Comparison of new MEMS accelerometers with commercial piezoelectric (PZT) state monitoring accelerometers

Today we will review some of the data to illustrate the state of the art and performance levels of MEMS technology and compare it to commercial piezoelectric (PZT) state monitoring accelerometers.

The investment in MEMS process technology coupled with design innovations has greatly improved MEMS performance, making MEMS a viable option for a wider range of condition monitoring applications. With specialized MEMS structures and process technologies, accelerometers with resonant frequencies up to 50 kHz and noise densities as low as 25 g/Hz are now available. The low noise advantages of these new accelerometers can be exploited through well-designed signal conditioning electronics.

Figure 1. Noise spectral density map of a new high-frequency accelerometer

To evaluate whether the latest MEMS accelerometers are suitable for condition monitoring applications, we performed a comparison with a commercial PZT state monitoring accelerometer. To ensure that the two sensors have similar mass and are subject to the same excitation signal, we attach the MEMS sensor to the outer casing of the PZT sensor. Like the PZT sensor, the single-supply analog output of the MEMS accelerometer is directly input to the analog input channel of the same data logger. A data acquisition instrument (DAQ) was used as the acquisition system for these experiments.

An actual scene is reconstructed on the vibration tester, such as the scene described in vibration based condition monitoring, to compare the devices with known excitation signals. This example shows the vibration level of a turbine operating at 5100 rpm (85 Hz) and a misaligned 3000 rpm (50 Hz) synchronous generator. This scenario illustrates the frequency and amplitude of the vibration system programmed when using the random vibration test mode. Table 1 lists the amplitude measurements of the two devices at the target frequency.

Table 1. Motor misalignment simulation set point

Figure 2 shows the spectrum measurements of a MEMS accelerometer with a 21 kHz resonant frequency and a PZT sensor with a 25 kHz resonant frequency. The rms output of the MEMS accelerometer in the 1 Hz to 1 kHz band is approximately 30 mg or 1.7% higher than the PZT accelerometer.

Figure 2. Noise density spectra of the PZT accelerometer (top) and MEMS accelerometer (bottom); at up to 10 kHz, the results are almost identical; the main difference is the low frequency response of the MEMS accelerometer.

Unlike PZT devices, MEMS devices have low-frequency response performance (measuring 1/f at 0.1 Hz); for ultra-low-frequency machines such as wind turbines, this is a concern (it also supports faster recovery from saturation). The frequency response of the vibration excitation system rolls off at very low frequencies, so the response of the two devices is tested by a "tap" test device and the response is captured. The recorded time domain measurements are then converted to the frequency domain. The result is shown in Figure 3. Note that MEMS accelerometers are capable of recording responses as low as DC.

Figure 3. Comparison of the response of two accelerometers when tapping

Compared to PZT sensors, MEMS sensors that directly drive DAQ with analog outputs achieve good results. This suggests that MEMS accelerometers are a suitable candidate for new state monitoring products that are reconfigured for output channels, especially for supporting new concepts based on semiconductor devices that operate from a single +5 V supply, such as wireless smart sensors.

On the surface, the first generation of accelerometers with high frequency response (22 kHz) and ±70 g, ±250 g, ±500 g wide full scale range (FSR) seem to be attractive for such applications. Unfortunately, its noise level is as high as 4 mg/Hz, which is unacceptable for most condition monitoring applications. The second-generation device was used in the comparison test, which was two orders of magnitude lower than the first-generation device, and the power consumption dropped to 40% of the first generation. Table 2 summarizes the performance comparisons of two generations of MEMS accelerometers and highlights performance improvements.

Table 2. Comparison of key specifications for first- and second-generation MEMS accelerometers for condition monitoring

The combination of electrical signal conditioning experience and the development of high-resolution MEMS accelerometers has enabled the performance of MEMS accelerometers to meet the requirements of state monitoring applications. High-frequency accelerometers with low physical noise levels, coupled with high-performance, low-noise, high-stability signal processing design techniques, overcome the fundamental limitations that previously prevented MEMS from providing performance comparable to PZT state-monitoring sensors.

Want to learn about the power, noise, bandwidth and temperature specifications of the ADI MEMS accelerometer family? Want to know if ADI can accurately detect and measure acceleration, tilt, shock and vibration MEMS accelerometer products in performance-oriented applications?