How can MEMS-based solutions be created more easily?

Among other things, the vibration modes of the machine are analyzed. The vibrations caused by the gearbox are reflected in the frequency domain as multiples of the shaft speed. Anomalies such as wear, imbalance or loose parts at different frequency points. We usually use MEMS (micro-electromechanical systems) based accelerometers to measure the frequency. Compared with piezoelectric sensors, they have higher resolution, excellent drift characteristics and sensitivity, and higher signal-to-noise ratio (SNR), and can also detect very low frequency vibrations almost close to the DC range.

Long press to watch the MEMS sensor technology training course~

This article describes a highly linear, low noise, broadband vibration measurement solution based on the ADXL1002 MEMS accelerometer. This solution can be used for bearing analysis or engine monitoring and is suitable for all applications that require a dynamic range up to ±50 g and a frequency response from dc to 11 kHz.

An example circuit is shown in Figure 1. The analog output signal from the ADXL1002 is fed through a 2nd-order RC filter to the AD4000 successive approximation register (SAR) analog-to-digital converter (ADC), which converts the analog signal into a digital value for further signal processing.

The ADXL1002 is a high frequency, single-axis MEMS accelerometer that provides an output signal passband that extends well beyond the sensor resonant frequency range. With this device, frequencies outside the 3 dB bandwidth can also be monitored. To implement this monitoring, the output amplifier of the ADXL1002 needs to support a 70 kHz small signal bandwidth. Capacitive loads up to 100 pF can also be directly driven using the output amplifier of the ADXL1002. To achieve loads higher than 100 pF, a series resistor of at least 8 kΩ is required.

An external filter is required at the output of the ADXL1002 to remove aliased noise from the output amplifier and other internal noise components of the ADXL1002, such as noise from coupling the internal 200 kHz clock signal. Therefore, the filter bandwidth needs to be adapted accordingly. With the dimensions shown in Figure 1 (R1 = 16 kΩ, C1 = 300 pF, R2 = 32 kΩ, C2 = 300 pF), approximately 84 dB of attenuation is achieved at 200 kHz. In addition, the ADC sampling rate should be selected to be higher than the bandwidth of the amplifier (for example, 32 kHz).

For the ADC, the supply voltage of the ADXL1002 is chosen as its reference voltage source because the output amplifier is ratiometric to the supply voltage. In this case, the tolerance and voltage temperature coefficient of the supply voltage (typically connected to an external regulator) are between the accelerometer and the ADC, so the implicit errors associated with the supply and reference voltages can be cancelled.

The frequency response of the accelerometer is the most important characteristic of the system and is shown in Figure 2. At frequencies above approximately 2 kHz to 3 kHz, the gain increases. For the resonant frequency (11 kHz), a maximum gain value of approximately 12 dB (a factor of 4) is produced in the output voltage.

To indicate a range overshoot (overrange), the ADXL1002 is equipped with a corresponding output (OR pin). In the event of a significant overrange event, an integrated monitor will generate an alarm.

Special attention should be paid to properly placing the accelerometer. The accelerometer should be mounted close to a rigid mounting point on the board to avoid any vibrations in the board itself and measurement errors due to undamped board vibrations. This placement ensures that any board vibrations experienced by the accelerometer will be at a frequency above the resonant frequency of the mechanical sensor and therefore effectively invisible to the accelerometer. Multiple mounting points, close proximity to the sensor and thicker boards also help reduce the effects of system resonances on sensor performance.

Using the circuit shown in Figure 1, it is relatively easy to build a MEMS-based solution that can detect vibrations in the dc range up to 11 kHz, a range often required for condition monitoring of rotating machinery.