Rotating machines rarely fail without warning. Long before a bearing seizes or a coupling breaks, the machine starts to vibrate differently. Predictive maintenance vibration sensors pick up these early changes, so maintenance teams can plan repairs before a breakdown stops production.
This guide covers the basics: why vibration matters, which sensors to use, what to measure, and how to start.
From Reactive to Predictive Maintenance
There are three broad maintenance strategies:
- Reactive: run the machine until it fails, then repair it. Simple, but failures are unplanned and often cause secondary damage.
- Preventive: replace parts on a fixed schedule. More predictable, but parts are often changed while still healthy, and failures can still occur between intervals.
- Predictive (condition-based): monitor the actual condition of the machine and intervene only when the data shows a developing fault.
Vibration analysis is one of the most established predictive techniques for rotating equipment such as motors, pumps, fans, compressors and gearboxes.
What Vibration Reveals
Each common mechanical fault produces a typical vibration signature:
- Unbalance: high vibration at 1× running speed, mainly radial.
- Misalignment: high vibration at 1× and 2× running speed, often with a strong axial component.
- Mechanical looseness: many harmonics of running speed.
- Bearing defects: high-frequency vibration at specific bearing fault frequencies, which depend on bearing geometry and speed.
- Gear problems: vibration at gear mesh frequency and its sidebands.
- Cavitation in pumps: broadband, random high-frequency vibration.
By tracking these patterns over time, technicians can identify the fault and judge how quickly it is progressing.
Types of Vibration Sensors
Piezoelectric accelerometers
The most widely used sensor for machine monitoring. A piezoelectric crystal produces a charge proportional to acceleration. Industrial versions usually include built-in electronics (often called IEPE or ICP-type) and provide a wide frequency range, which makes them ideal for detecting bearing and gear faults.
MEMS accelerometers
Micro-electromechanical sensors are small, low-cost and low-power. They are very common in wireless and battery-powered vibration monitors. Their frequency range and noise level are generally more limited than piezoelectric sensors, but they are well suited to trend monitoring of many machines.
Velocity sensors
Velocity sensors output a signal proportional to vibration velocity. Velocity is the parameter used in common vibration severity guidelines such as ISO 10816 and its successor ISO 20816, which makes these sensors convenient for overall machine health checks.
Proximity probes
Non-contact eddy-current probes measure the relative movement of a shaft inside its bearing. They are used on large machines with sleeve (fluid-film) bearings, such as turbines and large compressors.
Vibration transmitters
A vibration transmitter combines a sensor with processing electronics and outputs an overall value, typically as 4-20 mA or RS485 Modbus. It connects directly to a PLC or SCADA system and is a simple way to add continuous protection.
Comparison Table
| Sensor type | Measures | Frequency range | Best for |
|---|---|---|---|
| Piezoelectric accelerometer | Acceleration | Wide | Bearings, gears, detailed analysis |
| MEMS accelerometer | Acceleration | Moderate | Wireless monitoring, trending |
| Velocity sensor | Velocity | Low to medium | Overall severity, unbalance, misalignment |
| Proximity probe | Shaft displacement | Low | Turbines, sleeve bearings |
| Vibration transmitter | Overall value | Depends on model | PLC integration, alarms |
Key Measurements
- Overall velocity (RMS, mm/s): a general health indicator, compared against severity zones.
- Acceleration (g): sensitive to high-frequency faults such as bearing wear.
- Envelope or demodulated acceleration: isolates repetitive impacts from bearing defects.
- Frequency spectrum (FFT): shows which frequencies contain energy, so you can identify the fault type.
- Temperature: many sensors combine vibration and temperature, adding another useful indicator.
Mounting Matters
How a sensor is mounted directly affects data quality:
- Stud mounting gives the best frequency response and is preferred for permanent installations.
- Adhesive mounting is a good option where drilling is not possible.
- Magnetic mounting is convenient for route-based portable measurements but reduces high-frequency response.
Mount the sensor as close as possible to the bearing, on a rigid part of the housing. Avoid thin covers and fan shrouds. Keep the same location and direction for every measurement so trends remain comparable.
How to Start a Vibration Monitoring Programme
- List critical machines. Focus on equipment whose failure stops production or is expensive to repair.
- Choose online or route-based monitoring. Permanent sensors suit critical assets; portable collectors suit large numbers of less critical machines.
- Record a baseline while the machine is in good condition.
- Set alarm levels using severity guidelines and your baseline.
- Trend the data and look for changes rather than single readings.
- Act on alarms with a clear workflow: verify, diagnose, plan, repair.
- Review and refine thresholds as you learn the behaviour of each machine.
Wireless sensors and IoT gateways make it easier to add monitoring to machines that are hard to reach or far from control cabinets.
Building a Reliable Programme
Predictive maintenance vibration sensors turn invisible mechanical changes into actionable data. Start with a few critical machines, choose the sensor type that matches the faults you want to catch, mount it correctly and trend the results.
Explore our vibration and motion sensor category to compare accelerometers, transmitters and wireless monitors, or contact our team for help designing your first monitoring setup.