Understanding Band-Pass Filters
Definition: What is a Band-Pass Filter?
Band-pass filter (BPF) is a frequency-selective signal processing element that allows vibration components within a specified frequency band to pass through while attenuating components both below and above that band. It combines the characteristics of a high-pass filter (blocking low frequencies) and low-pass filter (blocking high frequencies) to create a “window” that passes only a selected middle frequency range. Band-pass filters are defined by their center frequency, bandwidth, and filter order/steepness.
In vibration analysis, band-pass filters are essential for envelope analysis (isolating bearing impact frequencies), focused diagnostics (examining specific frequency ranges), and eliminating unwanted vibration outside the frequency band of interest to improve signal-to-noise ratio and measurement clarity.
Filter Parameters
Center Frequency (f0)
- Middle of the passband
- Frequency of maximum filter response
- Selected based on frequency content of interest
- Typically chosen to match resonance or fault frequency
Bandwidth (BW)
- Definition: Frequency range between -3 dB points (f_high – f_low)
- Narrow Band: BW < 10% of center frequency (highly selective)
- Wide Band: BW > 50% of center frequency (less selective)
- Q Factor: Q = f0 / BW (higher Q = narrower, more selective)
Filter Characteristics
- Lower Cutoff (f_low): Frequency where lower slope reaches -3 dB
- Upper Cutoff (f_high): Frequency where upper slope reaches -3 dB
- Shape Factor: Ratio of stopband to passband widths (measure of selectivity)
Applications in Vibration Analysis
1. Envelope Analysis (Primary Application)
Critical first step in bearing defect detection:
- Band Selection: 500 Hz – 10 kHz or 1 kHz – 20 kHz typical
- Purpose: Isolate high-frequency bearing resonances excited by impacts
- Process: BPF → envelope detection → FFT of envelope
- Result: Enhanced bearing fault frequencies clearly visible
2. Resonance Band Analysis
- Filter around structural or bearing resonance frequency
- Isolate energy at resonance from other frequencies
- Assess excitation and response at specific mode
- Useful for resonance troubleshooting
3. Frequency Range Isolation
- Focus on specific diagnostic frequency range
- Example: 10-100 Hz for low-frequency analysis
- Removes low-frequency drift and high-frequency noise
- Improves clarity for frequencies of interest
4. Gear Mesh Isolation
- BPF centered at gear mesh frequency
- Passes mesh frequency and sidebands
- Blocks other gear stages and bearing frequencies
- Enables focused gear analysis
Band-Pass Filter Design
Cascaded Low-Pass and High-Pass
Most common implementation:
- High-pass filter blocks frequencies below f_low
- Low-pass filter blocks frequencies above f_high
- Series combination creates band-pass
- Each filter contributes to total selectivity
Direct Band-Pass Design
- Optimized as single filter rather than cascade
- More complex but can achieve better characteristics
- Used in specialized applications
Practical Considerations
Bandwidth Selection Trade-offs
Narrow Bandwidth
- Advantages: Better selectivity, stronger rejection of adjacent frequencies
- Disadvantages: May miss frequency variations, requires precise tuning
- Use: When exact frequency known and stable
Wide Bandwidth
- Advantages: Captures frequency variations, less critical tuning
- Disadvantages: Less rejection of nearby unwanted frequencies
- Use: When frequency varies or range of frequencies needed
For Envelope Analysis
- Typical Bands: 500-2000 Hz, 1000-5000 Hz, 5000-20000 Hz
- Selection: Choose band with good bearing resonance excitation
- Verify: Check raw acceleration spectrum to identify resonance
- Optimize: Adjust to maximize bearing defect signal
Filter Effects on Signals
Time Waveform Effects
- Filtered waveform shows only frequencies in passband
- Appears as modulated carrier (if narrow band)
- Removes low-frequency variations and high-frequency noise
- Can simplify waveform interpretation
Spectrum Effects
- Passband amplitudes preserved
- Stopband amplitudes reduced (40-80 dB typical)
- Cleaner spectrum focusing on band of interest
- Noise floor lowered if noise outside passband
Digital vs. Analog Band-Pass Filters
Analog Filters
- Hardware implementation in signal path
- Real-time operation
- Fixed characteristics once designed
- Used in anti-aliasing and signal conditioning
Digital Filters
- Software processing after digitization
- Adjustable parameters
- Can be applied/removed post-collection
- Modern analyzers offer extensive digital BPF options
Common Applications by Frequency Range
Low-Frequency Band-Pass (10-200 Hz)
- Unbalance and misalignment analysis
- Low-speed machinery monitoring
- Foundation and structural vibration
Mid-Frequency Band-Pass (200-2000 Hz)
- Gear mesh frequencies
- Blade/vane passing frequencies
- Lower bearing fault frequencies
High-Frequency Band-Pass (2-40 kHz)
- Bearing defect envelope analysis
- High-frequency impacts
- Ultrasonic frequencies
- Bearing resonance excitation
Band-pass filters are versatile signal processing tools that enable focused analysis of specific frequency ranges while rejecting unwanted low and high-frequency components. Mastering band-pass filter selection and application—particularly for envelope analysis and frequency range isolation—is essential for advanced vibration diagnostics and effective extraction of diagnostic information from complex vibration signatures.
EN 
AR 
AZ 
BG 
ZH 
HR 
CS 
DA 
NL 
ET 
FI 
FR 
DE 
KA 
EL 
HE 
HI 
HU 
ID 
IT 
JA 
KO 
LV 
LT 
MS 
NB 
FA 
PL 
PT_BR 
PT 
RO 
SR 
SK 
SL 
ES 
SV 
TH 
TR 
UR 
UK 
VI 