In one factory, we identified a distinctive vibration signature in a servo motor.
The vibration analysis pointed to a specific bearing-related condition. The case was investigated, documented, and the findings became part of our technical knowledge.
Some time later, at another production plant, we encountered something remarkably similar.
The same motor model.
The same bearing location.
A very similar spectral signature.
At that point, the second investigation was no longer starting from zero.
The first failure had become a reference for the second.
This is where condition monitoring becomes more than fault detection. It becomes a process for building reusable diagnostic knowledge.
1.-Industrial servo motor investigated through vibration analysis
Servo Motor Vibration Analysis: From an Isolated Failure to a Recurring Pattern
In the first factory, vibration measurements revealed a characteristic spectral pattern associated with the servo motor.
The frequency spectrum indicated that the anomaly was not simply an increase in overall vibration. Specific frequency content suggested that a more detailed investigation was required.
That first case established a reference.
Not just a spectrum, but a combination of the motor configuration, bearing location, operating conditions, vibration behaviour, diagnostic observations, and findings from the subsequent investigation.
When a second motor at another plant produced a remarkably similar vibration signature, that information became much more valuable.
The comparison between the two cases suggested that we were no longer looking only at two individual machines.
A recurring pattern was emerging.
Factory A: Pattern identified
Factory B: Pattern verified
The similarity between the two spectra provided a better starting point for the second investigation.
But similarity alone is not enough.
A recurring vibration
The second investigation did not start from zero
This is perhaps the most important part of the case.
When the vibration signature appeared for the second time, the previous investigation allowed us to ask a more focused question.
Instead of starting with:
What is happening inside this motor?
we could ask:
Are we seeing the same failure mechanism that we observed previously?
That changes the diagnostic process.
Previous knowledge can help focus the investigation on the most relevant frequencies, components and mechanical locations, while the new case still has to be independently verified.
The objective is not to assume that two similar spectra automatically have the same root cause.
The objective is to use previous evidence to build a stronger diagnostic hypothesis and then test it.
Same motor. Same bearing location. Similar spectral signature.
This was what made the comparison particularly interesting.
The measurements came from different production plants and were taken at different points in time, yet the vibration behaviour showed a recognizable similarity.
A single case can tell us a great deal about one motor.
A recurring signature across comparable motors can potentially give us something even more useful:
a diagnostic reference for future investigations.
When the same behaviour appears again, the previous case provides context.
Which frequencies were relevant?
Which component was involved?
What was found during inspection?
Did the physical evidence support the vibration diagnosis?
These connections transform historical measurements into engineering knowledge.
From vibration signature to physical evidence
Vibration analysis becomes particularly valuable when the signal can be connected to the physical condition of the machine.
Once a suspected mechanical condition has been identified, inspection after disassembly provides an opportunity to compare the vibration diagnosis with the actual condition of the component.
In this investigation, the bearing was examined after disassembly, allowing the spectral observations to be compared with physical findings.
This step is fundamental.
An FFT peak can indicate where to investigate.
A recurring spectral pattern can strengthen what we suspect.
Physical inspection can provide the evidence needed to determine whether that hypothesis is consistent with the actual component condition.
The diagnostic sequence therefore becomes:
Detection → Comparison → Hypothesis → Inspection → Evidence → Knowledge
And once that knowledge has been properly documented, it can be reused.
Knowledge Reused. Knowledge Extended.
This is where the value moves beyond repairing an individual servo motor.
The first investigation generated knowledge.
The second investigation reused that knowledge.
And the findings from the second case allowed that knowledge to be extended.
The process becomes a continuous diagnostic loop:
Case 1 → Learn → Document → Case 2 → Compare → Verify → Learn More
Every investigation can therefore improve the next one.
Over time, recurring vibration signatures can become valuable diagnostic references for specific motor models, bearing positions and operating conditions.
These references should never replace vibration analysis or engineering judgement.
They should make both more informed.
Why documenting recurring patterns matters
In industrial maintenance, valuable knowledge is often generated during individual interventions but never fully retained.
A bearing is replaced.
The motor returns to operation.
The problem disappears.
The intervention may be successful, but an important opportunity can be lost.
If the vibration data, operating conditions, diagnosis and physical findings are not connected and documented, much of the engineering knowledge generated during the investigation disappears with the repair.
A properly documented case is different.
Months later, potentially at another production facility, that information can become the starting point for another investigation.
The question then changes from:
How do we diagnose this failure?
to:
Have we seen this behaviour before?
That is an important shift.
Because the objective is no longer only to solve individual failures.
It is to learn from them.
From predictive maintenance to knowledge-based diagnostics
Predictive maintenance is generally associated with identifying developing machine conditions before functional failure occurs.
But detecting a problem is only one part of the value that condition monitoring can generate.
The information obtained from one machine can also improve how the next machine is investigated.
When recurring vibration signatures are systematically documented and connected with inspection findings, historical vibration data becomes more than a collection of spectra.
It begins to form a diagnostic knowledge base.
This can help engineering and maintenance teams:
- recognize previously observed vibration behaviour;
- compare equivalent motor configurations;
- focus investigations on relevant components and frequencies;
- connect spectral information with previous physical findings;
- identify recurring behaviour across different machines or production sites;
- and progressively improve the understanding of recurrent failure mechanisms.
A stored spectrum alone is not knowledge.
The value comes from connecting the signal with its context, diagnosis, intervention and physical evidence.
The detection was not the conclusion. It was the starting point.
Detecting a bearing-related vibration signature is valuable.
Understanding what produced it is more valuable.
Recognizing a similar pattern in another machine takes that knowledge one step further.
But the greatest value comes when everything learned from those investigations improves the next one.
That is when individual failure analysis starts becoming reusable engineering knowledge.
For us, this is probably the most important lesson from this case:
The next investigation should not start where the previous one started.
It should start where the previous one ended.
Every motor analysed should leave behind more than a repaired machine.
It should leave behind knowledge.
