Vibration Analysis Basics: How to Read Machine Vibration (2026 Guide)

Vibration analysis is the process of measuring, recording, and interpreting the oscillatory motion of a machine component to determine its mechanical condition. By examining how a machine vibrates — how fast, how hard, and at what frequencies — engineers can detect developing faults such as bearing wear, rotor imbalance, shaft misalignment, and gear damage weeks or months before they cause an unplanned shutdown.

Every rotating machine produces a characteristic vibration signature. When that signature changes, something inside the machine is changing too. Vibration analysis makes those changes visible, turning raw sensor data into actionable maintenance intelligence.

This guide covers the complete vibration analysis basics: the physical quantities involved, time-domain and frequency-domain analysis, common fault signatures, severity standards, and — the piece most instrument-vendor guides skip — how vibration sensors wire into a PLC and how the controller uses that data to trigger alarms and maintenance work orders.

What Is Vibration Analysis?

Vibration analysis is a core technique within condition monitoring vs predictive maintenance programs. It works by attaching one or more accelerometers to a machine housing, collecting the sensor signal over time, and processing that signal to extract information about specific mechanical faults.

The fundamental principle is that mechanical faults generate forces at predictable frequencies. A rotor with residual imbalance creates a force at exactly one times the running speed. A rolling-element bearing with a damaged outer race creates a force at a frequency determined by its geometry. If you know what to look for, a vibration spectrum becomes a highly specific diagnostic tool.

Vibration analysis is used across virtually every rotating machine type: electric motors, pumps, fans, compressors, gearboxes, conveyor drives, and turbines. In modern industrial facilities it is one of the primary data streams feeding a PLC predictive maintenance program.

---

The Three Physical Quantities: Amplitude, Frequency, and Phase

Before reading a vibration spectrum you need to understand what is being measured and how the three fundamental quantities relate to each other.

Amplitude

Amplitude is how much the machine is vibrating — the magnitude of the oscillatory motion. Depending on which physical quantity you measure, amplitude is expressed in three ways:

Quantity	Unit	Typical Use Case
Acceleration	g or mm/s²	High-frequency faults: bearings, gear mesh
Velocity	mm/s (RMS)	Mid-frequency faults: imbalance, misalignment (ISO 10816 standard)
Displacement	µm (peak-peak)	Low-frequency, large machines: journal bearings, slow shafts

The relationship between the three is purely mathematical: integrate acceleration once to get velocity, integrate velocity once to get displacement. The choice of which to use depends on the frequency range of interest. For a detailed breakdown of when to use each, see acceleration vs velocity.

As a practical starting point, velocity (mm/s RMS) is the most widely used quantity for general machinery health because it is proportional to the energy of vibration across the frequency range where most common faults appear (10–1000 Hz).

Frequency

Frequency tells you how many times per second the machine completes one oscillation cycle. It is measured in hertz (Hz) or, in rotating machinery, often expressed relative to the shaft running speed using the notation 1X, 2X, 3X (where 1X = one cycle per revolution = running speed in Hz).

Frequency is the most diagnostically powerful parameter in vibration analysis because mechanical faults generate vibration at specific, calculable frequencies. Once you know the machine's running speed and bearing geometry, you can predict exactly where fault energy will appear in the spectrum.

Phase

Phase is the timing relationship between the vibration signal and a reference point on the shaft (usually measured with a once-per-revolution tachometer pulse). Phase analysis is particularly useful for distinguishing imbalance from misalignment and for balancing corrections. In automated PLC-based monitoring, phase is less commonly used than amplitude and frequency because it requires a synchronous tachometer signal; overall amplitude and spectral alarming cover the majority of practical applications.

---

Time Domain vs. Frequency Domain (FFT)

Vibration sensors produce a raw time-domain signal: voltage (proportional to acceleration) vs. time. The time-domain waveform shows you that something is happening, but it is difficult to identify what is happening from the raw waveform alone.

The Time Domain

The time-domain signal is useful for identifying:

• Impacting faults — bearing spalls or gear tooth damage produce sharp, periodic spikes that are clearly visible in the waveform.
• Overall vibration level — peak or RMS amplitude from the raw waveform gives a quick health check.
• Modulation patterns — amplitude modulation in the waveform indicates a fault frequency is being modulated by another frequency (common in bearing and gear faults).

The Frequency Domain — FFT

The Fast Fourier Transform (FFT) converts the time-domain waveform into a frequency spectrum: amplitude vs. frequency. The FFT decomposes the complex vibration signal into its individual sinusoidal components, each at a specific frequency and amplitude.

The result is a plot where each vertical peak (called a spectral line) represents a discrete vibration source. Reading a spectrum means identifying which spectral lines are present, what their frequencies are relative to running speed and bearing geometry, and whether their amplitudes are growing over time.

Key FFT parameters that affect spectrum quality:

• Lines of resolution (LOR): more lines = finer frequency resolution. 800–3200 lines covers most rotating machinery.
• Frequency range (Fmax): set at least 3–5× the highest fault frequency of interest. For bearing faults on a motor running at 50 Hz, an Fmax of 1000 Hz is typical.
• Averaging: 4–8 linear averages reduce noise and stabilize the spectrum.
• Window function: use a Hanning window for most general-purpose machinery analysis to reduce spectral leakage.

---

Acceleration, Velocity, and Displacement: A Brief Comparison

The choice of measurement parameter affects which faults you can see:

• Displacement emphasizes low frequencies. It is specified in API 670 (machinery protection) for large turbomachinery with fluid-film (journal) bearings, where shaft position relative to the bearing clearance is the critical parameter.
• Velocity provides a relatively flat response across 10–1000 Hz, making it the preferred parameter for general machinery. ISO 10816 and ISO 20816 severity limits are expressed in velocity (mm/s RMS).
• Acceleration emphasizes high frequencies. It is the best choice for detecting early-stage rolling-element bearing faults and gear mesh problems above 1000 Hz, because those faults generate high-frequency energy that is attenuated in the velocity and displacement parameters.

For a deeper treatment of when each parameter applies to specific machine types, see acceleration vs velocity.

---

Common Machine Fault Signatures

The diagnostic power of vibration analysis comes from the fact that each fault type generates vibration energy at a predictable frequency or set of frequencies.

Imbalance — 1X Running Speed

Imbalance is the most common rotating machinery fault. An uneven mass distribution on the rotor creates a centrifugal force that rotates at exactly the shaft speed, producing a dominant spectral peak at 1X (running speed frequency).

Key characteristics:

• Strong peak at 1X with harmonics at 2X and 3X (usually much smaller).
• Vibration increases with the square of running speed (doubling the speed quadruples the vibration).
• Phase is stable — typically a single-plane or two-plane relationship to the reference mark.
• Correction: rotor balancing in the field or on a balancing machine.

Misalignment — 2X and Axial

Misalignment between coupled shafts (angular or parallel) creates a forcing function at 1X and 2X running speed, with 2X often dominant for angular misalignment. Axial vibration elevated relative to radial is a strong misalignment indicator.

Key characteristics:

• 2X peak at or above the 1X amplitude in the radial direction.
• Elevated axial vibration at 1X and 2X.
• High vibration appears on both sides of the coupling.
• Correction: precision shaft alignment using laser alignment tools.

Rolling-Element Bearing Defects — BPFO, BPFI, BSF, FTF

Rolling-element bearings (ball bearings, roller bearings) have four calculable bearing defect frequencies based on bearing geometry and running speed:

Frequency	Full Name	Fault Location
BPFO	Ball Pass Frequency Outer race	Outer race defect
BPFI	Ball Pass Frequency Inner race	Inner race defect
BSF	Ball Spin Frequency	Rolling element defect
FTF	Fundamental Train Frequency	Cage defect

Bearing manufacturers publish the geometric constants (Bd, Pd, n, contact angle) needed to calculate these frequencies, or they can be found in bearing databases.

Early-stage bearing faults appear as small peaks at the defect frequency surrounded by sidebands separated by running speed. As the fault progresses, the peaks grow, harmonics appear, and eventually broad-band noise (the "noise floor") rises — indicating the bearing is entering the final failure stage.

Envelope analysis (demodulation) is the most sensitive technique for early bearing fault detection. It detects the amplitude modulation of high-frequency bearing noise at the defect frequency, making faults visible before they appear clearly in the standard velocity spectrum.

Gear Mesh Frequency — GMF

Gear faults produce vibration at the gear mesh frequency (GMF), which equals the number of teeth on the gear times the shaft running speed. Sidebands around the GMF at ±1X (or ±gear shaft frequency) indicate gear wear, eccentricity, or tooth damage.

Key characteristics:

• GMF peak with sidebands at running speed multiples.
• High GMF harmonics (2× GMF, 3× GMF) indicate advanced wear.
• Ghost frequencies (not related to current tooth count) may indicate manufacturing errors.

Looseness — Sub-harmonics and Truncation

Mechanical looseness (bearing cap loose, loose rotor on shaft, structural looseness) generates a distinctive pattern of sub-harmonics (0.5X, 1.5X, 2.5X) and a forest of harmonics, often with truncation visible in the time-domain waveform. The presence of 0.5X subharmonic is a strong looseness indicator.

---

Reading a Vibration Spectrum: The 1X/2X/3X Framework

When first examining an FFT spectrum, apply the 1X/2X/3X framework as a rapid triage:

1. Find 1X: Identify the peak at running speed. Its amplitude tells you the overall imbalance level. On most well-maintained machines, 1X is the dominant peak.
2. Check 2X relative to 1X: If 2X amplitude approaches or exceeds 1X, investigate misalignment or looseness.
3. Check for harmonics (3X, 4X, 5X…): A series of running-speed harmonics with decreasing amplitude indicates mechanical looseness.
4. Calculate bearing defect frequencies: Mark BPFO, BPFI, BSF, and FTF on the spectrum. Any peaks at those frequencies warrant attention.
5. Check high-frequency region: Peaks above 500 Hz not explained by GMF or bearing defect frequencies may indicate cavitation, electrical noise, or resonance.

A spectrum with only a moderate 1X peak and nothing else of significance indicates a healthy machine. A spectrum with elevated 2X, a rising noise floor, or confirmed bearing defect frequency peaks indicates a machine requiring attention.

---

Types of Vibration Analysis

Vibration analysis methods range from simple overall-level checks to sophisticated signal-processing techniques:

• Overall level monitoring: A single RMS or peak value from the broadband signal. Simple to implement in a PLC analog input. Good for alarming but provides no diagnostic specificity.
• Spectral analysis (FFT): Decomposes the signal into frequency components. The standard diagnostic tool for most applications.
• Envelope analysis (demodulation): Detects early-stage bearing faults by analyzing the high-frequency energy envelope. Requires more signal processing, typically done in a smart sensor or edge device.
• Time-synchronous averaging (TSA): Averages the vibration waveform synchronously with shaft rotation, removing non-synchronous noise. Particularly effective for gear fault detection.
• Orbit analysis: Uses two orthogonal displacement probes to plot shaft centerline motion. Standard for large turbomachinery with fluid-film bearings.
• Modal analysis: Measures structural resonances to separate forced vibration from resonant amplification.
• ODS (Operating Deflection Shape): Visualizes how a machine structure deforms during normal operation, useful for identifying structural looseness or resonance.

For most industrial PLC-based monitoring applications, overall level monitoring and FFT spectral analysis provide the best cost-benefit ratio.

---

Severity Standards: ISO 10816 and ISO 20816

Without a severity scale, a vibration measurement has no context. The ISO 10816 series (now being superseded by ISO 20816) provides internationally recognized severity limits expressed in velocity (mm/s RMS) for different machine classes.

ISO 10816-3 Severity Zones (Industrial Machines 15 kW – 15 MW)

Zone	Velocity (mm/s RMS)	Condition
A	0 – 2.3	New machine, excellent
B	2.3 – 4.5	Acceptable for long-term continuous operation
C	4.5 – 7.1	Tolerable short-term; investigate and schedule maintenance
D	> 7.1	Dangerous; risk of damage; take immediate action

These limits apply to the bearing housing vibration of machines in the 15 kW–15 MW range. ISO 10816-1 provides overarching guidance; individual parts (ISO 10816-3 for industrial machines, ISO 10816-7 for rotodynamic pumps) give machine-specific limits.

ISO 20816 maintains the same measurement approach and similar severity zones but updates and expands the covered machine types. For new monitoring system designs, use ISO 20816 as the reference standard.

Practical note: ISO limits are a baseline, not a target. Many facilities set internal alarm thresholds below the ISO zone boundaries based on their specific machine population and maintenance strategy. A machine that has historically run at 0.8 mm/s and suddenly rises to 2.0 mm/s warrants investigation even though it remains in zone A, because the trend is significant.

---

How Vibration Sensors Wire into a PLC

This is where most instrument-vendor guides stop. Understanding the control-layer integration is what separates a standalone vibration measurement from an automated condition monitoring system.

Sensor Types and Output Signals

IEPE accelerometers (also called ICP sensors) are the most common vibration transducer for industrial applications. They output a voltage proportional to acceleration, biased on a constant-current supply (typically 2–20 mA at 24–27 V). Key specifications for PLC integration:

• Sensitivity: typically 100 mV/g for general-purpose machines; 10 mV/g for high-vibration environments.
• Frequency range: 0.5–10,000 Hz covers most rotating machinery applications.
• Output: raw AC voltage signal requiring signal conditioning before connection to a PLC analog input.

4–20 mA vibration transmitters include the signal conditioning circuit inside the housing and output a 4–20 mA loop-powered signal proportional to overall vibration velocity or acceleration RMS. This format connects directly to a standard PLC analog input module (AI) without additional conditioning hardware.

• 4 mA = zero vibration (or a defined minimum)
• 20 mA = full-scale vibration (user-configurable, typically 25–50 mm/s for velocity transmitters)
• Loop-powered: draws power from the PLC AI module's internal supply

For a broader view of how different sensor types connect to PLC inputs, the types of industrial sensors guide covers wiring conventions and signal types across the full sensor spectrum.

IO-Link Vibration Sensors

IO-Link vibration sensors represent a significant advancement for PLC-integrated condition monitoring. Rather than a single analog value, an IO-Link vibration sensor streams structured process data — including RMS velocity, peak acceleration, temperature, and crest factor — directly to an IO-Link master port.

Key advantages for PLC integration:

• Multi-parameter output: a single sensor connection provides overall vibration level, bearing condition indicator, and temperature simultaneously — data that would require multiple transmitters with traditional 4–20 mA wiring.
• Parameterization from the PLC: alarm thresholds, measurement ranges, and averaging parameters can be written to the sensor over the IO-Link connection without physical access to the sensor.
• Event-based alarming: the sensor can send a process alarm flag directly to the PLC when a threshold is exceeded, enabling the controller to respond without waiting for the next polling cycle.
• Plug-and-play replacement: the IO-Link master stores the sensor configuration (IODD parameter record), so a replacement sensor automatically receives the correct settings.

IO-Link sensor data typically maps to the PLC as a structured input data object. The PLC program reads velocity (mm/s), acceleration (g), and temperature in a single instruction cycle, making it straightforward to trend these values and implement multi-parameter alarming.

PLC Analog Input Wiring: Practical Points

When wiring a 4–20 mA vibration transmitter to a PLC analog input:

1. Check the module's input impedance: most PLC analog input modules are suitable for 4–20 mA (250 Ω internal resistance), but confirm against the module datasheet.
2. Use shielded cable: vibration transmitter cables run near motor cables and VFDs, which are significant noise sources. Use shielded twisted-pair and ground the shield at one end only (at the panel).
3. Check the power supply: loop-powered 4–20 mA transmitters draw their operating current from the loop. The PLC AI module must supply 24 VDC to the loop, or an external supply is required.
4. Configure the EU (engineering unit) scaling: in the PLC program, scale the raw analog count (e.g., 0–27648 for Siemens S7, 0–32767 for Allen-Bradley) to mm/s using the transmitter's configured range. A 0–25 mm/s transmitter with a 4–20 mA output maps 4 mA → 0 mm/s and 20 mA → 25 mm/s.

Scaling Example (Ladder Logic Concept)

For a 4–20 mA transmitter configured for 0–25 mm/s output, connected to a PLC analog input that returns a raw integer 0–27648:

Raw count 0      → 4 mA  → 0.0 mm/s
Raw count 27648  → 20 mA → 25.0 mm/s

EU Value (mm/s) = (Raw Count / 27648) × 25.0

In practice, most PLC platforms provide a built-in scaling instruction (SCL in Allen-Bradley, SCALE in Siemens, SCALE_X in IEC 61131-3) that performs this conversion automatically from configured range parameters.

---

Pushing Vibration Trends to SCADA and Historian

A PLC that reads a vibration transmitter and triggers a local alarm is useful. A PLC that sends that same data upstream to a SCADA system and historian transforms it into a long-term condition monitoring asset.

The typical data flow:

1. PLC analog input reads the 4–20 mA or IO-Link vibration value every scan (typically 10–100 ms).
2. PLC data block or tag holds the current EU-scaled value and a sampled average (e.g., 1-minute rolling average computed in the PLC).
3. OPC UA server (hosted on the PLC or an edge gateway) exposes the vibration tags to the plant network. For OPC UA configuration details, the IIoT PLC integration guide covers the full protocol stack.
4. SCADA historian (OSIsoft PI, Ignition, or similar) polls the OPC UA server and stores the trend data with a timestamp.
5. Trend analysis tools at the SCADA or cloud layer plot vibration history, calculate rate-of-change, and generate long-term condition reports.

The PLC handles the deterministic control response (trip the machine if vibration exceeds the protection threshold). The historian handles the long-term trend storage and analysis. Each layer does what it does best.

Triggering a Maintenance Work Order from PLC Logic

The combination of multi-day trending data and pre-alert alarm logic allows the PLC to close the loop with the maintenance management system (CMMS):

1. Alert threshold (e.g., ISO zone B→C boundary, 4.5 mm/s): The PLC sets a bit that triggers a SCADA alarm. No production impact, but the event is logged. At this stage a maintenance planner reviews the trend.
2. Warning threshold (e.g., 6.0 mm/s, below zone D): The PLC activates a persistent output (indicator light, SCADA banner) and sends a CMMS integration trigger (via SCADA or MQTT message) that automatically creates a scheduled maintenance work order.
3. Trip threshold (e.g., 7.1 mm/s, ISO zone D): The PLC initiates a controlled machine shutdown sequence to prevent catastrophic damage. The CMMS work order is automatically escalated to urgent/immediate.

This multi-tier approach avoids false alarms from transient vibration events while ensuring that sustained high-vibration conditions result in a real maintenance response without requiring manual monitoring.

---

Frequently Asked Questions

What is vibration analysis?

Vibration analysis is the measurement and interpretation of the oscillatory motion produced by a rotating or reciprocating machine. By examining vibration amplitude, frequency, and phase, engineers determine machine health and identify developing faults — such as imbalance, misalignment, bearing wear, or gear damage — before they cause unplanned downtime. It is the primary diagnostic technique in most industrial condition monitoring programs.

What are the basics of vibration analysis?

The basics of vibration analysis are: (1) measuring the vibration signal using an accelerometer or vibration transmitter; (2) expressing the measurement in the appropriate physical quantity — acceleration (g) for high-frequency bearing and gear faults, velocity (mm/s RMS) for mid-frequency imbalance and misalignment, or displacement (µm) for slow shafts; (3) converting the raw time-domain signal to a frequency spectrum using the FFT; (4) identifying fault signatures by comparing spectral peaks to calculated fault frequencies (1X for imbalance, 2X for misalignment, BPFO/BPFI for bearings); and (5) comparing measured amplitude against severity standards such as ISO 10816/20816 to determine urgency.

What does a vibration spectrum show?

A vibration spectrum (FFT plot) shows vibration amplitude on the Y-axis versus frequency on the X-axis. Each peak in the spectrum represents a vibration source at a specific frequency. By identifying which frequencies carry significant energy — particularly at running speed harmonics (1X, 2X, 3X) and at calculated bearing defect frequencies (BPFO, BPFI, BSF, FTF) — an analyst can determine which specific components are generating elevated vibration and whether their amplitude has reached a level that requires maintenance action.

How do vibration sensors connect to a PLC?

Vibration sensors connect to a PLC in three main ways. First, a 4–20 mA vibration transmitter outputs an overall RMS vibration level as a standard analog current loop signal that connects directly to any PLC analog input module. Second, an IO-Link vibration sensor connects to an IO-Link master port and streams structured data (RMS velocity, peak acceleration, temperature) as a digital process data object. Third, a raw IEPE accelerometer can connect to a dedicated vibration monitoring module (such as Siemens SM 1238 or Allen-Bradley 1756-IRT8) that performs signal conditioning, filtering, and FFT computation on-board, delivering processed results to the PLC data table. The 4–20 mA and IO-Link approaches are most common for PLC-based monitoring because they use standard wiring and no special modules.