SBAA274A September   2018  – March 2023 ADS1118 , ADS1119 , ADS1120 , ADS112C04 , ADS112U04 , ADS1146 , ADS1147 , ADS1148 , ADS114S06 , ADS114S06B , ADS114S08 , ADS114S08B , ADS1219 , ADS1220 , ADS122C04 , ADS122U04 , ADS1246 , ADS1247 , ADS1248 , ADS124S06 , ADS124S08 , ADS125H02 , ADS1260 , ADS1261 , ADS1262 , ADS1263

 

  1.   A Basic Guide to Thermocouple Measurements
  2.   Trademarks
  3. 1Thermocouple Overview
    1. 1.1 Seebeck Voltage
    2. 1.2 Thermocouple Types
      1. 1.2.1 Common Thermocouple Metals
      2. 1.2.2 Thermocouple Measurement Sensitivity
        1. 1.2.2.1 Calculating Thermoelectric Voltage from Temperature
        2. 1.2.2.2 Calculating Temperature From Thermoelectric Voltage
      3. 1.2.3 Thermocouple Construction
      4. 1.2.4 Tolerance Standards
    3. 1.3 Thermocouple Measurement and Cold-Junction Compensation (CJC)
    4. 1.4 Design Notes
      1. 1.4.1 Identify the Range of Thermocouple Operation
      2. 1.4.2 Biasing the Thermocouple
      3. 1.4.3 Thermocouple Voltage Measurement
      4. 1.4.4 Cold-Junction Compensation
      5. 1.4.5 Conversion to Temperature
      6. 1.4.6 Burn-out Detection
  4. 2Thermocouple Measurement Circuits
    1. 2.1 Thermocouple Measurement With Pullup and Pulldown Bias Resistors
      1. 2.1.1 Schematic
      2. 2.1.2 Pros and Cons
      3. 2.1.3 Design Notes
      4. 2.1.4 Measurement Conversion
      5. 2.1.5 Generic Register Settings
    2. 2.2 Thermocouple Measurement With Biasing Resistors Attached to the Negative Lead
      1. 2.2.1 Schematic
      2. 2.2.2 Pros and Cons
      3. 2.2.3 Design Notes
      4. 2.2.4 Measurement Conversion
      5. 2.2.5 Generic Register Settings
    3. 2.3 Thermocouple Measurement With VBIAS for Sensor Biasing and Pullup Resistor
      1. 2.3.1 Schematic
      2. 2.3.2 Pros and Cons
      3. 2.3.3 Design Notes
      4. 2.3.4 Measurement Conversion
      5. 2.3.5 Generic Register Settings
    4. 2.4 Thermocouple Measurement With VBIAS For Sensor Biasing and BOCS
      1. 2.4.1 Schematic
      2. 2.4.2 Pros and Cons
      3. 2.4.3 Design Notes
      4. 2.4.4 Measurement Conversion
      5. 2.4.5 Generic Register Settings
    5. 2.5 Thermocouple Measurement With REFOUT Biasing and Pullup Resistor
      1. 2.5.1 Schematic
      2. 2.5.2 Pros and Cons
      3. 2.5.3 Design Notes
      4. 2.5.4 Measurement Conversion
      5. 2.5.5 Generic Register Settings
    6. 2.6 Thermocouple Measurement With REFOUT Biasing and BOCS
      1. 2.6.1 Schematic
      2. 2.6.2 Pros and Cons
      3. 2.6.3 Design Notes
      4. 2.6.4 Measurement Conversion
      5. 2.6.5 Generic Register Settings
    7. 2.7 Thermocouple Measurement With Bipolar Supplies And Ground Biasing
      1. 2.7.1 Schematic
      2. 2.7.2 Pros and Cons
      3. 2.7.3 Design Notes
      4. 2.7.4 Measurement Conversion
      5. 2.7.5 Generic Register Settings
    8. 2.8 Cold-Junction Compensation Circuits
      1. 2.8.1 RTD Cold-Junction Compensation
        1. 2.8.1.1 Schematic
          1. 2.8.1.1.1 Design Notes
          2. 2.8.1.1.2 Measurement Conversion
          3. 2.8.1.1.3 Generic Register Settings
      2. 2.8.2 Thermistor Cold-Junction Compensation
        1. 2.8.2.1 Schematic
        2. 2.8.2.2 Design Notes
        3. 2.8.2.3 Measurement Conversion
        4. 2.8.2.4 Generic Register Settings
      3. 2.8.3 Temperature Sensor Cold-Junction Compensation
        1. 2.8.3.1 Schematic
        2. 2.8.3.2 Design Notes
        3. 2.8.3.3 Measurement Conversion
        4. 2.8.3.4 Generic Register Settings
  5. 3Summary
  6. 4Revision History

2.1.3 Design Notes

The measurement circuit requires:

  • Pullup and pulldown resistors
  • AINP and AINN inputs
  • Internal reference or an external voltage reference
  • Isothermal cold-junction connection and measurement

Figure 2-1 shows the most common method of thermocouple biasing. Matched resistors are attached to either lead of the thermocouple to set the DC biasing for the input signal. A first resistor pulls the positive lead of the thermocouple to AVDD, and a second resistor pulls the negative lead of the thermocouple to AVSS. Because the measured thermocouple voltage is small, the bias current can be approximated as the supply voltage divided by the two biasing resistors. If the resistors are matched, the thermocouple voltage is biased to the mid-point of the analog supply. Setting the biasing near the mid-point of the supply ensures that the input voltage is within the input range of the PGA. Consult the ADC data sheet for specific PGA common-mode and absolute input ranges.

Resistor values are large to reduce the amount of current passing through the thermocouple and the thermocouple leads. Bias current reacting with long resistive leads create an additional voltage which is measured by the ADC as an error voltage. However, the bias current must be large enough so that the resistor current is significantly larger than the input current of the ADC. If the bias current is small or close to the level of the ADC input current, the DC bias of the thermocouple is offset from the mid-point of the supply. Biasing resistor values are typically from 500 kΩ to 10 MΩ.

An additional error in the measurement comes from the input current of the ADC. An extra voltage error is seen as the ADC input current reacts with the series input filter resistors and any series resistance associated with the input multiplexer. Because this current cannot be removed, it is important to select an ADC with a low input current and calculate the contribution of this error to the measurement.

The biasing resistors are also used for burn-out measurement. In the case of a burned out thermocouple, the positive input is pulled to AVDD while the negative input is pulled to AVSS. This creates a large voltage across the analog inputs, over-ranging the ADC. If the ADC is over-ranged, the ADC value would read 7FFFh (assuming a 16-bit bipolar ADC), showing a full-scale reading to indicate a burned out thermocouple. To ensure that the ADC reports a full-scale reading, verify that the biasing resistors are low enough so that they can pull against the input biasing current of ADC and yield a voltage larger than the input full-scale voltage. Because the biasing resistors are always in place, a separate burn-out measurement is not needed.

Unless the cold junction is at 0°C, there should be a separate cold-junction measurement. This measurement can be done through several different methods, using either an RTD, calibrated thermistor, or a variety of integrated circuit temperature sensors.