groundup.ai
30/9/2026When the outer race starts to spall, every rolling element that passes the defect lands an impact at one specific frequency. That frequency is not a guess. It is arithmetic. And once it appears in the data, a countdown has started that the overall vibration reading will not show you until it is almost over. Most routes catch a bearing at the end of that countdown, when it is already rumbling. The data offered a name and a runway long before. This is how to read it.
When the outer race starts to spall, every rolling element that passes the defect lands an impact at one specific frequency. That frequency is not a guess. It is arithmetic. And once it appears in the data, a countdown has started that the overall vibration reading will not show you until it is almost over.
Most routes catch a bearing at the end of that countdown, when it is already rumbling. The data offered a name and a runway long before. This is how to read it.
A rolling-element bearing carries load through balls or rollers running between an inner and an outer race. The outer race usually sits fixed in the housing, and the load zone sits in one place on it. That is where damage tends to start: repeated stress in the load zone drives subsurface fatigue, then a surface spall, then spreading damage across the raceway.
Common drivers are familiar, and most of them are upstream:
That last one matters for diagnosis. Outer-race failure is a defect on the raceway. Outer-race looseness is a fit problem in the housing. They can look similar at a glance and call for different fixes, which is exactly why a single vibration number is not enough.
The reason a bearing is so diagnosable is that its geometry sets the frequency of the impacts. For an outer-race defect, that frequency is the ball pass frequency, outer race, or BPFO.
BPFO depends on the number of rolling elements, the shaft speed, and the bearing geometry. A reliable rule of thumb for many bearings is:
A healthy bearing has no energy there. A spalling outer race puts a peak at 108 Hz, and then repeats it: harmonics at roughly 216 Hz, 324 Hz, and beyond. That series of evenly spaced peaks, sitting on a frequency that no shaft, vane, or gear in the machine produces, is the fingerprint. It is why a bearing fault is not just detectable. It is nameable.
The first impacts from a fresh spall are tiny, high-frequency, and short. Overall velocity RMS averages energy across roughly 10 to 1000 Hz, a band ruled by low-frequency events like imbalance and misalignment, so those small impacts barely move the number.
This is why a bearing can be several weeks into failure and still pass an overall check. The energy is real, but it is high in the spectrum and low in amplitude, exactly where a broadband average is least sensitive. Enveloping, which demodulates the high-frequency ringing the impacts excite, is what pulls the BPFO series into plain view while the overall reading still reads green.
Here is the diagnostic detail that separates a real bearing analyst from a threshold. The three bearing defects live at different frequencies, and they carry different sideband signatures.
An outer-race defect sits still in the load zone. Every rolling element strikes it with roughly the same force, so the BPFO series is usually clean, harmonics with little or no sidebanding.
An inner-race defect rotates with the shaft, moving in and out of the load zone once per revolution. That modulates the impact amplitude at shaft speed, so the BPFI series carries sidebands spaced at one times running speed.
A rolling-element defect is modulated by the cage, so its series carries sidebands at the cage frequency.
Read that way, the spectrum does not just say bearing. It says which part of the bearing, which is what turns a vague alarm into a specific work order.
Figure 1. The sideband tell. An outer-race defect sits still in the load zone and produces a clean BPFO harmonic series. An inner-race defect rotates through the load zone and adds sidebands at shaft speed.
An outer-race fault does not fail all at once. It moves through four stages, and each one shows up in a different part of the data. This is the runway.
Stage 1. The earliest. Energy appears only at ultrasonic and very high frequencies, picked up by acoustic emission or high-frequency techniques. There are no defect frequencies yet. This is the largest slice of remaining life, often weeks to months.
Stage 2. The BPFO series appears under enveloping, with early harmonics. The bearing is now audibly and measurably degrading to the right tools, though still quiet on overall vibration. Weeks of runway remain, and this is the stage you want to catch.
Stage 3. The defect frequencies and their harmonics grow, sidebands develop, and the energy becomes visible even in the standard velocity spectrum. Damage is now spreading across the raceway. Days to weeks.
Stage 4. The discrete defect frequencies blur into a rising broadband noise floor as the raceway breaks up. The BPFO peak can actually shrink or vanish, replaced by random vibration and a bearing you can hear across the room. Failure is imminent, often hours to days.
The overall vibration alarm most programs rely on tends to fire somewhere in late Stage 3 or Stage 4. By then the useful runway is gone.
Figure 2. The four stages of a bearing failure. The runway is in Stage 2, when the BPFO series first appears under enveloping. The overall alarm fires near the end, when little runway remains.
Not on its own.
A single peak near the calculated BPFO can be a coincidence, a harmonic of something else, or the signature of outer-race looseness rather than a raceway spall. The confirmation is in the pattern and the trend.
Maintenance teams should ask:
A harmonic BPFO series that is rising under enveloping is a failing outer race. One static peak is a reason to keep watching.
The load zone on an outer race is set by the direction of load, and both load and speed change the picture. Higher load sharpens the impacts, higher speed moves the defect frequencies, and a variable-speed drive moves them continuously. A fixed alarm cannot follow any of that.
The same is true across assets. Two identical bearings in two identical housings do not produce identical spectra once fit, load, and lubrication diverge. A per-asset baseline, read against operating state, is what lets a rising BPFO series mean something rather than nothing.
Both outer-race failure and outer-race looseness are among the root causes the Groundup.ai Asset Library identifies, alongside roller failure, gear failure, coupling failure, poor lubrication, seal failure, and looseness.
That is what turns a BPFO series into a decision. The Asset Library holds the confirmed bearing signatures across fleets and sectors, so the system does not just show a peak at 108 Hz and leave the interpretation to whoever is on shift. It recognises the harmonic pattern, reads the sidebands to tell outer race from inner race from rolling element, separates a raceway spall from a loose fit, places the bearing on its failure stage, and estimates the runway against the P-F interval.
That is the difference between an alert and #CognitiveMaintenance: reading the data, and understanding it.
Yes. Outer-race failure is one of the root causes the Groundup.ai Asset Library identifies, and it is one of the clearest, because the physics hands the system a nameable frequency to look for.
Rather than a single threshold, the system evaluates the BPFO fundamental and its harmonics, the enveloped high-frequency energy, the sideband structure, and the trend against the bearing's own baseline. A raised high-frequency level alone is ambiguous. A rising, harmonic BPFO series with the right sideband signature is a confident outer-race diagnosis, with a stage and a runway attached.
The data names the bearing and the stage. The physical inspection confirms the raceway.
What frequency indicates an outer-race bearing fault?
The ball pass frequency, outer race, or BPFO, roughly 0.4 times the number of rolling elements times shaft speed. A failing outer race shows a harmonic series at BPFO, clearest under enveloping.
How do you tell an outer-race fault from an inner-race fault?
By the sidebands. An outer-race defect sits in the load zone and produces a clean BPFO series with little sidebanding. An inner-race defect rotates through the load zone and produces sidebands spaced at shaft speed.
Why does overall vibration miss a failing bearing?
Early bearing impacts are high-frequency and low-amplitude, and overall velocity averages a lower-frequency band, so it stays normal until the damage is advanced. Enveloping reveals the defect frequencies far earlier.
What are the four stages of bearing failure?
Ultrasonic energy with no defect frequencies, then a BPFO series under enveloping, then defect frequencies with harmonics and sidebands in the velocity spectrum, then a rising broadband noise floor as the raceway breaks up.
Can Groundup.ai detect an outer-race bearing fault?
Yes. It is one of the root causes the Groundup.ai Asset Library identifies, using the BPFO harmonic series, sidebands, enveloped energy, and trend against the bearing's own baseline.
A bearing gives you four stages of warning. Most maintenance routes catch the fourth.
The outer race is the easiest fault in the plant to name, because the geometry writes the frequency for you. A harmonic series at BPFO, rising under enveloping, with clean sidebands, is a failing outer race, and it says so weeks before the bearing is loud enough for anyone to hear.
The only question is whether anything is set up to read the defect frequency at Stage 2, while there is still a runway, or whether the plant waits for the overall alarm at Stage 4, when the runway is gone and the choice is no longer when to replace the bearing, but what else it took with it.
#CognitiveMaintenance #PredictiveMaintenance #ConditionMonitoring #ReliabilityEngineering #Bearings
Bergabunglah dengan Wall of ❤️ kami — bagikan dan tag kami!
Kami menyoroti para pemimpin yang mendorong inovasi dan kinerja, memberi Anda visibilitas sambil membentuk masa depan AI
Artikel Terkait
