Thermocouple: Basic Principle and Design
Thermocouples have been used extensively for critical temperature measurements since the early 20th
century, especially in the very high temperature range. For many industrial and process critical
applications, T / C and RTD (Resistance Temperature Detector) have become the "gold standard" for
temperature measurement. Although the RTD has better accuracy and repeatability, but relatively
speaking, the thermocouple has the following advantages:
• Large range
• Fast response times
• Lower cost
• Better durability
• Self-powered (no excitation signal required)
• No self-heating effect
However, the use of thermocouples for high-precision temperature measurement may be more complex. You can optimize your measurement accuracy with solid circuit design and calibration, but understanding how the thermocouple works helps you design your circuit or use a thermometer.
When a voltage source is applied to a length of wire, the current flows from the plus side to the
minus side, and the wire heats up, causing a part of the energy loss. The Seebeck effect,
discovered by Thomas Seibeck in 1821, is a reverse phenomenon: when a temperature gradient is
applied to a piece of wire, a potential is created. This is the physical basis of the thermocouple
.
Where ∇V is the voltage gradient, ∇T is the temperature gradient, and S (T) is the Seebeck coefficient. The Seebeck coefficient is related to the material and is also a function of temperature. The voltage between two different temperature points on a piece of wire is equal to the integral of the Seebeck coefficient function over temperature.
For example, T1, T2, and T3 in Figure 1 represent the temperatures at different points on a piece of wire. T1 (blue) indicates the lowest temperature point, and T3 (red) indicates the highest temperature point. The voltage between T2 and T1 is:
Similarly, the voltage between T3 and T1 is:
According to the additive nature of the integral, V31 is also equal to:
We keep this in mind when discussing thermocouple voltage and temperature conversions.
Figure 1: According to the Seebeck coefficient, the temperature gradient produces a voltage across
the conductive metal.
The thermocouple consists of two different metals, the Seebeck coefficient S (T) of the wire
is generally different. Since a metal temperature difference can produce a voltage difference,
why must two metals be used? It is assumed that the wire in Fig. 2 is made of the material "A".
If a voltmeter probe is made of the material A, in theory, the voltmeter will not detect any
voltage.
Figure 2: Voltage measurement connection. When the material of the probe and the wire are the same,
there will be no potential difference.
The reason is that when the probe is connected to the end of the wire, the equivalent of extending
the wire. The two ends of the long wire are connected to the input of the voltmeter and have the
same temperature (TM). If the temperature of the two ends of the wire is the same, no voltage is
generated. To prove this mathematically, we calculate the accumulated voltage across the entire
metal ring from the positive terminal to the negative terminal of the voltmeter.
According to the additivity of the integral, the above equation becomes:
When the lower boundary and the upper boundary of the integral are the same, the result of integration is V = 0. If the probe material is B, as shown in Figure 3, then:
To simplify the above equation, we obtain:
Equation 9 shows that the measured voltage is equal to the integral of the difference between the Seebeck coefficient functions of the two materials. This is why thermocouples use two different metals.
Figure 3: Voltage measurement connection. The probe and wire are made of different materials,
illustrating the physical reality of the Seebeck coefficient.
Material A: Material A
Material B: Material B
Voltmeter: Voltmeter
Based on the circuit in Figure 3 and Equation 9, we still can not calculate the temperature (TH) of the hot side unless we know the temperature (TC) at the cold junction, given that SA (T), SB (T) and the measured voltage are known. In the early stages of the thermocouple, an ice-point oven with a temperature of 0 ° C was used as the reference temperature (the term "cold end"), because of its low cost, ease of implementation, and self-regulating temperature. The equivalent circuit is shown in Fig
Figure 4: The thermocouple requires a reference temperature of 0 ° C as shown in the figure to
calculate the unknown temperature TH.
Although we know the reference temperature of the circuit shown in Figure 4, it is not practical
to obtain TH by integration. So there is support for the common type of thermocouple standard
reference table, the table can be obtained by the corresponding voltage output of the corresponding
temperature. However, one must keep in mind that all standard thermocouple reference tables are
drawn at 0 ° C as the reference point.
Thermocouple system
A modern thermocouple consists of two different wires connected at one end (TH). The voltage was
measured at the open end of the wire pair. According to the equivalent circuit shown in Fig. 5,
VC is the same as Equation 9 in Fig.
Figure 5: Modern thermocouple configuration with cold junction compensation.
Cold junction compensation
Cold junction compensation The cold junction (TC) temperature can be set to 0 ° C for a freezing
point furnace, but in practice, we do not use the ice bucket as the reference temperature. Using
the CJC (cold junction compensation) method, the hot junction temperature can be calculated
without using a 0 ° C cold junction temperature. Even the cold junction temperature is not
necessarily constant. The method uses only a single temperature sensor to measure the TC point
temperature. If TC is known, TH can be obtained.
If we use the temperature sensor to measure the cold junction temperature, why not use this sensor
to measure the temperature of the hot side directly? As you can see, the cold junction temperature
range is much narrower than the hot junction temperature range, so the temperature sensor does not
need to support the extreme temperatures supported by the thermocouple.
Using CJC to calculate the hot end temperature
As mentioned above, all standard thermocouple reference tables are obtained at 0 ° C at the cold
junction. So how to use the reference table to get hot side temperature? Imagine extending the open
end of the thermocouple above and connecting the imaginary end to a junction with a temperature of
0 ° C (Figure 6). If we can calculate the V0 value, using the reference table is very easy to get
the corresponding hot-side temperature.
Figure 6: Connect the extended thermocouple to the 0 ° C junction to determine the unknown hot
junction temperature, TH.
Determine V0
Rearrange the above formula:
The first term in Equation 13 is identical to Equation 10 (obtained from Figure 5). The equivalent voltage output is VC, which is a known value because the temperature at the cold junction is measured by a voltmeter. The second term is equivalent to the output of the thermocouple when the hot junction temperature equals TC and the cold junction temperature equals 0 ° C. Since TC is also measured by a separate temperature sensor, we can use the standard reference table to find the corresponding Seebeck voltage (Vi) for the second term in Equation 13:
With this V0 value, the corresponding temperature at TH can be determined from the standard
reference table.
The use of cold junction compensation to calculate the temperature of the hot end of the process
is divided into the following steps:
• Measure the cold junction temperature (TC) with a temperature sensor.
• Measure the cold junction temperature.
• TC is converted to voltage (Vi) by means of a standard reference table.
• Calculate V0 = Vi + VC.
• Convert V0 to temperature TH using the standard reference table.
Refer to the NIST ITS-90 Thermocouple Database for standard thermocouple reference tables.
The NIST ITS-90 website also provides a set of formulas for each thermocouple type that can be
used for temperature-to-voltage conversion if the look-up table can not be implemented in
microcontrollers due to memory or other reasons.
Key points of system design
So far, the above discussion is limited to the theoretical knowledge of thermocouple. In order to
optimize the accuracy of the actual system, there are a few things to note. Each device in the
basic thermocouple signal chain (Figure 7) will affect the conversion accuracy and must be
carefully selected to minimize errors.
Figure 7: The basic elements of a thermocouple measurement system include an amplifier and ADC,
and a microcontroller that can calculate the unknown temperature later.
System board: System board
Amplifier: Amplifier
Temp sensor: Temperature sensor
From the left side of Figure 7, the thermocouple is connected to the connector of the system
circuit board. The thermocouple itself is also a sensor and may also be an error source. Longer
thermocouples can easily pick up the electromagnetic noise of the surrounding environment;
shielded wire can effectively reduce noise. The next element is the amplifier, which has a high
input impedance is very important because the amplifier input impedance and thermocouple
resistance to form a voltage divider. The higher the input impedance of the amplifier, the smaller
the error.
In addition, the amplifier increases thermocouple output, thermocouple output is usually millivolt
range. Although the amplifier's high closed-loop gain amplifies both signal and noise, adding a
low-pass filter to the ADC input eliminates most of the noise. Because the temperature does not
change very quickly, the ADC conversion rate for such applications is typically very low - it may
sample only a few times per second, so the low-pass filter is very efficient.
Finally, the onboard temperature sensor needs to be very close to the cold end connector (ideally
in contact with the end of the thermocouple wire, but in many cases the condition is not allowed)
to obtain the best cold junction temperature measurement. Any errors in cold junction measurement
will be reflected in the hot junction temperature calculation.
Examples of thermocouple circuits and test results
Whether designing your own thermocouple measurement circuit or using a reference design, you need
to verify its accuracy. The precision verification of the MAXREFDES67 # reference design
(Figure 8)
is described below.
In this case,
Figure 8: The MAXREFDES67 # is a reference design for thermocouples and RTDs that measure voltage
and current and measure temperature in the -40 ° C to 150 ° C temperature range.
To illustrate how to minimize measurement errors, we first take the thermocouple system as an
example, such as Maxim's MAXREFDES67 reference design. In order to verify the error of the
measurement system or of any measurement system, a known temperature and a reliable instrument are
required for comparison. In this example, we used three reference thermometers: the Omega HH41
thermometer (now replaced by HH42), the ETI reference thermometer, and the Fluke 724 temperature
calibrator. A Type K thermocouple connected to the MAXREFDES67 # is placed in a Fluke 7341
calibration oven and calibrated at 20 ° C. The blue dot data is referenced by the Omega HH41, and
the green dot data is referenced to the ETI device. The red dot data shows a maximum error of less
than 0.1 ° C, based on the Fluke 724 calibrator, but unlike the previous test, the Fluke 724 was
not used as a reference instrument. Simulate the ideal K-type thermocouple output, and connect the
input of the MAXREFDES67 # to the thermocouple extension cord. Figure 9 shows the test results.
Figure 9. Using the Omnitec EC3TC (type K thermocouple, calibrated at 20 ° C), evaluate the
MAXREFDES67 # error versus temperature and compare it to the other three reference thermometers.
The results show a very high accuracy.
to sum up
Thermocouples have many advantages in industrial temperature measurement applications, including
temperature range, response time, cost and durability. The thermocouple theory is a bit more
complex, but we must fully understand it so that we can make accurate measurements as well as
high-precision transitions from voltage to temperature. MAXREFDES67 # reference design using
MAX11254 and MAX6126 these two chips, particularly suitable for thermocouple temperature
measurement of this small signal-sensitive, high-precision measurement applications. The MAX11254
is a 6-channel, 24-bit, delta-sigma ADC that achieves low noise and accuracy while reducing power
dissipation by 10 times. The MAX6126 is a very low noise, ultra-high accuracy, low dropout tandem
voltage reference, Temperature coefficient of 3ppm / ° C (max), with an excellent ± 0.02% (max)
initial accuracy.
