Practical Temperature

Measurements*
Thermocouple RTD Thermistor I. C. Sensor

V R R V or I

RESISTANCE

RESISTANCE

or CURRENT
VOLTAGE
VOLTAGE

T T T T
TEMPERATURE TEMPERATURE TEMPERATURE TEMPERATURE
□ Self-powered □ Most stable □ High output □ Most linear
□ Simple □ Most accurate □ Fast □ Highest output
□ Rugged □ More linear than □ Two-wire ohms □ Inexpensive
□ Inexpensive thermocouple measurement
□ Wide variety
□ Wide temperature
range

Disadvantages

□ Non-linear □ Expensive □ Non-linear □ T<200°C

□ Low voltage □ Current source □ Limited temperature □ Power supply
□ Reference required required range required
Advantages

□ Least stable □ Small ∆ R □ Fragile □ Slow

□ Least sensitive □ Low absolute □ Current source □ Self-heating
resistance required □ Limited configurations
□ Self-heating □ Self-heating

Figure 1
TABLE OF CONTENTS
APPLICATION NOTES-PRACTICAL TEMPERATURE MEASUREMENTS Page
Common Temperature Transducers.................................................................................... Z-19
Introduction............................................................................................................................ Z-20
Reference Temperatures................................................................................................... Z-21
The Thermocouple................................................................................................................ Z-21
Reference Junction............................................................................................................ Z-22
Reference Circuit............................................................................................................... Z-23
Hardware Compensation................................................................................................... Z-24
Voltage-to-Temperature Conversion.................................................................................. Z-25
Practical Thermocouple Measurement................................................................................ Z-27
Noise Rejection.................................................................................................................. Z-27
Poor Junction Connection.................................................................................................. Z-29
Decalibration...................................................................................................................... Z-29
Shunt Impedance............................................................................................................... Z-29
Galvanic Action.................................................................................................................. Z-30
Thermal Shunting............................................................................................................... Z-30
Wire Calibration................................................................................................................. Z-30
Diagnostics........................................................................................................................ Z-31
Summary............................................................................................................................ Z-32
The RTD.................................................................................................................................. Z-33
History................................................................................................................................ Z-33
Metal Film RTD’s............................................................................................................... Z-33
Resistance Measurement.................................................................................................. Z-34
3-Wire Bridge Measurement Errors................................................................................... Z-35
Resistance to Temperature Conversion............................................................................ Z-35
Practical Precautions......................................................................................................... Z-36
* Copyright © 1997, 2000 Agilent Technologies, Inc. Reproduced with Permission.
Z-19
 TABLE OF CONTENTS
APPLICATION NOTES-PRACTICAL TEMPERATURE MEASUREMENTS (con’t)
The Thermistor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z-36
Linear Thermistors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z-37
Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z-37
Monolithic Linear Temperature Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z-37
Appendix A-The Empirical Laws of Thermocouples . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z-37
Appendix B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z-38
Thermocouple Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z-38
Base Metal Thermocouples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z-38
Standard Wire Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z-39
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z-40 Z

Synthetic fuel research, solar energy conversion and Florentine thermometer, which incorporated sealed
new engine development are but a few of the construction and a graduated scale.
burgeoning disciplines responding to the state of our In the ensuing decades, many thermometric scales
dwindling natural resources. As all industries place were conceived, all based on two or more fixed points
new emphasis o n energy efficiency, the fundamental One scale, however, wasn’t universally recognized until
measurement of temperature assumes new importance. the early 1700’s, when Gabriel Fahrenheit, a Dutch
The purpose of this application note is to explore the instrument maker, produced accurate and repeatable
more common temperature monitoring techniques and mercury thermometers. For the fixed point on the low
introduce pro cedures for improving their accuracy. end of his temperature scale, Fahrenheit used a mixture
We will focus on the four most common temperature of ice water and salt (or ammonium chloride). This was
transducers: the thermocouple, the RTD, the the lowest temperature he could reproduce, and he
thermistor and the integrated circuit sensor. Despite labeled it “zero degrees”. For the high end of his
the widespread popularity of the thermocouple, it is scale, he chose human blood temperature and called
frequently misused. For this reason, we will concentrate it 96 degrees.
primarily on thermocouple measurement techniques. Why 96 and not 100 degrees? Earlier scales had
Appendix A contains the empirical laws of been divided into twelve parts. Fahrenheit, in an
thermocouples which are the basis for all derivations apparent quest for more resolution divided his scale
used herein. Readers wishing a more thorough into 24, then 48 and eventually 96 parts.
discussion of thermocouple theory are invited to read The Fahrenheit scale gained popularity primarily
REFERENCE 17 in the Bibliography. because of the repeatability and quality of the
For those with a specific thermocouple application, thermometers that Fahrenheit built.
Appendix B may aid in choosing the best type Around 1742, Anders Celsius proposed that the
of thermocouple. melting point of ice and the boiling point of water be
Throughout this application note, we will emphasize used for the two benchmarks. Celsius selected zero
the practical considerations of transducer placement, degrees as the boiling point and 100 degrees as the
signal conditioning and instrumentation. melting point. Later, the end points were reversed and
the centigrade scale was born. In 1948 the name was
Early Measuring Devices - Galileo is credited with
officially changed to the Celsius scale.
inventing the thermometer, circa 1592.1, 2, 3 In an open
In the early 1800’s William Thomson (Lord Kelvin),
container filled with colored alcohol he suspended a
developed a universal thermodynamic scale based
long narrow-throated glass tube, at the upper end of
upon the coefficient of expansion of an ideal gas. Kelvin
which was a hollow sphere. When heated, the air in
established the concept of absolute zero and his scale
the sphere expanded and bubbled through the liquid.
remains the standard for modern thermometry.
Cooling the sphere caused the liquid to move up the
The conversion equations for the four modern
tube.1 Fluctuations in the temperature of the sphere
temperature scales are:
could then be observed by noting the position of the
°C = 5/9 (°F - 32) °F= 9/5 °C + 32
liquid inside the tube. This “upside-down” thermometer
was a poor indicator since the level changed K = °C + 273.15 °R= °F + 459.67
with barometric pressure and the tube had no scale. The Rankine Scale (˚R) is simply the Fahrenheit
Vast improvements were made in temperature equivalent of the Kelvin scale, and was named after
measurement accuracy with the development of the an early pioneer in the field of thermodynamics,
W.J.M. Rankine.
1, 2, 3
Refer to Bibliography 1,2,3.

Z-20
 Reference Temperatures
We cannot build a temperature divider as we can a Metal A
voltage divider, nor can we add temperatures as we +
would add lengths to measure distance. We must rely eAB
upon temperatures established by physical phenomena
which are easily observed and consistent in nature.
–
Metal B
The International Practical Temperature Scale (IPTS)
is based on such phenomena. Revised in 1968, it eAB = SEEBECK VOLTAGE
establishes eleven reference temperatures. Figure 3
eAB = Seebeck Voltage
Since we have only these fixed temperatures to use
as a reference, we must use instruments to interpolate
Figure 3 effect. The most
All dissimilar metals exhibit this
common combinations of two metals are listed in
between them. But accurately interpolating between Appendix B of this application note, along with
these temperatures can require some fairly exotic their important characteristics. For small changes in
transducers, many of which are too complicated or temperature the Seebeck voltage is linearly proportional
expensive to use in a practical situation. We shall limit to temperature:
our discussion to the four most common temperature
transducers: thermocouples, resistance-temperature ΔeAB = αΔT
detector’s (RTD’s), thermistors, and integrated Where α, the Seebeck coefficient, is the constant of
circuit sensors. proportionality.
Measuring Thermocouple Voltage - We can’t
measure the Seebeck voltage directly because we
IPTS-68 REFERENCE must first connect a voltmeter to the thermocouple,
TEMPERATURES and the voltmeter leads themselves create a new
EQUILIBRIUM POINT K 0
C thermoelectric circuit.
Triple Point of Hydrogen 13.81 -259.34
Let’s connect a voltmeter across a copper-constantan
Liquid/Vapor Phase of Hydrogen 17.042 -256.108
(Type T) thermocouple and look at the voltage output:
at 25/76 Std. Atmosphere
Boiling Point of Hydrogen 20.28 -252.87 J3
Boiling Point of Neon 27.102 -246.048
Cu Cu
Triple Point of Oxygen 54.361 -218.789 + +
Boiling Point of Oxygen 90.188 -182.962 v V1 J1
– C –
Triple Point of Water 273.16 0.01 Cu
Boiling Point of Water 373.15 100
J2
Freezing Point of Zinc 692.73 419.58
EQUIVALENT CIRCUITS
Freezing Point of Silver 1235.08 961.93
Freezing Point of Gold 1337.58 1064.43
Cu + – Cu Cu
Table 1 V3
+ J1
J3 + J1
V1 V1

THE THERMOCOUPLE Cu
+
V2
–
–
+ –
–

When two wires composed of dissimilar metals are C Cu V2 C

joined at both ends and one of the ends is heated, there J2 J2
is a continuous current which flows in the thermoelectric
circuit. Thomas Seebeck made this discovery in 1821. MEASURING
V3 = 0 JUNCTION VOLTAGE WITH A DVM
Figure 4

We would like the voltmeter to read only V1, but by

Metal A Metal C connecting the voltmeter in an attempt to measure
the output of Junction J1, we have created two more
metallic junctions: J2 and J3. Since J3 is a
copper-to-copper junction, it creates no thermal EMF
Metal B (V3 = 0), but J2 is a copper-to-constantan junction
THE SEEBECK EFFECT which will add an EMF (V2) in opposition to V1. The
Figure 2 resultant voltmeter reading V will be proportional to the
The Seebeck Effect temperature difference between J1 and J2. This says
Figure
If this circuit is broken 2 the net open
at the center, that we can’t find the temperature at J1 unless we first
circuit voltage (the Seebeck voltage) is a function of the find the temperature of J2.
junction temperature and the composition of the two
metals.
Z-21
 The Reference Junction

Cu Cu + + T
+ + V1 J1
v V1 J1 v –
– + – –
Cu Cu V2 C + –
–
V2
Voltmeter
J2 J2
T=0°C

Ice Bath
EXTERNAL REFERENCE JUNCTION Z
Figure 5 The copper-constantan thermocouple shown in
Figure 5 is a unique example because the copper wire
One way to determine the temperature of J2 is to is the same metal as the voltmeter terminals. Let’s
physically put the junction into an ice bath, forcing use an iron-constantan (Type J) thermocouple instead
its temperature to be 0˚C and establishing J2 as the of the copper-constantan. The iron wire (Figure 6)
Reference Junction. Since both voltmeter terminal increases the number of dissimilar metal junctions in
junctions are now copper-copper, they create no the circuit, as both voltmeter terminals become Cu-Fe
thermal emf and the reading V on the voltmeter is thermocouple junctions.
proportional to the temperature difference between J1
and J2.
V3
J3
Now the voltmeter reading is (see Figure 5): -+
+
V = (V1 - V2) ≅ α(tJ1 - tJ2) v V1
If we specify TJ1 in degrees Celsius: –
-+
TJ1 (˚C) + 273.15 = tJ1 V1 = V
Voltmeter V4 J4
if V3 = V4
then V becomes: i.e., if
TJ3 = TJ4
V = V1 - V2 = α [(TJ1 + 273.15) - (TJ2+ 273.15)]
= α (TJ1 - TJ2) = α (TJ1 - 0) JUNCTION VOLTAGE CANCELLATION
V = αTJ1 Figure 7

We use this protracted derivation to emphasize that If both front panel terminals are not at the same
the ice bath junction output, V2, is not zero volts. It is a temperature, there will be an error. For a more precise
function of absolute temperature. measurement, the copper voltmeter leads should be
By adding the voltage of the ice point reference extended so the copper-to-iron junctions are made on
junction, we have now referenced the reading V to 0˚C. an isothermal (same temperature) block:
This method is very accurate because the ice point Isothermal Block
temperature can be precisely controlled. The ice point J3
is used by the National Bureau of Standards (NBS) as Cu
Cu Fe
the fundamental reference point for their thermocouple
+
tables, so we can now look at the NBS tables and v T1
– Fe C
directly convert from voltage V to Temperature TJ1. V2
Cu
Voltmeter Cu
J3 J4
TREF
Fe
Cu
+
v J1
–
Ice Bath
Cu C
Fe REMOVING JUNCTIONS FROM DVM TERMINALS
J4 Figure 8
J2
The isothermal block is an electrical insulator but a
good heat conductor, and it serves to hold J3 and J4 at
Ice Bath the same temperature. The absolute block temperature
is unimportant because the two Cu-Fe junctions act in
IRON-CONSTANTAN COUPLE opposition. We still have
Figure 6 V = α (T1 - TREF)

Z-22
 Reference Circuit
Let’s replace the ice bath with another isothermal This is a useful conclusion, as it completely eliminates
block the need for the iron (Fe) wire in the LO lead:
Isothermal Block
Cu Fe
Cu Fe +
HI
v J3 J1
J1 Cu
J3 C
LO
–
Cu Fe C
Voltmeter J4 J REF J4
TREF Isothermal Block
TREF
ELIMINATING THE ICE BATH
Figure 9a EQUIVALENT CIRCUIT
Figure 11
The new block is at Reference Temperature TREF, and
because J3 and J4 are still at the same temperature, we Again, V = α (TJ1 - TREF), where α is the Seebeck
can again show that coefficient for an Fe-C thermocouple.
V = α (T1-TREF) Junctions J3 and J4, take the place of the ice bath.
These two junctions now become the Reference
This is still a rather inconvenient circuit because we
Junction.
have to connect two thermocouples. Let’s eliminate the
Now we can proceed to the next logical step: Directly
extra Fe wire in the negative (LO) lead by combining
measure the temperature of the isothermal block
the Cu-Fe junction (J4) and the Fe-C junction (JREF).
(the Reference Junction) and use that information to
We can do this by first joining the two isothermal
compute the unknown temperature, TJ1.
blocks (Figure 9b).
Block Temperature = TREF
Cu Fe
HI
Cu J3
J1
J3 Fe
LO + +
Cu Fe C v V1 J1
J4 – –
J REF J4 C
Voltmeter Cu
Isothermal Bloc k @ TREF RT
JOINING THE ISOTHERMAL BLOCKS
Figure 9b EXTERNAL REFERENCE JUNCTION-NO ICE BATH
Figure 12
We haven’t changed the output voltage V. It is still
V = α (TJ1 - TJREF ) A thermistor, whose resistance RT is a function
Now we call upon the law of intermediate metals of temperature, provides us with a way to measure
(see Appendix A) to eliminate the extra junction. This the absolute temperature of the reference junction.
empirical “law” states that a third metal (in this case, Junctions J3 and J4 and the thermistor are all assumed
iron) inserted between the two dissimilar metals of a to be at the same temperature, due to the design of
thermocouple junction will have no effect upon the the isothermal block. Using a digital multimeter under
output voltage as long as the two junctions formed by computer control, we simply:
the additional metal are at the same temperature: 1) Measure RT to find TREF and convert TREF
to its equivalent reference junction
Metal A Metal B Metal C Metal A Metal C
= voltage, VREF , then
2) Measure V and add VREF to find V1,
and convert V1 to temperature TJ1.
Isothermal Connection This procedure is known as Software Compensation
because it relies upon the software of a computer to
compensate for the effect of the reference junction.
Thus the low lead in Fig. 9b: Becomes:
The isothermal terminal block temperature sensor can
C Cu C be any device which has a characteristic proportional
Cu Fe = to absolute temperature: an RTD, a thermistor, or an
integrated circuit sensor.
TREF
It seems logical to ask: If we already have a
TREF device that will measure absolute temperature (like
an RTD or thermistor), why do we even bother with
LAW OF INTERMEDIATE METALS
a thermocouple that requires reference junction
Figure 10

Z-23
compensation? The single most important answer
to this question is that the thermistor, the RTD, and
the integrated circuit transducer are only useful over Fe

a certain temperature range. Thermocouples, on the

C
other hand, can be used over a range of temperatures,
and optimized for various atmospheres. They are much + HI

more rugged than thermistors, as evidenced by the fact –

RT
LO
that thermocouples are often welded to a metal part
or clamped under a screw. They can be manufactured Voltmeter
Pt
on the spot, either by soldering or welding. In short,
thermocouples are the most versatile temperature All Copper Wires

transducers available and, since the measurement

system performs the entire task of reference
Pt - 10% Rh
Z
Isothermal Block
compensation and software voltage to-temperature (Zone Box)

conversion, using a thermocouple becomes as easy as ZONE BOX SWITCHING

Zone Box Switching
Figure 13
connecting a pair of wires. Figure 13
Thermocouple measurement becomes especially
convenient when we are required to monitor a large
number of data points. This is accomplished by using
the isothermal reference junction for more than one
thermocouple element (see Figure 13). Hardware Compensation
A reed relay scanner connects the voltmeter to the Rather than measuring the temperature of the
various thermocouples in sequence. All of the voltmeter reference junction and computing its equivalent
and scanner wires are copper, independent of the type voltage as we did with software compensation, we
of thermocouple chosen. In fact, as long as we know could insert a battery to cancel the offset voltage of the
reference junction. The combination of this hardware
what each thermocouple is, we can mix thermocouple
compensation voltage and the reference junction
types on the same isothermal junction block (often
voltage is equal to that of a 0°C junction.
called a zone box) and make the appropriate
modifications in software. The junction block The compensation voltage, e, is a function of the
temperature sensor RT is located at the center of the temperature sensing resistor, RT. The voltage V is
block to minimize errors due to thermal gradients. now referenced to 0°C, and may be read directly and
Software compensation is the most versatile converted to temperature by using the NBS tables.
technique we have for measuring thermocouples. Many Another name for this circuit is the electronic ice point
thermocouples are connected on the same block, reference.6 These circuits are commercially available
copper leads are used throughout the scanner, and the for use with any voltmeter and with a wide variety of
technique is independent of the types of thermocouples thermocouples. The major drawback is that a unique
chosen. In addition, when using a data acquisition ice point reference circuit is usually needed for each
system with a built-in zone box, we simply connect the individual thermocouple type.
thermocouple as we would a pair of test leads. All of Figure 15 shows a practical ice point reference
the conversions are performed by the computer. The circuit that can be used in conjunction with a reed relay
one disadvantage is that the computer requires a small scanner to compensate an entire block of thermocouple
amount of additional time to calculate the reference inputs. All the thermocouples in the block must be of the
junction temperature. For maximum speed we can use same type, but each block of inputs can accommodate
hardware compensation. a different thermocouple type by simply changing gain
resistors.

Cu Fe + Cu Fe
+ Cu Fe
T T +
C T
= =
v C v
C
Fe
Cu Fe Cu Fe
– – Cu
– + Cu
– RT
+
Cu
-

0ϒC

HARDWARE COMPENSATION CIRCUIT

6
Refer to Bibliography 6. Figure 14

Z-24
OMEGA TAC-Electronic ice point™ and
Thermocouple Preamplifier/Linearizer Plugs
into Standard Connector
OMEGA Electronic ice point™ Built into Thermocouple Connector -”MCJ”

Cu Fe

OMEGA ice point™ Reference Chamber.

Electronic Refrigeration Eliminates Ice Bath
Cu C
RH

By examining the variations in Seebeck coefficient,

Integrated Temperature
we can easily see that using one constant scale
Sensor
factor would limit the temperature range of the system
PRACTICAL HARDWARE COMPENSATION and restrict the system accuracy. Better conversion
Figure 15 accuracy can be obtained by reading the voltmeter
and consulting the National Bureau of Standards
Thermocouple Tables4 on page Z-203 in this Handbook
The advantage of the hardware compensation circuit
or electronic ice point reference is that we eliminate - see Table 3.
the need to compute the reference temperature. This T = a0 +a1 x + a2x2 + a3x3 . . . +anxn
saves us two computation steps and makes a hardware where
compensation temperature measurement somewhat T = Temperature
faster than a software compensation measurement. x = Thermocouple EMF in Volts
  HARDWARE COMPENSATION SOFTWARE COMPENSATION a = Polynomial coefficients unique to each
Fast Requires more computer thermocouple
Restricted to one thermocouple manipulation time
type per card Versatile - accepts any thermocouple n = Maximum order of the polynomial
As n increases, the accuracy of the polynomial
TABLE 2
improves. A representative number is n = 9 for ± 1˚C
accuracy. Lower order polynomials may be used over
Voltage-To-Temperature a narrow temperature range to obtain higher system
Conversion speed.
We have used hardware and software compensation Table 4 is an example of the polynomials used to
to synthesize an ice-point reference. Now all we have convert voltage to temperature. Data may be utilized
to do is to read the digital voltmeter and convert the in packages for a data acquisition system. Rather than
voltage reading to a temperature. Unfortunately, directly calculating the exponentials, the computer is
the temperature-versus-voltage relationship of a programmed to use the nested polynomial form to save
thermocouple is not linear. Output voltages for the execution time. The polynomial fit rapidly degrades
more common thermocouples are plotted as a function outside the temperature range shown in Table 4 and
of temperature in Figure 16. If the slope of the curve should not be extrapolated outside those limits.
(the Seebeck coefficient) is plotted vs. temperature, 80
E
as in Figure 17, it becomes quite obvious that the Type
+
Metals
–
60
thermocouple is a non-linear device. K
Millivolts

E Chromel vs. Constantan

A horizontal line in Figure 17 would indicate a 40
J J
K
Iron vs. Constantan
Chromel vs. Alumel
constant α, in other words, a linear device. We notice R
R Platinum vs. Platinum
13% Rhodium
20
that the slope of the type K thermocouple approaches a S S Platinum vs. Platinum
10% Rhodium
constant over a temperature range from 0°C to 1000°C. T Copper vs. Constantan
0ϒ 500ϒ 1000ϒ 1500ϒ 2000ϒ
Consequently, the type K can be used with a multiplying
Temperature ϒC
voltmeter and an external ice point reference to obtain a
moderately accurate direct readout of temperature. That THERMOCOUPLE TEMPERATURE
is, the temperature display involves only a scale factor. vs.
This procedure works with voltmeters. VOLTAGE GRAPH
Figure 16
4
Refer to Bibliography 4.
Z-25
 100 mV .00 .01 .02 .03 .04 .05 .06 .07 .08 .09 .10 mV
TEMPERATURES IN DEGREES C (IPTS 1968)

0.00 0.00 0.17 0.34 0.51 0.68 0.85 1.02 1.19 1.36 1.53 1.70 0.00

Seebeck Coefficient V/ϒC

80
0.10 1.70 1.87 2.04 2.21 2.38 2.55 2.72 2.89 3.06 3.23 3.40 0.10
E
0.20 3.40 3.57 3.74 3.91 4.08 4.25 4.42 4.58 4.75 4.92 5.09 0.20
T J 0.30 5.09 5.26 5.43 5.60 5.77 5.94 6.11 6.27 6.44 6.61 6.78 0.30
60 0.40 6.78 6.95 7.12 7.29 7.46 7.62 7.79 7.96 8.13 8.30 8.47 0.40
Linear Region 0.50 8.47 8.63 8.80 8.97 9.14 9.31 9.47 9.64 9.81 9.98 10.15 0.50
(SeeText) 0.60 10.15 10.31 10.48 10.65 10.82 10.98 11.15 11.32 11.49 11.65 11.82 0.60
40
0.70 11.82 11.99 12.16 12.32 12.49 12.66 12.83 12.99 13.16 13.33 13.49 0.70
0.80 13.49 13.66 13.83 13.99 14.16 14.33 14.49 14.66 14.83 14.99 15.16 0.80
K 0.90 15.16 15.33 15.49 15.66 15.83 15.99 16.16 16.33 16.49 16.66 16.83 0.90
1.00 16.83 16.99 17.16 17.32 17.49 17.66 17.82 17.99 18.15 18.32 18.48 1.00
20 1.10 18.48 18.65 18.82 18.98 19.15 19.31 19.48 19.64 19.81 19.97 20.14 1.10
R 1.20
20.14 20.31 20.47 20.64 20.80 20.97 21.13 21.30 21.46 21.63 21.79 1.20
1.30
21.79 21.96 22.12 22.29 22.45 22.62 22.78 22.94 23.11 23.27 23.44 1.30
S 1.40
23.44 23.60 23.77 23.93 24.10 24.26 24.42 24.59 24.75 24.92 25.08 1.40
–500ϒ 0ϒ 500ϒ 1000ϒ 1500ϒ 2000ϒ
Temperature ϒC

SEEBECK COEFFICIENT vs. TEMPERATURE TYPE E THERMOCOUPLE

Figure 17 Table 3
TYPE E TYPE J TYPE K TYPE R TYPE S TYPE T
Z
Nickel-10% Iron(+) Nickel-10% Chromium(+) Platinum-13% Rhodium(+) Platinum-10% Rhodium(+) Copper(+)
Chromium(+) Versus Versus Versus Versus Versus
Versus Constantan(-) Nickel-5%(-) Platinum(-) Platinum(-) Constantan(-)
Constantan(-) (Aluminum Silicon)
-100˚C to 1000˚C 0˚C to 760˚C 0˚C to 1370˚C 0˚C to 1000˚C 0˚C to 1750˚C -160˚C to 400˚C
± 0.5˚C ± 0.1˚C ± 0.7˚C ± 0.5˚C ± 1˚C ±0.5˚C

9th order 5th order 8th order 8th order 9th order 7th order
a0
0.104967248 -0.048868252 0.226584602 0.263632917 0.927763167 0.100860910
a1
17189.45282 19873.14503 24152.10900 179075.491 169526.5150 25727.94369
a2 -282639. 0850 -218614.5353 67233.4248 -48840341.37 -31568363.94 -767345.8295
a3 12695339.5 11569199.78 2210340.682 1.90002E + 10 8990730663 78025595.81
a4 -448703084.6 -264917531.4 -860963914.9 -4.82704E + 12 -1.63565E + 12 -9247486589
a5 1.10866E + 10 2018441314 4.83506E + 10 7.62091E + 14 1.88027E + 14 6.97688E + 11
a6 -1. 76807E + 11 -1. 18452E + 12 -7.20026E + 16 -1.37241E + 16 -2.66192E + 13
a7 1.71842E + 12 1.38690E + 13 3.71496E + 18 6.17501E + 17 3.94078E + 14
a8 -9.19278E + 12 -6.33708E + 13 -8.03104E + 19 -1.56105E + 19
a9 2.06132E + 13 1.69535E + 20

TEMPERATURE CONVERSION EQUATION: T = a0 +a1 x + a2x + . . . +anx 2 n

NESTED POLYNOMIAL FORM: T = a0 + x(a1 + x(a2 + x (a3 + x(a4 + a5x)))) (5th order)
where x is in Volts, T is in °C
NBS POLYNOMIAL COEFFICIENTS
Table 4
The calculation of high-order polynomials is a time- All the foregoing procedures assume the
consuming task for a computer. As we mentioned thermocouple voltage can be measured accurately
before, we can save time by using a lower order and easily; however, a quick glance at Table 3 shows
polynomial for a smaller temperature range. In us that thermocouple output voltages are very small
the software for one data acquisition system, the indeed. Examine the requirements of the system
thermocouple characteristic curve is divided into eight voltmeter:
sectors, and each sector is approximated by a third-
THERMOCOUPLE SEEBECK DVM SENSITIVITY
order polynomial.* TYPE COEFFICIENT FOR 0.1˚C
(μV/˚C) @ 20˚C (μV)
Temp.

E 62 6.2
J 51 5.1
K 40 4.0
R 7 0.7
{

S 7 0.7
a T 40 4.0

Voltage REQUIRED DVM SENSITIVITY

Table 5
Even for the common type K thermocouple, the
voltmeter must be able to resolve 4 μV to detect
2 3 a 0. 1˚C change. The magnitude of this signal is
Ta = bx + cx + dx an open invitation for noise to creep into any system.
CURVE DIVIDED INTO SECTORS For this reason, instrument designers utilize several
Figure 18 fundamental noise rejection techniques, including
tree switching, normal mode filtering, integration and
* HEWLETT PACKARD 3054A.

Z-26
 PRACTICAL THERMOCOUPLE MEASUREMENT
Noise Rejection

C DVM
C HI
+ Tree
Signal (20 Channels)
– Switch1

+ +
= –
Signal
DVM
~
= Signal
DVM
–
Noise 20 C C HI C HI
~ C
Sour ce Noise ~ ~
Source
Next 20 Channels Stray capacitance to noise
source is reduced nearly
C 20:1 by leaving Tree
Tree Switch 2 open.
Switch2
TREE SWITCHING
Figure 19

Tree Switching - Tree switching is a method of Guarding - Guarding is a technique used to reduce
organizing the channels of a scanner into groups, each interference from any noise source that is common
with its own main switch. to both high and low measurement leads, i.e., from
common mode noise sources.
Without tree switching, every channel can contribute
noise directly through its stray capacitance. With tree Let’s assume a thermocouple wire has been pulled
switching, groups of parallel channel capacitances through the same conduit as a 220 Vac supply
are in series with a single tree switch capacitance. line. The capacitance between the power lines and
The result is greatly reduced crosstalk in a large data the thermocouple lines will create an AC signal of
acquisition system, due to the reduced interchannel approximately equal magnitude on both thermocouple
capacitance. wires. This common mode signal is not a problem in an
ideal circuit, but the voltmeter is not ideal. It has some
Analog Filter - A filter may be used directly at
capacitance between its low terminal and safety ground
the input of a voltmeter to reduce noise. It reduces
(chassis). Current flows through this capacitance and
interference dramatically, but causes the voltmeter to
through the thermocouple lead resistance, creating
respond more slowly to step inputs.
a normal mode noise signal. The guard, physically a
Integration - Integration is an A/D technique which floating metal box surrounding the entire voltmeter
essentially averages noise over a full line cycle; thus, circuit, is connected to a shield surrounding the
power line related noise and its harmonics are virtually thermocouple wire, and serves to shunt the interfering
eliminated. If the integration period is chosen to be less current.
than an integer line cycle, its noise rejection properties
are essentially negated.
Since thermocouple circuits that cover long distances
are especially susceptible to power line related noise,
VIN VOUT
it is advisable to use an integrating analog-to-digital
converter to measure the thermocouple voltage.
Integration is an especially attractive A/D technique in t t
light of recent innovations which allow reading rates of
48 samples per second with full cycle integration.
ANALOG FILTER
Figure 20

Z-27
 Distributed 220 VAC Line
Capacitance

HI

LO
Distributed
Resistance Without Guard
DVM

HI

LO Z
Without Guard Guard
DVM

GUARD SHUNTS INTERFERING WITH CURRENT

Figure 21
Each shielded thermocouple junction can directly affect the reading: HI
contact an interfering source with no adverse effects, RS
since provision is made on the scanner to switch the LO
guard terminal separately for each thermocouple
channel. This method of connecting the shield to guard Guard

serves to eliminate ground loops often created when Noise Current

the shields are connected to earth ground.
The dvm guard is especially useful in eliminating Figure 24
noise voltages created when the thermocouple junction
comes into direct contact with a common mode noise
source. Notice that we can also minimize the noise by
minimizing Rs. We do this by using larger thermocouple
wire that has a smaller series resistance.
240 VRMS
To reduce the possibility of magnetically induced
noise, the thermocouple should be twisted in a uniform
manner. Thermocouple extension wires are available
commercially in a twisted pair configuration.
Practical Precautions - We have discussed the
Figure 22 concepts of the reference junction, how to use a
In Figure 22 we want to measure the temperature at polynomial to extract absolute temperature data,
the center of a molten metal bath that is being heated and what to look for in a data acquisition system, to
by electric current. The potential at the center of the minimize the effects of noise. Now let’s look at the
bath is 120 V RMS. The equivalent circuit is: thermocouple wire itself. The polynomial curve fit relies
upon the thermocouple wire being perfect; that is, it
HI
RS must not become decalibrated during the act of making
LO
120VRMS a temperature measurement. We shall now discuss
some of the pitfalls of thermocouple thermometry.

Noise Current
Aside from the specified accuracies of the data
acquisition system and its zone box, most measurement
Figure 23
errors may be traced to one of these primary sources:
1. Poor junction connection
The stray capacitance from the dvm Lo terminal to
2. Decalibration of thermocouple wire
chassis causes a current to flow in the low lead, which
in turn causes a noise voltage to be dropped across the 3. Shunt impedance and galvanic action
series resistance of the thermocouple, Rs. This voltage 4. Thermal shunting
appears directly across the dvm Hi to Lo terminals and 5. Noise and leakage currents
causes a noisy measurement. If we use a guard lead
6. Thermocouple specifications
connected directly to the thermocouple, we drastically
reduce the current flowing in the Lo lead. The noise 7. Documentation
current now flows in the guard lead where it cannot
Z-28
 Robert Moffat in his Gradient Approach to
Poor Junction Connection Thermocouple Thermometry explains that the
There are a number of acceptable ways to connect thermocouple voltage is actually generated by the
two thermocouple wires: soldering, silver-soldering, section of wire that contains the temperature gradient,
welding, etc. When the thermocouple wires are and not necessarily by the junction.9 For example, if we
soldered together, we introduce a third metal into the have a thermal probe located in a molten metal bath,
thermocouple circuit, but as long as the temperatures there will be two regions that are virtually isothermal
on both sides of the thermocouple are the same, the and one that has a large gradient.
solder should not introduce any error. The solder
does limit the maximum temperature to which we can In Figure 26, the thermocouple junction will not
subject this junction. To reach a higher measurement produce any part of the output voltage. The shaded
temperature, the joint must be welded. But welding section will be the one producing virtually the entire
is not a process to be taken lightly.5 Overheating thermocouple output voltage. If, due to aging or
can degrade the wire, and the welding gas and the annealing, the output of this thermocouple were found
atmosphere in which the wire is welded can both 25˚C 100˚C
diffuse into the thermocouple metal, changing its
characteristics. The difficulty is compounded by the
very different nature of the two metals being joined. 200
300
Commercial thermocouples are welded on expensive 400
500
machinery using a capacitive-discharge technique to
insure uniformity.
500˚C
Fe Metal Bath

C Solder (Pb, Sn)

Junction: Fe - Pb, Sn - C = Fe - C GRADIENT PRODUCES VOLTAGE

Figure 26
SOLDERING A THERMOCOUPLE
Figure 25 to be drifting, then replacing the thermocouple junction
alone would not solve the problem. We would have to
A poor weld can, of course, result in an open replace the entire shaded section, since it is the source
connection, which can be detected in a measurement of the thermocouple voltage.
situation by performing an open thermocouple
check. This is a common test function available with Thermocouple wire obviously can’t be manufactured
dataloggers. While the open thermocouple is the perfectly; there will be some defects which will cause
easiest malfunction to detect, it is not necessarily the output voltage errors. These inhomogeneities can be
most common mode of failure. especially disruptive if they occur in a region of steep
temperature gradient. Since we don’t know where an
imperfection will occur within a wire, the best thing we
Decalibration can do is to avoid creating a steep gradient. Gradients
Decalibration is a far more serious fault condition can be reduced by using metallic sleeving or by careful
than the open thermocouple because it can result placement of the thermocouple wire.
in a temperature reading that appears to be correct.
Decalibration describes the process of unintentionally
altering the physical makeup of the thermocouple wire Shunt Impedance
so that it no longer conforms to the NBS polynomial High temperatures can also take their toll on
within specified limits. Decalibration can result from thermocouple wire insulators. Insulation resistance
diffusion of atmospheric particles into the metal caused decreases exponentially with increasing temperature,
by temperature extremes. It can be caused by high even to the point that it creates a virtual junction.7
temperature annealing or by cold-working the metal, an Assume we have a completely open thermocouple
effect that can occur when the wire is drawn through operating at a high temperature.
a conduit or strained by rough handling or vibration.
The leakage Resistance, RL, can be sufficiently low
Annealing can occur within the section of wire that
to complete the circuit path and give us an improper
undergoes a temperature gradient.
voltage reading. Now let’s assume the thermocouple is
not open, but we are using a very long section of small
5 Refer to Bibliography 5 diameter wire.
9 Refer to Bibliography 9
7 Refer to Bibliography 7

Z-29
 (Open)

To DVM RL

LEAKAGE RESISTANCE
Figure 27
RS RS

To DVM RL
T2

RS T1 RS

VIRTUAL JUNCTION
Figure 28
Z
If the thermocouple wire is small, its series resistance, wire. It generally is specified over a much lower
RS, will be quite high and under extreme conditions RL temperature range than premium grade thermocouple
< < RS. This means that the thermocouple junction will wire. In addition to offering a practical size advantage,
appear to be at RL and the output will be proportional to extension wire is less expensive than standard
T1 not T2. thermocouple wire. This is especially true in the case of
platinum-based thermocouples.
High temperatures have other detrimental effects on
thermocouple wire. The impurities and chemicals within Since the extension wire is specified over a narrower
the insulation can actually diffuse into the thermocouple temperature range and it is more likely to receive
metal causing the temperature-voltage dependence mechanical stress, the temperature gradient across
to deviate from published values. When using the extension wire should be kept to a minimum.
thermocouples at high temperatures, the insulation This, according to the gradient theory, assures that
should be chosen carefully. Atmospheric effects can be virtually none of the output signal will be affected by the
minimized by choosing the proper protective metallic or extension wire.
ceramic sheath
Noise - We have already discussed line-related
noise as it pertains to the data acquisition system.
Galvanic Action The techniques of integration, tree switching and
The dyes used in some thermocouple insulation guarding serve to cancel most line-related interference.
will form an electrolyte in the presence of water. This Broadband noise can be rejected with the analog filter.
creates a galvanic action, with a resultant output The one type of noise the data acquisition system
hundreds of times greater than the Seebeck effect. cannot reject is a dc offset caused by a dc leakage
Precautions should be taken to shield thermocouple current in the system. While it is less common to see
wires from all harsh atmospheres and liquids. dc leakage currents of sufficient magnitude to cause
appreciable error, the possibility of their presence
Thermal Shunting should be noted and prevented, especially if the
thermocouple wire is very small and the related series
No thermocouple can be made without mass. Since
impedance is high.
it takes energy to heat any mass, the thermocouple will
slightly alter the temperature it is meant to measure. If
the mass to be measured is small, the thermocouple Wire Calibration
must naturally be small. But a thermocouple made with Thermocouple wire is manufactured to a certain
small wire is far more susceptible to the problems of specification, signifying its conformance with the
contamination, annealing, strain, and shunt impedance. NBS tables. The specification can sometimes be
To minimize these effects, thermocouple extension wire enhanced by calibrating the wire (testing it at known
can be used. Extension wire is commercially available temperatures). Consecutive pieces of wire on a
wire primarily intended to cover long distances between continuous spool will generally track each other more
the measuring thermocouple and the voltmeter. closely than the specified tolerance, although their
Extension wire is made of metals having Seebeck output voltages may be slightly removed from the
coefficients very similar to a particular thermocouple center of the absolute specification.
type. It is generally larger in size so that its series If the wire is calibrated in an effort to improve its
resistance does not become a factor when traversing fundamental specifications, it becomes even more
long distances. It can also be pulled more readily imperative that all of the aforementioned conditions be
through a conduit than can very small thermocouple heeded in order to avoid decalibration.

Z-30
 Documentation - It may seem incongruous to
speak of documentation as being a source of voltage Zone Box Test - A zone box is an isothermal terminal
measurement error, but the fact is that thermocouple block of known temperature used in place of an ice
systems, by their very ease of use, invite a large number bath reference. If we temporarily short-circuit the
of data points. The sheer magnitude of the data can thermocouple directly at the zone box, the system
become quite unwieldy. When a large amount of data is should read a temperature very close to that of the zone
taken, there is an increased probability of error due to box, i.e., close to room temperature.
mislabeling of lines, using the wrong NBS curve, etc.
If the thermocouple lead resistance is much greater
Since channel numbers invariably change, data than the shunting resistance, the copper wire shunt
should be categorized by measure and, not just channel forces V = 0. In the normal unshorted case, we want to
number.6 Information about any given measure and, measure TJ, and the system reads:
such as transducer type, output voltage, typical value
V ≅ α (TJ - TREF)
and location, can be maintained in a data file. This
can be done under computer control or simply by But, for the functional test, we have shorted the
filling out a pre-printed form. No matter how the data is terminals so that V=0. The indicated temperature T’J is
maintained, the importance of a concise system should thus:
not be underestimated, especially at the outset of a 0 = α (T’J - TREF)
complex data gathering project. T’J = TREF
Thus, for a dvm reading of V = 0, the system will
Diagnostics indicate the zone box temperature. First we observe
the temperature TJ (forced to be different from TREF),
Most of the sources of error that we have mentioned
then we short the thermocouple with a copper wire
are aggravated by using the thermocouple near
TREF
its temperature limits. These conditions will be
Cu
encountered infrequently in most applications. Cu Fe
But what about the situation where we are using +
small thermocouples in a harsh atmosphere at high v Copper Wire Short TJ
– C
temperatures? How can we tell when the thermocouple Cu
is producing erroneous results? We need to develop a Voltmeter Cu
reliable set of diagnostic procedures. Zone Box
Isothermal Block
Through the use of diagnostic techniques, R.P. Reed
has developed an excellent system for detecting faulty SHORTING THE THERMOCOUPLE AT THE
TERMINALS
thermocouples and data channels.10 Three components
of this system are the event record, the zone box test,
and the thermocouple resistance history. and make sure that the system indicates the zone box
temperature instead of TJ.
Event Record - The first diagnostic is not a test at all,
but a recording of all pertinent events that could even This simple test verifies that the controller, scanner,
remotely affect the measurements. An example would voltmeter and zone box compensation are all operating
be: correctly. In fact, this simple procedure tests everything
but the thermocouple wire itself.
MARCH 18 EVENT RECORD
10:43 Power failure Thermocouple Resistance - A sudden change in
10:47 System power returned the resistance of a thermocouple circuit can act as a
11:05 Changed M821 to type K thermocouple
13:51 New data acquisition program
warning indicator. If we plot resistance vs. time for each
16:07 M821 appears to be bad reading set of thermocouple wires, we can immediately spot a
sudden resistance change, which could be an indication
Figure 29
of an open wire, a wire shorted due to insulation failure,
We look at our program listing and find that measurand changes due to vibration fatigue, or one of many failure
#M821 uses a type J thermocouple and that our new data mechanisms.
acquisition program interprets it as a type J. But from For example, assume we have the thermocouple
the event record, apparently thermocouple M821 was measurement shown in Figure 31.
changed to a type K, and the change was not entered into
the program. While most anomalies are not discovered We want to measure the temperature profile of an
this easily, the event record can provide valuable insight underground seam of coal that has been ignited. The
into the reason for an unexplained change in a system wire passes through a high temperature region, into a
measurement. This is especially true in a system cooler region. Suddenly, the temperature we measure
configured to measure hundreds of data points.
10 Refer to Bibliography 10
Z-31
 resistance of a battery. We can attack this problem
To Data with a technique known as offset compensated ohms
Acquisition T1 measurement.
System
As the name implies, the voltmeter first measures
the thermocouple offset voltage without the ohms
current source applied. Then the ohms current source
T = 1200˚C T = 300˚C is switched on and the voltage across the resistance is
BURNING COAL SEAM measured again. The voltmeter software compensates
Figure 31 for the offset voltage of the thermocouple and calculates
the actual thermocouple source resistance.
rises from 300°C to 1200°C. Has the burning section of
the coal seam migrated to a different location, or has
Special Thermocouples - Under extreme conditions, Z
we can even use diagnostic thermocouple circuit
the thermocouple insulation failed, thus causing a short
configurations. Tip-branched and leg-branched
thermocouples are four-wire thermocouple circuits that
R
allow redundant measurement of temperature, noise,

t1 Time
Leg-Branched Thermocouple
THERMOCOUPLE RESISTANCE vs. TIME
Figure 32

circuit between the two wires at the point of a hot spot?

If we have a continuous history of the thermocouple
wire resistance, we can deduce what has actually
happened. Tip-Branched Thermocouple

The resistance of a thermocouple will naturally Figure 34

change with time as the resistivity of the wire changes

TS T1 Summary
In summary, the integrity of a thermocouple system
Short can be improved by following these precautions:
CAUSE OF THE RESISTANCE CHANGE • Use the largest wire possible that will not
Figure 33 shunt heat away from the measurement area.
• If small wire is required, use it only in the region
due to varying temperature. But a sudden change in of the measurement and use extension wire for
resistance is an indication that something is wrong. the region with no temperature gradient.
In this case, the resistance has dropped abruptly, • Avoid mechanical stress and vibration which
indicating that the insulation has failed, effectively could strain the wires.
shortening the thermocouple loop. • When using long thermocouple wires, connect
the wire shield to the dvm guard terminal and use
The new junction will measure temperature Ts, not T1. twisted pair extension wire.
The resistance measurement has given us additional • Avoid steep temperature gradients.
information to help interpret the physical phenomenon
detected by a standard open thermocouple check. • Try to use the thermocouple wire well within its
temperature rating.
Measuring Resistance - We have casually • Use a guarded integrating A/D converter.
mentioned checking the resistance of the thermocouple • Use the proper sheathing material in hostile
wire as if it were a straightforward measurement. But environments to protect the thermocouple wire.
keep in mind that when the thermocouple is producing
• Use extension wire only at low temperatures and
a voltage, this voltage can cause a large resistance only in regions of small gradients.
measurement error. Measuring the resistance of
• Keep an event log and a continuous record of
a thermocouple is akin to measuring the internal
thermocouple resistance.

Z-32
 THE RTD
A more rugged construction technique is shown
History in Figure 37. The platinum wire is bifilar wound on a
The same year that Seebeck made his discovery glass or ceramic bobbin. The bifilar winding reduces
about thermoelectricity, Sir Humphrey Davy announced the effective enclosed area of the coil to minimize
that the resistivity of metals showed a marked magnetic pickup and its related noise. Once the wire
temperature dependence. Fifty years later, Sir William is wound onto the bobbin, the assembly is then sealed
Siemens proffered the use of platinum as the element with a coating of molten glass. The sealing process
in a resistance thermometer. His choice proved assures that the RTD will maintain its integrity under
most propitious, as platinum is used to this day as extreme vibration, but it also limits the expansion of
the primary element in all high-accuracy resistance the platinum metal at high temperatures. Unless the
thermometers. In fact, the Platinum Resistance coefficients of expansion of the platinum and the bobbin
Temperature Detector,15 or PRTD, is used today match perfectly, stress will be placed on the wire as
as an interpolation standard from the oxygen point the temperature changes, resulting in a strain-induced
(-182.96°C) to the antimony point (630.74°C). resistance change. This may result in a permanent
Platinum is especially suited to this purpose, as it can change in the resistance of the wire.
withstand high temperatures while maintaining excellent There are partially supported versions of the RTD
stability. As a noble metal, it shows limited susceptibility which offer a compromise between the bird-cage
to contamination. approach and the sealed helix. One such approach
The classical resistance temperature detector (RTD) uses a platinum helix threaded through a ceramic
construction using platinum was proposed by C.H. cylinder and affixed via glass-frit. These devices will
Meyers in 1932.12 He wound a helical coil of platinum on maintain excellent stability in moderately rugged
a crossed mica web and mounted the assembly inside vibrational applications.
a glass tube. This construction minimized strain on the
wire while maximizing resistance.

Typical RTD Probes

MYERS RTD CONSTRUCTION Thick Film Omega® Film Element

Figure 35
Thin Film Omega® TFD Element

Although this construction produces a very stable Glass sealed Biflar Winding

element, the thermal contact between the platinum and

TYPICAL RTD’s
the measured point is quite poor. This results in a slow Figures 36 and 37
thermal response time. The fragility of the structure
limits its use today primarily to that of a laboratory
standard.
Metal Film RTD’s
Another laboratory standard has taken the place
In the newest construction technique, a platinum or
of Meyers’ design. This is the bird-cage element
metal-glass slurry film is deposited or screened onto a
proposed by Evans and Burns.16 The platinum element
small flat ceramic substrate, etched with a
remains largely unsupported, which allows it to move
laser- trimming system, and sealed. The film RTD
freely when expanded or contracted by temperature
offers substantial reduction in assembly time and
variations.
has the further advantage of increased resistance for
Strain-induced resistance changes over time and a given size. Due to the manufacturing technology,
temperature are thus minimized, and the bird-cage the device size itself is small, which means it can
becomes the ultimate laboratory standard. Due to the respond quickly to step changes in temperature. Film
unsupported structure and subsequent susceptibility RTD’s are presently less stable than their hand-made
to vibration, this configuration is still a bit too fragile for counterparts, but they are becoming more popular
industrial environments. because of their decided advantages in size and
12 Refer to Bibliography 12 production cost. These advantages should provide the
15 Refer to Bibliography 15 impetus for future research needed to improve stability.
16 Refer to Bibliography 16

Z-33
 Metals - All metals produce a positive change in impedance can contribute a significant error to our
resistance for a positive change in temperature. This, temperature measurement.
of course, is the main function of an RTD. As we shall A ten ohm lead impedance implies 10/0.385 ≅ 26°C
soon see, system error is minimized when the nominal error in measurement. Even the temperature coefficient
value of the RTD resistance is large. This implies of the lead wire can contribute a measurable error. The
a metal wire with a high resistivity. The lower the classical method of avoiding this problem has been the
resistivity of the metal, the more material we will have to use of a bridge.
use.
Table 6 lists the resistivities of common RTD
Z
+ DVM
materials.
–
METAL RESISTIVITY OHM/CMF

_________ ___________________
(cmf = circular mil foot) RTD

Gold Au 13.00
Silver Ag 8.8
Copper Cu 9.26 WHEATSTONE BRIDGE


Platinum Pt
Tungsten w
59.00
30.00
Figure 39
Nickel Ni 36.00
Table 6 The bridge output voltage is an indirect indication
Because of their lower resistivities, gold and silver of the RTD resistance. The bridge requires four
are rarely used as RTD elements. Tungsten has a connection wires, an external source, and three
relatively high resistivity, but is reserved for very high resistors that have a zero temperature coefficient. To
temperature applications because it is extremely brittle avoid subjecting the three bridge-completion resistors
and difficult to work. to the same temperature as the RTD, the RTD is
Copper is used occasionally as an RTD element. separated from the bridge by a pair of extension wires:
Its low resistivity forces the element to be longer
than a platinum element, but its linearity and very
low cost make it an economical alternative. Its upper + DVM

temperature limit is only about 120˚C. – RTD

The most common RTD’s are made of either
platinum, nickel, or nickel alloys. The economical nickel
derivative wires are used over a limited temperature
Figure 40
range. They are quite non-linear and tend to drift
with time. For measurement integrity, platinum is the These extension wires recreate the problem that we
obvious choice. had initially: The impedance of the extension wires
affects the temperature reading. This effect can be
Resistance Measurement minimized by using a three-wire bridge configuration:
The common values of resistance for a platinum RTD
range from 10 ohms for the bird-cage model to several
thousand ohms for the film RTD. The single most A
DVM
common value is 100 ohms at 0°C. The DIN 43760 C
standard temperature coefficient of platinum wire is α
= 0.00385. For a 100 ohm wire, this corresponds to +
B
0.385 ohms/°C at 0°C. This value for α is actually the
average slope from 0°C to 100°C. The more chemically 3-WIRE BRIDGE
pure platinum wire used in platinum resistance
Figure 41
standards has an α of +0.00392 ohms/ohm/°C.
Both the slope and the absolute value are small If wires A and B are perfectly matched in length, their
numbers, especially when we consider the fact that impedance effects will cancel because each is in an
the measurement wires leading to the sensor may opposite leg of the bridge. The third wire, C, acts as a
be several ohms or even tens of ohms. A small lead sense lead and carries no current.
R=5�
Lead
The Wheatstone bridge shown in Figure 41 creates
100 � RTD a non-linear relationship between resistance change
and bridge output voltage change. This compounds the
Lead already non-linear temperature-resistance characteristic
R=5�
of the RTD by requiring an additional equation to convert
EFFECT OF LEAD RESISTANCE bridge output voltage to equivalent RTD impedance.
Figure 38
Z-34
 4-Wire Ohms - The technique of using a current in series with both Rg and R3:
source along with a remotely sensed digital voltmeter Again we solve for Rg:
alleviates many problems associated with the

( ) ( )
Vs - 2Vo 4Vo
Rg = R3 ———— - RL ————
+ VS + 2Vo Vs + 2Vo
i =0
i 100 W RTD
DVM The error term will be small if Vo is small, i.e., the
i =0
–
1�
4-WIRE OHMS MEASUREMENT 200�
- +
Figure 42 6V
+
- 3V
VO
2.0066V
1�
bridge. The output voltage read by the dvm is directly 100�

proportional to RTD resistance, so only one conversion

Figure 45
equation is necessary. The three bridge-completion
resistors are replaced by one reference resistor. The
digital voltmeter measures only the voltage dropped bridge is close to balance. This circuit works well with
across the RTD and is insensitive to the length of the lead devices like strain gauges, which change resistance
wires. value by only a few percent, but an RTD changes
resistance dramatically with temperature. Assume the
The one disadvantage of using 4-wire ohms is that we
RTD resistance is 200 ohms and the bridge is designed
need one more extension wire than the 3-wire bridge.
for 100 ohms:
RTD = Rg
R1
Since we don’t know the value of RL, we must use
VS + - + equation (a), so we get:
- VO

( )
R2
R3 R = 100 —————
6 - 1.9868 = 199.01 ohms
g 6 + 1.9868
Figure 43
The correct answer is of course 200 ohms. That’s a
This is a small price to pay if we are at all concerned
with the accuracy of the temperature measurement. Resistance to Temperature
Conversion
3-Wire Bridge Measurement Errors The RTD is a more linear device than the
If we know VS and VO, we can find Rg and then solve for thermocouple, but it still requires curve-fitting. The
temperature. The unbalance voltage Vo of a bridge built Callendar-Van Dusen equation has been used for years
with R1 = R2 is: to approximate the RTD curve:11, 13 3

R3 1
(
VO= VS ——— – VS —
R3 + Rg ) (2)

RT=R0+R0 α

[ (
T ——
T-δ —— -1
100 100

100
) ( ) (
T -β —— -1
T —— T
100
) ( ) ]
If Rg = R3, VO= 0 and the bridge is balanced. This can
be done manually, but if we don’t want to do a manual Where:
bridge balance, we can just solve for Rg in terms of VO: RT = Resistance at Temperature T
Ro = Resistance at T = 0˚C

(V————
+ 2V )
VS - 2VO
Rg = R3 α = Temperature coefficient at T = 0˚C
S O
(typically +0.00392Ω/Ω/˚C)
δ = 1.49 (typical value for .00392 platinum)
RL
Rg β = 0 T > 0
+ - +
VS
- VO
0. 11 (typical) T < 0

R3
RL The exact values for coefficients α , β , and δ are
determined by testing the RTD at four temperatures and
Figure 44 solving the resultant equations. This familiar equation
DVM was replaced in 1968 by a 20th order polynomial in
This expression assumes the lead resistance is zero. order to provide a more accurate curve fit.
If Rg is located some distance from the bridge in a The plot of this equation shows the RTD to be a more
3-wire configuration, the lead resistance RL will appear linear device than the thermocouple:
11, 13 Refer to Bibliography 11 and 13.

Z-35