Diagnosing Bearing Defects

1. The Nature of Bearing Defects

Rolling-element bearings are fundamental components in most rotating machinery, but they are also a common point of failure. A typical bearing consists of an outer race, an inner race, a set of balls or rollers, and a cage to maintain their spacing. A “defect” is a microscopic or macroscopic flaw, such as a crack, spall, or pit, on one of these surfaces.

As a rolling element passes over a defect, it generates a small, high-frequency impact or “click.” While a single impact is low in energy, these impacts occur repetitively. Vibration analysis is exceptionally effective at detecting these repetitive, periodic impacts long before the bearing begins to overheat or make audible noise.

2. The Four Fundamental Fault Frequencies

The key to diagnosing bearing defects is knowing that for a given bearing geometry and rotational speed, the impacts will occur at very specific, predictable frequencies. These are known as the Fundamental Fault Frequencies:

• BPFO (Ball Pass Frequency, Outer Race): The rate at which the rolling elements pass a single point on the outer race. This is the most common bearing defect frequency.
• BPFI (Ball Pass Frequency, Inner Race): The rate at which the rolling elements pass a single point on the inner race. Since the inner race is rotating, this frequency is higher than BPFO.
• BSF (Ball Spin Frequency): The frequency at which a rolling element spins on its own axis.
• FTF (Fundamental Train Frequency): The rotational frequency of the bearing cage. This is a very low frequency, typically less than 0.5X the running speed.

These frequencies can be calculated based on the bearing’s dimensions (such as pitch diameter and ball diameter) and the shaft’s rotational speed. Vibration analysis software typically includes a large database of bearings and can calculate these frequencies automatically.

3. How Bearing Defects Appear in the Spectrum

When a bearing defect is present, its fault frequency will appear in the FFT spectrum in a characteristic pattern:

• High-Frequency Peaks: The fault frequency itself (e.g., BPFO) will appear as a peak in the high-frequency range of the spectrum.
• Harmonics: There will often be several harmonics (multiples) of the fault frequency.
• Sidebands: This is a critical diagnostic marker. The bearing fault frequency peak will often have sidebands spaced at 1X the running speed. For example, a BPFO peak with 1X sidebands is a classic sign of an outer race defect. An inner race defect (BPFI) almost always has 1X sidebands because the defect is rotating in and out of the machine’s load zone.

In the early stages, these peaks may be buried in the noise floor of the spectrum. This is why specialized techniques are often used for detection.

4. Envelope Analysis for Early Detection

Envelope Analysis (also called demodulation) is the most powerful technique for detecting early-stage bearing defects. It is a signal processing method that filters out the low-frequency, high-energy vibration (from unbalance, etc.) and focuses only on the high-frequency, low-energy impacts generated by the bearing flaw.

The envelope spectrum is very “clean” and clearly shows the bearing fault frequencies and their harmonics, allowing for detection months or even years before the bearing would otherwise fail.

← Back to Main Index