Installing a thermocouple inside a furnace to measure the temperature of products undergoing heat treatment seems like an obvious approach. However, this involves not only deciding on the type of sensor but also on its precise location. Even a high-precision thermistor installed in a corner of the furnace can only detect the temperature around its perimeter. If the temperature distribution within the furnace is not uniform, the temperature at that point may or may not represent the accurate temperature within the furnace. This is a typical mistake made by home heating contractors when installing thermostats. For ease of wiring, thermostats are often installed very close to the central air conditioning unit. However, if the thermostat is installed in a hallway or enclosed space, it cannot control the temperature in other parts of the room and can only maintain the required temperature in its immediate vicinity. As a result, other areas of the room end up being very hot or cold. "Shortening the distance between the ultrasonic sensor and its measured object by half can increase the intensity of the returned sound waves fourfold. Therefore, even a slight reduction in the distance can significantly increase the probability of the sensor successfully detecting the measured object." To avoid the aforementioned drawbacks, an instrument engineer must consider where the required data is (and whether it is) accurate, and select the sensor installation location accordingly. Most ideal installation locations are quite obvious, but there are also some very subtle factors that affect sensor performance. These include noise, data transmission, and some unique characteristics of ultrasonic sensors. Measurement Noise Problems can also arise if too much redundant data exists at the set sensor installation location. This is especially true when electrical grounding loops, mechanical vibrations, radio frequency interference (RFI), and other environmental factors cause measurement noise. RFI noise is prevalent in factories using cordless phones, pagers, wireless networks, and sparks caused by the use of electromechanical contacts. [align=center] Figure 1: An optional thickness gauge is installed downstream of the roller to measure the thickness of a steel plate. If the thickness gauge is installed far from the rollers, the controller will become impatient and exacerbate the problem because it takes a long time to correct for thickness measurement errors. If the thickness gauge is installed near the AC power supply circuit, a 60Hz noise signal can interfere with the instrument's measurement and cause fluctuations in the steel plate thickness. For example, some proximity sensors may generate a false positive signal when near a relay. A simple solution is to relocate susceptible proximity sensors away from switch boxes and relays. Alternatively, if relocation is not possible, some electromechanical equipment in the plant must be shielded or replaced with solid-state devices to eliminate the source of RFI. In flow measurement applications, turbulence caused by bends, connections, and valves in the pipe is a fundamental cause of measurement noise. Turbulence has a particularly significant impact on magnetic flow meters. A simple solution is to install all magnetic flow meters on straight sections of the pipe or use other flow measurement technologies, such as vortex flow meters, which can correct for turbulence. A serious problem that cannot be ignored is noise—especially RFI noise caused by the AC power supply circuit. 60Hz AC power, due to its sufficiently slow current oscillations, can significantly impact some production processes. As illustrated in the "thickness gauge" diagram: hot-rolled steel is flattened into a uniform thickness using two rollers rotating in opposite directions. A thickness gauge, mounted downstream of the rollers, measures the steel plate thickness and sends feedback to a controller to increase or decrease pressure to compensate for deviations beyond a specified thickness. A 60Hz noise signal overlapping the thickness gauge's output signal can cause the roller pressure to oscillate at 60Hz via the controller. If the steel plate is placed on rollers moving at 6 feet per second, this oscillation can create protrusions on the steel plate surface every tenth of an inch. The severity of this defect depends on the amplitude of the initial noise signal, the inertia of the rollers, and the tuning of the controller. Regardless, it is prudent to mount the thickness gauge away from the AC power supply whenever possible. Data transmission distance . The thickness gauge example also illustrates what can happen when a sensor is mounted far from the data source. Theoretically, the thickness gauge should be installed close to the rollers to reduce the time required between pressure changes on the rollers and the resulting changes in thickness measurement. However, this would prevent the controller from detecting any errors, resulting in inconsistent thicknesses across steel plates. If the thickness gauge is installed downstream of the rollers, the transmission time for thickness data to the gauge worsens the situation, causing the controller to become impatient. Seeing no results after the initial control action, the controller continues to operate until the sensor measurement begins to change. By then, the controller's cumulative effect has overcompensated for the original error, creating a reverse error. This results in constant fluctuations in roller pressure, causing significant damage to steel plates due to lateral wavy deformation. Example of an ultrasonic sensor : When an ultrasonic proximity sensor is installed, distance also becomes an issue. This involves detecting reflected sound pulses from the object being measured. The time it takes for the pulse signal to travel from the object to the sensor indicates the distance between the object and the sensor. However, the speed of the pulse signal must be known, and the distance between the object and the sensor cannot be too far; otherwise, the reflected pulse signal will be too weak to be detected. The problem becomes particularly serious when the object is too small, allowing only a portion of the sound waves to be reflected back. The smaller the object, the closer the ultrasonic wave must be to the sensor to detect it. Fortunately, the distance between the sensor and the object is four times the intensity of the returned sound wave, so even a slight reduction in distance can significantly increase the probability of successful detection. When the object is far from the sensor, the air temperature changes significantly, making the determination of the ultrasonic pulse speed problematic. The sensor can only measure the temperature of the surrounding air to determine the speed of sound propagation. However, air temperatures vary over long distances, and the sensor cannot account for these temperature differences to alter the ultrasonic wave's propagation speed. It can only assume that all air temperatures are the same. Therefore, if the actual temperature changes during pulse transmission, the sensor will incorrectly calculate the total distance the pulse signal travels, leading to inaccurate positioning of the object. The same problem exists with ultrasonic level sensors when they are installed far from the liquid surface inside the tank. See Figure “Temperature Gradient”. [align=center] Figure 2: The pulse signal emitted from an ultrasonic level sensor installed at the top of the tank travels faster as the gas temperature increases during its journey to the hot liquid at the bottom of the tank. This shortens the time it takes for the pulse signal to reflect back to the sensor, resulting in the sensor transmitting a higher level data than the true value. Installing the sensor closer to the liquid surface reduces the likelihood of a temperature gradient forming below the sensor. [/align] [align=center] Figure 3: When a proximity sensor at the bottom is installed too close to the object being measured, the pulse signal emitted by the sensor may reflect twice. Therefore, the sensor will simultaneously detect the true distance to the object and a phantom object at twice the true distance.[/align] Ultrasonic proximity sensors also have problems measuring the distance to objects that are too close. Consider the situation described in the “Secondary Echo” figure. In this case, the distance from the bottom sensor to the bottom object is exactly half the distance from the top sensor to the top object. The transmission time of the ultrasonic pulse signal between the sensor and the object being measured at a shorter distance should be exactly half that of the longer distance. However, if the distance between the bottom sensor and the bottom object being measured is very small, the sound wave reflected by the bottom object will also be reflected once by the bottom sensor, thus doubling the travel distance of the sound wave. The time required for the second echo to return to the sensor should be exactly the same as the time measured by the top sensor. As a result, the bottom sensor will measure two objects: one is the real object, and the other is a virtual image at twice the distance. [i] 5 Basic Questions to Answer When Installing Sensors Is there any useful information about the sensor installation location? When the sensor is installed in the reservoir, the size of the shadow on the target being measured can be observed visually, as obstructions between the sensor and the data source will greatly reduce the strength and accuracy of data acquisition. Is there any additional useful information? Unnecessary sources of information, such as radio frequency interference, can add a lot of noise to the desired data, rendering the sensor measurement useless. Can the sensor accurately read the data? In some extreme conditions, such as turbulence, excessive vibration, or high temperature, some sensors may fail or be completely damaged. Is the received information the same as the information transmitted from the location where the measurement is needed? Incorrect installation may cause the sensor to record virtual signals of auditory and visual feedback. Are there any advantages to acquiring data from a single point quickly enough? If the data acquisition time is too long, the acquired data will already be outdated for the current situation.