The wearable sensor market is growing rapidly at a CAGR of 17.8%. However, sensor technology also faces many challenges, particularly the increasing demands for miniaturization and low power consumption. Several key parameters are crucial when measuring various types of sensors. This article will explore the sensor field and convey the following information:
The various types of sensors on the market and how sensors are changing the world we live in.
The key parameters of the sensor are unaffected by various physical measurement methods.
Benefits of characterizing and linearly transforming sensors
When characterizing and linearly transforming sensors, what factors should be considered when selecting appropriate testing instruments?
introduction
Sensors power the world. Whether at home, at work, in a car, or elsewhere, sensors are embedded in the electronic devices we use. It's hard to imagine life without mobile devices, and it's sensor technology that underpins these devices.
Today, people crave instant information, and sensors play a crucial role in this process. For example, if a user wants to check the weather on their phone, they might need to use a biometric sensor to unlock it. The weather app then uses data collected from millions of sensors (wind speed, direction, humidity, and temperature) located in different geographical locations to report the weather conditions.
The sensor market is growing at a CAGR of 11%. In emerging markets such as wearable IoT devices and medical devices, the CAGR for sensors is even higher at 17.8% (2020-2027). They are virtually ubiquitous, including smart headphones, smart clothing, head-mounted displays, fitness trackers, smartwatches, and portable medical devices that continuously monitor patients' vital signs.
Sensors can be categorized into many types, each containing several subtypes. Figure 1, for example, lists a series of sensor types in the middle section. Temperature sensors are one such type, and they can be further divided into many subtypes, such as resistive temperature sensors, thermocouples, and thermistor sensors. The circles surrounding the sensor types represent the key markets driving sensor usage and growth.
1: Benefits of Sensor Optimization
Sensors in their natural state may exhibit varying degrees of sensitivity to external parameters such as light, temperature, pressure, force, and humidity. The sensitivity of a sensor depends on its material, manufacturing method, and application. For example, some sensors possess capacitive or resistive characteristics, which change under varying degrees of influence from external parameters (such as force or temperature). To obtain a useful readable voltage or current output from the sensor, the designer needs to calibrate the sensor's current bias accordingly.
The benefits of sensor optimization are obvious:
Accuracy, sensing range, and sensitivity can be improved several times over. Some sensors output measurements on a logarithmic scale. Therefore, the potential performance gains from optimization can reach tenfold or even a hundredfold.
The sensor can also be adjusted for integration with the overall system. Operation becomes more efficient because less processing is required to convert the sensor output into useful information.
2: How to optimize sensors
Several methods can be used to optimize sensors, including characterization, linear transformation, dynamic error compensation, and signal conditioning.
2.1: Sensor Characterization
Because sensor sensitivities can vary, manufacturers characterize their sensors and publish technical specifications and data to help users better utilize them. Manufacturers can guarantee accuracy in measurement parameters such as dynamic range, bandwidth, response time, and accuracy. For specific sensors and measurement types, manufacturers provide recommended voltage or current biases to ensure proper sensor operation.
Sensitivity is one of the most important characteristics of a sensor, determining its performance in capturing or measuring subtle changes in physical parameters. Figure 2 shows two sensors with different sensitivities. Assuming both sensors have linear outputs, the gradient of the output represents the sensor's sensitivity. For example, the ratio of Δa to Δb is the gradient of the red line. The sensor producing the red line output has higher sensitivity than the sensor producing the blue line output. A typical sensor may not produce a linear response across its entire range. Therefore, sensitivity may vary across different ranges.
There are many ways to improve sensor sensitivity. For example, in a photodiode, increasing the gain will increase the sensor's small signal output, thereby reducing noise entering the sensor or noise inside the sensor, and allowing for a more sensitive readout circuit design.
Figure 3 shows the output response of a sensor excited by a step change in input parameters. The horizontal axis in the figure represents the measurement time of the sensor response. When the typical value of the sensor output corresponds to a step change in the input parameters, the output switches to the new level. However, after reaching the new level, overshoot and undershoot occur, and it takes some time for the output to stabilize at the new level.
2.2: Sensor Linear Conversion
Sensors are typically non-linear. Linear transformation is an important sensor optimization process, often used to fit curves or lines mixed with edge noise to a straight line. Sensors that have undergone linear transformation can be easily integrated into product design systems. Through linear transformation, data processing becomes simpler and more efficient.
Figure 4 shows the output diagrams of the two sensors. The blue line represents the nonlinear output of the sensor; the measurement accuracy deteriorates significantly at both ends of the sensor's range. The red line represents the ideal linear output of the sensor. As shown in the figure, the linearity error is the difference in F(x) between the blue and red lines.
2.3: Sensor Dynamic Error Compensation
In most cases, the data provided in the manufacturer's datasheet is sufficient to integrate sensors into a user's product or system. However, manufacturer datasheets may not be specific enough to meet critical implementation requirements. In such cases, product design engineers need to characterize the sensor and its product across multiple dimensions. For example, depending on the type of sensor used, the sensor's response can change with temperature fluctuations. If the product needs to operate precisely over a wide temperature range, then multi-dimensional characterization and analysis of the sensor is crucial. Figure 5 below shows a sensor model characterized from three dimensions.
When products or systems require precise displacement or control, design engineers may have to deal with issues such as sensor hysteresis. Certain types of sensors, such as temperature sensors, exhibit hysteresis during measurement. For example, we can observe hysteresis when measuring a known temperature point in a controlled oven from low to high, and then measuring it again from high to low. The difference between the two temperature measurements represents the temperature hysteresis error. Hysteresis appears as if the sensor is obstructing or lagging behind. This hysteresis depends on the inherent properties of the sensor material or the design of the sensing element.
System design engineers have ways to accurately model hysteresis and implement feedback and feedforward control to dynamically compensate for errors in real time.
2.4: Sensor Signal Conditioning
The output of raw sensor signals is typically weak and contains significant noise. These signals need to be conditioned to a form suitable for system measurements. Signal conditioning components or circuits can be integrated into the system to adjust the raw sensor signal. Such components include signal preamplifiers, noise filters, attenuators, or predistortion circuits.
3: Test instruments used to characterize sensors
When selecting appropriate testing instruments for sensor measurements, two key technical specifications need to be considered: accuracy and resolution. Accuracy measures the quality of the measured value, while resolution represents the level of detail that can be measured or the number of significant digits displayed by the testing instrument.
Modern test instruments typically use built-in analog-to-digital converters to digitize and process measurement data. The linearity of the test instrument is crucial when characterizing highly sensitive data sensors.
The stability of the testing instrument is also crucial for sensor characterization. Testing instruments such as digital multimeters use a reference voltage to improve measurement accuracy. If the reference value drifts, the measurement accuracy will also drift. Therefore, it is essential to choose testing instruments with self-calibration capabilities to reduce or eliminate such drift.
Sensors are precision components. When selecting testing instruments for sensors, it is essential to choose high-quality instruments that minimize interference with the sensor's measurements. Select testing instruments that do not introduce environmental noise or their own inherent noise into the measurement process.
Summarize
The benefits of optimizing sensors are significant, including improved accuracy, sensing range, and sensitivity. Characterizing sensors, or performing linear transformations on them, facilitates their integration into larger control systems, enhancing their efficiency.
Choosing appropriate testing instruments to characterize the sensor is crucial. It is essential to ensure that the testing instruments possess measurement accuracy, resolution, linearity, stability, and extremely low interference to meet the sensor optimization requirements.
