Ultrasound detects the friction and impact sounds of early bearing defects at 20–100 kHz, before these defects are large enough to generate measurable vibration at bearing defect frequencies. Vibration analysis identifies the specific defect location (outer race, inner race, ball) once the defect has grown enough to produce detectable energy at BPFO, BPFI, or BSF. Both detect the same deteriorating bearing, at different stages of its progression.
What Each Technique Actually Measures
The difference between ultrasound and vibration analysis is not merely one of frequency range. The two methods measure fundamentally different physical phenomena.
Acoustic Monitoring: High-Frequency Acoustic Emission
Airborne ultrasound instruments detect acoustic emission in the 20–100 kHz range. At a deteriorating bearing, asperity contact between the rolling element and raceway surface generates friction and micro-impact events. These events radiate acoustic energy at high frequencies, well above the bearing defect frequencies that vibration analysis targets.
The ultrasound instrument does not need to resolve a specific defect frequency. It measures overall acoustic emission amplitude and, in more advanced implementations, the statistical character of the signal (dBuV, crest factor, kurtosis). A rise in amplitude against a baseline indicates degraded lubrication or surface contact that should not be occurring.
Vibration Analysis: Mechanical Force at Defect Frequencies
An accelerometer mounted on a bearing housing measures mechanical vibration transmitted from the bearing through the housing structure. When a rolling element strikes a defect on the outer race, the impact generates a repeatable force at a predictable frequency: the outer race ball pass frequency (BPFO). The same logic applies for inner race defects (BPFI), ball defects (BSF), and cage defects (FTF).
Most bearing fault detection in vibration analysis occurs in the 100–5,000 Hz range, though high-frequency acceleration envelope analysis (HFEA) extends this into the 5,000–20,000 Hz range for specific applications. The key requirement is that the defect must be large enough to generate a discrete, repeatable impact detectable above the noise floor.
Fault Stage Sensitivity: Where Each Method Excels
Bearing deterioration progresses through identifiable stages. The sensitivity advantage of each technique shifts as the fault develops.
Stage 1: Lubrication film begins to degrade. Metal surfaces approach contact but no discrete surface defect has formed yet. Ultrasound amplitude rises as friction increases. Vibration at bearing defect frequencies remains at baseline because there is no impact to generate them. Ultrasound is the more sensitive indicator at this stage.
Stage 2: A discrete surface defect (pit, spall, or crack) has formed on one of the races or rolling elements. The defect is small, but large enough to generate a repeatable impact at the corresponding defect frequency. Vibration analysis begins to detect elevated energy at BPFO, BPFI, or BSF. Ultrasound continues to show elevated amplitude. Both techniques now indicate a problem, but vibration analysis identifies its location.
Stage 3: The defect has grown. Sideband frequencies appear around the defect frequency in the vibration spectrum. Harmonics become visible. Overall vibration level rises. Ultrasound amplitude is also elevated, often with a higher crest factor indicating more severe impacting. Vibration analysis now provides the more detailed picture of severity and progression rate.
Stage 4: Failure is imminent. Overall vibration levels are high across the spectrum. Defect frequencies may actually become less distinct as damage is widespread. Both techniques indicate a critical condition. The bearing should already have been replaced.
Slow-Speed Bearing Sensitivity
This is where ultrasound has a decisive practical advantage over standard vibration analysis.
Bearing defect frequencies are proportional to shaft speed. At 15 RPM, the BPFO for a typical six-element bearing falls below 1 Hz. Standard accelerometers have frequency response floors of 1–10 Hz. Even where the hardware could theoretically detect such low frequencies, the signal is buried in environmental noise and structural vibration that dominates the low-frequency spectrum. Envelope analysis techniques help, but the fundamental challenge remains.
Ultrasound does not face this limitation in the same way. The acoustic emission from asperity contact at 20–100 kHz is generated by the physical contact event, not by the rotation rate. A bearing running at 10 RPM with a degraded lubrication film still produces friction-generated acoustic emission at high frequencies. The ultrasound instrument detects it regardless of how slow the shaft is turning.
Applications where this matters: kilns, slow-speed paper machine rolls, large gearboxes operating below 50 RPM, conveyor drive bearings, and any rotating equipment where defect frequencies fall below reliable accelerometer detection thresholds.
Lubrication Monitoring and Feedback
Ultrasound provides something vibration analysis cannot: real-time lubrication condition feedback during greasing.
When a bearing is underlubricated, friction increases and ultrasound amplitude rises above baseline. As grease is added, the lubricant film is restored, friction decreases, and the ultrasound reading returns toward the established baseline. The technician stops adding grease at that point, having added exactly what was needed and nothing more.
Over-greasing is a common and underestimated cause of premature bearing failure. Excess grease causes churning, heat buildup, and seal damage. Real-time ultrasound feedback during relubrication prevents both under- and over-greasing in a way that no post-hoc vibration measurement can replicate.
Vibration analysis can detect a severely underlubricated bearing once surface deterioration has begun, but it provides no useful feedback during the greasing process itself.
Fault Type Specificity
Vibration analysis identifies fault type with a precision that ultrasound cannot match.
A frequency spectrum showing elevated energy at 3.1x running speed, with sidebands at 1x intervals, is consistent with an outer race defect on a bearing with a specific geometry. That specificity allows the maintenance team to plan a targeted repair rather than a precautionary replacement. Trending the amplitude at BPFO over successive routes quantifies progression rate and informs the replacement window.
Ultrasound gives overall acoustic condition, often expressed in dBuV or as a ratio against established baseline. It tells you the bearing is deteriorating, and it can quantify relative severity, but it does not tell you whether the defect is on the outer race, inner race, or rolling element. For planning purposes, that distinction matters.
Beyond bearings, vibration analysis detects faults that ultrasound cannot: shaft imbalance (1x running speed), misalignment (1x, 2x, and sometimes 3x), gear mesh defects (gear mesh frequency and sidebands), structural looseness (sub-synchronous frequencies and harmonics), and resonance conditions. These require frequency-domain analysis of the vibration signal, not acoustic emission measurement.
Non-Bearing Applications
Both techniques have applications beyond bearing condition monitoring, and these applications do not overlap.
Ultrasound instruments detect pressurized gas leaks (compressed air, steam, nitrogen) through the turbulent flow signal they generate at high frequency. The same instrument used for bearing surveys doubles as a leak detection tool on a compressed air survey. Electrical discharge (corona, tracking, arcing) in switchgear and motor windings also generates ultrasonic acoustic emission that instruments in contact or airborne mode can detect.
Vibration analysis has no equivalent application for leak detection or electrical discharge. Its non-bearing applications are mechanical: gear condition, pump cavitation (through vibration signature), structural resonance mapping, and motor electrical frequency analysis through motor current or vibration spectrum.
Instrument Cost and Analysis Complexity
A portable ultrasound instrument costs less than a full-featured portable vibration analyzer with data collector, analysis software, and bearing frequency database. Entry-level ultrasound instruments are available for a fraction of the cost of entry-level vibration data collectors with spectral analysis capability.
The analysis workflow also differs. Ultrasound routes compare current readings against baselines and alarm thresholds. The decision logic is straightforward. Vibration analysis involves spectrum interpretation, envelope analysis, harmonic cursor identification, and trend judgment that requires more specialized training.
This cost and complexity difference makes ultrasound accessible to a broader range of maintenance personnel. A technician can be productive with ultrasound-based bearing screening in hours. Reliable vibration spectrum interpretation typically requires structured training and months of supervised practice.
Comparison Table
| Dimension | Acoustic Monitoring (Ultrasound) | Vibration Analysis |
|---|---|---|
| Physical phenomenon measured | Acoustic emission from asperity contact and friction (20–100 kHz) | Mechanical force transmitted to housing at defect frequencies (0–20,000 Hz) |
| Primary frequency range | 20–100 kHz | Typically 1–20,000 Hz; bearing fault detection most common in 100–5,000 Hz |
| Earliest fault stage detected | Stage 1: lubrication film degradation, before discrete surface defect forms | Stage 2: discrete surface defect large enough to generate repeatable impacting at defect frequency |
| Slow-speed bearing sensitivity | Strong: friction-generated emission independent of shaft speed | Limited: defect frequencies below 1 Hz at low RPM; unreliable below ~50 RPM for many bearings |
| Lubrication feedback | Yes: real-time amplitude change during relubrication confirms correct grease volume | No: no real-time lubrication condition signal during greasing |
| Fault type specificity | Overall condition only; cannot distinguish outer race from inner race or ball defect | Identifies specific defect location via BPFO, BPFI, BSF, FTF frequency analysis |
| Non-bearing applications | Compressed air leaks, steam trap condition, electrical discharge (corona, arcing, tracking) | Imbalance, misalignment, gear mesh defects, looseness, structural resonance, cavitation |
| Instrument cost | Lower; entry-level instruments practical for screening routes | Higher; full-featured data collector with spectral analysis and bearing database required |
| Analysis complexity | Lower; baseline comparison and threshold alarms; shorter training requirement | Higher; spectrum interpretation, envelope analysis, harmonic identification; specialized training required |
| Role in combined program | Initial screening, lubrication confirmation, slow-speed assets, leak and electrical survey | Fault type identification, severity progression trending, frequency-specific alarm management |
When to Choose Ultrasound Over Vibration Analysis
Several scenarios favor ultrasound as the primary or sole tool.
Slow-Speed Rotating Equipment
Any asset running below 50 RPM where bearing defect frequencies fall below reliable accelerometer detection. Kilns, large agitators, slow-speed gearboxes, and main conveyor drive bearings are common examples. Ultrasound detects friction from surface degradation regardless of shaft speed.
Real-Time Lubrication Confirmation
When the maintenance program includes precision lubrication routes, ultrasound provides the only real-time feedback that confirms correct grease quantity has been applied. This replaces time-based or quantity-based greasing schedules with condition-based ones.
Initial Asset Screening at Scale
For a large bearing population where resources do not support vibration routes on every asset, ultrasound provides a faster, lower-cost initial screening pass. Assets that flag on ultrasound screening move to vibration analysis for detailed fault characterization. Assets that remain at baseline are deprioritized.
Physically Inaccessible Mounting Points
Accelerometer mounting requires physical access to the bearing housing. Where that access requires scaffolding, confined space entry, or removal of guards, the route cost increases substantially. Ultrasound instruments can often assess bearings from a greater standoff distance using a parabolic reflector or contact probe on an extension, without the same access requirements.
Compressed Air and Steam System Surveys
An ultrasound instrument already in hand for bearing surveys doubles for leak detection on compressed air headers, steam trap surveys, and valve seat integrity checks. Vibration analysis has no role in these applications.
Electrical Discharge Detection
Ultrasound detects corona discharge, tracking, and arcing in medium-voltage switchgear, motor windings, and bus duct through the acoustic emission these phenomena generate. This allows electrical inspections without infrared camera equipment in many cases. Vibration analysis does not detect electrical discharge.
When to Choose Vibration Analysis Over Ultrasound
Other scenarios require vibration analysis capabilities that ultrasound cannot provide.
Fault Type Identification for Maintenance Planning
When screening indicates a deteriorating bearing and the maintenance team needs to know whether the defect is on the outer race, inner race, or rolling element before scheduling a replacement, vibration spectral analysis is required. That specificity informs whether the bearing can run to the next planned outage or requires immediate attention.
Imbalance and Misalignment Diagnosis
Imbalance produces a dominant vibration peak at 1x running speed. Misalignment produces elevated energy at 1x and 2x, sometimes 3x, with characteristic phase relationships between measurement planes. These are mechanical faults in the rotor system, not bearing surface defects. Ultrasound instruments are not designed to resolve these frequency components.
Frequency-Specific Alarm Management
A mature predictive maintenance program sets alarm thresholds at specific frequency bands, not just overall vibration level. A 1x alarm catches imbalance. A BPFO band alarm catches outer race defects. This precision reduces false alarms caused by normal overall vibration variation. Ultrasound programs use amplitude thresholds against baseline, which is less specific for assets where normal operating conditions vary.
High-Speed Bearings with Complex Spectra
At high shaft speeds, multiple bearing defect frequencies, harmonics, and sidebands appear simultaneously in the vibration spectrum. Separating these components and identifying which bearing in a multi-bearing machine is defective requires full spectral analysis. Ultrasound gives a combined signal from all acoustic sources in range; it is less effective at isolating one bearing among several in close proximity.
Gear Mesh Analysis
Gearbox condition assessment through gear mesh frequency, sideband analysis, and cepstrum analysis requires vibration data. Ultrasound can detect severe gear damage through the friction and impact signals it generates, but it does not resolve the frequency structure needed for gear fault characterization or early-stage gear surface fatigue detection.
Using Both in a Combined Program
The most effective bearing monitoring programs use ultrasound and vibration analysis in defined roles, not as alternatives.
Ultrasound routes cover the full bearing population at higher frequency. They catch Stage 1 deterioration and confirm lubrication condition. Assets that alarm on ultrasound move to a vibration analysis route for fault characterization and severity trending. Slow-speed assets stay on ultrasound-only routes where vibration analysis is unreliable.
Vibration analysis routes cover critical and complex assets at higher cost and lower frequency. They provide fault type identification, severity progression data, and detection of mechanical faults (imbalance, misalignment, gear defects) that ultrasound does not address.
The combination gives earlier detection, more complete fault coverage, and better allocation of the higher-cost vibration analysis resource to the assets where it adds the most value.
Frequently Asked Questions
Can ultrasound replace vibration analysis for bearing monitoring?
No. Ultrasound detects early-stage bearing deterioration and provides lubrication feedback that vibration cannot, but it does not identify whether the defect is on the outer race, inner race, or rolling element. Vibration analysis resolves fault type and severity through frequency-specific indicators (BPFO, BPFI, BSF, FTF). The two techniques are complementary, not interchangeable.
Why is ultrasound better for slow-speed bearings?
Bearing defect frequencies scale with rotation speed. At 15 RPM, the outer race defect frequency (BPFO) for a typical bearing falls below 1 Hz. Standard accelerometers cannot reliably detect signals at that frequency, and the signal is buried in low-frequency noise. Ultrasound detects the friction and impact energy generated by asperity contact regardless of shaft speed, making it effective where vibration-based fault frequency analysis is not.
How does ultrasound confirm that bearing lubrication is correct?
During relubrication, an ultrasound instrument monitoring the bearing in real time produces a distinct amplitude change as the lubricant film is restored. The technician stops adding grease when the ultrasound reading returns to baseline. This prevents both under-greasing and over-greasing, which are leading causes of premature bearing failure. Vibration analysis gives no equivalent real-time lubrication condition signal.
At what bearing fault stage does vibration analysis become more useful than ultrasound?
Vibration analysis becomes the more informative tool from Stage 2 onward, when a discrete surface defect has grown large enough to generate repeatable impacting at a calculable defect frequency. At this stage, spectral analysis can identify whether the defect is on the outer race, inner race, or rolling element, and amplitude trending at those frequencies allows severity tracking over time. In Stage 1, when only lubrication film degradation is present, ultrasound is more sensitive.
What faults can vibration analysis detect that ultrasound cannot?
Vibration analysis detects rotor imbalance, shaft misalignment, gear mesh defects, structural resonance, and looseness through frequency-domain analysis. These faults produce measurable vibration at predictable frequencies (1x, 2x, gear mesh frequency, sidebands) that ultrasound instruments are not designed to resolve. Ultrasound operates in a frequency band above these mechanical frequencies and is optimized for friction, impacting, and high-frequency discharge events.
Continuous Bearing Monitoring Without Manual Routes
Tractian's condition monitoring sensors combine vibration and acoustic data continuously on every connected asset, detecting bearing deterioration at Stage 1 without scheduled inspection routes.
See Mechanical Monitoring