VectoraLab Lab

What is FFT? The Fast Fourier Transform (FFT) converts a time-domain vibration signal into a frequency-domain spectrum. It reveals which frequencies are present in the signal and at what amplitudes. Mathematically, FFT computes the Discrete Fourier Transform (DFT) efficiently: X[k] = Σ x[n] · e^(-j2πkn/N) Key FFT Parameters: • Sample Rate (Fs): Must be > 2 × max frequency (Nyquist theorem) • Record Length (N): More samples = better frequency resolution Δf = Fs/N • Frequency Resolution: Δf = Fs/N (Hz per bin) • Max Frequency (Fmax): = Fs/2 (Nyquist limit) Window Functions: • Rectangular: Best frequency resolution, worst leakage • Hann: Good all-around, most common • Hamming: Similar to Hann • Blackman: Excellent leakage reduction, wider main lobe • Flat-top: Best amplitude accuracy for discrete tones Interpreting the Spectrum: • Peaks at 1X, 2X, 3X of shaft speed: mechanical faults • Sharp isolated peak: discrete frequency source • Broad hump: structural resonance or random noise • Sidebands: modulation (fault interacting with carrier) Common Mistakes: • Spectral leakage from wrong window choice • Aliasing from insufficient sample rate • Averaging too few records (poor averaging) • Not accounting for machine speed variation Important: FFT shows possible fault frequencies. Always combine with trending, phase analysis, and physical inspection for reliable diagnosis.