Envelope analysis is a signal processing technique for detecting early-stage bearing faults — by demodulating high-frequency vibration signals to reveal the low-frequency impact pulses characteristic of BPFO, BPFI, BSF, and FTF.
Also called High Frequency Resonance Technique (HFRT) or demodulation analysis. The principle: when a bearing has a minor defect, the impact pulses are too weak to appear directly in the standard vibration velocity spectrum (10–1,000 Hz). For a full overview of condition monitoring strategies, see bearing vibration monitoring and browse deep groove ball bearings recommended for predictive maintenance programs. But each impact excites the mechanical resonance of the bearing and housing at high frequencies (2–40 kHz). Envelope analysis bandpass-filters around this resonance frequency, takes the signal envelope, then applies FFT — revealing the low-frequency modulating signals (BPFO, BPFI, BSF, FTF) that the direct velocity spectrum cannot detect.
Four-Step Envelope Analysis Procedure
Standard procedure per ISO 13373-3:
Step 1 — Select filter band: Identify the bearing/housing resonance frequency from the raw spectrum — typically 2–20 kHz. Select a bandpass filter 500–2,000 Hz wide centered on the resonance peak.
Step 2 — Filter and rectify: Apply the bandpass filter to the raw time-domain signal, then take the absolute value (rectification).
Step 3 — Lowpass filter: Apply a lowpass filter to retain only the envelope signal — the slowly varying amplitude of the high-frequency content.
Step 4 — FFT: Compute the FFT of the envelope signal → envelope spectrum. Identify peaks at BPFO, BPFI, BSF, FTF and their harmonics.
Earliest detection: Stage 1–2 (80–95% of remaining life left) — before the direct velocity spectrum shows any change.
Practical Example: Early Detection on 6308 C3
At a food processing plant in Binh Duong, a pump motor running at 2,900 rpm used a 6308 C3 bearing (d=40, D=90, Z=8 balls, BPFO_factor=3.57). The direct velocity spectrum (10–1,000 Hz) was normal; overall vibration 1.8 mm/s — below the ISO 10816 alert threshold. Envelope analysis at the 8–12 kHz band revealed a clear peak at 172.5 Hz (BPFO = 3.57 × 48.33 Hz) and 345 Hz (2×BPFO). The bearing was replaced after 2 weeks — physical inspection confirmed Stage 2 spalling on the outer raceway. The direct velocity spectrum had not yet increased because the defect was still small.
| Detection method | Fault stage detected | Lead time before failure | Key advantage |
|---|---|---|---|
| Overall vibration (ISO 10816) | Stage 3–4 | 2–6 weeks | Simple, low-cost |
| Velocity spectrum (FFT) | Stage 2–3 | 4–12 weeks | Identifies fault type |
| Envelope analysis | Stage 1–2 | 3–6 months | Earliest detection |
| Oil analysis | Stage 2–3 (oil systems) | 4–8 weeks | Complements vibration |
Selecting the Filter Band
The bandpass filter selection is the most critical step. A poorly chosen filter band will produce a clean envelope spectrum with no peaks — not because there is no fault, but because the resonance is outside the filter. Three approaches for finding the correct resonance frequency:
Impact test: Tap the bearing housing lightly with a test hammer and measure the spectrum — the resonance peak appears clearly in the response.
Manufacturer data: Some OEM equipment documentation publishes bearing housing resonance frequencies.
Default band: Use 2–10 kHz as a starting point for most low-to-medium speed industrial applications. For high-speed equipment (>3,000 rpm), extend to 10–30 kHz.