Temperature Measurement Measurement Principles3-6

Temperature measurement is one of the most common applications in venti­lation systems. The air temperature in several places in the ventilation system and in the ventilated space is usually of primary’ interest. The surface temperature of enclosures is also of interest in some cases. A rare application in connection with ventilation is the measurement of temperatures inside a structure. All these tasks can be solved by using the contact measurement principle. Surface temperatures can be measured using the noncontact principle as well.

Temperature Measurement

Contact temperature measurement is based on a sensor or a probe, which is in direct contact with the fluid or material. A basic factor to understand is that in using the contact measurement principle, the result of measurement is the temperature of the measurement sensor itself. In unfavorable situations, the sensor temperature is not necessarily close to the fluid or material temper­ature, which is the point of interest. The reason for this is that the sensor usu­ally has a heat transfer connection with other surrounding temperatures by radiation, conduction, or convection, or a combination of these. As a conse­quence, heat flow to or from the sensor will influence the sensor temperature. The sensor temperature will stabilize to a level different from the measured medium temperature. The expressions “radiation error” and “conduction er­ror” relate to the mode of heat transfer involved. Careful planning of the mea­surements will assist in avoiding these errors.

Instruments based on the contact principle can further be divided into two classes: mechanical thermometers and electrical thermometers. Mechanical thermometers are based on the thermal expansion of a gas, a liquid, or a solid material. They are simple, robust, and do not normally require power to oper­ate. Electrical resistance thermometers utilize the connection between the elec­trical resistance and the sensor temperature. Thermocouples are based on the phenomenon, where a temperature-dependent voltage is created in a circuit of two different metals. Semiconductor thermometers have a diode or transistor probe, or a more advanced integrated circuit, where the voltage of the semi­conductor junctions is temperature dependent. All electrical meters are easy to incorporate with modern data acquisition systems. A summary of contact thermometer properties is shown in Table 12.3.

The noncontact measurement principle, usually called optical or radiation temperature measurement, is based on detecting electromagnetic radiation emitted from an object. In ventilation applications this method of measure­ment is used to determine surface temperatures in the infrared region. The ad­vantage is that the measurement can be carried out from a distance, without contact with the surface, which possibly influences the heat balance and the temperatures. The disadvantages are that neither air (or other fluid) tempera­ture nor internal temperature of a material can be measured. Also the temper —

HH TABLE 12.3 Contact Thermometers and Useful Measuring Ranges

Type of thermometer

Temperature range (°C)


Speed of response


-60 to 510

Medium to high



-40 to 650



Vapor actuated

-40 to 320


Medium (bare bulb)

Gas actuated

-85 to 540

Medium to high

Fast(bare bulb)


-40 to 540 (430 constant)

Low to medium

Medium to slow


-70 to 540


Medium to fast


-120 to 400

Medium to high

Fast (bare)


-18 to 180



Ature of a small area is difficult to measure, as the radiation thermometer is usually receiving the radiation from a larger area. Finally, the surface emissiv — ity is often required as an input value to obtain a good result, even though some instruments are able to make some form of emissivity correction. Mechanical Thermometers 3-6

Liquid-in-Glass Thermometers

Liquid-in-glass thermometers measure the thermal expansion of a liquid, which is placed in a solid container, on a length scale. The mercury thermome­ter is one example of liquid thermometers. Alcohol is also used with this type of instrument. The temperature range is -80 to +330 °C depending on the liquid. The quality, stability, and accuracy vary considerably. The advantages are a simple construction and low price. A disadvantage is that they are not compatible for connection to monitoring systems.

Filled Thermometers

In filled thermometers the thermal expansion of a gas or a liquid is trans­mitted through a thin capillary tube to a bellows or helix, where the deforma­tion indicates the temperature. The temperature range of filled thermometers is very wide, approximately -200 to +700 °C. They are extremely robust but are not very high in accuracy. The application is mainly for process instrumen­tation and as stand-alone control devices.

Bimetallic Thermometers

When two strips of different metals having different thermal expansion coefficients are rigidly attached to each other, the material will bend or straighten according to its temperature. This deformation is by means of cogs and levers converted to the movement of a pointer. The measurement range of bimetallic thermometers depends on the materials used. The range between -50 and +550 °C can be spanned with this type of instrument. The accuracy is low but the simplicity and low cost are an advantage. Electrical Temperature Measurement

Resistance Thermometers

Resistance thermometers are made of a pure metal, such as platinum, nickel, or copper. The electrical resistance of such a material is almost linearly dependent on temperature. Resistance thermometers are stable, having a small drift. A widely used and the best-known resistance probe is the PT-100 probe, which is platinum, having a resistance of 100 ohms at the temperature of 0 °C. Other resistance values for PT probes are available. The resistance versus temperature values as well as tolerances for platinum probes are standardized.7’8 The shape and size of a resistance probe can vary consid­erably, resulting in changes in probe dynamics.

The probe resistance can be measured either directly by passing a small constant current through the probe and measuring the voltage drop, or by con­necting the probe to a bridge. In any case, the current should be low enough to avoid self-heating of the probe. To compensate for the resistance of the measur­ing wires, the resistance probe can be connected to the system by a three-wire or a four-wire connection. If only two wires are used, the connecting wires must not be too long in order to keep the resistance as low as possible.

The measurement range for platinum is -200 to +800 °C, for nickel -50 to +250 °C, and for copper -50 to +200 °C. The advantages are good accu­racy, almost linear characteristics, and stability. A disadvantage is the small change of resistance with temperature, which requires a high sensitivity from die rest of the measurement equipment.


Thermistors are temperature-dependent resistances, normally constructed from metal oxides. The resistance change with temperature is high compared with the metallic resistances, and is usually negative: the resistance decreases with temperature increase. The temperature characteristics are highly nonlin­ear. Such thermistors, having a negative temperature coefficient, are called NTC thermistors. Some thermistors have a positive temperature coefficient (PTC), but they are not in common use for temperature measurement.

The resistance of a thermistor is achieved in a similar manner to the mea­surement of metallic probes. The advantage of thermistors is that their resis­tance is usually high, and the compensation for the measurement wire resistance is not so critical.

The measurement range of a thermistor is dependent on the probe type, typically -100 to +300 °C. The stability is not as good as that of metallic re­sistances. Thermistors are not standardized like some of the metallic probes. The thermistor has the advantage of a high change of the resistance with tem­perature. A very wide variety of sizes and shapes and a low price makes them attractive in relation to the metrological performance.


Thermocouples are primarily based on the Seebeck effect: In an open cir­cuit, consisting of two wires of different materials joined together at one end, an electromotive force (voltage) is generated between the free wire ends when sub­ject to a temperature gradient. Because the voltage is dependent on the tempera­ture difference between the wires (measurement) junction and the free (reference) ends, the system can be used for temperature measurement. Before modern electronic developments, a real reference temperature, for example, a water-ice bath, was used for the reference end of the thermocouple circuit. This is not necessary today, as the reference can be obtained electronically. Thermo­couple material pairs, their temperature-electromotive forces, and tolerances are standardized.10-13 The standards are close to each other but not identical. The most common base-metal pairs are iron-constantan (type J), chomel-alumel (type K), and copper-constantan (type T). Noble-metal thermocouples (types S, R, and B) are made of platinum and rhodium in different mixing ratios.

The measurement ranges for the base-metal thermocouples are -40 to +750 °C (type J), -200 to +1200 °C (type K), and -200 to +350 °C (type T). The noble-metal thermocouples can be used at higher temperatures up to 1700 °C. The dynamic response of sheathed thermocouples is not very fast; however, a probe made from bare, thin wires can have very fast dynamic properties. One of the best features of thermocouples is the simplicity of mak­ing new probes by soldering or welding the ends of two wires together. Infrared Thermometers’4-15

There are several types of optical thermometers: total radiation pyrometers, brightness pyrometers, two-color pyrometers, and infrared radiation thermome­ters. The measurement range and applications of these instruments differ from each other. The main principle is the same: The measurement is based ori ob­serving the electromagnetic radiation emitted by the surface or gas. As a conse­quence, the second common feature is the noncontact-type measurement, where the measurement is carried out at a distance from the object without touching it with a probe. Different pyrometer-type instruments are used for high tempera­tures, covering the range between 300 and 6000 °C. In ventilation applications the relevant temperatures are, however, considerably, lower. For this reason only the infrared radiation thermometer is of interest.

In the infrared thermometer, long-wave radiation is focused on the detec­tor with a lens or mirror system. Lenses must be made of a glass capable of not absorbing too much radiation. The detector, which converts the radiation to an electrical signal, can be a thermal detector such as a thermopile, a photo­voltaic detector, or a photomultiplier. The focusing system can be connected to the detector through an optical fiber, which gives flexibility in placing the dif­ferent parts of the instrument. The detector signal is amplified and treated to give a proper output for the display.

The measurement range is dependent on the instrument but can cover the range -50 to +500 °C. The accuracy is not as high as the best contact thermom­eters. One reason for this is that the emissivity of the surface has an effect on the measurement result, and an emissivity correction is necessary for most instru­ments. The positive features are noncontact measurement and very fast dynam­ics, which enable a rapid scan of surface temperatures from a distance; this is convenient when carrying out, for example, thermal comfort measurements. Measurement Errors

In temperature measurement there are always several factors that intro­duce errors into the results. Here the main focus is on the external or environ­mental factors outside the measuring instrument.

With contact temperature measurement, placing the measurement probe in contact with the object of measurement (duct, surface, etc.) produces an ad­ditional route for heat conduction to or from the object. This perturbation error9 changes the initial temperature field in the vicinity of the contact point and creates measurement errors.

The fouling of the probe when inserted into a duct or pipe acts as an isolating layer and increases the measurement error. To avoid this conduction error, the probe should be a poor heat conductor. In measuring surface temperatures, the probe should not have an insulating effect, as this will change the temperature in the measuring point.

Also, the radiation exchange between the probe and surrounding surfaces can be a cause of measurement error. For example, measuring the air tempera­ture near to a hot heating coil can produce errors if no radiation shield is used around the probe. In general, conduction and radiation errors can usually be eliminated to a sufficient degree. This just requires some knowledge of the dif­ferent inodes of heat transfer, and applying this knowledge to the problem.

Another error related to the use of resistance-type probes is the self-heat­ing error. To carry out resistance measurement of the probe, a current must be introduced to the probe. The current generates electrical power in the probe, heating it up and causing a self-heating error. The magnitude of the self-heat­ing error can be predicted using the heat balance of the probe. In stationary situations the heating power generated in the probe must equal the power dis­sipated from the probe to the environment. From this, the self-heating error is

Where / is the measurement current, R is the probe resistance, B is the heat transfer coefficient between the probe and the environment, and A is the probe surface area. If radiation effects are important, they can be taken into account by the linearization of the radiation term and introducing a radiation heat transfer coefficient. In this case the term H in the above equation also contains the radiation heat transfer coefficient. Calibration16

The calibration of thermometers is based on the International Temperature Scale of 1990, ITS-90.17 This scale became an internationally recognized standard on January 1, 1990. It defines temperatures between 0.65 K and 5 K in terms of the vapor-pressure temperature relations of 3 He and 4 He. Above 5 K, the ITS-90 assigns temperatures to 16 fixed points, which are triple-points, melting points, or points of solidification of chosen substances. In the low-temperature end, different gases such as hydrogen, noble gases, and oxygen are used. The range -38 to +156 °C is obtained using mercury, water, gallium, and indium. Tin, aluminium, silver, and other metals are utilized at the high end of the scale.

Between the fixed points, temperatures on the ITS-90 are obtained by inter­polation using standard instruments and assigned formulae. These standard in­struments are the helium gas thermometer (3 K to 24.5 K), the platinum resistance thermometer (13.8 K to 1235 K), and the optical thermometer (above 1235 K).

From the ventilation point of view, the fixed points -38.83 °C (triple-point of mercury), 0.010 °C (triple-point of water), 29.76 °C (melting point of gallium), and 156.60 °C (freezing point of indium) are of relevance. The triple-point of wa­ter is relatively simple to achieve and maintain with a triple-point apparatus.18 Some freezing point cells are covered in standards.19 In practical temperature cali­bration of measuring instruments, the ITS-90 fixed points are not used directly.

The so-called standard instrument is used for interpolation between the fixed points and for the calibration of other thermometers lower in the metro­logical hierarchy. The standard instrument in the moderate temperature range is a special platinum resistance probe, as it has to fulfill set requirements. It is important in all calibration that traceability to a primary normal, here the fixed-point ITS-90 scale, exists.