Understanding Cross-Talk (Cross-Axis Sensitivity) in Vibration Measurement

Devices

Portable balancer & Vibration analyzer Balanset-1A

€1,975.00 + VAT (if applicable)

Parts

Vibration sensor

€90.00 + VAT (if applicable)

Parts

Optical Sensor (Laser Tachometer)

€124.00 + VAT (if applicable)

Devices

Balanset-4

€6,803.00 + VAT (if applicable)

Parts

Magnetic Stand Insize-60-kgf

€46.00 + VAT (if applicable)

Parts

Reflective tape

€10.00 + VAT (if applicable)

Devices

Dynamic balancer “Balanset-1A” OEM

€1,735.00 + VAT (if applicable)

Cross-talk — more formally cross-axis sensitivity or transverse sensitivity — is a measurement error inherent to vibration transducers, and especially to accelerometers. It is the transducer’s tendency to produce an output in response to vibration that is perpendicular to its intended measurement axis. In an ideal world, an accelerometer built to measure vertical motion would respond only to vertical motion and ignore everything horizontal or axial. In the real world, microscopic asymmetries in the sensing element give it a small but non-zero response to those “off-axis” inputs — and that unwanted output is cross-talk.

1. Definition: What is Cross-Talk?

Every practical accelerometer has one nominal sensitive axis along which it is calibrated. Cross-axis sensitivity describes how much the same sensor responds to motion at right angles to that axis. The imperfection arises from tiny misalignments between the seismic mass, the piezoelectric crystal, and the mounting base — see the piezoelectric accelerometer for the underlying mechanism. Because the error is built into the sensing element, it cannot be tuned out in the field; it can only be specified, minimised at manufacture, and managed by good measurement practice.

It is worth distinguishing cross-talk from electrical channel-to-channel interference. Here the term refers to mechanical cross-axis response within a single sensor, not signal leakage between cables or analyser inputs.

2. Why is Cross-Talk a Problem?

Cross-talk contaminates vibration data and can lead directly to diagnostic errors, because vibration from one direction “leaks” into the measurement of another. Consider a machine with very high horizontal vibration but low vertical vibration. A vertically mounted accelerometer with significant cross-axis sensitivity picks up a fraction of that strong horizontal motion and adds it to its own output. The reading then shows more vertical amplitude than truly exists — and an analyst may chase a vertical-direction fault that is not there.

This becomes particularly troublesome when:

• Performing modal analysis or Operating Deflection Shape (ODS) analysis, where accurate measurements in all three axes (X, Y, Z) are essential to animate the machine’s motion correctly. Errant cross-axis energy distorts the computed mode shapes.
• Diagnosing faults in complex machinery where the directional signature is the key to the root cause — for example, separating the directional behaviour of different types of misalignment from genuine unbalance.
• Performing high-precision balancing — especially on a balancing machine, where plane-separation accuracy depends on clean, direction-true signals.

3. Quantifying Cross-Talk

Cross-axis sensitivity is usually quoted by the sensor manufacturer as a percentage of the primary-axis sensitivity. A good industrial accelerometer might specify less than 5%; precision laboratory units do considerably better. A 5% figure means that for every 1 g of vibration applied perpendicular to the main axis, the sensor outputs a signal equivalent to less than 0.05 g in the primary direction.

The total cross-talk error you actually see depends on two factors working together:

1. The inherent cross-axis sensitivity of the sensor itself.
2. The ratio of the transverse vibration magnitude to the magnitude being measured along the primary axis.

The second factor is easy to underestimate. Even a sensor with low cross-axis sensitivity can produce a significant error when the off-axis vibration is far larger than the signal of interest.

Worked example: a sensor rated at 4% cross-axis sensitivity, mounted to read a 1.0 mm/s vertical level while 10 mm/s exists horizontally, can pick up roughly 0.04 × 10 = 0.4 mm/s of spurious signal — a 40% error on the quantity you care about.

Best- and worst-case errors also depend on the angular orientation of the dominant transverse motion, since cross-axis sensitivity itself varies with direction around the sensor.

4. Minimising the Effects of Cross-Talk

• Use high-quality sensors: The most direct defence is a precision-engineered accelerometer with a low specified cross-axis sensitivity. A shear-mode accelerometer typically offers better transverse rejection than a compression design.
• Mount it properly: Poor mounting exaggerates cross-talk. The sensor must sit flat and perpendicular to the surface so its primary axis is truly aligned with the intended direction; a tilted sensor effectively redefines its own axes. Follow ISO 5348 for mechanical mounting.
• Use triaxial accelerometers: Where accurate multi-axis data is required, a triaxial sensor — three orthogonal sensing elements in one block, factory-calibrated to minimise inter-axis cross-talk — is often the better choice and removes orientation guesswork.
• Check the calibration certificate: A traceable calibration certificate states the measured transverse sensitivity for the individual unit, not just the type-test maximum.

In the field, the practical answer is usually disciplined mounting. A two-channel field instrument such as the Balanset-1A relies on each accelerometer reading its own plane cleanly; a sensor stud-mounted square to a machined pad gives a direction-true 1× amplitude and phase, whereas a magnet perched on a curved, painted housing invites both cross-talk and mounting resonance to corrupt the measurement.

← Back to Main Index