What is FFT frequency resolution?

FFT frequency resolution (Δf) is the smallest frequency difference that can be distinguished in a spectrum. It equals Fmax divided by the number of lines: Δf = Fmax / Lines.

How many FFT lines do I need?

More lines give finer resolution but require longer acquisition time. 400 lines is common for general monitoring, 800-1600 for diagnostics, 3200-6400 for detailed bearing or gearbox analysis.

What is the relationship between sampling rate and Fmax?

Most analyzers use a sampling rate of 2.56 × Fmax. The factor 2.56 (rather than 2.0 from Nyquist) provides margin for the anti-aliasing filter.

Can I resolve bearing defect frequencies with 400 lines?

It depends on the Fmax. For example, 400 lines with Fmax=1000 Hz gives Δf=2.5 Hz. Typical BPFO frequencies for a shaft at 1500 RPM are around 100-150 Hz — resolvable, but sidebands at ±25 Hz spacing would need finer resolution to separate clearly.

Free Engineering Tool

FFT Resolution Calculator

Calculate frequency resolution Δf, time record length, sampling rate, and Nyquist frequency from FFT lines and Fmax. Check bearing defect frequency resolvability.

FFT Lines Δf Resolution Nyquist

Number of FFT Lines

Fmax (Hz)

OR Sampling Rate (Hz, optional — overrides Fmax)

Shaft Speed (RPM, for bearing check)

Quick presets

Results

Frequency Resolution Δf

—

Time Record Length T

—

Sampling Rate fs

—

Nyquist Frequency

—

Total FFT Points (N)

—

Bandwidth

—

Bearing Defect Frequency Resolvability

Frequency Resolution

The frequency resolution of an FFT spectrum is the smallest frequency increment between adjacent spectral lines:

Where Fmax is the maximum analysis frequency and Lines is the number of spectral lines (400, 800, 1600, etc.).

Time Record Length

The time record required to achieve a given frequency resolution is the reciprocal:

More lines or lower Fmax means a longer time record and finer resolution.

Sampling Rate & Nyquist

Most vibration analyzers use a sampling rate factor of 2.56 (not 2.0) to allow for anti-aliasing filter roll-off:

The Nyquist frequency is f_s/2. Frequencies above Nyquist will alias into the spectrum.

FFT Lines Reference Table

Lines	Fmax=1000 Hz	Fmax=2000 Hz	Fmax=5000 Hz	T record
400	2.50 Hz	5.00 Hz	12.50 Hz	0.40 s
800	1.25 Hz	2.50 Hz	6.25 Hz	0.80 s
1600	0.625 Hz	1.25 Hz	3.125 Hz	1.60 s
3200	0.3125 Hz	0.625 Hz	1.5625 Hz	3.20 s
6400	0.15625 Hz	0.3125 Hz	0.78125 Hz	6.40 s

Practical Example

Example — Motor Bearing Analysis

Given: 400 lines, Fmax = 1000 Hz, shaft speed = 1500 RPM (25 Hz)

Δf = 1000 / 400 = 2.5 Hz

T = 1 / 2.5 = 0.4 s

f_s = 2.56 × 1000 = 2560 Hz

Typical BPFO ≈ 25 × 5.2 = 130 Hz → easily resolvable (Δf = 2.5 Hz)

Sideband spacing at 1× = 25 Hz → resolvable (2.5 Hz ≪ 25 Hz)

⚠️ Note: For reliable detection, Δf should be at least 3–5× smaller than the frequency of interest. For sideband analysis, Δf should be ≪ shaft speed in Hz.

FFT Resolution Calculator | Lines, Bandwidth, Fmax

Published by Nikolai Shelkovenko on February 15, 2026 February 15, 2026