Understanding Band-Pass Filters
A band-pass filter (BPF) is a frequency-selective signal-processing element that lets vibration components inside a chosen frequency band pass through while attenuating everything both below and above that band. It is, in effect, the combination of a high-pass filter (which blocks low frequencies) and a low-pass filter (which blocks high frequencies), forming a “window” that admits only a selected middle range. Every band-pass filter is described by three numbers: its centre frequency, its bandwidth, and its order or steepness. In vibration work the BPF is indispensable for envelope analysis, for focused diagnostics on a particular range, and for lifting weak signals out of noise by rejecting everything outside the band of interest. It is one of the most-used tools in the wider toolbox of signal filtering.
1. Filter Parameters
Centre Frequency (f₀)
- The middle of the passband and the point of maximum filter response.
- Chosen to match the frequency content of interest — typically a known resonance or fault frequency.
Bandwidth (BW)
- Definition: the frequency span between the −3 dB points, fhigh − flow.
- Narrow band: BW < 10% of f₀ — highly selective.
- Wide band: BW > 50% of f₀ — less selective.
- Q factor: Q = f₀ / BW; a higher Q means a narrower, more selective filter.
Filter Characteristics
- Lower cutoff (flow): where the lower skirt falls to −3 dB.
- Upper cutoff (fhigh): where the upper skirt falls to −3 dB.
- Shape factor: the ratio of stopband to passband width — a measure of how sharply the filter cuts off.
2. Applications in Vibration Analysis
2.1 Envelope Analysis — the Primary Use
The band-pass filter is the critical first step in detecting rolling-element bearing defects:
- Band selection: typically 500 Hz–10 kHz or 1 kHz–20 kHz.
- Purpose: isolate the high-frequency structural resonances that bearing impacts excite.
- প্রক্রিয়া: BPF → envelope detection (demodulation) → FFT of the envelope.
- Result: the bearing fault frequencies stand out clearly in the resulting envelope spectrum.
2.2 Resonance Band Analysis
Filtering tightly around a structural or bearing resonance isolates the energy at that mode from all other frequencies, letting you assess the excitation and response at a specific resonance — a powerful aid in resonance troubleshooting.
2.3 Frequency-Range Isolation
A BPF can focus on a chosen diagnostic range — say 10–100 Hz for low-frequency work — stripping away low-frequency drift and high-frequency noise to clarify the components you care about.
2.4 Gear Mesh Isolation
Centring the band on the gear mesh frequency passes that peak and its sidebands while rejecting other gear stages and bearing frequencies, enabling focused gear analysis. Where the goal is to follow a varying speed rather than a fixed band, a tracking filter performs the same isolation referenced to shaft order.
3. Band-Pass Filter Design
Cascaded Low-Pass and High-Pass
The most common implementation simply chains the two simpler filters:
- A high-pass section blocks everything below flow.
- A low-pass section blocks everything above fhigh.
- In series they form the band-pass, each section contributing to the overall selectivity.
Direct Band-Pass Design
Alternatively the filter is optimised as a single stage rather than a cascade. This is more complex to design but can achieve better characteristics, and it is reserved for specialised applications. A close relative is the notch filter, which does the inverse job — rejecting one narrow band while passing everything else.
4. Practical Considerations
Bandwidth Trade-Offs
Narrow bandwidth gives better selectivity and stronger rejection of adjacent frequencies, but it may miss frequency drift and demands precise tuning — best when the frequency of interest is known and stable. Wide bandwidth captures frequency variation and is far less fussy to tune, at the cost of weaker rejection of nearby unwanted content — best when the frequency wanders or a whole range matters.
Choosing the Band for Envelope Analysis
- Typical bands: 500–2,000 Hz, 1,000–5,000 Hz, and 5,000–20,000 Hz.
- Selection: pick the band with the strongest bearing-resonance excitation.
- Verify: check the raw acceleration spectrum to locate that resonance first.
- Optimise: adjust the band to maximise the bearing-defect signal.
5. Filter Effects on the Signal
Time-Waveform Effects
A band-pass-filtered time waveform shows only the passband content. With a narrow band it appears as a modulated carrier; low-frequency variations and high-frequency noise are gone, which can greatly simplify interpretation.
Spectrum Effects
In the spectrum, passband amplitudes are preserved while stopband amplitudes are cut by a typical 40–80 dB. The result is a cleaner display focused on the band of interest, with a lowered noise floor wherever the noise lay outside the passband.
6. Digital vs. Analog, and Bands by Frequency Range
Digital vs. Analog Filters
এনালগ band-pass filters are implemented in hardware in the signal path, operate in real time, have fixed characteristics once built, and are used in anti-aliasing and signal conditioning. Digital filters process the signal in software after digitisation, offer adjustable parameters, and can be applied or removed even after the data is collected — which is why modern analysers provide extensive digital BPF options.
Common Bands by Range
- Low-frequency (10–200 Hz): unbalance and misalignment analysis, low-speed machinery, and foundation or structural vibration.
- Mid-frequency (200–2,000 Hz): gear mesh frequencies, blade and vane passing frequencies, and the lower bearing fault frequencies.
- High-frequency (2–40 kHz): bearing-defect envelope analysis, high-frequency impacts, and bearing-resonance excitation.
7. Band-Pass Filtering in the Field
In practice, the band-pass filter is rarely used alone — it is a stage inside a measurement chain that also samples, windows, and transforms the signal, so the chosen band must sit within the instrument’s sampling bandwidth. A portable two-channel analyser such as the ব্যালানসেট-১এ measures vibration across roughly 5 Hz to 1 kHz and resolves the 1× amplitude and phase needed for on-site balancing; band-pass and envelope techniques then complement that workflow when an engineer needs to confirm whether a high-frequency bearing defect, rather than simple unbalance, is the true source of the trouble. When setting up such an analysis, the FFT Resolution Calculator helps match the line count and bandwidth to the band you intend to examine, so closely spaced fault lines and sidebands are not smeared together. Mastering band-pass selection — above all for envelope analysis and frequency-range isolation — is essential to extracting clear diagnostic information from a complex vibration signature.
English 
العربية 
Azərbaycan dili 
বাংলা 
Български 
简体中文 
Hrvatski 
Čeština 
Dansk 
Nederlands 
Eesti 
Suomi 
Français 
ქართული 
Deutsch 
Ελληνικά 
עִבְרִית 
हिन्दी 
Magyar 
Bahasa Indonesia 
Italiano 
日本語 
한국어 
Latviešu valoda 
Lietuvių kalba 
Bahasa Melayu 
Norsk bokmål 
فارسی 
Polski 
Português do Brasil 
Português 
Română 
Русский 
Српски језик 
Slovenčina 
Slovenščina 
Español 
Svenska 
ไทย 
Türkçe 
Українська 
اردو 
Tiếng Việt 