The Gear Had Broken Teeth—Why Did Vibration Remain Normal?

It is generally believed that once a gear suffers tooth breakage, vibration will increase significantly, with acceleration and velocity vibration amplitudes rising accordingly and triggering threshold alarms, thereby drawing the attention of the maintenance team.

But what if a gearbox develops a single or even multiple broken teeth, while the vibration amplitude remains largely unchanged?

Why does it happen? How should we detect the fault?

Case Background: The Gear Was Damaged, Yet Vibration Remained “Normal”

This case comes from a belt conveyor gearbox at a cement plant, where the fourth-shaft (low-speed shaft) gear had multiple broken teeth.

Six vibration measurement points were deployed to continuously monitor the motor and gearbox bearings, with RONDS Remote Diagnostic Center providing 24/7 condition monitoring and analysis.

Based on the equipment structure, operating conditions, and characteristic frequency analysis, the fault was ultimately localized to measurement point 6H at the output end of the gearbox fourth shaft.

Throughout the entire damage progression, both acceleration and velocity vibration trends remained relatively stable, with no obvious abnormal alarm.

If only regular overall vibration values are considered, the equipment could easily be regarded as operating normally.

Why Did the Overall Vibration Remain Stable Even After Tooth Breakage?

One key reason is the low rotational energy of the low-speed shaft.

In this case, the fourth-shaft rotational frequency was only 0.763 Hz.

For low-speed gearboxes, even when localized pitting, cracking, or tooth breakage occurs, the impact energy generated by abnormal gear meshing may be insufficient to cause a noticeable change in the overall vibration level.

Therefore:

A stable overall vibration level does not necessarily mean that the gear is in a healthy condition.

This is a typical blind spot in low-speed gear fault monitoring.

Rather than relying solely on overall acceleration and velocity vibration values, greater attention should be paid to:

• GMF and its harmonics

• The number and energy of sidebands around the GMF and its harmonics

• Time-domain impacts after narrowband filtering around GMF-related frequency bands

• The low-speed shaft rotational frequency and its harmonics in the envelope spectrum

In other words, for this type of low-speed gear fault, “Has the vibration increased?” is not the only question that needs to be answered.

More importantly:

What is happening around the GMF?

From Minor Pitting to Severe Tooth Breakage: Reconstructing the Full Gear Damage Process

Based on continuously collected monitoring data, the entire fault development process was divided into three stages through analysis of the GMF and its harmonics, the number and energy of sidebands, and narrowband-filtered time-domain waveforms.

Stage 1: Early Pitting or Initial Crack (November 8–10, 2021)

• Overall vibration: Acceleration and velocity remained stable.

• Narrowband time domain: After bandpass filtering in the 30–175 Hz range, weak impacts spaced at the fourth-shaft rotational frequency could be observed, although the features were still diffuse.

• Spectrum: The number and energy of sidebands spaced at the fourth-shaft rotational frequency around 1X–3X the third-stage GMF continuously increased.

• Envelope spectrum: The energy at the fourth-shaft rotational frequency and its harmonics began to increase.

At this stage, a single gear tooth may have developed minor pitting or an initial crack. The damaged tooth generated relatively weak and irregular impacts during meshing. The overall energy was insufficient to change the overall vibration level, but the sidebands around the GMF had already begun to reflect load modulation caused by the localized damage.

The overall vibration level remained unchanged, but fault characteristics had already begun to emerge.

Stage 2: Crack Propagation (November 10–12, 2021)

• Overall vibration: Remained stable.

• Narrowband time domain: Periodic impacts occurring at the fourth-shaft rotational frequency became clearly visible.

• Spectrum: The number and energy of sidebands around 1X–3X the third-stage GMF, spaced at the fourth-shaft rotational frequency, increased significantly.

• Envelope spectrum: The energy at the fourth-shaft rotational frequency and its harmonics increased significantly.

As the crack continued to propagate, the periodic impact generated by the damaged tooth each time it entered the meshing zone became stronger, resulting in more pronounced modulation characteristics.

At this stage, although the overall vibration level still did not trigger an alarm, the narrowband time-domain features and detailed spectral characteristics had already shown significant abnormal changes.

The overall vibration still showed no alarm, but the fault characteristics had become significantly more pronounced.

Stage 3: Tooth Fracture and Detachment (November 12–26, 2021)

• Overall vibration: Remained stable, with no triggered threshold alarm.

• Narrowband time domain: The impact characteristics became diffuse again.

• Spectrum: The number of sidebands around 1X–3X the third-stage GMF, spaced at the fourth-shaft rotational frequency, continued to increase, while sideband energy decreased compared with Stage 2.

• Envelope spectrum: The energy at the fourth-shaft rotational frequency and its harmonics decreased compared with Stage 2.

A phenomenon that is easy to overlook appeared at this stage:

The number of sidebands continued to increase, while the sideband energy decreased.

This does not mean that the fault was becoming less severe.

As the damaged tooth finally fractured and detached, the original impact mechanism during gear meshing changed. The missing tooth created a gap during the meshing cycle, resulting in more irregular and lower-energy disturbances. Consequently, the energy of individual sidebands decreased, while the impact characteristics in the narrowband time-domain waveform became less distinct.

Although the original impact mechanism changed after tooth breakage, the periodic disturbance associated with the damaged gear could still modulate the GMF, so the sidebands remained prominent and their number continued to increase.

At this point, the gear damage had become irreversible, yet the overall vibration level still showed no obvious abnormality.

Field Inspection: Diagnosis Confirmed

During subsequent field inspection, multiple broken teeth were found on the output-shaft gear of the gearbox, consistent with the diagnostic analysis.

What Does This Case Tell Us?

(1) Low-speed gear faults cannot be identified by overall vibration alone

The impact energy generated by faults in low-speed gear stages can be relatively low and may not be sufficient to cause a significant change in overall acceleration or velocity vibration amplitude.

Stable overall vibration ≠ Normal Gear Condition

(2) Pay close attention to sidebands around the GMF and its harmonics

Key indicators include:

• Number of sidebands

• Sideband energy

In general:

Sideband number and energy increasing together → Pay attention to continued fault development.

However when:

The number of sidebands continues to increase while their energy decreases significantly → Suspect tooth fracture, detachment, or a change in the damage mechanism.

(3) Narrowband time-domain analysis can reveal low-speed impacts hidden in the original signal

By applying narrowband filtering around GMF-related frequency bands, weak low-speed impacts that are not obvious in the original time-domain waveform can be extracted.

Changes in impact patterns across different stages can further help assess the progression of gear damage.

(4) Combining multiple indicators is essential for reconstructing the complete fault evolution

Any single indicator may lead to an incomplete or misleading interpretation.

However, by combining:

Overall vibration + GMF and harmonics + sideband number and energy + narrowband time domain + envelope spectrum

the complete evolution of a low-speed gear fault—from early-stage damage to final tooth breakage—can be reconstructed more effectively.

This is also the value of detailed spectral and filtered time-domain analysis in low-speed gear fault monitoring.

Conclusion

Gear fault monitoring is not simply about asking:

“Has the vibration amplitude increased?”

For low-speed gears, it is equally important to examine:

• Have the number and energy of sidebands around the GMF and its harmonics changed?

• Are periodic impacts corresponding to the low-speed shaft rotational frequency emerging in the narrowband time-domain waveform?

• Is the modulation energy in the envelope spectrum changing?

When the overall vibration amplitude still appears “normal,” detailed spectral and time-domain features may already be telling us—

The fault is already there—and it is progressing.