Vibration measurement was at the heart of this ECK consulting engagement for a predictive motor maintenance system, combining fast acquisition, embedded processing, battery operation and LoRa communication.

The goal was to create a compact, battery-powered device that could be mounted directly on the motor, collect vibration data, do part of the processing locally and send relevant information to an analysis system.

The aim was to reduce reliance on corrective maintenance and help spot signs of wear before a failure caused an unplanned production stop.

Vibration measurement and unplanned downtime

Electric motors are subject to wear in bearings, bushings and other mechanical components.

When maintenance happens only after a failure, the cost can be high, not just for replacing the part but mainly because of the unexpected interruption of operations.

On a production line, an unplanned stop can mean lost productivity, the need for contingency stock, rearranged processes and an impact on customer delivery.

On the other hand, replacing parts preventively without a real indication of need can also drive up maintenance costs.

That is why condition monitoring systems aim to provide information that supports better-informed decisions.

Measuring motor vibration

Motor vibration can be monitored with different types of sensors.

The alternatives evaluated included piezoelectric solutions and high-speed accelerometers.

For this project, an accelerometer capable of fast acquisition was chosen, suited to capturing the vibration characteristics the application needed.

The sensor was selected with technical support from Analog Devices, looking for a solution that offered enough performance without resorting to very expensive specialized accelerometers.

The data volume challenge

One of the main challenges in this type of application is the amount of data generated.

High-speed vibration measurement can produce a large number of samples.

Continuously sending all of this data to a server would significantly increase the required bandwidth, power consumption and complexity of the communication system.

So the architecture adopted called for part of the processing to be done locally.

The embedded system would handle acquisition, a pre-processing stage and extraction of relevant information before transmission.

This way, instead of continuously sending the entire raw signal, the device could transmit parameters describing the vibration behavior over time, such as maximum and minimum values and other information defined by the application.

Embedded processing and LoRa communication

For processing and communication we used the AcSiP ST50H, a module based on the STM32WLE5.

The solution combined processing capability with LoRa communication in a compact platform.

The STM32WLE5 uses an ARM Cortex-M4 core, with enough capacity to perform the initial processing of the acquired data before transmission.

Choosing a module, rather than using the bare chip directly, also helped reduce the effort and risk associated with radio frequency development.

This covered aspects such as layout, transmission line, matching, component selection, assembly and radio validation.

Although the ST50H cost more than using the STM32WLE5 directly, it simplified the LoRa implementation and reduced the technical risk of the development.

Why LoRa?

LoRa was chosen mainly because of the need for low power consumption and adequate range in an industrial environment.

The idea was for the sensor to stay mounted on the motor for long periods, running on battery and requiring as little intervention as possible.

This mattered because a power source is not always available at the point where the device will be installed.

Long-range communication also made it possible to send the information to a central point without new cabling in the plant.

Simple installation on the motor

From the start, the architecture took into account the need to make installation easy.

One option evaluated was magnetic mounting, allowing the device to be placed on the motor quickly.

This would make it easy to install, remove and reposition the device during testing, maintenance or a change of measurement point.

Local processing before transmission

The device mounted on the motor was not just there to acquire the signal.

Part of the processing would be done locally to reduce the amount of information transmitted.

The device could, for example, log maximum and minimum vibration values and other parameters over a given period.

This data would then be sent to an external system responsible for analyzing the motor’s condition.

This architecture split the problem into two parts:

• acquisition and initial processing close to the motor;
• analysis and interpretation of the data in an external system.

The analysis system

The external system would be responsible for interpreting the collected information and helping identify changes in the motor’s behavior.

Based on historical data and the criteria defined for the application, it would be possible to identify situations that justified an inspection or maintenance action.

The specific motor analysis and diagnostic algorithms were the responsibility of the client, who had specialized knowledge in this area.

ECK’s role in the project

In this development, ECK worked mainly on the hardware engineering and electronic architecture of the system.

The work involved evaluating vibration sensor alternatives, choosing the processing platform, defining the LoRa communication, analyzing power consumption and supporting hardware development.

ECK also developed drivers and the basic support required by the embedded software.

Implementation of the specific vibration analysis and motor condition diagnostic algorithms was the client’s responsibility.

Combining electronics and application know-how

This project is a good example of how an industrial product often depends on combining different areas of expertise.

ECK contributed the electronic architecture, signal acquisition, embedded processing, communication and power consumption expertise.

The client contributed specific knowledge of motor behavior, analysis criteria and interpretation of vibration data.

This split made it possible to develop a technically sound architecture without concentrating all the knowledge in a single team.

Outcome of the consulting work

The project resulted in an architecture able to combine:

• high-speed vibration acquisition;
• local processing;
• LoRa communication;
• low power consumption;
• battery operation;
• simple installation on the equipment;
• integration with an external analysis system.

The main challenge was balancing measurement quality, processing capacity, data volume, communication range and battery life.

This kind of architecture decision is key to turning a monitoring idea into a solution that works off the bench and under real industrial operating conditions.

Learn more about ECK’s electronic engineering consulting.