We use sensors quite frequently, and it is common to notice sensor drift after they have been in use for some time. What causes sensor drift? How can we eliminate pressure sensor drift during the design process?
Due to the inherent asymmetry of component parameters in the sensor bridge circuit, and the differing temperature coefficients and thermal expansion coefficients of the materials used in the elastic elements and the resistance strain gauges, as well as the inconsistent lengths of the bridge leads, the overall temperature coefficients of adjacent arms in the sensor bridge end up being slightly different. When the temperature changes, the resistance in the adjacent arms changes differently, causing the bridge to produce an unbalanced output, which results in zero drift.
For smart linear displacement sensors, encoders, electronic scales, and imported displacement sensors, time drift refers to the aging effect on the system as time progresses, leading to structural changes in the system and resulting in drift. Temperature drift, on the other hand, is caused by temperature-induced zero point instability. It is evident that temperature effects are the primary cause of zero drift and are also the most challenging to control.
Causes of Sensor Drift
Sensor drift refers to the phenomenon where the output values of a sensor change over time. This drift can lead to inaccuracies in the sensor’s measurement results, affecting its reliability and stability in practical applications. The causes of sensor drift are varied, and the following will introduce them one by one.
- Temperature Changes
Temperature changes are one of the common causes of sensor drift. Variations in temperature can cause the materials inside the sensor components to expand or contract, which in turn affects the sensor’s mechanical structure and electrical conductivity, leading to drift in the output values. For example, an increase in temperature can cause the resistance of a resistive sensor to increase, resulting in higher output values.
- Power Supply Variations
The output value of a sensor is influenced by the supply voltage. When the supply voltage changes, the sensor’s output value may also change. This occurs because variations in the supply voltage can alter the operating conditions of the sensor’s internal circuitry, thereby affecting the amplitude and stability of the output signal.
- Long-Term Use
Long-term use is another significant cause of sensor drift. During operation, the sensor may be affected by mechanical stress, chemical exposure, or thermal expansion and contraction, leading to changes in its internal structure and resulting in drift of the output values. Additionally, the sensor may also be impacted by external environmental factors such as vibration and shock, which can further exacerbate the drift phenomenon.
- Sensor Aging
Over time, the performance of a sensor may gradually degrade, leading to drift. This occurs because the materials and components inside the sensor can age as the duration of use increases, causing changes in their physical properties. For example, the electrolyte inside the sensor may gradually deplete, resulting in a decrease in its sensitivity and stability, which in turn causes drift in the output values.
- Environmental Impact
Sensor drift can also be influenced by environmental factors. For instance, changes in air pressure, humidity, and light exposure can lead to drift in the sensor’s output values. This is because variations in environmental conditions can alter the interaction between the sensor and the measured object, thereby affecting the sensor’s measurement accuracy and stability. In the early development of pressure sensors, glass powder was used to seal the diffusion silicon chip and the metal base. However, this method had the drawback of creating significant stress around the pressure chip.
Even after annealing, this stress could not be completely eliminated. When the temperature changes, due to the differing thermal expansion coefficients of the metal, glass, and diffusion silicon chip, thermal stress is generated, causing the sensor’s zero point to drift. This explains why the zero-point thermal drift of the sensor is much greater than that of the chip alone.
Additionally, if silver paste and terminal posts are not properly welded, it can lead to unstable contact resistance. This instability becomes more pronounced with temperature changes, contributing to significant zero-point drift and temperature drift in sensors.
Semiconductor Theoretical Analysis of Zero-Point Thermal Drift Causes
Zero-point thermal drift is minimized, and the zero-point output voltage of the bridge is reduced, only when the doping concentration and resistance values of the resistors are consistent.
This is highly beneficial for improving the sensor’s performance. However, achieving uniform doping distribution during the diffusion process is challenging, making it essential to keep the piezoresistors as close to each other and as short as possible.
Circuit Analysis of Zero-Point Thermal Drift Causes
Under normal conditions, the resistance values of the four diffused resistors that form the Wheatstone bridge should be equal. The occurrence of zero-point thermal drift is due to the changes in the resistance values of the diffused resistors as the temperature varies. Within a certain temperature range, the resistance values increase as the temperature rises, indicating that the temperature coefficient RR of the diffused resistors is positive.
Solutions to Sensor Drift Issues
The solutions for compensating zero-point drift in pressure sensors can be broadly categorized into hardware compensation and software compensation.
1. Hardware Zero-Point Compensation Methods
① Using Appropriate Constant Resistors in Series or Parallel with the Bridge Arms:
- Bridge Arm Thermistor Compensation: This method uses thermistors in the bridge arms for temperature compensation.
- External Thermistor Compensation: Thermistors are connected externally in series or parallel to the bridge for compensation.
- Dual Bridge Compensation Technology: Utilizes two bridges for more precise compensation.
- Transistor Compensation Technology: Transistors are employed to enhance the compensation effects.
② Optimizing the Circuit:
- Circuit Design Optimization: Reasonable circuit design can reduce the impact of sensor drift. For example, using a temperature compensation circuit can correct the effect of temperature changes on the sensor’s output values, improving measurement accuracy and stability.
③ Other Circuit Design Techniques:
- Filtering and Amplification: Techniques like filtering and amplification can be employed in the circuit design to eliminate the effects of power supply variations and environmental interference on the sensor.
2. Software Zero-Point Compensation Methods
④ Software Compensation for Zero-Point Drift:
During signal acquisition, at times when no triggering signal has occurred or after the acquisition has ended, the input signal should ideally be zero, but the output signal may not be zero due to drift. This drift data is often random noise, which is meaningless for subsequent data processing. We define the signal values collected during these periods as zero-point drift. The software methods employed include:
a. Polynomial Fitting Method: In practical measurements, physical quantities like temperature and pressure measured by the pressure sensor may not have a strictly linear relationship with the output values. Therefore, the functional relationship is often expressed as a polynomial. Polynomials can be used to fit nonlinear signals, with the key being to solve for the coefficients of the polynomial.
b. RBF Neural Network Method: This method is based on Radial Basis Function (RBF) neural networks. The basic principle is that traditional zero-point temperature compensation algorithms are often complex, and their fitting accuracy can be limited. Artificial neural networks offer advantages such as requiring fewer samples, being simple to implement, having the ability to approximate arbitrary functions, and having a promising application outlook.
c. Other Software Methods: Additional software techniques include look-up table methods, interpolation methods, and more. These methods are used to further refine and correct sensor output, ensuring greater accuracy and stability in the presence of drift.