A degrading bearing tells you it is degrading long before it fails. The vibration signature it produces evolves through identifiable stages, and a handheld vibration meter — combined with the right interpretation — gives you weeks or months of advance warning. Here are the five patterns to recognise.
Pattern 1 — High-frequency stress waves (Stage 1)
The earliest sign of subsurface fatigue damage shows up in the high-frequency range, typically 20-60 kHz. Standard vibration meters miss it; dedicated stress-wave or acoustic emission instruments catch it. At this stage the bearing has weeks to months of remaining life. The corrective action: schedule replacement during the next planned maintenance window, no rush.
Pattern 2 — Bearing defect frequencies (Stage 2)
As fatigue progresses, vibration energy appears at discrete defect frequencies that depend on the bearing geometry: BPFO (ball pass frequency, outer race), BPFI (inner race), BSF (ball spin), FTF (cage). These frequencies are calculable from the bearing dimensions and the shaft speed. Their appearance is the unambiguous sign that a localised raceway defect has formed.
Diagnostic tip: divide the shaft speed by the bearing pitch diameter and the rolling element count using the formulas — modern vibration analysers do this automatically when you input the bearing model.
Pattern 3 — Sidebands around defect frequencies (Stage 3)
When the defect has grown enough to be visible to the rolling elements every pass, sidebands appear around the defect frequency, spaced at the cage rotation frequency. At this stage the bearing has weeks of remaining life and audible noise often becomes detectable with a stethoscope.
Pattern 4 — Broadband noise (Stage 4)
The defect frequencies become buried in a rising broadband vibration noise floor. The overall RMS vibration level may be 2-3x its baseline. The bearing is structurally damaged and failure is imminent — days, not weeks. Audible noise is now obvious.
Pattern 5 — Loss of trend (Stage 5)
Counterintuitively, vibration sometimes drops just before catastrophic failure: the bearing’s internal clearance has grown so much that the rolling elements skid rather than roll, dampening vibration. Temperature, on the other hand, is climbing fast. If your vibration trend is dropping while temperature is rising, stop the machine immediately.
The practical inspection protocol
- Establish a baseline vibration reading on each critical bearing when the machine is healthy.
- Measure monthly (or weekly on critical assets) at the same point, the same load condition.
- Trend the overall RMS and any defect-frequency peaks.
- Triple the inspection frequency once you see Stage 2.
- Plan replacement during Stage 3 — do not push into Stage 4.
Tools you actually need
You do not need a $10,000 analyser to do this. A $200 handheld vibration meter that reads overall RMS plus envelope is enough for trending. A $1,000 unit with FFT and bearing-defect calculator handles 95% of industrial cases. With IoT vibration nodes now under $50 each, continuous monitoring is also realistic for any plant with critical rotating equipment.
The FFT spectrum interpretation in depth
Beyond the five high-level patterns, the FFT spectrum from a vibration meter contains diagnostic information at multiple frequencies that together tell the bearing’s full story. The fundamental shaft speed (1×) appears as a peak — its amplitude reflects unbalance. 2× shaft speed reflects misalignment. The bearing defect frequencies (BPFO, BPFI, BSF, FTF) appear as peaks at calculated frequencies derived from the bearing geometry and shaft speed. Sidebands around the defect frequencies tell the story of how mature the defect is.
For practical interpretation: use the bearing manufacturer’s defect frequency calculator (built into most modern analysers) to compute the expected frequencies. Compare against the spectrum captured. Any peak at a defect frequency above twice the noise floor is significant. Sidebands appearing around a defect frequency indicate the defect has matured into a localised raceway feature visible to the rolling elements every pass.
Order tracking vs frequency tracking
Variable-speed equipment (motors driven by inverters, machine tool spindles, wind turbines) makes frequency-based analysis difficult: the shaft speed varies, so the defect frequencies vary with it, and a simple FFT smears the peaks across multiple bins. Order tracking solves this by re-sampling the vibration signal in shaft revolutions rather than time. The result: defect “orders” (multiples of shaft speed) appear as sharp peaks regardless of speed variation.
For maintenance teams analysing variable-speed equipment, order-tracking-capable analysers are worth the additional cost. The diagnostic accuracy improvement on inverter-driven motors and machine tool spindles is substantial.
The acoustic emission alternative for Stage 1 detection
For the earliest detection — Stage 1 subsurface fatigue damage before any defect frequency appears in the vibration spectrum — acoustic emission (AE) sensors deliver capability that standard vibration analysis cannot match. AE captures high-frequency stress waves (typically 100-500 kHz) generated by microscopic crack initiation. The signal appears weeks to months before defect frequencies appear in the vibration spectrum.
AE monitoring is more expensive than vibration monitoring and produces noisier data that requires more sophisticated interpretation. For critical assets where the earliest possible warning is worth the cost, AE complements vibration monitoring; for general industrial assets, vibration alone is sufficient.
Continuous monitoring vs walk-around inspection
The choice between continuous IoT-based vibration monitoring and periodic walk-around inspection depends on asset criticality, plant size, and maintenance team structure. Continuous monitoring catches degradation faster and trends the data automatically; walk-around inspection costs less in capital but more in skilled labour and produces less consistent data.
For the typical European industrial plant, the breakeven shifts toward continuous monitoring on the top 20-50 critical assets and walk-around inspection on the remainder. The 2026 sub-$50 IoT sensor pricing makes this hybrid approach economically realistic for a much wider range of plants than was possible in 2019.
What pattern recognition actually catches
AI-based pattern recognition trained on labelled bearing failure data catches subtler patterns than human analysts typically identify. Specific examples: gradual increase in noise floor around a defect frequency before the peak itself becomes significant, correlated changes in temperature and vibration that point at lubrication issues, and characteristic patterns of misalignment-driven damage that develop before vibration amplitude crosses traditional thresholds.
For maintenance organisations adopting AI-based monitoring in 2026, the practical workflow change is significant. The AI handles routine alert generation; the skilled vibration analyst handles the cases AI flags as ambiguous. Productivity per analyst rises substantially.
The vibration analysis training pathway
Skilled vibration analysts are increasingly hard to find. The European industrial reliability industry trains analysts through internationally-recognised certification programmes: ISO 18436 Category 1-4, ASNT Level I-III, Mobius Institute training. Each level represents increasing depth of theoretical knowledge and practical experience. Category 1 analysts handle routine inspections; Category 4 analysts lead complex failure investigations.
For maintenance organisations building vibration analysis capability internally, the recommended pathway: Category 1 certification for technicians performing routine inspection, Category 2 for those interpreting basic FFT spectra, Category 3 for the team lead, and external Category 4 support for complex root-cause investigations. The training investment is significant but the productivity payback is substantial.
The maintenance shift toward condition-based intervention
Vibration-driven condition-based intervention represents a fundamental shift from time-based or run-to-failure maintenance philosophies. The maintenance organisation moves from a calendar-driven schedule to a data-driven response. The transition requires CMMS integration, alert workflow design, and team culture adjustment.
Successful transitions follow a phased approach: deploy monitoring infrastructure first, collect baseline data over 4-6 weeks, then progressively shift maintenance triggers from calendar to condition. The full transition typically takes 12-24 months for a mid-size industrial plant. The productivity benefits compound throughout the transition and beyond.
What the trend data actually reveals
Looking at vibration trend data across years rather than days reveals patterns that no single inspection can capture. Bearings that fail prematurely often show a characteristic slow drift in baseline vibration over months before the formal Stage 2 defect frequencies appear. Lubrication interventions show measurable improvement in baseline. Misalignment-driven damage progresses faster than other failure modes.
These macro-level patterns inform improvements in maintenance practices that single-event analysis cannot identify. Maintenance organisations that systematically review trend data quarterly or annually identify systemic patterns and implement improvements that improve fleet reliability across all assets.
Industry context and supplier alignment
The European bearing industry continues to consolidate around fewer larger suppliers, more sophisticated technology platforms, and tighter integration between bearing supply and reliability services. For customers, the practical implication is supplier selection becoming a longer-term strategic decision rather than a transactional cost optimisation. The supplier relationship in 2026 carries forward a multi-year roadmap of product evolution, technology integration, and engineering partnership.
Customers who build deliberate, multi-source supplier relationships position themselves to navigate this consolidation effectively. The ability to substitute between suppliers — supported by clean cross-reference data and qualified engineering equivalence — protects against any single supplier’s strategic missteps and captures the competitive value of supplier rivalry while it persists.
Related guides on Eurobearing
- Reducing Machine Downtime through Proper Bearing Maintenance
- Predictive Maintenance: How Sensors Help
- Invisible Bearing Damage
Need help choosing the right bearing for your application? Our technical team can support you with selection, cross-references, and lead-time information.
Need help with bearing selection? Contact our technical team for a free consultation.
Related articles