Glossary term

Mechanical Looseness

Vibration fault condition caused by insufficient preload, loose supports, worn fits, cracked foundations or nonlinear contact that changes machine stiffness during operation.

Branch: Mechanical Engineering
Glossary type: phenomenon
Content: Glossary term
Updated: Jun 25, 2026
Revision: v1.0.0 · reviewed

Definition

phenomenon

Mechanical looseness is a vibration condition where insufficient preload, worn fits, cracked supports, loose fasteners or intermittent contact allow a machine to move with nonlinear stiffness.

Mechanical looseness occurs when a load path that should remain clamped, seated or constrained begins to open, slip, impact or change stiffness during operation. It can involve loose anchor bolts, soft foot, cracked bases, worn bearing housings, fretted fits, loose guards, damaged mounts, coupling looseness or foundation defects. The vibration evidence often includes harmonics, unstable phase, broadband energy, amplitude-dependent stiffness and poor repeatability.

Mechanical looseness is a vibration condition caused by a load path that no longer behaves like a stable, clamped or seated connection. A machine may have loose anchor bolts, soft foot, a cracked base, worn bearing fits, fretted mounting surfaces, loose guards, degraded mounts or a support that opens and closes during each cycle.

The key feature is nonlinearity. A tight structure has approximately repeatable stiffness for small vibration. A loose structure can change stiffness with direction, load, displacement, temperature or speed. That is why looseness often creates harmonics, unstable phase and poor repeatability instead of a clean single-frequency response.

Common Sources

Source	Mechanism	Typical evidence
loose anchor bolt	base slips or impacts under dynamic load	harmonics, fretting dust, changing phase
soft foot	frame twists when bolted down	alignment repeatability problems
worn bearing fit	housing or outer race moves under load	temperature, noise and vibration changes
cracked support	local stiffness drops under load	shifted natural frequency and local strain
loose guard or bracket	lightweight part rattles near machine	high-frequency broadband vibration
degraded isolator	support stiffness and damping change	transmissibility and base-motion changes

Looseness is not a single component failure. It is a boundary-condition failure: a connection that should constrain motion no longer does so consistently.

Preload Screen

A simple preload margin can be written as:

\displaystyle \eta_F=\frac{F_p}{F_{req}}

where $F_p$ is available clamp preload and $F_{req}$ is the preload required to prevent slip, opening or loss of seating under the operating load.

If:

\eta_F<1

then the joint does not have enough clamp margin in the simplified screen.

Example: a motor base requires estimated clamp preload:

F_{req}=42\ \text{kN}

A torque audit and bolt condition review estimate remaining preload:

F_p=31\ \text{kN}

The preload margin is:

\displaystyle \eta_F=\frac{31}{42}=0.74

That does not prove the exact vibration mechanism, but it supports a looseness hypothesis when vibration evidence also shows harmonics, phase instability or fretting marks.

Stiffness-Loss Screen

Looseness can reduce effective stiffness. For a simplified single-degree-of-freedom screen:

\displaystyle f_n=\frac{1}{2\pi}\sqrt{\frac{k_{eff}}{m}}

If the modal mass is unchanged, the stiffness ratio can be estimated from frequency shift:

\displaystyle \frac{k_2}{k_1}=\left(\frac{f_2}{f_1}\right)^2

Suppose a base mode measured during commissioning was:

f_1=34\ \text{Hz}

After months of service, the same mode is measured near:

f_2=29\ \text{Hz}

The effective stiffness ratio is:

\displaystyle \frac{k_2}{k_1}=\left(\frac{29}{34}\right)^2=0.728

That implies about $27.2\%$ loss of effective stiffness in this simplified comparison. The next step is not to tune the spectrum; it is to inspect the load path.

Spectral and Phase Evidence

Mechanical looseness often appears as more than one spectral line. A useful harmonic ratio is:

\displaystyle R_{2x/1x}=\frac{A_{2x}}{A_{1x}}

If a machine has:

A_{1x}=4.0\ \text{mm/s RMS}

and:

A_{2x}=3.2\ \text{mm/s RMS}

then:

\displaystyle R_{2x/1x}=\frac{3.2}{4.0}=0.80

A high 2x ratio does not prove looseness, but it makes a pure unbalance diagnosis weak. The analyst should check alignment, soft foot, base bolts, bearing fits, cracked supports and whether phase is stable between runs.

Distinction from Similar Faults

Mechanical looseness is not unbalance response. Unbalance is a synchronous 1x force from mass eccentricity. Looseness can create 1x, 2x, 3x, broadband energy and changing phase because the boundary condition is nonlinear.

Mechanical looseness is not misalignment. Misalignment is a shaft or coupling geometry problem. It can cause looseness over time, and looseness can make alignment readings repeat poorly, but the corrective action is different.

Mechanical looseness is not rotor rub. Rub is contact between rotating and stationary parts. Looseness can allow rub by increasing motion or shifting position, but the evidence must show contact plausibility before calling it rub.

Validation and Release

A defensible looseness diagnosis uses more than a spectrum. Useful evidence includes bolt torque or tension audit, soft-foot measurement, base flatness, shim condition, fretting dust, cracked paint, witness marks, thermal growth, alignment repeatability, phase stability, operating load, vibration trend and inspection photographs.

Release should be withheld when phase is unstable, harmonics grow rapidly, a support crack is suspected, preload margin is below the required value or vibration changes after a bolt-tightening pass. Release can be considered after the load path is corrected and a repeat run shows stable phase, lower harmonics, acceptable temperature and no new broadband impact evidence.

Inspection Sequence

A practical inspection sequence starts with the safest external checks and moves toward disassembly only when the evidence justifies it:

compare vibration amplitude and phase against the previous baseline;
inspect anchor bolts, shims, grout, base cracks, fretting dust and witness marks;
check soft foot and frame distortion before changing alignment;
repeat the measurement after a controlled bolt-tightening or shim correction;
inspect bearing fits, coupling hubs, guards and brackets if symptoms remain;
preserve before/after spectra, phase, photographs and torque or tension records.

The most useful result is not simply “looseness found.” The report should identify which load path was loose, what evidence showed it, what correction restored stiffness or preload and which repeat measurement proved that the fault did not move somewhere else.

Common Mistakes

Do not balance a machine to hide looseness. A balance correction may reduce 1x vibration while the loose joint continues to fret, crack or impact.

Do not diagnose looseness from 2x alone. Misalignment, shaft bow, ovality, electrical forcing and signal distortion can also create harmonic content.

Do not assume a tight torque wrench reading proves the joint is healthy. Threads, washers, embedment, corrosion, soft foot and surface damage can make torque a poor proxy for real clamp load.

Limits

The preload and stiffness screens are simplified. Real joints have friction scatter, bolt relaxation, thermal expansion, gasket or shim compression, base distortion and load sharing between fasteners. Modal frequency can also shift because operating condition, sensor location, temperature or support boundary changed.

The practical value of the diagnosis is decision quality: stop unnecessary balancing, inspect the load path, correct the boundary condition and verify the result with repeatable vibration and physical evidence.

REF

Disciplines