2nd Balkan Region Conference on Engineering Education
Sibiu, Romania, 16 - 19 September, 2003
Studying and Calibrating Thermocouples: A Laboratory Exercise at the Technological
Educational Institute of Crete
J.P. Makris1), I.O. Vardiambasis1), V. Saltas2) & N. Petrakis1)
1) Technological Educational Institute of Crete, Department of Electronics
2) Technological Educational Institute of Crete, Department of Natural Resources & Environment
Chania, Crete Island, Greece
ABSTRACT: Thermocouples stand for the most common sensors for the temperature measurement based on thermoelectric effects.
The objective of this work is to present a laboratory exercise suitable for an introductory study of the thermocouples’ behaviour and
calibration procedure. Various types of thermocouples (K, N, J, Τ) are examined simultaneously in the temperature range 25°C to
100°C. Reference temperature values are obtained using a high-quality mercury-in-glass thermometer. The thermocouple circuit
applied is that of cold junction compensation (reference junction). Furthermore, the mechanical thermometer is compared against
the reference junction compensation using type T thermocouple and thermistor. The data are collected in a PC which is
communicating through its serial port with a programmable data logger thus familiarizing the students with low level programming,
as well as with the principles of digital measurement and data acquisition procedures. Data analysis consists of calibration curves,
assessment of uncertainties, polynomial fitting, study of the thermoelectric power (sensitivity) against temperature, design of
algorithm to calculate temperatures from thermocouple voltages and comparison of the results with NIST standard tables.
The Laboratory of Electronic Tests & Measurements and Data
Acquisition Systems of Electronics Department at the
Technological Educational Institute of Crete is focusing its
technological and scientific research on the principles of
sensors, measuring and data acquisition systems studying their
structure, features and specifications. Special attention is given
on equipment calibration procedures with respect to
traceability to international standards and the calculation of
uncertainties when measurements are conducted. In the
aforementioned issues are also developed the educational
activities of the Laboratory.
In the second law of thermodynamics, temperature is related to
heat transfer and the internal energy of the materials. Statistical
physics correlates temperature with average kinetic energy of
the molecules of ideal gases and with energy levels in liquids
and solids [1-3]. Since pressure, volume, electrical properties
etc., are all related to temperature through the fundamental
molecular structure, the measurement of temperature and its
variations is inevitable as well as important and in many cases
crucial for almost any technological application and
measurement of other physical quantities. Furthermore, the
calibration of temperature sensors and related measuring
devices must be performed regularly through the comparison
with established standards (ITS-90) [4,5].
The above discussion makes imperative, from educational
point of view, a laboratory exercise for engineers on the
measurement of temperature and its most common sensors i.e.,
THERMOELECTRIC EFFECTS (THERMOCOUPLES)
Two dissimilar metals or alloys, A (positive) and B (negative)
“thermoelements”, joined to form a circuit constitute a
thermocouple sensor (TC). T.J. Seebeck [1-3] first discovered
in 1822 that a thermocouple would produce a current in closed
circuit having two junctions at different temperatures due to an
electromotive force (emf), EAB, and this is the reversible
Seebeck effect. Seebeck coefficient, SAB, is defined as the
temperature derivative of EAB:
where SA and SB are the absolute thermoelectric power for A
and B respectively (see Table I). Two additional reversible
thermoelectric effects are not of prime concern on a
Table I. Thermoelectric power against Platinum of a
thermoelement made of listed materials, µVoC-1 (reference
junction at 0 oC).
material S (µVoC-1)
The Peltier effect [1-3] discovered by J.C.A. Peltier in 1834 is
an exchange of heat between a junction and its environment
directly proportional to the current flow into the TC-circuit. W.
Thomson (later Lord Kelvin) discovered in 1847-54 that heat is
evolved or absorbed reversibly when current is drawn through
a homogeneous conductor in a temperature gradient [1-3]. The
net thermocouple emf is a line integral on the path from
positive to negative thermoelement:
v⋅∇⋅−= ∫ (2)
dependent on the path if the thermocouple is inhomogeneous
(variation in chemical composition or physical state of
thermoelements, strain, handling etc.) as is always the case.
The ideal thermocouple behavior is summarized with the
three laws of thermoelectric circuits . (i) Law of
homogeneous metals: a thermoelectric current cannot be
sustained in a circuit of a single homogeneous material by the
application of heat alone. (ii) Law of intermediate metals: the
algebraic sum of the emfs in a circuit composed of any number
of dissimilar materials is zero if all of the circuit is at uniform
temperature. (iii) Law of successive or intermediate
temperatures: if two dissimilar homogeneous metals produce
an emf E1, when the junctions are at temperatures T1 and T2,
and a thermal emf of E2, when they are at T2 and T3, the emf
generated when the junctions are at T1 and T3, will be E1+E2.
From the hundreds of different types of TCs identified and
studied, only a few types have been standardized, because of
their favourable characteristics (low resistivity, high
temperature coefficient, high linearity, environmental strength
etc.) and the fact that they are the ones most commonly used.
They are identified by a letter code (ANSI-MC-96.1, 1982;
ASTM 230-87)  that is still in use although has been
replaced with a new specification colour code standard
BS4937 Part 30 1993 which conforms to IEC 584-3 1989 .
The choice of a thermocouple is strongly influenced by the
temperature range of the measurement. Furthermore, TCs are
frequently used in transient temperature measurements due to
their advantages of small transient response time, no self-
heating of the bias electric current, etc . However, the TCs
must be calibrated in advance because they are always
different from one another for the reason of the internal stress
inside the materials and inhomogeneity of the chemical
composition of the thermoelements.
The educational objectives that the laboratory exercise design
intends to meet are: (i) carrying out temperature and
temperature variation measurements, (ii) study the
characteristics of different types of thermocouples and their
conventional method of circuitry (establishing reference
temperature), (iii) perform a first calibration of the examined
TCs and calculate the uncertainties involved and (iv)
introducing to low level programming of measuring systems
and data acquisition features. The experimental setup of the
laboratory exercise is depicted in Fig. 1 and described in detail
following in this section.
Four different types of base-metal thermocouples were selected
to be tested during the exercise, i.e., K, N, T and J. They were
supplied from RS-Components and their specification details
can be found in the relevant data sheets . Types K (Nickel
Chromium/Nickel Aluminum), N (Nicrosil/Nisil) and T
(Copper/Constantan) are welded tips, PTFE (Poly Tetra Fluoro
Ethylene) insulated for chemical resistance, having lm length,
covering the temperature range -40°C to +200°C with an
average sensitivity of ~40µV/°C. Type J (Iron/Constantan) is a
non-oxidizing welded tip, glass fiber insulated, 2m long, with a
sensitivity rising to 55µV/°C in the range -50°C to +400°C .
Figure 1. Experimental setup of the laboratory exercise for the
study and calibration of thermocouples.
Usually, the thermoelectric emf is expressed in terms of the
potential generated with a reference junction at 0°C. This
method is commonly referred as reference (cold) junction TC-
circuit and it was adopted in the experimental setup of the
present exercise for the thermocouples K, N and T (Figs 1 and
2). In such a case the output voltage can be approximated with
901900AB tttE α+α+α≈ (3)
where t90 is the thermodynamic International Celsius
temperature (ITS-90) :
15.273K/TC/t 9090 −°
(see also Table II). Then, the thermoelectric power, or
sensitivity of a thermocouple is given by:
For the J-type it is used the hardware-compensation device .
This is an alternative circuit to alleviate the problem of the
reference junction (see Fig. 3). A thermistor is in thermal
contact with terminal strip of the datalogger to which the J-
thermocouple is connected. The voltage υb and thermistor’s
temperature coefficient are adjusted so that VC will match the
TC-temperature coefficient. The value of Rx is regulated so
that the ∆V readout is zero at 0°C.
Figure 2. Cold (0
C) junction technique for establishing
reference temperature in thermocouple circuit.
As a system with temperature varying in time, it is used 200ml
distilled water contained in a pyrex vessel, in room temperature
(t90~25°C) that is heated, with no constant rate, up to boil point
(t90=100°C) using a heating resistance. The standard
temperature values necessary for the calibration are provided
through the analog readings on high quality accredited
mercury-in-glass mechanical thermometer with traceability to
ITS-90 (see Fig. 2).
Table II. Thermoelectric emf (in mV) for commonly used
thermocouples, at various thermodynamic International
Celsius temperatures, according to ITS-90 (reference junction
0 0 0 0 0
25 0,992 1,277 1,000 0,402
50 2,036 2,585 2,023 0,836
75 3,132 3,918 3,059 1,297
100 4,279 5,269 4,096 1,758
Thermocouple reference junctions are maintained at 0°C in an
ice bath (Fig. 2), which consists of a mixture of finely divided
pure ice and distilled water contained in a Dewar flask.
As a millivoltometer and data acquisition system it is selected
the Micrologger 21X of Campbell Scientific (see Fig. 1). The
21X combines precision measurement with processing and
control capabilities in a rechargeable battery operated system
. In its wiring terminal strips are connected the TCs and an
insulated cover of the analog strips provides diminution of
temperature gradients across the input terminals. The 21X is
programmed to make differential measurement to the used
inputs (1-3, 5) which means that the voltage on the H(igh)
terminal is measured with respect to the voltage of the L(ow)
terminal of every input (input noise voltage 0.1µVrms,
CMRR>140dB, input resistance 200GΩ). TC-voltage
measurements are performed by integrating the input signal for
a fixed time (20ms), thus removing noise such as 50Hz from ac
power, and then holding the integrated value for the A/D
conversion which is made with a 14bit successive
approximation technique that resolves the signal to
approximately 1/15,000 of the full scale range on the
differential measurement, i.e., 1/15,000 × 5mV = 0.33µV.
Furthermore, the 21X Micrologger is capable to perform, after
suitable programming, automatic temperature measurements
using an attached to it thermocouple and its panel temperature
as the reference junction . Firstly, the reference junction
temperature is measured with a built-in thermistor mounted
under the analog input terminal strip. The accuracy of this
measurement is a combination of the thermistor’s
interchangeability, specification and precision of the bridge
resistors and any difference in temperature between the
thermistor and the actual reference junction (the input to which
the TC is attached). The latter influence is drastically reduced
by using the isolating terminal cover and connecting the
selected T-thermocouple at input 4 which is located exactly
above the thermistor. The 21X calculates the voltage that the
T-thermocouple would output at the reference junction
temperature if its reference junction was at 0°C, and adds this
voltage to the measured TC-voltage. The temperature of the
measuring junction is then calculated from a polynomial
approximation of the National Institute of Standards and
Technology (NIST Monograph 175 ) TC calibrations. The
aforementioned procedure is also realized in the presented
exercise, in order to have additional and independent
temperature measurements of the water under heating.
The internal memory of 21X is divided into five areas: (i) the
input storage where the measurements are held, (ii) the
intermediate storage for data processing, (iii) the final storage
which contains the processed data ready to transfer to the PC,
(iv) the system memory, and (v) the user program memory. In
the latter the students have to enter the program, created by
themselves after consulting the relevant sections of the 21X’s
manual, that performs and controls all the datalogger’s
measuring and data processing operations. For a detailed
description, the reader is referred to the exercise booklet and
the 21X’s manual [9, 12].
Figure 3. Hardware-compensation device for reference
junction compensation using thermistor.
The communication of the datalogger with data collection PC
is established through an optically isolated RS232 interface
. The thermocouple data are retrieved from the datalogger
to the PC using the provided software bundle. The PC208W
datalogger support software package provides a graphical user
interface that is servicing exchange of data, programs, and
commands between 21X and the computer . Furthermore,
it allows the real-time monitoring of the measured TC-voltages
versus the time and continuously displays the time of the
EXPERIMENTAL PROCEDURE AND DATA ANALYSIS
For the execution of the laboratory exercise, the synergy of
three students is requested. Initially, the students must have
thoroughly study the about 10 pages booklet  that already
own, in order to realize the basic theory and the
instrumentation, as well as the steps undertaken during the
execution in the laboratory and the relevant homework (data
analysis and interpretation).
During the measurements each student is assigned a specific
task but their collaboration and coordination as well as their
concentration are requested in the superlative degree. This is
very important from educational and pedagogical point of view
and it was accounted during the design of the exercise. One
student is responsible carefully read the indications of the
mercury-in-glass thermometer. The second is responsible to
write down the readings of the first. The third student checks
continuously the time on the PC’s monitor and instructs the
first student when to make a thermometer reading. The
programmed datalogger takes measurements from all of its
inputs every one second but due to the fact that this is a very
high sampling rate for the experimentalists, they measure the
temperature of the water under heating every 5 seconds. The
complete time is noted only for the first measurement in order
to correspond with their measurements, these of the datalogger.
The experiment lasts about 15-20 minutes i.e., until to bring
the water temperature from its room-value (~25°C) to the boil
point (100°C). Then, the data are collected through the serial
connection from the datalogger to the PC and stored in a file
with comma separated ASCII format. The students receive the
created file in a diskette in order to process data at home. The
data can be easily imported in spreadsheet software (e.g.,
Microsoft Excel) for further analysis. Of course, it is necessary
to resample the data accordingly, in order to correspond with
the manually obtained measurements.
20 30 40 50 60 70 80 90 100
temperature ( oC)
output voltage (mV)
Figure 5. Observed calibration curves for the thermocouples
tested in the frame of the laboratory exercise.
Among the various steps of data analysis and interpretation and
questions that have to be answered, which contained in the
exercise booklet , we following quote the most important:
(i) calibration curve (thermocouple emfs vs temperature) for
each thermocouple under study in the temperature range 25°-
100°C (see Fig. 4), (ii) linear regression analysis of each
calibration curve and calculation of the thermocouple
sensitivity; comparison with the declared values from RS-
Components, (iii) third-order polynomial fitting of the
calibration curves and determination of the relevant
coefficients, (iv) for thermocouples J, K and T, calculation of
temperature from TC-voltages by forming a polynomial
21090 vc...vcvcct ++++= (6)
where t90 the temperature in °C, v the TC-voltage (with
reference junction at 0°C) and ci the polynomial coefficients
that have to be compared with those depicted on NIST
standard table , (v) comparison of the manual temperature
measurements using the mechanical thermometer with the
temperature values provided automatically from the datalogger
using a T-thermocouple with the reference junction in the
21X’s panel temperature. Finally, a summary containing all the
comments, remarks and conclusions drawn from the
elaboration of the exercise must be the epilogue of the report
that the students will submit.
In this work it was presented a laboratory exercise on studying
and calibrating thermocouples, designed and developed at the
Laboratory of Electronic Tests & Measurements and Data
Acquisition Systems of Electronics Department at the
Technological Educational Institute of Crete.
The laboratory exercise is aiming to introduce the students to
the principles of the temperature measurement based on the
thermoelectric effects and using the most spread sensors, i.e.,
the thermocouples. The basic and essential steps in the
calibration of a sensor were strived to be depicted in the frame
of the exercise. Furthermore, the students have the opportunity
to acquire some experience on low level programming of
measuring systems and data acquisition.
Different types of thermocouples are tested and studied and
various techniques of TC-circuits were adopted. An
educational advantage of the presented exercise is the existence
both of analog and digital measurements and devices, which
helps the students to clarify the difference between the two
concepts. Also, a first level familiarization with standards
(ITS-90, NIST table) and uncertainties assessment and their
major role in engineering is fairly achieved. The extended data
analysis and interpretation requested as homework and the
submission of a full report with the results by the students
underlines the importance and the cohesion of both the
experimental work and the data processing.
The presented exercise demands the activation, participation,
concentration and alertness of every student and cultivates
their collaboration skills, fundamental qualifications of the
future engineers. Furthermore, although the exercise does not
take a lot of laboratory time to complete a good preparation is
need for a successful execution.
Finally, there are a lot of ideas to evolve and enrich the
exercise. As indicative examples we just mention here aspects
that already we are working on: (i) thermopiles, (ii)
thermocouple response and compensation and (iii) comparison
of thermocouples with other temperature sensors.
1. Holman, J.P., Experimental Methods for Engineers. 7th
International Edition, N.Y.: McGraw-Hill, 355-419
2. Pallas-Areny, R. and Webster, J.G., Sensors and Signal
Conditioning. N.Y., Chichester, Brisbane, Toronto,
Singapore: John Wiley & Sons, 233-247 (1991).
3. Burns, G.W. and Scroger, M.G., The Calibration of
Thermocouples and Thermocouple Materials. Natl. Inst.
Stand. Technol. (NIST) Spec. Publ. 250-35, 201p. (1989).
4. Preston-Thomas, H., The International Temperature Scale
of 1990 (ITS-90). Metrologia, 27, 3-10, (1990).
5. Burns, G.W., Scroger, M.G. and Strouse, G.F.,
Temperature Electromotive Force Reference Functions
and Tables for the Letter Designated Thermocouple Type
Based on ITS-90. Natl. Inst. Stand. Technol. (NIST)
Monograph 175 (1993).
6. Manual on the Use of Thermocouples in Temperature
Measurements. Amer. Soc. for Test. & Materl. (ASTM)
Spec. Publ. 470B (1981).
7. Thermocouples Data Sheet 240-0070, RS-Components,
8. Zhang, P., Xu, Y.X., Wang, R.Z., and Murakami, M.,
Fractal study of the ﬂuctuation characteristic in the
calibration of the cryogenic thermocouples. Cryogenics,
43, 53–58 (2003).
9. 21X Micrologger, Instruction Manual, Campbell
Scientific, Inc. (1992).
10. SC32A Optically Isolated RS232 Interface, Reference
Manual, Campbell Scientific, Inc. (1994).
11. PC208W, Datalogger Support Software, User Guide,
Campbell Scientific, Inc. (1996).
12. Booklet of the laboratory exercise “Comparative Study
and Calibration of Thermocouples”, Laboratory of
Electronic Tests & Measurements and Data Acquisition
Systems, Electronics Department, Technological
Educational Institute of Crete, 2002.