Beruflich Dokumente
Kultur Dokumente
3
Thermoelectric Thermometers
3.1 Physical Principles
dE = 6A (6)d6 (3.1)
EA(01,62)=EA(61)-EA(62) (3 .3)
or
In summing up, it should be emphasised that the resultant thermoelectric force of the
circuit of Figure 3 .1 is only a function of the types of metal A and B and of the
temperatures, 191 and t~2 .
From the principles ofthe double subscript notation it is known that:
or finally :
Equation (3 .10) indicates that the emf of the circuit of two metals A and B is a function of
the measured temperature, t91 , of the measuring junction, which is junction 1 in Figure 3 .1 .
It is of cardinal importance, as it is the foundation of modern thermoelectric thermometry
E=eAB(Or)+eBC(or)+eCA(6r)=0 (3 .12)
7
B C
mV
a ~1
1~ t 2 ~z
<7
eael ,8,~l~~e sa I''fzl A C
The law of the third metal is based upon the result given in equation (3 .14) . This law may
be stated as :
Introducing a third metal, C, into a circuit of the two metals A and B, does not alter
the resulting emf in the circuit, provided that both ends of the third metal C are at the
same temperature .
The third metal, C, may be introduced into a circuit at any point . For example, if metal B
is cut and the indicating instrument, which is regarded as the third metal C, is connected at
that point, the resulting emf in the circuit shown in Figure 3 .3 . would be :
E= eAB(01)+eBC(V2)+eCB(62)+eBA(13r) (3 .15)
As eBC (62) = - eCB (62) this equation may be rewritten in the form :
Figure 3 .4 illustrates the situation when the junctions of metals B and C are at the
respectively different temperatures of V2 and 293 . In that case the resulting emf is :
E' = eAB (Bl) + eBC (02) + eCB (03) + eBA (&r) (3 .17)
-51
B B
C C
MV ~ 12 MV
C C
B B
4
Figure 3 .3 Thermoelectric circuit of metals, A Figure 3.4 Thermoelectric circuit of metals, A
and B, with third metal, C, inserted across the and B, with third metal, C, inserted at any
ends of metal B, cut in two point, with its ends at different temperatures
PHYSICAL PRINCIPLES 41
A comparison of equations (3.16) and (3.17) indicates that the difference between the emf
in the two cases under consideration is given by:
Consequently the temperatures at both junctions of the metals B and C should be equal to
ensure that the circuit gives the correct value of emf.
Table 3.1 Value ofthe emf for different metals referred to platinum at 100 C and reference
at 0 C (Roeser and Wensel, 1941 ; Lieneweg, 1975)
30
20 R
C,pQeF, 1F0N
~~o~~~
> 10 ~" PtRh10
I
PLATINUM
1010
a NICKEL AILUIy
10
-zo
CO'YSr9 I
~rq
PACTA01UM
-30
-200 0 200 400 600 800 1000 1200 1400
TEMPERATURE .~9. , C
Figure 3 .5 Thermal emfs as a function of temperature for various metals using platinum as the
reference metal
- (3 .19b)
ECA ( 6 1 , 0r) = eCA (f91) eCA (6r)
-
EBA( 6iA) - ECA( 61, 6r) = eBAWI) eBAWr) - eCA(A)+eCA( 6r) (3 .20)
Consider the thermoelectric circuit given in Figure 3 .7 . If the three junctions of the
metals A, B and C are at the same temperature 01 then
or
Replace the temperature, t91 , by the reference temperature, 6, . Equation (3 .21) may
now be used to obtain :
00( A
C
31
A C
13r
A B
or finally:
C
-1-11
C
Asi
A C B
This equation demonstrates another important thermoelectric law known as the law of
consecutive metals, which may be stated as :
.- thermoelectric . .-
- ..
- r r . r
..
- (61 ,63 ) r r
When equation (3 .24b) is subtracted on both sides of equation (3 .24a) it is seen that :
-
- r r (63,62) r r
44 THERMOELECTRIC THERMOMETERS
I ~ I I
(W (b) I I A
E
,9i EAB ( .3 4)
43 ~~
~~I I B
~ A
~E AB ( A '~2 ) I I
A BA C BA B i $ I I aB( ~ 4 1
A
B
I I 'A8f' ? '3r I
.92 '_,
I
.9
I I
B
n
n4
$
P2 PA
Z
Figure 3 .8 Law of consecutive metals; (a) thermoelectric circuits, (b) determination of resulting emf
by linear superposition
or finally
The emf of a circuit whose measuring junction is at a temperature 01, and whose
reference junction is at a temperature equals the difference in the emfs of this circuit
when its reference junction is at a temperature of and of a circuit whose measuring
junction temperature is and its reference junction temperature is .
3.2 Thermocouples
3 2.1 General information
The combination of two dissimilar conductors, which may be metals, alloys or non-metals,
connected at one end is known as a thermocouple . In the notation shown in Figure 3 .9, their
point of connection is called the measuring junction and their free ends are referred to as
the reference junction .
MEASURING CONDUCTORS
JUNCTION 1!
REFERENCE
I JUNCTION
" "-
continuous " possibly dependence " "-
In practice
" the commonly used thermocouple materials make a compromise between the
different demands quoted above.
The emf versus temperature values of the more commonly used thermocouples are
shown in Figure 3 . 10 . In this figure, and throughout this book, the convention of quoting
the positive conductorfirst in the specific name of the thermocouple is used. The majority
of the thermocouples which will be described below are given in the standard IEC 584, Part
" f the International Electro-technical Commission . Moreover they also conform to the
national standards of many other countries such as France, Germany, Italy, Japan, Poland,
UK, USA and Russia.
80
70
60
1 ~ """"p'
.
30 --
201
1 ~/
"~
0. 0 200 400 600 800 1000 1200 1400 16DO 18W
TEMPERATURE 3 *C
The reference calibration tables of standardised thermocouples are given at the end of
the book in Tables II to IX and the tolerances upon their output values in Table X. Table XI
gives the polynomials for calculating the reverse functions t9o = f (E) for computer
applications. Typical data for standardised thermocouple wires, with commonly used
diameters are in Table XIII, where the maximum permissible working temperatures in pure
air are also quoted.
Properties of metals and alloys for thermocouples are given in Table 3.2.
T+ J-* J+ K-
Chemical 55 % Cu 95 % Ni,
Cu Fe
composition 45 % Ni 2 % Mn,
Density
8 g/cm 3 .9 1 8.85 1 7.85 1 8.7
Resistivity x10-6 S2m 0.017 4 0.49-0.51 4 0 .11-0.13 4 0.20-0.:
Thermal coefficient , 4
9 4.3 4 0.05 4 1 .8
of resistance, mean X10 -3 1/ C
value in range C 20-600 20-600 20-600 20-1!
Thermal 389 at 20'C 41 .8 75 .3 at 20 C 58 .1
conductivity W/m 'C 356 at 500C 0-300C 33 .5 at 800 C 20-701
Specific oC 481 at 20 C 418.7 461 at 20C 544
J/kg
heat 435 at 500'C 0-300'C 712 at 800 C 20-401
Coefficient of
X10 -6 1/C 1 .8 16 .8 14 .6 16
linear thermal
expansion, mean
~C 20-600 20-600 20-600 20-E
value in range
Melting
temperature C 1083 3 1270 3 1536 3 1400
B- S+ R+ B+ Kanthal P + Ka
Chemical 94 % Pt 90 % Pt 87 % Pt 70 % Pt 90 % Ni 98
composition 6 % Rh 10 % Rh 13 % Rh 30 % Rh 10 % Cr 2.
20
Density g/cm3 .5 3 20 1 19 .55 3 17 .6 2 8 .68 5
Resistivity x10-6 Urn 0.185 4 0.20 4 0.189 3 0.21 4 0.69 5
Thermal coefficient x10-3 1/C 1 .4 4
1 .8 4 1 .6 1 .3 4 0.5 5
of resistance, mean
value in range C 0-100 20-1600 0-100 0-100 20-100
Thermal 30 .2 at 23 .8 at
conductivity W/m C 20 C 100 C
Specific 146 .5 at 128 at
J/kg
heat C 0 C 20'C
Coefficient of
linear thermal x10-6 1/C 9.0 14
expansion, mean
value in range C 20-800 20-100
+ is the positive wire ; - is the negative wire ; * also known as constantan ; x ,Approximate values
Sources: I DIN 43712; 2 Bums and Hurst (1975); 3 Omega Engineering catalogue; 4 Westhoff(1965) ; 5 1
50 THERMOELECTRIC THERMOMETERS
Typical data for standardised thermocouple wires, with commonly used diameters are in
Table XIII, where the maximum permissible working temperatures in pure air are also
quoted.
Section 3.1 .2. A large number of sources of information exist on measuring junctions
(Baker et al., 1953 ; Considine, 1957 ; Kortvelyessy, 1987 ; Zanstra, 1975)
Rare metal thermocouples are usually electric-arc welded, without the use of welding flux.
The ac electric arc is formed between two graphite electrodes ensuring that there is no direct
physical contact between the thermocouple materials and the electrodes. Conductors should
be free from any mechanical defects such as indentations, kinks and non-uniformity of
diameter. Before welding, both conductors are twisted together to form about two turns. A
correctly formed junction should have the shape of a spherical bead with a smooth surface.
Overheating of thejunction should be avoided as this may cause surface porosity.
The Type J thermocouple is formed by twisting the conductors to form 2 to 3 turns after
cleaning by rubbing with fine emery paper. It is usually welded in non-reducing oxy-
acetylene flame. Before welding, both ends have to be heated to a red colour in a neutral
flame, followed by immersion in borax . The conductor with the higher melting point should
be heated first during welding . After welding, residual flux has to be removed by washing .
It is also possible to manufacture this type of thermocouple using either electric-arc welding
or hard soldering. In the case ofelectric-arc welding the thermocouple acts as one electrode
whilst the other is graphite. Hard-soldered thermocouples are only suitable for use in
measuring temperatures not exceeding 600 C.
Type K thermocouples may be manufactured using the same process as described above.
Spot or contact welding may also be used.
Type T thermocouples may be fabricated using soft or hard soldering depending upon the
maximum temperature . Units for use up to 600 C may be hard soldered whereas those for
up to 200 C may be soft soldered . Arc-welding with no flux may also be used. Typical hard
and soft solders and their corresponding properties are given in Table XIV .
In some instances the need arises to lengthen the thermocouple conductors . This is
particularly the case in rare-metal units where even very short lengths must be used to keep
the unit cost reasonable. To avoid the generation of parasitic emfs, it is recommended that
these pieces be inserted at those parts of the structure furthest from large temperature
gradients. This implies as far away from the measuring junction as possible.
After welding, all thermocouples should be stress-relieved to stabilise their properties
and neutralise any inhomogeneities . These may arise from the welding formation of
junctions, in the further steps of the fabrication process or as a result of transportation .
i
i
i je
There are many different opinions concerning the best manner of stress-relieving .
Nevertheless, it is commonly accepted that the thermocouple should be maintained at its
highest permissible working temperature for some hours .
TERMINAL HEAD
CONNECTING SCREWS
GLAND
HEAD
ADAPTER
v
FLANGE CERAMIC
SINGLE HOLE
INSULATORS
PROTECTING THERMOCOUPLE
SHEATH LEADS
MEASURING
JUNCTION
mechanical strength, temperature tolerance, electrical resistance and thermal inertia of the
wire . Thermocouple wires should be protected from bending and indentation . They should
also be kept clean to prevent the generation of parasitic emfs . Electrodes may also be made
in the form of bands which are used for non-rare metal thermocouples . Their dimensions are
usually 0 .2x3 .0 ; 0 .3x4 .0 ; 0 .5x2 .0 ; 1 .0x3 .0 and 1 .0x4 .0 mm . Thin film thermocouples are
also used (Kreider, 1992) .
Ceramic wire insulators separate the conductors from each other and from the metal
sheath. They are produced as insulating beads or as insulating tubes with one, two, four or
more bores . An essential requirement of these insulators is the ability to withstand the
highest thermocouple working temperature without losing their insulating properties or
poisoning the thermocouple . More information about different insulating materials can be
found in Section 3 .3 .2 .
Sheath materials influence the stability of calibration during long service times as well
the dynamic properties of the thermocouple . Hence, it is important to make the correct
choice of material . Typical sheath materials for temperature sensors, which are presented in
Table 3 .3, may be mainly metallic or ceramic .
Metal sheaths provide good protection for all base metal thermocouples used at
temperatures up to 1100 C . As rare metal thermocouples are liable to be contaminated by
metal vapours, an additional inner ceramic sheath is necessary . The application ranges of
metal sheaths can be considerably increased by applying an additional protecting external
layer . For example the Kawecki Chemical Co . (USA) produces tantalum sheaths protected
by a plasma deposited layer of zirconium diboride. These sheaths can stand in molten
aluminium at temperatures around 760 C for up to 14 000 hours . Metal protection tubes
used for temperature measurement in flowing media, have to stand not only the static
pressure but also dynamic forces, causing vibrations . Conically shaped wells are therefore
preferable to cylindrical wells, especially at higher flow velocities . Protecting metallic
sheaths intended for operation in air at atmospheric pressure, can be manufactured either as
seam welded or seamless tubes with a welded bottom. For temperature measurement in
molten metals or salts, as well as for use under high pressures, a type of well, which is
machined from solid bars, will be necessary . When long sheaths are required, savings in the
Table 3.3 Properties, applications and operating characteristics of materials for protective sheaths and thermowell
Applications
Material Resistance to
Temperature
Maximum
operating Oxid- Reduc-Sulphur-
temperature ising ing bearing Gases
General Specific in air atmos- atmos- atmos- with NZ Carbur-
group
Noation type (OC)
pheres pheres pheres low in OZ Gases and ste
1 2 3 4 5 6 7 8 9 10
1 Copper 150 f g p Steam, light du
2 Brass * p f p environments
4 z Iron 550 p p p f
jtechnically l
pure
c0
5 Cast iron 870 p f p Reducing
atmospheres
up to 870
Ceramic sheaths, used for temperatures over 1100 C, do not possess such good resistance
to mechanical and thermal influences as metallic sheaths. However, ceramic materials are
usually impervious to gases . As they are susceptible to thermal shocks, they require some
protection . Adequate thermal shock resistance may be provided by a second sheath with
good resistance to thermal shock.
Porous ceramics can only be used as primary sheaths, which enclose a secondary
impervious tube. An example of a regularly applied combination is an impervious mullite
sheath protected on the outside by a silicon carbide sheath. Mullite is suitable for use with
rare metal thermocouples at high temperatures. In the case of impervious aluminous
porcelain the use of a double sheath is also recommended. Recrystallised alumina, which is
impervious, can be used as a single sheath, and also for insulating wires . As mutual reaction
may occur between the sheath material and the insulation at high temperature, they must be
correctly chosen. In principle, it is thus advisable to make the wire insulation and the
protection sheath from the same material .
Cermets or ceramic-metals are regarded as the optimum insulating materials since they
possess a combination of the best properties of both ceramics and metals . They are
characterised by high mechanical strength, high resistance to thermal shock and high
operating temperatures . As they are mostly impervious they do not need any protective
sheath . For example, the external sheath of molybdenum and zirconium from Metalwerk
Plansee (Austria) with an A12 03 internal tube have been successfully used in molten steel .
Protective sheaths, based upon cermets with a constitution of chromium and A1203 from
Union Carbide Corporation (USA) are impervious, hard, abrasion resistant and have a
thermal conductivity equal to that of stainless steel. They are recommended for measuring
the temperature of molten copper up to 1150'C, and liquid aluminium and sulphur
containing gases up to 1095 C. They also find application in chemical reactors up to
1375 C. The proprietary cermet, called Metamic, from Morgan Refractories Ltd (USA), is
composed of molybdenum and aluminium oxide . When used in the temperature
measurement of liquid steel up to 1900 C, a second internal sheath ofA1203 must be added
to protect the thermocouple from poisoning . Cermets of chromium, A1203 and Si02, which
can be used up to 1450 C, necessarily require an additional inner sheath when used with
rare metal thermocouples .
- Metal primary, 3, 6, 8,
z--~ X
mineral insulated 9, 10, 11
w
Metal sheathed 6, 9, 11
~
.y
rr
b
Metal sheathed 4, 5, 6, 7, 9, 10, 11
o
Ceramic or 12, 13, 14, 15,
graphite sheathed 16, 17, 18
Straight drilled or 2, 6, 7, 9, 10
welded thermowell,
-
3o with thermocouple
insert
-d'
,~ Conical drilled 2, 6, 7, 9, 10
E" thermowell, with
thermocouple insert
THERMOCOUPLE SENSORS 59
Right-angle (corner) thermocouples are used for measuring the temperature of molten
metals or salt baths . Their shape ensures that their terminal head is not placed directly over
the high temperature medium. In this manner it is protected against both high temperature
radiation and any aggressive vapours . Their flanges, or other installation fittings, are
mounted so that they may be suitably moved, either on the vertical part to be immersed or
on the horizontal supporting part . Protecting sheaths for this type of unit are mostly
constructed using two different materials . The horizontal supporting part is made of
constructional steel, while the immersible vertical part is made of material capable of
withstanding immersion in the molten medium. Uniform sheaths, made of a single material
are often built of pure iron, which constitutes the positive conductor of an iron/ copper-
nickel thermocouple (a Type J unit) . The copper-nickel lead is then welded into the bottom
of the iron sheath giving the sensor a very low thermal inertia . Although uniform
thermocouple sensors are more expensive, they have the advantage of a longer working life .
High-pressure thermocouple sensors, shown in Figure 3 .14, are composed of what has
been called a thermocouple insert placed in a thermocouple well . Thermocouple inserts, like
those manufactured by TC Ltd for example, normally contain a short mineral insulated (Ml)
metal sheathed thermocouple attached to a standard terminal block. Inserts, which are now
popular easily exchanged replacement sub-assemblies, are pressed into the thermocouple
well and fitted with standard terminal heads . As they are spring loaded, they have improved
dynamic behaviour with good resistance to vibration.
The permissible temperature, pressure and flow velocity in different media, all of which
depend upon the constructional details, dimensions and well material may be determined
using special diagrams such as those in DIN 43763 .
Cylindrical screw-in thermocouples are normally mounted in tank or pipe-line walls
using welded fixing attachments similar to that in Figure 3 .15(a) . The washer secures the
tightness of the screw connections which may be either straight as in Figure 3 .15(a) or
conical, directly in the tank wall as in Figure 3 .15(b) .
(al (b)
METAL TUBE
SEAL
MOUNTING
FLANGE
FURNACE
WALL
Figure 3 .14 High pressure Figure 3 .15 Mounting of high pressure thermocouples
thermocouple
Conical well thermocouples, which are machined from solid bars in a variety of
materials and are usually mounted as shown in Figure 3 .15(c), may be used for high-
pressure measurements . They can be welded to or screwed into the walls .
Mineral insulated (MI) thermocouples, which have a metal sheath, were originally
designed in 1957 for temperature measurement in gas-cooled nuclear reactors . As they have
important properties, they are now widely applied in almost all other locations where their
properties may be of benefit. A considerable advantage is their thin, robust, flexible
structure which permits easy bending and twisting. Their high resistance to corrosion and
vibration, considered by Barzanty and Hans (1979), Bliss (1972) and Thomson and Fenton
(1975), combined with a short response time, has made them indispensable in both research
and industry .
Figure 3 .16 illustrates the axial cross-section of different forms of mineral insulated
thermocouples while Figure 3 .17 gives the lateral cross-section . A continuous protective
metallic sheath contains the wires of the thermocouple, which are placed in a densely
compacted mineral insulating powder . Variations may contain single, as in Figure 3 .16(a),
double, as shown in Figure 3 .16(b), or sometimes triple Ml thermocouples . The sheaths may
have diameters of between 0 .2 mm and 8 mm while they may be many metres long .
Concentric MI thermocouples, which are shown in Figure 3 .16(c) and Figure 3 .17(e) have
the terminals on both sheath ends . They have the following advantages when compared with
the normal types . As their insulation layer is thicker they possess a higher electrical
resistance between the sheath and the conductor . Longer working life and a higher
resistance to corrosion is ensured by the thicker walls of the sheath. This is also responsible
for increased thermal conductivity and electrical resistance, due to a possibly higher
pressure on the insulation layer . Smaller physical size is accompanied by two important
benefits . Firstly, there is greater ease in hermetically sealing the ends. In addition, the low
resulting thermal inertia gives a short response time . A final benefit of this form of MI
thermocouple, pointed out by Thomson and Fenton (1975), is the possibility which it
1
11 1
.
i Illll~~y~
i ~
1 ~y
1 ~
1 I I ~
1
j 1, I I j ~ 1~~"
1 1 I I I 1 1 1
.
.I = 1 1 .I1 1
I
I
i I
1 1 1
1 1
1 0 j ~ ~ I
1 " 1 1
:t t :1 1
THERMOCOUPLE SENSORS 63
95%W,5%Re / 74%W,26%Re to 2000 C. Bentley and Morgan (1986) also report the
application of Nicrosil-Nisil MI thermocouples up to about 1000 C where they still exhibit
good stability .
Compressed MgO and A1203 are usually used as insulation for MI thermocouples . The
resistance of MgO between the thermocouple and the metal protective sheath may be over
10 12 S2/m at 20 C with the ability to withstand some 250 V. Grain size for insulation need
to be in the order of 1 prn. As the hardness of A1203 makes working to this grain size
difficult it is not so popular . In exceptional circumstances, when the temperatures to be
measured are up to 2500 C, BeO and Th02 are used as the insulating medium.
Moisture causes a reduction of the resistance of insulating materials as well as exerting
an accelerating influence on wire corrosion . Thus, as Kinzie (1973) states, the insulation of
all MI thermocouples should be protected from it. Even short term exposure, for as little
time as some minutes, to humid atmospheres may considerably shorten the life-time of Ml
thermocouples . It is also impractical to attempt to dry the insulation after such exposure . For
this reason MI thermocouples have both terminations hermetically sealed to prevent the
ingression ofundesirable moisture.
Metal sheaths for use in MI thermocouples may be made of Cr, Ni and Ti stabilised
stainless steel for temperatures up to 800 C. Temperatures up to about 1100C require
inconel sheaths whilst Nicrosil may be used for Nicrosil-Nisil thermocouples up to 1250 C.
At higher temperatures Pt, Pt10%Rh, Ni, Ta, Rh and Mo are used. Additional protection of
metal sheaths may be arranged when thin outer layers of Cr or A1203 are deposited using
the plasma method . Alternatively, increased thermocouple lifetime is achieved by using
two-layer metal sheaths .
The minimum bending radius is usually 10 to 12 times the sheath diameter. Some
producers, such as TC Ltd, permit the bending radius of MI thermocouples to be as small as
four times sheath diameter, if carefully handled .
In the process of manufacture the diameter of the thermocouple sheath is reduced by
drawing. Simultaneously, this also proportionately reduces the diameter of the conductors.
To ensure that the conductors maintain their concentric position within the sheath, two, four
or six bore ceramic insulating tubes are used . During the drawing process, the insulating
tubes are crushed and compressed . A commensurate increase in their thermal conductivity
takes place. Any reduction in diameter should not be greater than about 20 to 1 . Exceeding
this ratio causes wide non-uniformity of conductor cross-section and also a 'spreading' or
distortion of the measuring junction . Concentric MI thermocouple units are much more
easily drawn and have a longer life.
Ml thermocouples may be used in a wide variety of applications. Non-exclusive
examples occur in nuclear reactors, in industries for the manufacture and production of
automobiles, aircraft, glass and chemicals and in medicine as well as in research
laboratories of all kinds . Typical proportions of the components of MI thermocouples are
given in Figure 3 .16 . The link between the thermoelectric wires and extension cables is
accomplished with a range of different plug-and-socket connectors as shown in Figure 3 .18 .
It should be noted that Type R thermocouples of extremely small diameter of about 0.08
to 0.13 mm and placed in Pt10%Rh sheaths of 0.5 to 0.8 mm in the neighbourhood of steel
or inconel elements may be a cause of important variations of their thermoelectric
characteristics . This happens after long operation at temperatures over 950 C . So far, the
mechanism ofthese variations has not been clarified.
64 THERMOELECTRIC THERMOMETERS
_ +'+
Needle thermocouples are made, for example, of enamel insulated thin Cu l CuNi
thermocouple soldered on the end of a stainless steel tube. The thermocouple is connected
to extension wires inside the head as shown in Figure 3 .19 . This type of unit may also be
made from NiCr tube serving as a positive conductor and an insulated CuNi wire soldered
into the tube . They are used predominantly for the temperature measurement of plastic,
liquid and granular substances, in medicine and research The needles are 40 to 60 mm long
and have diameters of 1 to 2 mm . They are being replaced to an increasing extent by MI
units.
DETAIL A
THERMOCOUPLE SOLDER
STEEL TUBE
TERMINAL HEAD A
COMPENSATING CABLE
MEASURING
JUNCTION TERMINAL HEAD
W/Mo and W/MoW thermocouples are characterised by high working temperatures and
low emf values. Although W / Mo types cannot be used in oxidising atmospheres they are
suitable for use in reducing environments up to about 2400C (Caldwell, 1962) . As a
reaction between Mo and Zr, Mg or A1 2 03 above 1600 C gives severe emf variations, the
50
E ."
W 40 ". o VJ 2go/oRe
~A to
5
~. .GJolo~
0
30
20
T-400/0?'b
10 Q.
" 1000
,_
0 '' 500 . . ....
. 1500 2000 2500 3000
TEMPERATURE, fl C
inner and outer insulation has to be chosen carefully . The improved resistance of W / MoW
is accompanied by lower mechanical strength, which may lead to easy breaking . In addition,
instability of emf values, with polarity reversal in the temperature range 1400-2600 C, also
limits their application (Caldwell, 1962; Kinzie, 1973 ; Lachman and McGurty, 1972) .
W / WRe and WRe /Re thermocouples have recently become more widely employed due to
the better mechanical properties of Re-alloys than W and Mo . Different alloy compositions
are used such as W / 74%W-26%Re, 95%W-5%Re / 74%W-26%Re and 97%W-
3%Re / 75%W-25%Re. Although all have similar emf temperature dependence, the latter
two are more frequently employed as they are more linear at lower temperatures and more
ductile . Replacing the W-electrode with one containing Re provides the opportunity to
stabilise the resulting thermocouple by a short duration heating process, which causes no
embrittlement of the positive conductor . Burns and Hurst (1972) have stated that this
conductor may be made still more ductile if its alloy is enriched with traces of other
elements which inhibit grain growth. Units of this type, which may be used up to
temperatures of 2500 C, resist poisoning in both neutral and hydrogen containing
atmospheres as well as in vacuum. However, Anderson and Bliss (1972) have pointed out
that they cannot tolerate the presence of oxygen, H2O, CO or CO2. Further details of the
properties of W and Re alloys are given by Vertogradsky and Chekhovskoy (1972) .
As insulating materials such as MgO and A1203 may only be employed up to 1500 C
their use in thermocouple construction is restricted to those parts ofthe wire structure whose
temperature never exceeds this value . Although BeO can be applied up to 2000 C it is
poisonous to staff (Anderson and Bliss, 1972). Details of external protective sheaths, which
are normally made ofWRe and IrRh alloys, appear in Anderson and Bliss (1972) .
Typical applications of high temperature thermocouples made of W and WRe include
the temperature measurement of the thermal insulation of spaceships, nuclear reactor cores
and the exhaust gases of gas turbines. Thermocouples from W and Re alloys, which are
considered to give the best types at high temperature units, are also manufactured as MI
types as exemplified by the Philips Thermocoax proprietary form. These MI thermocouples,
which, although rather expensive, are not very flexible, have wire insulation of MgO, BeO
and Hf0 2 and sheath diameter at 1.5 to 6 mm.
Graphite / W as a thermocouple, which was first described by Watson and Abrams in 1925,
is further detailed in Zysk and Robertson (1972). In vacuum, it can withstand temperatures
up to 1800C, whereas it can withstand reducing environments at temperatures up to
2500 C. Under the influence of moisture, its calibration undergoes some variations, which
can be restored to its original value by annealing. As shown in Figure 3 .23(a) they are
constructed as a tungsten wire in a graphite tube, which serves as both the positive electrode
COMPENSATING CABLES 67
GRAPHITE STOPPER AI,O, INSULATING TUBES W-WIRE SiC CEMENT SiC TUBE METALLIZED TERMINALS
and the sheath (Caldwell, 1962; Franks, 1962; Kinzie, 1973). They are applied in the
measurement of the temperature of molten metals, and also in carborising atmospheres . A
significant feature is polarity reversal at a temperature of about 1000 C.
Graphite /graphite with 0 .1% to 0 .2% Be, gives a thermocouple which can be used for spot
readings up to 2600 C or continuously up to 2000 C in reducing or neutral atmospheres
without any protective sheath . At 2000 C its emf is about 50 mV, which is unfortunately
different from batch to batch. Constructed with the same screwed rod and tube form as the
previous type, its resistance is about 0.5 S2 (Zysk and Robertson, 1972), while its emf is a
function of both its exact chemical composition and the technology used in its manufacture .
A great diversity ofrelevant data appears in the literature (Caldwell, 1962; Burns and Hurst,
1975) .
Other high temperature thermocouples, not described above, may be constructed from a
wide range of other elements such as W, Mo, Rh, Re, Pt, Ir, Pd, C, SiC, Au, Nb, Te, Ta and
Zr. More comprehensive descriptions and data can be found in the works by Anderson and
Bliss (1972), Bums and Hurst (1972), Caldwell (1962), Franks (1962), Kinzie (1973),
Lachman and McGurty (1972), Vertogradsky and Chekhovskoy (1972) and Zysk and
Robertson (1972) . A list of different high temperature thermocouples may be found in the
texts by Kinzie (1973) and K6rtvelyessy (1987) .
When the compensating leads are made from the same material as the thermocouple
wires, they are called extension leads . Under these conditions there is no emf generated in
the thermocouple junctions formed by connecting the extension leads to the thermocouple
wires . However, because of the cost, which may be involved in rare metal units, it is more
normal for the leads to be of a material different from that of the thermocouple. Another
reason may be the high resistance of the extension leads. As the leads are mostly of different
materials, it is essential, as a rule, to ensure that both connection points are at the same
temperature, 62 , shown in Figure 3 .24 . This guarantees that the resulting thermal emfs
generated at the connection points will cancel, as they will be of the same magnitude but
opposite polarity .
Examination of Figure 3 .24 shows that the emf, E, across the terminals of the
compensating leads is given by :
Equation (3 .28) shows that the resulting emf equals that value of emf exhibited by a
thermocouple having its reference junction at the temperature, 6, . , in a similar manner as
exposed by equation (3 .6) . Hence it is apparent that the temperature, 0, . , may be called the
reference temperature in this case also .
When using compensating cables, attention must be paid to ensure that they are
connected to the thermocouple leads with the correct polarity . Reversed polarity will lead to
far larger errors than would occur in the case when no compensating cables are used .
Moreover, Weichert et al. (1976) have shown that false matching between the compensating
cables and the thermocouple may also cause large errors . For example, consider a Type K
thermocouple matched with Type S compensating cable of correct polarity. When
TERMINAL REFERENCE
MEASURING HEAD JUNCTION
JUNCTION , ~2 .~.r _
r B'
$
A L-J ~-J
Q~'
Colour coding for compensating cables. The type of compensating cable can be identified
based on the colour coding . The internationally standardised colour code, where the
insulation colour is the same as for the positive lead, is given in Table 3 .6 . In some
Table 3 .5 Errors, A9 C, resulting from using copper wires instead of special compensating cables for
Type B thermocouples, The assumed nominal reference temperature is 9m = 20 C
Table 3 .6 Colour code for compensating cables based upon the Standard Project IEC 60584-3
countries colour marks, such as a colour thread, a line or points, are placed on the external
cable surface . In other countries, such as USA, the colour coding involves the outer cable
surface as well as the insulation of both conductors .
The compensating cables, for Type K thermocouple may be composed of different
metals or alloys and are differently marked.
National colour codes, still used, for the insulation of compensating and extension
cables are given in Table XV . It should be noted that the codes for the positive lead are not
universal, different colours being adopted in different countries .
Design of compensating cables uses the techniques associated with the technology of
pairs of solid or stranded conductors with the total cross-sectional area of about 0 .2 to
2 .5 mm2 . Solid conductors are normally used for permanent installations while stranded
cables are eminently suitable for portable instruments . The electrical insulation is made of
PCV, teflon, silicon caoutchouc, glass fibre, ceramic fibre, depending on necessary
operating temperature and humidity as detailed in Table 3 .7 . Protection of the cables may
use stainless steel wire armour against mechanical damage or screened with A1 tape with a
Cu earth wire against electromagnetic or electrostatic interference . Multi-pair cables with up
to 24 pairs, collectively or individually armoured and screened are available .
In an industrial environment the surrounding temperature may be as high as 500 C as in
the case, say, of measuring the temperature of molten steel using dip probes . Consequently,
for Type S and -R thermocouples, it is recommended by Bugden et al . (1975) and Kinzie
(1973), that three lead cables are used. The positive lead is a special Cr / Ni / Fe alloy
conductor while the negative one is a stranded conductor composed of two different
numbers of wires of two different Cr / Ni / Fe alloys . If the thermocouple is Type S this
form of compensation keeps the compensating errors well below 2 C in the whole range
of reference temperature variation between 0 C and 500 C . One of the advantages of this
method of compensation is the high resistance to oxidising atmospheres possessed by the
alloys used . This idea, which is not very popular in practice as yet, could be used to solve
many non-typical problems .
Thermocouple connectors, which are available for most standardised thermocouples, are
used for reliable quick connections between thermocouples and compensating or extension
cables . They replace the terminal head in many applications, especially those involving
many sensors . The pins and sockets of the connectors are made of alloys which are matched
to the specific type of thermocouple so that interconnection errors are prevented. Pins may
be round or flat with corresponding sockets which are standard colour coded against
thermocouple type .
Different polarity pins have different sizes and are additionally marked for polarity. For
temperatures around 210 C connector envelopes are made from glass reinforced nylon .
Higher temperatures, up to 650 C, require connector envelopes made from ceramic
material (Omega Engineering Inc ., USA) . Connector panels for up to 40 circuits, provided
with matching colour codes, are also produced . When circumstances dictate, it may be more
convenient to use terminal strips with plugs made from thermocouple grade alloys and
marked with the code and polarity of the compatible units .
Table 3.7 Thermocoup le exte nsion and compensating cable
Operating
Overall insulation
temperature Conductor Resistance to
(C) insulation Optional overbraiding Moisture Abra
<480 Glass fibre Glass fibre silicon varnished Very good Very
silicon varnished
3 .6 Reference Temperature
3 .6 .1 Calculation of correction
The correction, AE1, may be read from the thermocouple characteristics for the
temperature difference Orl > Om Figure 3 .25 (a) demonstrates how the true temperature,
6 t , may be derived from the same characteristic . Similarly, when the reference temperature
is too low, corresponding to dri > orn the corrected emf will be :
E = E'+AEZ (3 .30)
Again the correction, AEZ , may be derived as shown in Figure 3 .25(b) . If the connected
measuring instrument is calibrated in C at the nominal reference temperature, Om , then for
any other reference temperature say, 0,1 , the true measured temperature, Ot, may be found
as follows . The emf, E' , corresponding to the indicated temperature, tai , as well as the
error in the emf, AE , for the difference in reference temperature, Wrl - Orn), may both be
found from the thermocouple characteristics as in Figure 3 .25 . Combining these two values
(a) (b)
2<
E n
P E'
E
W
W
a Ea
J
J
W
S
2
H
AE,
AE2
0
0 .9i ,ft 1A2^y'rn $
TEMPERATURE 4 TEMPERATURE ,!
Figure 3.25 Graphical method to determine the correction, AE, to the thermocouple emf, E, when the
reference temperature, 9r, differs from its nominal value, 9rn .
REFERENCE TEMPERATURE 73
gives the true value of the emf as E = E'+ AE , from which the corresponding true value of
temperature, Ot, may be found. In a similar manner the corrections may be found from the
thermocouple tables .
Simplified formulae also exist for the calculation of the true emfvalues . These formulae,
which are valid in the temperature range 0 < 6, <50 C have the universal form :
where E is the emf at i~m = 0 C, is the emf at Or # 0 C and Or is the actual value of the
reference temperature in C . The coefficients k1 and k2 have values which depend upon the
specific thermocouple . They have the values given below.
S 0.0054 0.012x10-3
B 0 0
K 0.0404 0
J 0.0532 0
T 0.0406 0
It should be noted that calculation of the corrections is only used in the laboratory
measuring environment or in the case of short time industrial measurements .
Numerical example
A measuring instrument has been calibrated for a Type K thermocouple at the nominal reference
temperature, 9rn = 0 C . At the reference temperature, 9r = 30 C, the indicated temperature was,
,91 = 830 C . Using tables, find the value of emfcorresponding to 9r. Hence deduce the corrected
emf and find the corrected true temperature from tables .
Use the simplified formulae to estimate the correction and comment
Solution :
From thermocouple tables, at a temperature 91 = 830 C, E'= 34 .5 mV
HOUSING MEMBRANE
INSULATION
A' g, SWITCH
COPPER
BLOCK =t
WATER &
A g
ICE MIXT = = CAPILLARY
THERMOCOUPLE
Cu-CONDUCTORS Cu-CONDUCTORS
_ _r
I - I TO MEASURING TO MEASURING
INSTRUMENT INSTRUMENT
,.9
B A I 1-HEATING ELEMENT
42 2-TEMPERATURE
CONTROLLER
A,e-THERMOCOUPLE
A',B'-COMPENSATING
2 LEADS
A B A B I, _
shown in Figure 3 .27(b), all the terminals of the copper conductors should be maintained
at the same but optionally chosen temperature . Both circuits produce equivalent results .
5. Heatpipe thermostat. The highest possible long time temperature stability is achieved in
heat-pipe thermostats, reaching even 0 .03 C in 24 hours or 0 .05 C in three months .
Wang and Mo (1996) report a thermostat using a semi-continuous electronic resistance
bridge controller, which gave uniformity of temperature distortion of better than
0 .02 C . Typical application range of this thermostat is in laboratory measurements .
All the devices described so far can be built for one or many thermocouples . It is also
possible to eliminate the effect of fluctuating reference temperature on the readings from
thermocouple thermometers by the use of one of three main kinds of correcting devices.
These may be incorporated into thermocouple terminal heads, into measuring instruments or
may form a separate unit .
Correcting bridge circuit. This circuit is connected in series with the thermocouple
leads . Correction is accomplished by generating an additional voltage which is a function
of the actual value of the reference temperature . The circuit of Figure 3 .28 is a
representation of the method . It consists of a do resistance bridge which is balanced at
the assumed nominal reference temperature (mostly +20 C) . The resistors, marked 2 in
the circuit, are made of constantan or manganin whereas the resistor, marked 1, which is
of Cu or Ni, has a temperature dependent value . If a simultaneous deviation of the
reference temperature, 6, , and of the bridge temperature occurs, relative to the assumed
reference value, orn, an off-balance voltage will be generated . When this voltage is
added in series with the thermocouple emf it reaches its nominal value, corresponding to
)r = 6m . A corresponding reversal in the polarity of this generated voltage depends on
7
whether the actual reference temperature is below or above the nominal value . To adapt
the circuit to different types of thermocouple, the current limiting resistor and supply
voltage are modified . The temperature sensitive resistor must be placed so that it always
adopts the temperature of the terminals of the compensating leads . In most cases, at
20 C, all of the bridge resistors have a resistance of 1 S2 . A precision of about 0 .5 C
in the correction of the readings is achievable using this circuit. More details may be
76 THERMOELECTRIC THERMOMETERS
Cu-CONDUCTORS
A'
4
COMPENSATING 1 TO MEASURING
LEADS 12 INSTRUMENT
B'
2 25,
2K L CURRENT LIMITING
RESISTOR
found in Hunsinger (1966) . There are also miniaturised bridge circuits ('Electronic Ice
Point' by Omega Engineering Inc .) . They include a self-contained battery, which is built
into the thermocouple connectors, giving over 2500 hours continuous operation . These
units correspond to the same colour codes as the connectors .
1 . Electronic correcting devices . As expected, electronic technology offers an attractive
solution to the problem of correction for all these thermocouple circuits . An operational
amplifier is used to amplify the emf of the thermocouple into direct signals with a span of
10 V . This is eminently suitable, as it allows compatible operation with modem
computer processing techniques as well as contemporary signal communication and
transmission circuits . The actual reference temperature is measured by an additional
temperature sensor, which may be, for example, a transistor . After suitable conditioning
of the transistor derived temperature signal a correcting signal is fed to the operational
amplifier. These circuits can provide a precision of about 0 .8 C within the reference
temperature range of around 5 to 45 C . A temperature of 0 C is assumed as the nominal
reference .
2 . Correcting devices built into the measuring instruments . In this frequently used type, the
terminals of the compensating leads must then be connected directly to the instrument
terminals . Comparisons of different stabilisation or correction methods are given in
Table 3 .8 .
3 .7 Measuring Circuits
3.7 .1 Introduction
Thermoelectric thermometers use the following measuring circuits :
Only the first two types are described below as they are the only ones used exclusively
for thermocouple sensors. The other types, which can be applied to measure the electrical
output signals of other sensors are described in Chapter 12 .
REFERENCE ADJUSTABLE
JUNCTION RESISTOR
M
A, r_ 15-1
Ra (~)
V=E RM = RM (3 .32)
R RT + Rc + Rc + R a + R M I + RM
where E is the electromotive force (more precisely thermoelectric force) at the measured
temperature, Ot, and reference temperature, 0,- , and R l is the loop resistance of the circuit
excluding RM .
Loop resistance, which is the sum of the resistances of each element in the loop external
to the measuring instrument, is given by
R, = R T + R c + R c + Ra (3 .33)
If the circuit also comprises a correcting bridge circuit, its resistance must also be
considered and included in equation (3 .33) . The nominal standardised value of the loop
resistance is, in most cases, 20 S2 . As the thermocouple resistance, R T , is a function of the
measured temperature and the resistances, R C and Rc , are functions of ambient
temperature, it is never possible to precisely satisfy the condition R1 =RIn . Finally, the
reference temperature, 6, , must equal the nominal value, 0" , for which the instrument has
been calibrated .
In deflection type circuits, moving-coil millivoltmeters, calibrated in C (or in F) are
used . Their internal resistance should be as high as possible, preferably between 80 S2 and
600 S2 . They should be vibration proof with a light and consequently low inertia moving
element . Industrial types, which belong to class 1 or 1 .5, may have a measuring range as
small as 5 to 10 mV . As they are calibrated at 0, 20 or 50 C nominal reference
temperatures, they should be properly chosen for use in these actual operating conditions .
Unless the instrument has a built-in correcting device, the pointer position for zero emf
should agree with the actual reference temperature . Directly coupled amplifiers are being
increasingly employed to boost electro-thermal signals before they are measured by
deflection type instruments .
In instruments with built-in correcting devices, the ends of the compensating leads
should be connected directly to the instruments terminals .
3 .7.3 Potentiometers
As shown in Figure 3 .30 the principle of operation of the potentiometer method is the
comparison of voltages . In this case the thermocouple thermoelectric force, EX, which is to
be measured, is compared with the adjustable voltage, Vc , of opposite polarity . This variable
MEASURING CIRCUITS 79
voltage is generated in the potentiometer instrument. From Figure 3 .30 the balance
condition corresponds to :
E X -Vc = 0 (3 .34)
This balance condition occurs when a zero current is indicated by the zero current
detecting galvanometer . The value of the emf is indicated by the slider position of the
resistor, As potentiometer type indications are independent of the circuit resistance, they
RP'
attain an accuracy which is higher than that of deflection type instruments . A limitation in
loop resistance, which may be high, is only imposed by the necessary sensitivity . The two
different types of potentiometers are variable voltage potentiometers, sometimes called the
Poggendorf type, and variable current potentiometers, also called the Lindeck-Rothe type .
EX -IC R C = 0 (3 .35)
During the measurement the current; Ic , should have a constant, precisely prescribed value.
Thus the potentiometer has a stabilised voltage source, fed either from mains or battery . The
error limit of the potentiometer is from 0.1 to 0 .3 % and its measurement range, divided into
many subranges, is mostly zero to 50 mV . Manually operated potentiometers, which are
usually portable, find common use in laboratories and for more precise short-term or
industrial control measurements . Some potentiometers also have a temperature scale for
standardised thermocouples. In this case the actual ambient temperature, which is equal to
the reference temperature, is set on a separate scale based on indications from a built in
mercury-in-glass thermometer. Such a potentiometer may have an additional voltage output
for checking or calibrating the measuring instruments of thermoelectric thermometers .
SLIDE-WIRE
POTENTIOMETER
my
_ R,
~+
THERMOCOUPLE ~,
V~
A A +
i _Ex, i
B B'
Cu
COMPENSATING LEADS ZERO DETECTOR
V_=CONST . 41
+ - A B
R, R Ir =CONST
SLIDE-WIRE POTENTIO-
S V METER
MV A B'
SHUNT
Rs f ZERO
DETECTOR
- (--7
+ L-%~(41
V
REFERENCE
TEMPERATURE STABILIZED 0. C.
COMPENSATION VOLTAGE SOURCE
_
I Rs { a Ro c
. + I
I _ E. I _
~~---j
A, B. Rs
r R, R
CURRENT
A B Rv LIMITING
ZERO RESISTOR
SUPPRESSION b
~t RESISTOR
74Tr SLIDE-WIRE SERVO MOTOR C
S POTENTIOMETER z
R c . This voltage drop is adjusted by changing the compensating current Ic using a slide-wire
resistor, Rr, satisfying the condition :
EX - ICRc = 0
CURRENT SETTING
RESISTOR
E \ Rr r
+
A B
INDICATOR CALIBRATED
IN mV OR C mA
R, A' B'
-V~-.__ Cu
lr
i iE~
ZERO DETECTOR
3 .8 References
Anderson, T.M . and Bliss, P. (1972) Tungsten-rhenium thermocouples. Summary Report appearing
in Temperature: Its Measurement and Control in Science and Industry, 4(3), Instrument Society of
America, Pittsburgh, 1735-1746 .
Baker, H .D ., Ryder, E.A . and Baker, N.H . (1953) Temperature Measurement in Engineering, 1, John
Wiley and Sons, New York .
Barzanty, J. and Hans, R. (1979) Fehlerfrueherkennung an Mineralstoff isolierten Mantel-
thermoelementen and Mantelmessleitem, Techn . Messen, 46(6), 239-248.
Bentley, R.E. and Morgan, T.L . (1986) Ni-based thermocouples in the mineral insulated metal
sheathed format: thermoelectric instabilities to 1100 C . J. Phys. E . Sc. Instr., 19, 262-268 .
Bliss, P . (1972) Fabrication of high-reliability sheathed thermocouples; Temperature: Its
Measurement and Control in Science and Industry, 4(3), Instrument Society of America,
Pittsburgh, 1797-1803 .
Brookes, C ., Chandler, T.R .D . and Chu,B. (1985) Nicrosil-nisil : a new high stability thermocouple
for the industrial user. Measurement and Control, 18(9), 245-248 .
Bugden, W.G ., Tomlinson, J.A . and Selman, G.L . (1975) Improved compensating lead system for
platinum-base thermocouples . Temperature Measurement: Conference Series No 26 . Institute of
Physics, London, 1975, 181-187.
Burley, N.A. (1987) A novel advanced integrally sheathed Type N thermocouple of ultrahigh
thermoelectric stability. Proc . TEMPEKO 87, 3-rd Int Conf Temperature Measurement in Science
and Industry, Sheffield, Institute of Physics, London, 115-125.
Burley, N.A ., Burns, G.W . and Powell, R.L . (1975) Nicrosil and Nisil, their development and
standardisation . Temperature Measurement: Conference Series No . 26. Institute of Physics,
London, 1975, 162-171 .
Burley, N.A . and Jones, T.P . (1975) Practical performance of Nicrosil-Nisil thermocouples .
Temperature Measurement: Conference Series No . 26 . Institute of Physics, London, 1975, 172-
180.
Burns, G.B . (Editor) (1992) New reference functions for Platinum- 13%Rhodium versus Platinum
(Type R) and Platinum-30%Rhodium versus Platinum-6%Rhodium (Type B) thermocouples based
on ITS-90 . Temperature: Its Measurement and Control in Science and Industry, American Institute
of Physics, New York, 559-564.
REFERENCES 83
Burns, G.W. and Hurst, W.S . (1972) Studies of performance of W-Re type thermocouples.
Temperature : Its Measurement and Control in Science and Industry, 4(3), Instrument Society of
America, Pittsburgh, 1751-1766.
Burns, G.W. and Hurst, W.S . (1975) Thermocouple thermometry. Temperature Measurement:
Conference Series No. 26 . Institute of Physics, London, 1975, 144-161 .
Caldwell, F.R . (1962) Thermocouple materials. Temperature : Its Measurement and Control in
Science andIndustry, 3(2), Reinhold Publ . Co ., New York, 81-134.
Considine, D. (Ed.) (1957) Process Instruments and Controls Handbook. McGraw-Hill, New York .
Franks, E. (1962) High-temperature thermocouples using non-metallic members. Temperature : Its
Measurement and Control in Science andIndustry, 3(2), Reinhold Publ . Co ., New York, 189-194.
Groot de, M.J . (1992) Measurements towards the development of new standards tables for Platinum-
Rhodium / Platinum thermocouples, Temperature : Its Measurement and Control in Science and
Industry, 6(1), American Institute of Physics, New York, 553-557.
Hall, J.A. and Barber, C.R. (1964) Calibration of temperature measuring instruments . Ed 3 Notes on
Applied Science No . 12, 1964, National Physical Laboratory, London .
Hunsinger, W. (1966) Temperaturmessung : Handbuch der Physik. Springer-Verlag, Berlin .
Kinzie, P.A. (1973) Thermocouple Temperature Measurement, John Wiley and Sons, New York .
K6rtvelyessy, L. (1987) Thermoelement Praxis, 2nd edition. Vulkan-Verlag, Essen.
Kreider, K.G . (1992) Thin film thermocouples, Temperature : Its Measurement and Control in-
Science and Industry, 6(1), American Institute of Physics, New York, 643-647.
Lachman, L.C . and McGurty, L. A. (1972) The use of refractory metals for ultra-high temperature
thermocouples, Temperature : Its Measurement and Control in Science and Industry, 4(2),
Reinhold Publ . Co ., New York, 177-187.
Lieneweg, H. (1975) Handbuch, Technische Temperaturmessung, F. Vieweg, Braunschweig .
Omega Engineering Inc., USA (1999) Temperature Measurement Handbook.
Reed, R.P . (1992) Absolute Seebeck thermometric characteristics - principles, significance and
applications, Temperature : Its Measurement and Control in-Science and Industry, Vol. 6, Part 1,
American Institute of Physics, New York, 503-508.
Roeser, W.F . and Wensel, H.T . (1941) Appendix, Temperature : Its Measurement and Control in-
Science and Industry, Reinhold Publ . Co ., New York, 1308-1310.
Sutton, G.R . (1975) Thermocouple referencing, Temperature Measurement : Conference Series
No 26, Institute of Physics, London, 1975, 188-194.
Thomson, A. and Fenton, A.W . (1975) High-integrity, small-diameter mineral-insulated thermo-
couples, Temperature Measurement: Conference Series No 26, Institute of Physics, London, 1975,
195-202.
Vertogradsky, V.A . and Chekhovskoy, V.Ya. (1972) Thermal properties of Tungsten-Rhenium alloys
used in high-temperature thermocouples. Temperature: Its Measurement and Control in Science
andIndustry 4(3), Instrument Society of America, Pittsburgh, 1747-1749.
Wang Zhensen and Mo Honggi (1996) A 0 C thermostat using freon heat-pipe technology for
realizing reference junction temperature in thermocouple calibration, Proc TEMPMEKO 96,
Torino, Italy, 213-217.
Weichert, L. (Editor) (1976) Temperaturmessung in der Technik-Grundlagen and Praxis . Lexika-
Verlag, Grafenau .
Westhoff, G. (1965) Grundlagen and Praxis der Temperaturmessung and Temperaturregelung.
Vulkan-Verlag, Essen.
White, W.P . (1941) Potentiometers for thermoelectric measurements, Temperature: Its Measurement
and Control in Science andIndustry, Reinhold Publ . Co ., New York, 265-278 .
84 THERMOELECTRIC THERMOMETERS
Zanstra, P.E . (1975) Welding uniform sized thermocouple junctions from thin wires, J. Phys E: Sc .
Instr, 8, 262-263
Zysk, E.D . and Robertson, A.R . (1972) Newer thermocouple materials. Temperature : Its
Measurement and Control in Science and Industry, 4(3), Instrument Society of America,
Pittsburgh, 1697-1734.