Understanding Aliasing in Vibration Analysis

Definition: What is Aliasing?

Aliasing is a critical signal processing error that can occur during the digital analysis of vibration data. It happens when a signal is sampled at a rate that is too low to accurately capture its highest frequency components. As a result, these high frequencies “fold down” or “impersonate” lower frequencies in the resulting FFT spectrum, creating false frequency peaks that can lead to a serious misdiagnosis of the machine’s condition.

The Nyquist Theorem and Sampling Rate

To understand aliasing, one must first understand the Nyquist Theorem (also known as the Nyquist-Shannon sampling theorem). This fundamental principle of digital signal processing states:

To accurately represent an analog signal in digital form, the sampling frequency (Fs) must be at least two times the highest frequency component (Fmax) present in the signal.

This minimum sampling rate (2 * Fmax) is called the Nyquist rate. In vibration analysis, the highest frequency of interest that can be accurately measured is therefore half the sampling rate (Fmax = Fs / 2). This Fmax is often referred to as the Nyquist frequency.

How Does Aliasing Occur?

Imagine a high-frequency vibration signal being measured by a digital analyzer. The analyzer takes discrete samples (snapshots) of the signal at a fixed rate (the sampling frequency).

• If the sampling rate is high enough (well above the Nyquist rate), the analyzer captures a sufficient number of points to accurately reconstruct the waveform.
• However, if the sampling rate is too low, the analyzer “misses” what happens between samples. The few points it does capture can be connected to form a completely different, lower-frequency sine wave. This new, false low frequency is the “alias.”

For example, if a signal contains a 900 Hz component but the analyzer’s Fmax is set to 500 Hz (meaning a sampling rate of 1000 Hz), the 900 Hz component cannot be measured correctly. It will be “aliased” and appear as a peak at a lower frequency (specifically at Fs – 900 Hz = 1000 – 900 = 100 Hz), potentially being mistaken for a 1X running speed vibration.

Preventing Aliasing: The Anti-Aliasing Filter

It is impossible to know in advance all the high-frequency content (e.g., from ultrasonic noise, impacts, or radio frequency interference) that might be present in a signal. Therefore, relying on simply setting the sampling rate high enough is not a practical solution.

The solution used in all modern digital vibration analyzers is the anti-aliasing filter. This is a steep low-pass filter that is placed in the signal path *before* the analog-to-digital converter (ADC). Here’s how it works:

1. The user sets the desired maximum frequency (Fmax) for their analysis.
2. Based on this Fmax, the analyzer automatically sets the anti-aliasing filter’s cut-off frequency slightly above Fmax.
3. The analog signal from the sensor passes through this filter, which removes or strongly attenuates all frequencies above the cut-off point.
4. Only the filtered, “clean” signal is then sent to the ADC for sampling.

By removing the high frequencies that the chosen sampling rate cannot handle, the anti-aliasing filter makes it physically impossible for aliasing to occur. It is one of the most critical components of a digital signal analyzer, ensuring that the resulting FFT spectrum is a true and accurate representation of the machine’s vibration within the chosen frequency range.

← Back to Main Index