In vibration analysis, considerable attention is given to selecting the correct accelerometer, setting the appropriate measurement parameters, and interpreting spectra. However, one of the most important parts of the measurement chain is sometimes treated as an afterthought: how the sensor is mounted to the machine.

At RMT, we experienced multiple situations where the quality of data is directly affected by how the sensor is mounted to the machine.

A high-quality accelerometer connected to an advanced data-acquisition system can still produce poor or misleading data if the mechanical connection between the sensor and the machine is inadequate.

The basic principle is simple:

The accelerometer can only measure the vibration that is successfully transmitted from the machine casing into the sensor.

Anything between the vibrating surface and the sensing element—including mounting studs, magnets, adhesive, paint, dirt, uneven surfaces, mounting pads, and even poor installation technique—becomes part of the measurement system.

For rotating machinery diagnostics, sensor mounting can influence measured amplitude, frequency response, phase, repeatability, and ultimately the diagnosis itself.

The Sensor Is Part of a Mechanical System

An accelerometer mounted on a machine should not be considered an isolated instrument. Once installed, the sensor and its mounting arrangement form a mechanical system with their own stiffness, mass, damping, and natural frequency.

Ideally, the sensor should move exactly with the machine surface.

If the machine casing accelerates by a certain amount, the accelerometer body should experience essentially the same acceleration.

In practice, the mounting interface has finite stiffness. The combination of:

• sensor mass,
• mounting stiffness,
• contact area,
• mounting material,
• mounting geometry, and
• damping

creates a mechanical resonance.

Below this mounting resonance, vibration transmission from the machine to the sensor is generally good. As the measurement frequency approaches the mounting resonance, however, the measured vibration can become amplified or distorted. Above this region, the measured amplitude may decrease significantly and no longer represent the true machine vibration.

This is why the same accelerometer can have very different usable frequency ranges depending on how it is mounted.

Why Mounting Matters in Rotating Equipment

Many conventional rotating-equipment faults occur at relatively low frequencies.

Examples include:

• unbalance at 1× running speed,
• misalignment at 1× and 2× running speed,
• looseness,
• resonance,
• oil whirl,
• structural vibration, and
• many shaft-related phenomena.

For these measurements, several mounting methods may provide acceptable results.

The situation becomes much more critical when diagnosing high-frequency faults such as:

• rolling-element bearing defects,
• gear-mesh problems,
• lubrication problems,
• impacting,
• cavitation,
• electrical discharge damage, and
• early-stage bearing deterioration using acceleration enveloping or demodulation techniques.

These signals can contain significant vibration energy in the kilohertz range. A poor mounting arrangement can attenuate, amplify, or completely distort these high-frequency components.

As a result, sensor mounting can determine whether an early bearing fault is detected or missed.

1. Stud Mounting

For permanent installations and high-quality diagnostic measurements, stud mounting is generally the preferred method.

The accelerometer is mechanically attached to the machine using a threaded connection, normally through a prepared mounting surface or mounting pad.

A properly installed stud provides:

• high mounting stiffness,
• excellent mechanical coupling,
• good repeatability,
• wide usable frequency response, and
• reliable phase measurements.

Because the mechanical connection is stiff, the mounting resonance is generally much higher than with less rigid mounting methods.

Surface Preparation

The mounting surface should be:

• flat,
• clean,
• smooth,
• free from rust,
• free from paint at the contact surface when required, and
• large enough to fully support the accelerometer base.

Poor surface contact reduces mounting stiffness.

For example, mounting an accelerometer directly onto a rough casting may create only a few small contact points between the sensor and machine. The mechanical connection can then behave differently from a properly machined mounting location.

Mounting Torque

Correct tightening is also important.

An under-tightened sensor may have insufficient stiffness and poor repeatability. Excessive torque can damage the accelerometer, stud, mounting thread, or machine surface.

Manufacturer-recommended mounting torque should therefore be followed whenever available.

2. Adhesive Mounting

Adhesives are widely used where drilling and tapping the machine casing is undesirable.

Common arrangements include:

• accelerometer directly bonded to the machine,
• adhesive-mounted threaded pads, and
• epoxy-mounted sensor bases.

A well-installed adhesive mounting pad can provide very good results.

However, the adhesive layer introduces another material into the vibration transmission path.

Its effect depends on:

• adhesive stiffness,
• adhesive thickness,
• contact area,
• curing quality, and
• surface preparation.

A thin, stiff adhesive layer generally provides better high-frequency transmission than a thick, flexible layer.

Too much adhesive can effectively create a compliant cushion between the sensor and the machine.

Surface preparation is therefore critical. Oil, dust, corrosion, loose paint, or contamination can reduce both bond strength and vibration transmission.

3. Magnetic Mounting

Magnetic bases are extremely convenient for route-based condition monitoring.

They allow a vibration technician to quickly take measurements from multiple machines without installing permanent sensors.

For many routine measurements, magnetic mounting can provide excellent results.

However, magnets typically have lower mounting stiffness than a properly installed stud.

The quality of magnetic mounting is strongly influenced by:

• magnet type,
• magnet strength,
• sensor mass,
• surface curvature,
• paint thickness,
• surface roughness,
• corrosion,
• debris, and
• the amount of contact area.

A magnet mounted on a flat, clean steel surface will normally perform significantly better than the same magnet installed over thick paint on a rough or curved bearing housing.

Rocking Motion

One particular concern with magnetic mounts is rocking.

If the magnetic base does not sit firmly and evenly against the machine, the sensor can experience small rocking motions. This can introduce additional resonances and distort high-frequency measurements.

A two-pole magnet on a curved surface, for example, may behave very differently from a flat magnet on a machined pad.

4. Handheld Probe Measurements

Handheld vibration probes provide speed and convenience, but they usually offer the poorest repeatability.

The measurement depends strongly on the operator.

Variables include:

• probe pressure,
• probe angle,
• exact measurement location,
• hand stability, and
• probe-tip contact.

Even small changes can alter the frequency response.

Handheld probes may be suitable for:

• quick screening,
• low-frequency vibration checks,
• locating a vibration source, or
• measurements where installing a sensor is impractical.

They are less suitable when accurate trending or high-frequency fault detection is required.

If a vibration trend changes by 20%, the analyst should be confident that the machine changed—not simply the operator’s probe pressure.

5. Surface Condition Can Be as Important as Mounting Type

Consider two magnetic measurements.

Measurement A

The magnet is installed on:

• a machined steel pad,
• with no paint,
• on a flat surface,
• with full magnetic contact.

Measurement B

The same accelerometer and magnet are installed on:

• a rough casting,
• covered with several layers of paint,
• close to a curved edge,
• with only partial magnetic contact.

Although both records may be labelled “magnetic accelerometer measurement,” their effective frequency responses can be substantially different.

For repeatable condition monitoring, dedicated measurement pads can therefore be extremely valuable.

A simple flat steel target can improve:

• positioning,
• coupling,
• repeatability, and
• data quality.

6. Sensor Location and Mounting Are Closely Related

A perfectly mounted sensor in the wrong location can still produce poor diagnostic information.

Vibration travels through the machine structure, and different structural paths have different stiffness and transmissibility.

For rolling-element bearing monitoring, the accelerometer should normally be located as close as practical to the bearing load zone with a short, stiff mechanical transmission path.

Avoid mounting across unnecessary interfaces such as:

• thin guards,
• inspection covers,
• loose bearing covers,
• sheet metal,
• flexible brackets,
• cooling fins, and
• unsupported structures.

For example, measuring vibration on a motor fan cover is not equivalent to measuring directly on the bearing housing.

The fan cover may have its own resonances and can significantly amplify certain frequencies.

The vibration spectrum may therefore describe the behaviour of the cover rather than the bearing.

7. Measurement Direction Matters

Sensor orientation is another important part of mounting.

Typical rotating-equipment measurements are taken in:

• horizontal,
• vertical, and
• axial directions.

Different faults may appear more strongly in different directions.

For example, axial vibration can be particularly useful when investigating certain alignment or thrust-related problems, while horizontal and vertical measurements often provide important information about radial machine behaviour.

If the sensor orientation changes significantly between measurement rounds, the trend may no longer represent an equivalent measurement.

This is especially important when using single-axis accelerometers.

Even with permanently mounted sensors, orientation should be clearly defined during installation.

8. Cable Installation Can Affect Measurements

Sensor mounting does not end at the accelerometer body.

The cable can also influence measurement quality.

Common cable-related problems include:

• loose cables striking the machine,
• excessive cable movement,
• poor connector installation,
• cable strain on the accelerometer,
• electrical noise pickup, and
• triboelectric or motion-related cable noise in certain sensor systems.

The cable should normally be secured close enough to the sensor to prevent excessive movement while avoiding significant mechanical loading of the sensor.

This becomes especially important when measuring low-level vibration or high-frequency signals.

9. Sensor Mass Loading

Accelerometers are not massless.

When attached to a structure, the sensor adds mass. On large bearing housings this effect is generally negligible.

On lightweight components, however, sensor mass can alter the vibration characteristics of the structure being measured.

This is known as mass loading.

The effect can become important when measuring:

• thin panels,
• small pipes,
• lightweight brackets,
• small motors,
• cooling fins,
• sheet-metal structures, or
• components with high-frequency structural resonances.

In such applications, a smaller and lighter accelerometer may provide more representative data.

10. Mounting Can Change the Spectrum

One dangerous aspect of poor mounting is that the resulting spectrum does not necessarily look obviously wrong.

The analyst may still see peaks and broadband vibration.

The problem is that their amplitudes may no longer accurately represent the machine.

A mounting resonance can create an apparent increase around a particular frequency range.

Poor coupling can also attenuate frequencies above a certain point.

Consequently, two sensors measuring the same machine can produce noticeably different spectra simply because they were mounted differently.

This is especially significant when comparing:

• historical data,
• measurements from different technicians,
• permanent versus portable sensors,
• online versus route-based monitoring systems, or
• readings from different sensor types.

Before concluding that the machine condition has changed, confirm that the measurement conditions have not changed.

11. The Effect on Bearing Condition Monitoring

Bearing diagnostics provide one of the clearest examples of why sensor mounting matters.

An early bearing defect often produces short-duration impacts.

These impacts excite structural resonances in the bearing housing and sensor mounting system. Enveloping techniques use this high-frequency response to recover the repeating impact pattern associated with the bearing defect.

If the sensor is poorly coupled to the housing, the high-frequency impact energy may be severely attenuated.

The consequence can be:

Poor mounting → weak impact transmission → low envelope amplitude → missed bearing fault.

The opposite problem is also possible.

A mounting resonance can artificially amplify vibration over a particular frequency range, potentially creating an apparent increase in condition severity.

For this reason, consistent sensor installation is essential when trending acceleration enveloping measurements.

12. Repeatability Is Often More Important Than Absolute Accuracy

Condition monitoring is heavily dependent on trending.

Suppose a bearing vibration level changes as follows:

Measurement Round	Overall Vibration
Month 1	2.1 mm/s
Month 2	2.2 mm/s
Month 3	2.3 mm/s
Month 4	4.1 mm/s

Such a trend would normally attract attention.

But imagine that during Month 4 the technician moved the sensor from a prepared measurement pad to a nearby painted casting.

Is the machine deteriorating, or has the measurement path changed?

Without measurement consistency, it can be difficult to know.

Reliable trending requires consistency in:

• sensor type,
• sensor location,
• orientation,
• mounting method,
• data acquisition settings, and
• machine operating condition.

A slightly less sophisticated measurement taken consistently is often more useful for trending than a theoretically superior measurement taken differently every time.

13. Comparing Common Mounting Methods

The following general comparison illustrates the relative characteristics of common accelerometer mounting arrangements.

Mounting Method	Repeatability	High-Frequency Response	Installation Convenience	Typical Application
Stud mounted	Excellent	Excellent	Low	Permanent monitoring and detailed diagnostics
Adhesive pad	Very good	Very good if properly installed	Medium	Permanent/semi-permanent monitoring
Magnetic base	Good	Good to moderate depending on installation	Excellent	Portable route measurements
Handheld probe	Moderate to poor	Limited	Excellent	Screening and quick checks

These classifications are general rather than absolute. Actual performance depends on the accelerometer, mounting hardware, surface preparation, machine structure, and frequency range of interest.

14. Practical Example

Consider a motor operating at 1,500 rpm.

Its rotational frequency is:

1,500 / 60 = 25 Hz

Suppose the machine has an unbalance problem producing strong vibration at 25 Hz.

A magnetic-mounted accelerometer and a stud-mounted accelerometer may both measure this component reasonably well because 25 Hz is relatively low.

Now assume the same motor develops an early rolling-element bearing defect whose impacts excite a structural resonance around several kilohertz.

The difference between the two mounting arrangements may become much more significant.

The stud-mounted sensor may transmit the high-frequency impacts clearly, while a poorly installed magnet on a painted or uneven surface may attenuate them.

The analyst could therefore reach two completely different conclusions about the bearing condition despite measuring the same machine.

15. Good Practice for Vibration Sensor Installation

For reliable rotating-equipment vibration measurements:

1. Use stud mounting whenever practical for permanent sensors and high-frequency diagnostics.
2. Prepare the mounting surface correctly.
   Ensure the mounting location is clean, flat, and mechanically sound.
3. Avoid thick paint or soft material between the sensor and machine.
4. Use dedicated measurement pads for portable monitoring routes where possible.
5. Maintain the same sensor location and orientation during every measurement round.
6. Mount as close as practical to the bearing being monitored.
7. Avoid flexible covers, guards, thin plates, and unsupported brackets.
8. Follow the sensor manufacturer’s recommended installation and torque requirements.
9. Secure sensor cables appropriately and avoid cable strain.
10. Consider sensor mass when measuring lightweight structures.
11. Treat unexpected spectral changes cautiously if the mounting arrangement has changed.
12. For high-frequency bearing or gear diagnostics, pay particular attention to mounting stiffness.

A Useful Diagnostic Question

Whenever a vibration spectrum looks unusual, one of the first questions should be:

“Am I seeing a change in the machine, or a change in the measurement?”

Before diagnosing a new fault, check:

• Was the sensor placed at exactly the same location?
• Was the same mounting method used?
• Is the magnetic base sitting correctly?
• Is there paint, rust, grease, or debris under the sensor?
• Has the mounting pad become loose?
• Is the sensor orientation correct?
• Is the cable secure?
• Has another sensor with different characteristics been used?

These simple checks can prevent significant diagnostic errors.

Conclusion

Sensor mounting is not merely an installation detail. It is part of the vibration measurement chain.

The mounting arrangement determines how effectively vibration is transmitted from the machine into the accelerometer and can therefore influence amplitude, frequency response, phase, repeatability, and diagnostic sensitivity.

At low frequencies, differences between mounting methods may appear relatively small. At higher frequencies—particularly during bearing, gear, impact, and envelope analysis—the mounting arrangement can become critical.

For condition monitoring programs, consistency is essential. Sensor location, orientation, surface condition, mounting technique, and acquisition settings should all be controlled so that changes in the vibration data represent changes in the machine rather than changes in the measurement method.

A vibration analyst should never evaluate the spectrum without considering the measurement chain that produced it.

Ultimately:

Good vibration analysis starts before the data reaches the analyzer. It starts at the sensor-machine interface.