and life
expectancy of the sensor. The wire wound element of an RTD, for example, is very
fragile. Therefore, installing an RTD in a location where there is substantial vibration
could potentially shorten the life of the sensor. It should be noted that when an RTD
fails, it typically fails in an “open” (infinite resistance) condition. In a commercial
building environment, the sensor may not need to be installed in a stainless steel sheath
and be equipped with a weather-proof, explosion-rated termination head. However,
municipal buildings or industrial facilities where acidic or caustic materials may be
present, where vibration may be significant, and where the risk of physical damage is
potentially high, the added expense of protecting the sensor may pay off in extended life
expectancy and reliability.
Proper wiring, shielding and grounding can prevent occurrences of ground loops or
imposed signal error on the sensor wires. Ground loop conditions occur when the ground
conductor of signal wires are grounded at more than one end of the conductors. The
ground potential can be different at each end of a long run of wire. By grounding the
conductor at each end, a circuit is completed and the voltage differential between the two
ground points is imposed on the ground conductor or shield. Since the ground reference
is no longer at ground for a given location, the difference in ground potential becomes the
error (also referred to as “offset” of the signal).
A ground loop occurs when there is more than one ground connection path between two
pieces of equipment. The duplicate ground paths form the equivalent of a loop antenna
which very efficiently picks up interference currents. Lead resistance transforms these
currents into voltage fluctuations. As a consequence of ground-loop induced voltages, the
ground reference in the system is no longer a stable potential, so signals ride on the noise.
The noise becomes part of the signal.
Proper selection of a sensor requires factoring in the following things during the selection
process:
Note that the cost consideration is at the bottom of the list. As important as cost may be,
the cost will largely be driven by the type of sensor that is needed to properly fit the
application. Little money is saved on a sensor that is cheaper but inadequate for the
application and fails within a matter of days, weeks or months of being installed and
perhaps was never accurate in the first place.
Accuracy is a term used to specify how close the measurement value will be to the actual
temperature of the fluid or surface that is being measured. You will often see accuracy
ratings that are in a plus and minus range. For example: the specification of a particular
sensor has a stated accuracy that is ±1 percent of the range. So for a sensor having a
range of 100 degrees to 600 degrees F its stated accuracy will be ±5 degrees F. That may
or may not be accurate enough for a given application. Accuracy can vary due to any
number of factors that include: the type of sensor used; sensor placement; and signal error
introduced by the signal wiring between the sensor and display (or recording device –
such as a data logger) and the error attributable to the signal transducer (if needed) and
the display or recording device itself.
Repeatability is synonymous with precision - the ability of the sensor to exhibit the same
value each time that it is sensing the same temperature. Repeatability is independent of
accuracy. If the repeatability of a sensor is very good and the sensor happens to be, say
inaccurate by 2 degrees F, then it will always provide the same measurement and be
consistently inaccurate by 2 degrees. This repeatable inaccuracy is sometimes referred to
as the “offset”.
Pyrometers
There are numerous types of thermocouple style pyrometers commercially available that
can be used to accurately measure temperatures as high as 1500 °C. A typical modern
day digital thermocouple-style pyrometer is shown in Figure No. 13.
A much more recent method of temperature measurement, and the only “non-contact”
type of temperature measurement that we will discuss in this course, is infrared
temperature detection; also referred to as infrared thermometers. As its name suggests,
this technology employs measurement of infrared radiation produced by an object when
that object is warmer than absolute zero. This technology utilizes an optical lens to focus
the infrared (IR) energy onto an infrared detector, which converts the detected energy to
an industry standard electrical signal that can be displayed in standard temperature units.
The quality of these instruments has improved substantially while their cost has
decreased substantially in recent years.
Infrared thermometers were initially used for research and industrial applications that
required relatively accurate measurement without any direct contact of the sensor with
the object or substance of which the temperature was being measured. Often this was to
avoid potential contamination of a product; destruction of a standard “contact” type of
sensor – which would begin to melt at very high temperatures even when sheathed in
stainless steel; or badly corrode (oxidize) due to aggressive acids or caustics. They were
also used to measure the surface temperature of things such as steam pipes that were not
easily accessible without a powered lift or ladder. Their use has broadened as many
forensic engineers and service technicians now use these instruments largely for
obtaining accurate surface temperature measurements of such things as light fixtures,
electrical panels, junction boxes, plugged steam traps, vessels, conduits and pipe lines
that are located in relatively inaccessible locations.
Figure No. 14 – Examples of Handheld Portable and Stationary Infrared Thermometers
Stationary or fixed position infrared detectors are used in situations where the product
whose temperature is to be measured passes in front of the instrument on a material
handling system such as a conveyor, such as the instrument shown on the right in Figure
No. 14.
One type of change-of-state indicator label technology uses multiple waxed layers, with
the upper layers formed using a mixture of viscous material and powdered petroleum wax
that melts at a selectively preset temperature to expose the bottom, “indicator” layer that
is in contact with the object being “sensed”.
In summary, the technology of temperature measurement has been around for several
centuries and has advanced considerably in terms of sensing accuracy, reliability, and
methodology. It is interesting that most of these temperature measurement devices are
still commercially available; attesting to their unique usefulness in a wide range of
applications. There are a number of parameters to consider when selecting the
appropriate temperature measurement device. Determine exactly what it is you want to
measure, identify the optimum measurement location, make note of the conditions
(environment) to which the device will be exposed and how accurate and repeatable it
must be to meet your needs, and then select the most cost-effective device for your
application.