Based on the FXLS8974CF, a 3-axis low-g accelerometer from NXP, this Click board offers a combination of performance and versatility suitable for developing and training machine learning (ML) models for vibration analysis.
According to Nebojsa Matic, CEO of MIKROE, “A new member of our company’s 1750-strong mikroBUS -enabled Click board family of compact add-on boards, ML Vibro Sens Click can be used to collect data for training ML models to recognise different types of vibrations, and to monitor the health of machines and industrial equipment based on vibration patterns. It can also be used to track motion and activity in wearable devices, and to detect vibrations caused by earthquakes or other seismic events.”
The FXLS8974CF offers ultra-low-power operation alongside high-performance modes, ensuring efficient use in diverse scenarios. Its integrated digital features simplify data collection and reduce system power consumption, while its performance over extended temperature ranges enhances reliability in demanding applications, including industrial diagnostics, wearable technology, and environmental monitoring.
This Click board incorporates two DC motors to simulate vibration stimuli for machine learning. The BALANCED motor generates steady ‘nominal’ vibrations, serving as a baseline signal for training ML models in a ‘healthy’ state. The UNBALANCED motor is designed to provide customizable vibration signals, ranging from low-intensity to specific frequency-based vibrations.
The FXLS8974CF accelerometer captures detailed data from the balanced and unbalanced motors, enabling the differentiation between healthy baseline states and anomalous conditions.
It communicates with the host MCU via a standard 2-wire I2C interface.
