Piezoelectric accelerometers are the most important sensors for vibration testing. Although piezoelectric accelerometers have a wide measurement range, the large variety of these accelerometers on the market makes correct selection somewhat difficult. As a general principle for selecting vibration sensors, correct selection should be based on the analysis and estimation of the following three aspects of the measurement signal: a. The...
Piezoelectric accelerometers are the most important sensors for vibration testing. Although piezoelectric accelerometers have a wide measurement range, the large variety of these accelerometers on the market makes correct selection somewhat difficult. As a general principle for selecting vibration sensors, correct selection should be based on the analysis and estimation of the following three aspects of the measurement signal.
a . The magnitude of the measured vibration
b . Frequency range of the measured vibration signal
c . Vibration test site environment
The following section will further discuss the specific selection of sensors based on the three aspects mentioned above and with reference to relevant technical specifications.
Sensor sensitivity and measurement range
Sensitivity is one of the most fundamental indicators of a sensor. The magnitude of sensitivity directly affects the sensor's measurement of vibration signals. It's easy to understand that the sensor's sensitivity should be determined based on the magnitude of the measured vibration (acceleration value). However, since piezoelectric accelerometers measure the acceleration value of vibration, and under the same displacement amplitude, the acceleration value is proportional to the square of the signal frequency, the magnitude of acceleration signals varies greatly across different frequency bands. The acceleration value of low-frequency vibrations in large structures may be quite small; for example , a 1mm displacement signal at a frequency of 1Hz has an acceleration value of only 0.04m /s² ( 0.004g ). However, for high-frequency vibrations, a 0.1mm displacement signal at a frequency of 10kHz can reach an acceleration value of 4 x 10⁵ m/s² (40000g). Therefore, although piezoelectric accelerometers have a large measurement range, for vibration signals used to measure both high and low frequencies, a sufficient estimate of the signal strength should be made when selecting the sensitivity of the accelerometer. The sensitivity of the most commonly used vibration measurement piezoelectric accelerometer is 50~100mV/g for the voltage output type (IEPE type) and 10~50pC/g for the charge output type.
The measurement range of an accelerometer refers to the maximum value that the sensor can measure within a certain range of nonlinear error. The nonlinear error of most general-purpose piezoelectric accelerometers is 1%. As a general principle, higher sensitivity results in a smaller measurement range, and vice versa.
The measurement range of an IEPE voltage-output piezoelectric accelerometer is determined by the maximum output signal voltage allowed within the linear error range, which is typically ±5V. The maximum range of the sensor can be calculated as the ratio of the maximum output voltage to the sensitivity. It's important to note that the range of an IEPE piezoelectric sensor is affected not only by the magnitude of nonlinear error but also by the supply voltage and the sensor's bias voltage. When the difference between the supply voltage and the bias voltage is less than the range voltage specified in the sensor's technical specifications, the maximum output signal will be distorted. Therefore, the stability of the bias voltage of an IEPE accelerometer not only affects low-frequency measurements but can also cause signal distortion. This phenomenon requires special attention during high and low temperature measurements. When the sensor's internal circuitry is unstable under conditions other than room temperature, the bias voltage may drift slowly and continuously, causing the measurement signal to fluctuate.
The measurement range of charge output type sensors is limited by the mechanical stiffness of the sensor. Under the same conditions, the maximum signal output of the sensing element, constrained by the nonlinearity of the mechanical elastic range, is much larger than that of IEPE type sensors. This value often needs to be determined experimentally. Generally, when the sensor sensitivity is high, the mass of its sensing element is larger, and the sensor's range is relatively smaller. Simultaneously, because the mass is larger, its resonant frequency is lower, making it easier to excite the resonant signal of the sensing element. As a result, the resonant wave is superimposed on the measured signal, causing signal distortion. Therefore, when selecting the maximum measurement range, the frequency composition of the measured signal and the sensor's own natural resonant frequency must be considered to avoid the generation of resonant components. At the same time, there should be sufficient safety margin in the range to ensure that the signal is not distorted.
The calibration method for accelerometer sensitivity typically employs a comparative method. The ratio of the output of the sensor under test vibrating at a specific frequency (usually 159Hz or 80Hz) to the acceleration value read by a standard sensor is the sensor sensitivity. For impact sensors, the sensitivity is determined by measuring the output response of the sensor under test to a series of different impact acceleration values. This establishes the correspondence between the input impact acceleration value and the electrical output within the sensor's measurement range. Numerical calculations then yield the straight line with the smallest difference between each point; the slope of this straight line represents the sensor's impact sensitivity.
The nonlinear error of an impact sensor can be expressed in two ways: full-range deviation or segmented linear error. The former refers to the percentage error based on the sensor's full-range output; regardless of the measured value, the error is calculated as a percentage of the full range. Segmented linear error is calculated in the same way as full-range deviation, but the reference is not the full range; instead, the error value is calculated based on segments of the range. For example, a sensor with a range of 20,000g has a full-range deviation of 1%, resulting in a linear error of 200g across the full range. However, if the sensor's linear error is measured in segments of 5,000g, 10,000g, and 20,000g, and the error remains 1%, then the linear errors in the three different range segments are 50g, 100g, and 200g, respectively.
Sensor measurement frequency range
The frequency measurement range of a sensor refers to the range of frequencies that the sensor can measure within a specified frequency response amplitude error (±5% , ±10% , ±3dB). The upper and lower limits of the frequency range are called the high and low frequency cutoff frequencies, respectively. The cutoff frequency is directly related to the error; a larger allowable error range results in a wider frequency range. As a general principle, the high-frequency response of a sensor depends on its mechanical characteristics, while the low-frequency response is determined by the combined electrical parameters of the sensor and its subsequent circuitry. Sensors with high high-frequency cutoff frequencies are inherently small and lightweight, while high-sensitivity sensors used for low-frequency measurements are relatively large and heavy.
High-frequency measurement range of the sensor
The high-frequency measurement parameters of a sensor are typically determined by its high-frequency cutoff frequency, and a given cutoff frequency is related to its corresponding amplitude error. Therefore, when selecting a sensor, one cannot only consider the cutoff frequency; the corresponding amplitude error value must also be understood. A small frequency amplitude error in a sensor not only improves measurement accuracy but, more importantly, reflects the sensor's ability to control installation accuracy deviations during manufacturing. Furthermore, because the vibration signal frequency band of the measured object is wide, or the sensor's inherent resonant frequency is not high enough, the excited resonant signal wave may superimpose on the signal within the measurement frequency band, causing a significant measurement error. Therefore, when selecting the high-frequency measurement range of a sensor, in addition to the high-frequency cutoff frequency, the influence of the resonant frequency on the measurement signal should also be considered; of course, such signals outside the measurement frequency band can be eliminated by using filters in the measurement system.
Generally, the high-frequency cutoff frequency of a sensor is independent of the form of its output signal (i.e., charge-type or low-impedance voltage-type); however, it is closely related to the sensor's structural design, manufacturing process, installation method, and installation quality. The table below provides a qualitative classification of the high-frequency responses of different types of accelerometers .
