Internet Explorer는 TI.com에서 지원되지 않는 브라우저입니다. 최상의 환경을 위해 다른 브라우저를 사용하십시오.

비디오 시리즈

Precision Labs 시리즈: 온도 센서

TI 정밀 랩은 아날로그 엔지니어를 위한 전자 업계 최대의 종합 온라인 강의실입니다. 주문형 코스 및 튜토리얼은 이론과 응용 실습을 접목하여 전문 엔지니어의 기술 전문성을 심화하고 입문 시기의 경력 개발을 가속화합니다. 이 모듈식 주문형 커리큘럼에는 온도 센서 설계 고려 사항을 온라인 강의, 퀴즈 및 실습과 함께 다루는 실습 교육 비디오가 포함되어 있습니다. 온도 센서 커리큘럼에는 온도 센서의 기본 사항, 용어, 설계를 위한 주요 사양, 애플리케이션 팁 등을 다루는 짧은 교육 비디오가 포함되어 있습니다! 이 시리즈에 새로운 콘텐츠가 계속 추가되므로 이 페이지에서 최신 온도 센서 강의를 확인하세요!

Video Player is loading.
Current Time 0:00
Duration 0:00
Loaded: 0%
Stream Type LIVE
Remaining Time 0:00
 
1x
  • Chapters
  • descriptions off, selected
  • subtitles off, selected

      발표자

      Hello. And welcome to the TI Precision Lab video discussing temperature sensor accuracy, error, and repeatability. In this video, we'll discuss the meaning of accuracy and error in the context of temperature sensors and go through examples to explain how these terms are used to characterize a system's performance. Finally, we'll discuss repeatability and how this could affect the system's overall accuracy.

      One of the most important characteristics of a sensor is its accuracy, because the accuracy affects the overall temperature measurement in the system. Different air sources cause the accuracy of a sensor to be affected. In other words, for a system to have very high accuracy, it must have very low error. These two terms may be interchangeably used when specking a temperature sensor.

      As an example, the three thermometers shown are used to measure the temperature of the same source. Two of these thermometers read the same value, while the third reads a different value. So the natural question here is, which one is the accurate reading?

      Just because two thermometers, as shown in the example, read the same temperature does not necessarily make its reading accurate. We will come back to answering this question once we understand what accuracy and repeatability are. To better understand accuracy and error, let's take a look at some of the ways an error may be defined for a temperature sensor.

      Error may be defined as a percentage of the measurement, a percentage of the full scale, or as an absolute value. Extreme care must be taken when interpreting the error of a lower value of relative terms may mean a higher absolute error. As an example, let's take two temperature sensors which have a temperature span of 0 degrees Celsius to 120 degrees Celsius.

      One temperature sensor is specified with an error of 1% of the measurement at 25 degrees Celsius, while another is specified with an error at 0.5% of the full scale. When converted to absolute error value, 1% of the measurement gives an error at 0.25 degrees Celsius, while 0.5% of the full scale makes the error 0.6 degrees Celsius. Clearly, a lower value in percentage may not be a lower value in absolute terms.

      In the data sheet for temperature sensors, the sensor is most commonly specked as an absolute value. This makes it simpler to compare temperature sensors when accuracy is critical to the application. In the electrical characteristics of a digital temperature sensor data sheet, the accuracy is often specified in a table as shown. This is also reflected in the temperature accuracy graph.

      As can be seen, the accuracy of the temperature sensor is specified over a specific temperature range. For example, in the full temperature range from negative 55 degrees Celsius to 150 degrees Celsius, the typical temperature error will be plus or minus 0.1 degrees Celsius. The typical values may be based on measurements done on specific parts. However, the maximum value falls between plus or minus 0.3 degrees Celsius across the entire operating range.

      Analog sensors like NTC thermistors specify the accuracy in terms of the resistance and beta value tolerance. These are often specified in a percentage and require conversion to absolute scale. Additionally, the non-linear behavior may lead to absolute error much larger in a specific temperature range. Although not all systems require such a high degree of accuracy or linearity, it is important to understand the specifications and how additional error may still affect the overall temperature measurement.

      Even though this sensor can have high accuracy within 0.1 degrees Celsius, typically, there are external effects which may introduce errors in the accuracy. It is important to be able to distinguish these errors when testing a temperature sensor in a system. There are three main sources of errors when sensing temperature. The first is the measurement system; secondly, the operating environment or measurement conditions; and lastly, the golden reference.

      The reference refers to the source of measurement which the temperature sensor is being compared against. This is typically a high accuracy RTD, which is used for evaluating other sensors' accuracy. And in some systems, it is used as the temperature sensor. We will consider the reference to be ideal for ease of understanding. But in practice, there could be some error associated with it.

      Let's take a look at the measurement system. Depending on the type of temperature sensor, whether it has a digital or analog output, the signal condition and surrounding circuitry will be different. The error of the overall measurement may be affected by some or all of these additional components.

      The temperature sensor, as we discussed, has some error of its own. The output may be far away from the processing circuit, and thus may require some additional signal conditioning, like an amplifier, to transmit the output across the circuit board. If the output of the sensor is analog, then data conversion is necessary from analog to digital signal to be read by the controller. This usually has its own error due to quantization noise. The error from the ADC may also arise from the error of the voltage reference.

      Finally, there could be error due to the interface or processor itself. Needless to say, within the measurement system itself, there may be many sources of error. This is why it's important to evaluate the entire signal chain to determine where the error sources are located in order to improve overall accuracy. With some kind of compensation and calibration, the effect of error sources may be reduced or eliminated to improve the overall accuracy of the system.

      To better understand the effect of measurement system and operating environment on the accuracy of a sensor, we will use a simple example. In this example, we will consider a heat source whose temperature is being measured through a heat sink. The environment in which the sensor operates is called the operating environment, which includes the ambient air and the heat sink. The heat source is shown in the center of the heat sink, and the temperature sensor is located nearby.

      A reference probe, which is considered to be ideal, is used to measure the temperature to compare to the sensor's reading. The reference probe reads 80 degrees Celsius, while the sensor reads 78 degrees Celsius. This results in an absolute error of negative 2 degrees Celsius.

      Now we move the temperature sensor closer to the heat source, which results in two possible outcomes. In the first case, the sensor still reads 78 degrees Celsius, while the reference probe reads 80 degrees Celsius. In this case, we can conclude that the error is due to the measurement system.

      In the second case, the sensor reads 79 degrees Celsius, while the reference probe reads 80 degrees Celsius. In this case, we can conclude that there is a temperature gradient across the heat sink resulting in a non-uniform temperature. This is an example of error due to the operating environment and not just the sensor.

      So far we have assessed accuracy under the steady state operating condition. And we have ignored any effect of the sensor and referenced thermal time constants. These sources of errors are referred to as measurement condition errors.

      As shown in the image, temperature does not change instantly. And thus, there is a time dependency factor that may affect accuracy. This is called the temperature step response and is generally defined as the time for the output to reach 63% of the steady state value. This typically is impacted by the sensor's package, layout, type of PCP, and the testing environment.

      When the time constants of the sensor and the reference are not equal, there will be a time-dependent transient error. When the time constant of the reference is much shorter than the sensor, a measurement may result in negative error. Similarly, if the time constant of the reference is much longer than the sensor, a measurement may result in positive error. Thus, in order to compare the accuracy of the measurement system to the reference, the thermal step response of both the reference and measurement system must be taken into account.

      So far, we have seen how accuracy is specified in a typical temperature sensor data sheet. Also, we have reviewed the various sources of error outside of the sensor itself, which includes the measurement system, operating environment, measurement conditions, and how to isolate each of them in terms of accuracy. Another important parameter to be considered when making highly accurate measurements is repeatability.

      Repeatability is defined as how constant a sensor is against itself. It can be used to describe the ability of a sensor to provide the same result under the same circumstances over and over again. For instance, the temperature sensor shown here has a repeatability of plus or minus 1 LSB. Since the temperature resolution of the sensor is 7.8125 millidegrees Celsius, this means that when measurements are made under the same exact test conditions, the results should be within 7.8125 millidegrees Celsius of the mean of those measurements.

      It is important to recognize the difference between repeatability and accuracy. Repeatability is closely related to precision. It is an important value in temp sensor specifications, because a sensor must be repeatable in order to be determined as accurate. Furthermore, the sources of error that affect repeatability are not the same as accuracy. Repeatability is affected due to random noise and aging, which are often outside of the user control.

      To better understand the concept of repeatability and accuracy, we can use a dartboard in which the bull's eye represents the accuracy limits of a sensor. Now let us make five temperature measurements with the sensor. In case one, the values read from the sensor are spread all over the dartboard. Such a sensor has a non-repeatable and inaccurate output.

      In case two, the values read from the sensor are close to each other but outside the center circle. Such a sensor has a repeatable but inaccurate output. In case three, values read from the sensor are within the center circle but with a large spread. Such a sensor has a non-repeatable but accurate output.

      Lastly, in case four, the values read from the sensor are within the circle and with a small spread. Such a sensor has a repeatable and accurate output. To ensure a temperature sensor is best suited for the application, is important to ensure that the sensor is accurate. And when tested under the same environmental and operating conditions, the output must remain as close as possible to the accurate value.

      In summary, this training video shows what accuracy and repeatability means for a temperature sensor, how external parameters affect the accuracy of a temperature sensor beyond its own specification, and why repeatability is important in making accurate measurements. Just because two thermometers read the same value does not make them accurate and one thermometer reads the correct value does not make it precise. Thank you for your time.

      View series

      Precision Labs 시리즈: 온도 센서