As I understand it, both thermistors and thermocouples are temperature sensors. So what are the advantages/disadvantages of using one over the other to measure the temperature? What are the specific applications for either of the sensors?


4 Answers 4



  • wide range of temperature sensing (Type T = -200-350°C; Type J = 95-760°C; Type K = 95-1260°C; other types go to even higher temperatures)
  • can be very accurate
  • sensing parameter = voltage generated by junctions at different temperatures
  • thermocouple voltage is relatively low (4.3mV for Type T thermocouple with one end at 0 C, other at 100 C, so that's 43uV/C tempco)
  • mostly linear


  • more narrow range of sensing (Quality Z thermistors spec'd at -55 to +150 C)
  • sensing parameter = resistance
  • usually very nonlinear
  • NTC thermistors have a roughly exponential decrease in resistance with increasing temperature
  • good for sensing small changes in temperature (unless you are careful in your signal conditioning, it's hard to use a thermistor accurately and with high resolution over more than a 50 C range).
  • sensing circuit doesn't need amplification & is very simple (voltage divider with reference resistor tied to reference voltage usually is sufficient) – see my blog for more information about signal conditioning.
  • accuracy is usually hard to get better than 1°C without calibration
  • 2
    \$\begingroup\$ And there is no need for calibration with a thermocouple? I can buy one from the shop and just trust the reading? \$\endgroup\$ Commented Oct 10, 2014 at 20:01
  • \$\begingroup\$ And how do prices compare for the two? \$\endgroup\$ Commented Oct 10, 2014 at 20:03
  • \$\begingroup\$ good q's... thermocouples do have tolerances depending on what you buy, which type, and which temperature range you are using them. Type T is the most accurate near room temperature. The error comes from variation in the metallurgy of the wires (both during manufacturing, and over time as the wires may react with air or undergo metallurgical changes (annealing/quenching/work hardening etc.) \$\endgroup\$
    – Jason S
    Commented Oct 10, 2014 at 20:47
  • \$\begingroup\$ prices: Thermistors are less expensive, unless you're talking about short lengths of plain old unshielded sheathed 30AWG wire, which look like they are $0.20-0.50/ft. Thermocouple probes are typically in the $15-$150 range. \$\endgroup\$
    – Jason S
    Commented Oct 10, 2014 at 20:53
  • 1
    \$\begingroup\$ This answer is imprecise and biased towards thermocouples, Need to add three of their major drawbacks: Reduced Accuracy, Calibration Drift with time (ref.), and they are mainly Non Linear (ref.) \$\endgroup\$
    – Brethlosze
    Commented Apr 7, 2020 at 4:42

A thermistor is a temperature-sensitive resistor, whilst a thermocouple generates a voltage proportional to the temperature. Thermocouples can work at much higher temperatures than thermistors.

They are commonly used for temperature control in heating systems.


Different Principles A thermistor is an electrical resistance, made of semiconducting material, that can be wired into a circuit. The semiconducting material is usually made of manganese oxides and nickel oxides. The thermistor works based on the principle that the electrical resistance of this material changes with temperature. A thermocouple, on the other hand, is made of two wires of different metals, such as copper and iron. The equal length wires are connected electrically together at one end and open at the other end. The principle is that if the open ends of the wires are at a fixed temperature and you change the temperature at the connected end, this generates a voltage between the two wires at the open end of the thermocouple.

Measuring Temperature With a thermistor, you could use an instrument that measures electrical resistance and connect it across the thermistor. You will measure a resistance change with a change in temperature. If you then refer to a table that lists temperature change versus resistance, you can find out the temperature from this table. In the case of a thermocouple, you will use an electrical circuit to measure the difference in voltage between the two wires at the open end and use this to measure the temperature difference between the two ends of the wires.

Thermistor or Thermocouple? In general, thermocouple readings are more precise than thermistor readings. However, they react more slowly to changes in temperature. Thermistors are also more expensive generally than thermocouples, due to the need for an external power source and the device's circuitry. The decision to use one or the other will depend on your specific needs.


The resistance of a thermistor varies with temperature. putting it in series with a resistor to form a voltage divider and you can measure temperature with a voltmeter. For even more accurate measurements it can be part of a Wheatstone bridge. They are often used as an over temperature sensor.

A thermocouple is a pair of dissimilar wires connected together. When heat is applied a small current is generated. The amount of current is roughly proportional to temperature. They work at higher temperatures than thermistors. A common use for a thermocouple is a detector on a gas pilot light. If the pilot light goes out, the thermocouple cools down and the gas supply is turned off.

  • \$\begingroup\$ Isn't the voltage proportional to the temperature? \$\endgroup\$
    – m.Alin
    Commented Jan 12, 2012 at 21:07
  • 1
    \$\begingroup\$ A bimetallic strip is more common as a detector for a pilot light. It can turn off the supply without any other components. \$\endgroup\$ Commented Jan 12, 2012 at 21:09
  • \$\begingroup\$ "For even more accurate measurements it can be part of a Wheatstone bridge." Not really. That's true for strain gages where matched elements take advantage of Wheatstone bridge configuration. Thermistors don't have that ability. \$\endgroup\$
    – Jason S
    Commented Jan 12, 2012 at 21:14
  • 1
    \$\begingroup\$ @m.Alin: "Isn't the voltage proportional to the temperature" -- thermocouple voltage is proportional to temperature difference. (approximately -- there is some nonlinearity) But you're right, it's voltage, not current. \$\endgroup\$
    – Jason S
    Commented Jan 12, 2012 at 21:15
  • 1
    \$\begingroup\$ @Leon: I've never heard of using bimetalic strips. A thermocouple can be used completely passively for this purpose by running the generated current through a coil which holds the valve open against a small spring. Simple and inexpensive, without needing any external power. \$\endgroup\$ Commented Jan 12, 2012 at 21:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.