Motor Bearing Vibration Analysis and Fault Signature Testing: A Predictive Maintenance Test Bench Guide

Why Bearing Fault Testing Is Different from NVH Testing

Motor NVH (hałas, wibracja, and harshness) testing — well-covered in the EV powertrain world — asks the question: “Is this motor quiet enough for the passenger cabin?” Bearing fault signature analysis asks an entirely different question: “Is this bearing healthy, and how quickly will it fail?” The two disciplines share the same sensor (an accelerometer on the motor housing) but diverge in everything else — the frequency range of interest, the analysis algorithm, the acceptance threshold, and the engineering objective.

NVH testing typically examines the 0–20 kHz acoustic range with emphasis on tonal content at motor electrical orders. Bearing fault analysis uses high-frequency resonance excitation in the 5–40 kHz range, where bearing defect impulses ring structural resonances that a narrow-band order-tracking analysis would completely miss. For a motor manufacturer, a bearing fault test bench is a quality gateway: it catches assembly defects, zanieczyszczenie, and pre-damaged raceways before the motor reaches the field — where the cost of a bearing failure in a 500 kW industrial pump or a rail traction motor is orders of magnitude higher than the motor replacement cost.

The Physics of Bearing Fault Signatures

When a bearing defect — a spall, a pit, or a flat on a rolling element — strikes a mating surface, it produces a short-duration impulse of mechanical energy. This impulse excites the natural resonant frequencies of the bearing outer ring, inner ring, or the housing structure. The impulse repeats at a characteristic frequency determined by the geometry of the bearing and the shaft speed: the Ball Pass Frequency Outer race (BPFO), Ball Pass Frequency Inner race (BPFI), Ball Spin Frequency (BSF), and Fundamental Train Frequency (FTF). These four bearing defect frequencies — derivable from the bearing geometry data sheet — are the fingerprints of specific defect locations.

Because the impulse energy is spread across a broad frequency band via resonance excitation, detecting it requires envelope analysis (also called resonance demodulation): bandpass filtering around a structural resonance, full-wave rectifying, and low-pass filtering to extract the amplitude-modulated envelope, whose spectral content reveals the bearing defect frequencies.

Podstawowe elementy testowe

1. BPFO/BPFI/BSF/FTF Frequency Signature Detection

The test runs the motor at multiple discrete speeds (np., 600, 1,200, 1,800, 3,000 obr./min) while recording tri-axial vibration at each bearing housing. An FFT spectrum and Cepstrum analysis identifies peaks at BPFO and BPFI harmonics. The speed-normalised representation (in orders rather than Hz) allows comparison across speed points and provides a clean separation of bearing defect frequencies from electrical harmonics (which also scale with speed) and gearmesh frequencies (which depend on tooth count).

Kluczowe wskaźniki: BPFO amplitude limit < OEM threshold (np., < 0.5 g at 3,000 rpm for a motor intended for quiet operation); BPFI/BSF peaks absent or below 20% of BPFO limit; sidebands at ±1× rotating frequency around BPFI confirm inner-race defect location.

2. Envelope Analysis and High-Frequency Resonance Demodulation

A charge amplifier-coupled piezoelectric accelerometer with a flat frequency response to 40 kHz captures the bearing impulse energy. The signal is bandpass filtered (typically 8–20 kHz for steel-housing motors) around the strongest structural resonance frequency identified during a prior impact hammer test. The envelope spectrum reveals bearing defect frequencies at amplitudes far below the noise floor of a standard low-frequency vibration measurement. This technique is sensitive enough to detect a single 0.2 mm spall on a bearing raceway that would be invisible to standard ISO 10816 vibration measurement.

Kluczowe wskaźniki: Envelope spectrum floor < 0.01 g²/Hz in the bearing defect frequency band; BPFO and BPFI peaks < 3× floor for a healthy bearing; any peak > 6× floor flags for investigation or rejection.

3. Shock Pulse Measurement

The shock pulse method (SPM) uses a special sensor tuned to 32 kHz and measures the decibel shock value (dB(sVm)) as a single-number bearing health indicator, normalised for speed and bearing bore diameter. The method is particularly useful for production-line screening where a pass/fail gate is needed without a spectral analysis expert. The SPM “carpet value” (background noise level) I “max value” (peak shock level) together indicate both lubrication adequacy and bearing surface condition.

Kluczowe wskaźniki: dB(sVm) max value − carpet value < 20 dB for a “good” łożysko; 20–35 dB indicates “reduced” condition requiring monitoring; > 35 dB indicates “bad” condition, reject or re-inspect.

4. High-Frequency Accelerometer Calibration and Mounting Repeatability

Because bearing fault analysis depends on absolute amplitude levels in the 5–40 kHz range, accelerometer calibration and mounting are critical. An accelerometer screwed directly into a tapped hole in the housing (stud mounting) provides accurate response to 40+ kHz; a magnet-mounted accelerometer resonates at 3–7 kHz and is suitable only for routine screening, not for the detailed envelope spectrum. The test bench fixture must ensure consistent, repeatable mounting torque at the bearing measurement point — a ±2 dB amplitude variation from mounting inconsistency would falsely classify 30% of borderline bearings.

Kluczowe wskaźniki: Calibration uncertainty < ±1 dB at the bearing defect frequency band centre; mounting resonance frequency > 30 kHz for stud-mounted accelerometer; cross-axis sensitivity < 5% of main axis.

5. Baseline and Run-Out Mapping at Multiple Speeds

A new bearing, freshly greased and properly fitted, produces a characteristic “noise floor” baseline that is recorded during acceptance testing. This baseline, stored in the motor’s digital birth certificate alongside the bearing serial number and lot code, becomes the reference for in-service condition monitoring — allowing field measurements years later to be compared against the known factory baseline. Run-out mapping (measuring the shaft eccentricity at slow roll: 10–30 rpm) separates geometric imperfections of the shaft and housing from dynamic bearing signals at operating speed.

Kluczowe wskaźniki: Slow-roll shaft run-out < 0.025 mm TIR; baseline 1/3-octave band RMS levels recorded at 600, 1,200, 1,800, I 3,000 rpm at each bearing housing and stored to motor traceability record.

6. Accelerated Ageing: Grease Degradation Under Load

A compact accelerated bearing life test (ABLT) rig applies the motor’s rated axial and radial bearing loads at elevated temperature (90°C grease temperature) while monitoring the shock pulse level as a function of operating hours. The test plots a degradation curve — dB(sVm) versus time — to validate the bearing and grease selection for the rated regreasing interval (np., >20,000 hours grease life at 60°C continuous). This test also validates the effectiveness of sealed-for-life versus regreaseable bearing arrangements for a given application duty cycle.

Kluczowe wskaźniki: dB(sVm) max value must remain < 35 dB for the full rated relubrication interval hours at the accelerated temperature factor; no catastrophic surface fatigue (spalling) before L10 life hours.

Co to oznacza dla wyboru stanowiska testowego

A bearing fault analysis test bench requires instrumentation that most general-purpose motor test facilities do not carry: 40 akcelerometry kHz, resonance demodulation signal processing, shock pulse meters calibrated to the SPM standard, and an ABLT rig with programmable load and thermal control. For motor manufacturers implementing IEC 60068-2 (environmental testing) or aiming to supply to rail (W 50155), wind energy (IEC 61400-4), or oil and gas (API 541, API 547) markets, bearing health validation is a mandatory qualification element, not optional.

Our test engineering team integrates bearing vibration diagnostics into motor acceptance test sequences for new motor qualification programmes. Skontaktuj się z nami to discuss embedding bearing fault screening into your production test flow.