Humpday | Grace Technologies Blog

From 35 Possibilities to 1 Root Cause: How Vibration Analysis Works

Written by Alyssa Rice | Mar 4, 2026 7:30:00 PM

How Good Vibration Analysis Eliminates Thirty Faults to Find Five Causes

Managers need to approve expensive repairs based on your vibration analysis, but most can't distinguish good methodology from guesswork. They hear your conclusions but don't understand how you got there.

Here's the framework that changes that: systematic elimination from 35 possible machinery faults to 5 common root causes. When managers understand this process, they ask better questions, trust your recommendations, and make faster decisions.

 

The Framework: How 35 Becomes 5

Professional vibration analysts work from about 35 possible machinery faults. Good analysts show their work. Poor ones just show conclusions.

Here's how the process should work:

Step 1: Identify the Machine Type (35 → ~23 faults)

Before looking at any vibration data, a competent analyst asks what type of equipment they're analyzing.

This is like a mechanic asking if you have front-wheel drive or rear-wheel drive before diagnosing a handling problem; certain issues simply can't exist on certain configurations.

Example machine: A center-hung cooling tower fan with an AC induction motor, direct-coupled drive, and rolling element bearings.

Based solely on knowing this is a direct-coupled fan with rolling bearings, a good analyst immediately eliminates:

  • DC motor electrical problems (AC motor)
  • All gearbox-related faults (direct-coupled, no gearbox)
  • Belt drive problems (no belts)
  • Oil whirl and oil whip (rolling element bearings, not fluid film)
  • Overhung rotor unbalance issues (center-hung configuration)

That's roughly 12 faults eliminated without collecting a single vibration measurement

 

Step 2: Consider the Frequency Grouping (23 → 9 faults)

Now the analyst examines your vibration spectrum and identifies the dominant frequency pattern. This is like identifying that your car only vibrates at highway speeds versus at idle. In vibration analysis, "when" means "at what frequency."

Five main categories:

  • Synchronous (at running speed): Unbalance, eccentricity, bent shaft, some misalignment types
  • Harmonic (exact multiples): Certain misalignment types, looseness, blade pass, gear mesh
  • Sub-synchronous (below running speed): Some looseness, oil whirl, certain gear problems
  • Non-synchronous (not exact multiples): Bearing defects, some electrical issues
  • Modulation (pulsating combinations): Advanced bearing defects, gear tooth problems

In the example: Vibration data shows a dominant peak at 1x RPM (1,785 cycles per minute), the exact running speed. This is synchronous.

This single observation eliminates all sub-synchronous, non-synchronous, and pure harmonic faults.

 

Step 3: Consider the Main Fault Direction (9 → 5 faults)


Different machinery faults have directional preferences, like how a wheel bearing problem shows up on one side of a car, not evenly across the vehicle. Some faults appear primarily in horizontal or vertical measurements (radial). Others show up along the shaft axis (axial). Some have no clear preference.

Proper analysis requires measurements in all three directions at each bearing location.

In the example: The data shows:

  • Vertical: 0.42 inches/second (high)
  • Horizontal: 0.08 inches/second (low)
  • Axial: 0.03 inches/second (minimal)

This is radial dominant, specifically vertical. This pattern eliminates all axial-dominant faults like angular misalignment and certain thrust bearing problems.

You're now down to 5 remaining candidates: unbalance, eccentricity, looseness type A (soft foot), radial misalignment, or natural frequency resonance.

 

Step 4: Run Tests to Confirm (5 → 1 root cause)

With five faults remaining, the analyst runs specific tests to identify which one matches all the evidence, just like how a mechanic might jack up your car and spin the wheel by hand to confirm a bearing issue rather than just listening to it while driving.

  • Phase analysis measures timing relationships between measurement points—distinguishes unbalance from misalignment from looseness
  • Time waveform analysis shows vibration patterns over time, revealing repetitive impacts from looseness or gears
  • Transient analysis tests machine response during startup or coastdown to identify natural frequencies

In the example: Phase analysis shows a 90-degree phase shift between horizontal and vertical at most bearings—normal. However, at bearing 3, the vertical response is exceptionally high with phase instability. Cross-coupling measurements show the signature of structural looseness.

Conclusion: Looseness Type A, most likely a soft foot condition, where one of the motor's mounting feet has improper shimming or a loose bolt.

The Bottom Line

Vibration analysis is systematic elimination from 35 possibilities to 1 root cause. Understanding this framework lets you evaluate any analyst, spot weak methodology, and make confident decisions.

For teams looking to improve visibility into equipment health and catch issues earlier, solutions like GraceSense provide continuous condition monitoring that helps turn data into actionable insights. 

 

To safer, smarter operations,


connect with us