4 Stages of Bearing Failure: How to Catch a Defect Months Before It Fails

A bearing doesn't die suddenly

The most dangerous misconception in operations is believing that a bearing "worked fine, then suddenly failed." In reality, from the first microscopic defect to catastrophic failure, several months typically pass. The bearing is signalling the whole time — it's just that different diagnostic methods "hear" those signals at different stages.

Understanding these stages is the foundation of predictive maintenance. It answers the key question every reliability department asks: "How much time do I have before this bearing becomes a problem?"

The P-F curve: your window for action

The P-F curve concept describes how any component degrades over time:

• P (Potential failure) — the point of potential failure, the moment a defect first becomes detectable.
• F (Functional failure) — the point of functional failure, when the component stops doing its job (a breakdown).

The interval between P and F is the P-F interval — your window for planning a repair. The earlier the diagnostic method you use, the further left point P sits on the curve, and the more time you have to react. Ultrasound pushes point P as far left as possible — it detects the defect first.

The 4 stages of rolling bearing failure

Stage 1: defect initiation (ultrasound)

At the microscopic level, fatigue changes begin: the first subsurface microcracks, breakdown of the oil film, the onset of wear. Vibration is still normal, temperature is still normal, the unit runs perfectly.

What detects it: only ultrasound. Defects and insufficient lubrication generate high-frequency acoustic emission (around 36 kHz), which the SDT340 instrument picks up long before vibration appears. At this stage, metered lubrication under ultrasonic guidance is often enough to halt the degradation — see more in the article "Why does a motor bearing overheat".

Time to failure: months.

Stage 2: defects appear on the raceways (ultrasound + high-frequency vibration)

Microcracks reach the surface, the first pits appear. The ultrasound signal grows, and modulated peaks at bearing defect frequencies (BPFO, BPFI, BSF, FTF) start to appear in the high-frequency vibration band.

What detects it: ultrasound, confidently, plus vibration analysis in the acceleration/envelope band. Fixed Bently Nevada systems start to see the defect in the envelope spectrum.

Time to failure: weeks to a couple of months.

Stage 3: developed defects (vibration — spectrum and harmonics)

Pits grow larger and the defects become visible in the classic vibration spectrum: clear peaks at the defect frequencies and their harmonics, sidebands. A faint rumble appears, sometimes a barely noticeable rise in heat. At this stage the defect is already obvious to any vibration analyst.

What detects it: vibration analysis — this is its home territory. We covered how to read this kind of spectrum and distinguish between defects in "How to read an FFT spectrum", and an unusual case in “The mysterious 0.6× frequency”.

Time to failure: days to weeks. Time to plan replacement.

Stage 4: imminent failure (vibration, temperature, noise, the human ear)

The bearing is breaking down: overall vibration levels rise, temperature climbs, there's audible noise and vibration you can feel by hand. The spectrum becomes "noisy" — energy smears across the entire frequency range.

What detects it: by now, everything detects it, including your own ears and a pyrometer. But this is the last warning — only hours or days may remain before seizure.

Time to failure: hours to days. Emergency shutdown.

Why no single method covers every stage

Stage	Ultrasound	Vibration (envelope)	Vibration (spectrum)	Temperature
1. Initiation	✅	—	—	—
2. First defects	✅	✅	—	—
3. Developed defects	✅	✅	✅	±
4. Imminent failure	✅	✅	✅	✅

The conclusion is clear: the earlier a method "switches on," the more time you have and the cheaper the intervention. Ultrasound gives the earliest warning and simultaneously lets you lubricate the bearing correctly, preventing it from ever reaching stage 2. Vibration confirms and accurately classifies the defect across stages 2–4. Temperature is just a statement of fact after the event.

How to build a bearing monitoring program

1. Route-based ultrasonic inspection of the whole fleet — catches stages 1–2 while also checking lubrication.
2. Periodic vibration monitoring of critical assets — classifies defects at stages 2–3.
3. Fixed systems on your most critical equipment — continuous monitoring and protection.
4. Trends, not single readings — decisions should be based on signal dynamics, not a single measurement.

This multi-layered approach turns a bearing from "a source of sudden failures" into a predictable component with a clearly understood remaining life.

Conclusion

A bearing goes through 4 stages of failure, and you have months to react — if you use the right method at the right stage. Ultrasound opens the widest window (point P as far left as possible on the P-F curve), vibration analysis refines the diagnosis, and temperature only confirms what's already too late to fix easily. Combining methods is what mature predictive maintenance actually looks like.

KEG TRK helps Kazakhstan enterprises build a multi-layered bearing-fleet monitoring program — from route-based ultrasonic instruments to fixed vibration monitoring systems. Submit a request — we'll select a program for your equipment fleet.