Resistance-Temperature Sensors, RTDs and Thermistors
I was at an event recently and one of the topics was sensor basics, brilliant, right in my wheelhouse. The presenter started a section on thermo-couples by discussing the range of sensor types available and listed both RTDs (Resistance Temperature Detectors) and Thermistors (Thermal-resistors) as competing sensor types to Thermo-couples and then began to explain the use of thermo-couples. It appears I’m not the only engineer who thinks they’re fascinating.
But what about the other two types? RTD’s and Thermistors. They pose an interesting question, if both RTDs and Thermistors have resistance-temperature effects at the core of their function, why do they belong in two different classes of sensor? What makes them tick, and why are they useful.
On the surface, this is a straightforward question to answer, RTDs typically use the resistance-temperature response of pure metals, and thermistors use the same effect but in semi-conductors – simple. But what insight into the sensor does information actually give us? Let’s dive a little deeper.
The material being used in the device is essentially irrelevant, the key question is how the material affects the behaviour of the device and how we use them differently because of that behaviour. If the the response metals and semiconductors were the same, then we wouldn’t have two different categories. So how do semiconductors behave? What even are they? There are many types of semiconductors, materials that do not conduct electricity aswell as pure conductors, but are not insulators either. One you may have heard of is silicon, which we use for computer chips, but other metalloids often behave as semi-conductors. Thermistors are typically made of metal oxides, another type of semi-conductor, but metalloids are sometimes used also. The resistance of semiconductors is typically high compared to metals, the core material of RTDs. Some semiconductors have responses where they get less resistive in response to increased temperature, as more electrons are raised into conductive energy bands by heat energy. Others get more resistive in response to temperature increases. The paths for electrons through the material are removed by various effects as atoms get more excited by the increased heat. In some materials, these and other effects can be approximately balanced, and these are the materials we use for resistors, where we want the resistance to be approximately the same across a range of temperatures this means that, depending on the semi-conductor, we can tailor the resistance-temperature relationship to one that is useful to us.
For thermistors we typically want the response to be very strong, either positive or negative, it’s even better if we can keep it linear across a useful temperature range. Because the effect is strong, we can use thermistors with much lower quality A to D converters, as the larger deviations in resistance are much easier to detect and the effect of lead resistance is small in comparison, this opens up the thermistor as an option for cheap consumer electronics being both low in cost and demanding low-cost electronics to process them. Medical Thermometers usually use a thermistor at their core, as the range of temperature of a human body we would expect, perhaps 37 +/- 5 degrees can be accurately measured by a cheap thermistor. Thermistors are also quite tough, making them ideal for use in a hand-held device.
Resistance Temperature Detectors, or RTDs, are usually made of metal, often platinum. Platinum is used because it doesn’t corrode at high temperature, other metals can be used with shielding to prevent oxidisation and the accompanying resistance changes this would produce. This is usually supported by a ceramic to avoid changes in the shape of the metal leading to strain-related changes in resistance. Depending on the metal, RTDs have a relatively small change in resistance in response to temperature, however, using a Wheatstone bridge and high-quality electronics we can detect this. Now, many many materials have a resistance response to temperature, the key is whether or not this response is near linear or can easily be modeled with a second or third-order correction and is repeatable. A sensor is no good if the response only works for one cycle. The metals used in RTDs have excellent repeatability and linearity across a very wide range, partially because of the effects we mentioned earlier. However, because of the materials used for them and the high-quality electronics necessary to support them, they are high cost to both purchase and use. This is further complicated, because the smaller resistance of the RTD is far closer to the lead resistance of the wires than that of a thermistor, and so we usually need to introduce techniques to mitigate lead resistance. This is why we usually use three wire systems for RTD which essentially moves one lead to the other side of the Wheatstone bridge cancelling out the lead resistance. Whilst this is complex, the excellent range and accuracy of an RTD often make the added effort and expense worth it for the results they yield.
Throwing in the thermo-couple, we have a family of temperature sensors. Thermistors are low cost with high accuracy often within 0.1 degree of true value, but only function over a limited temperature range (still usually with a range of a hundred degrees or so with some types). RTD’s have similarly excellent accuracy, an order of magnitude better than thermocouples, 0.1 of a degree or so. With good conditioning and electronics, both thermistors and RTDs can achieve at least an order of magnitude better performance. Thermocouples complete the set with the largest range of operation with the highest maximum temperature, whilst largely having the lowest cost.
| Type | Cost | Range | Accuracy |
| Thermocouple | Low | Excellent (Typically -200C – 1300C or wider) | Low (+/- 1 degree C) |
| Thermistor | Medium | Limited (+/- 50 deg C) | High (+/- 0.1 deg C or better) |
| RTD | High | Good (Typically -100C – 400C but wider is possible) | High (+/- 0.1 deg C or better) |
Hopefully, this answers the question I posed at the start, thermistors and RTDs, despite both utilising resistance response to temperature at their core, behave very differently and are both useful additions to the temperature measurement toolbox.
Do you have a job that requires the accurate measurement of temperature? At Flintmore we have the expertise and experience required to design and implement an instrumentation solution, even as a one-off, get in contact today.
Further Reading
Both thermistors and RTD’s are very well understood with a lot of high quality material available on them.
Wikipedia returns to form with a great article for both:
https://en.wikipedia.org/wiki/Resistance_thermometer
https://en.wikipedia.org/wiki/Thermistor
Omega as ever have great articles from an industry leader in sensor design and production, although if they have an article on RTD’s I havent found it yet:
https://www.omega.co.uk/prodinfo/thermistor.html
In the case of these widely used sensors, a lot of information and feel for them can actually be discerned just from looking around suppliers websites. I recommend a quick search for the available sensors from RS or OMEGA.
