Glossary term

Electrical Runout

Apparent displacement variation in a proximity-probe signal caused by shaft material, surface or electromagnetic effects rather than true shaft motion.

Branch: Mechanical Engineering
Glossary type: phenomenon
Content: Glossary term
Updated: Jun 25, 2026
Revision: v1.0.1 · reviewed

Definition

phenomenon

Electrical runout is apparent displacement variation in a proximity-probe signal caused by target material, surface, magnetic or electrical effects rather than actual geometric motion of the shaft.

In rotating machinery, electrical runout appears when an eddy-current proximity probe sees variations in shaft electrical conductivity, magnetic permeability, residual magnetism, plating, surface finish, seams, scratches or local material properties as the shaft rotates. It can create a once-per-revolution or harmonic signal even at slow roll, when dynamic forces are too small to explain the measured vibration.

Electrical runout is an apparent displacement signal produced by the interaction between a proximity probe and a rotating shaft target. The shaft may not be moving by that amount. Instead, the probe output changes because the target material, surface condition or electromagnetic properties change with angular position.

For an eddy-current proximity probe, this matters because the sensor is not measuring geometry alone. It is measuring an electromagnetic interaction affected by probe gap, target conductivity, permeability, residual magnetism, plating, shaft surface finish, local repairs, scratches, seams and calibration material.

A useful measurement model is:

x_{meas}(\theta)=x_{dyn}(\theta)+x_{mech}(\theta)+x_{elec}(\theta)+e(\theta)

where $x_{meas}$ is the probe-scaled displacement signal, $x_{dyn}$ is true dynamic shaft motion, $x_{mech}$ is mechanical or geometric runout, $x_{elec}$ is electrical runout and $e$ is residual measurement error. The angular coordinate $\theta$ usually comes from a tachometer, keyphasor or encoder reference.

Engineering Role

Electrical runout is most visible during slow roll, coastdown, barring-gear rotation or low-speed checks, where dynamic vibration forces are small. If a once-per-revolution signal is already present at slow roll, the same angular pattern can contaminate operating-speed vibration, orbit plots and 1x order amplitudes.

Engineers use electrical-runout checks to decide whether a proximity-probe channel is fit for:

shaft-relative vibration acceptance;
critical-speed and coastdown interpretation;
orbit-plot diagnosis;
balance correction based on 1x amplitude and phase;
trip or alarm settings based on displacement;
shaft centreline and bearing-clearance assessment.

The practical question is not whether the slow-roll trace is perfectly zero. The practical question is whether the residual runout is small, stable, documented and treated consistently enough for the intended diagnostic or protection decision.

Residual Runout and Acceptance Use

Electrical runout becomes a release problem when it is large compared with the decision limit. A simple screen is:

\displaystyle R_{er}=\frac{|X_{sr}|}{X_{limit}}

where $|X_{sr}|$ is the slow-roll runout vector magnitude in the same peak, peak-to-peak or RMS convention as the limit. If $R_{er}$ is large, the measured 1x vector may be dominated by probe-target effects rather than true dynamic shaft motion.

After compensation, the residual should also be checked:

\displaystyle R_{res}=\frac{|X_{res}|}{X_{limit}}

where $X_{res}$ is the remaining repeatable runout after the chosen correction. A low residual does not prove the machine is healthy. It only supports that the proximity-probe channel is fit for the intended orbit, shaft centerline, balancing or trip decision.

The acceptance record should state whether compensated or uncompensated data are used for alarms, trips, balancing vectors and trending. Mixing those conventions can make a machine look improved or degraded when only the signal processing changed.

Slow-Roll Compensation

If the shaft rotates slowly enough that dynamic motion is negligible, the measured slow-roll signal can approximate the combined mechanical and electrical runout pattern:

x_{sr}(\theta)\approx x_{mech}(\theta)+x_{elec}(\theta)

The operating signal may then be corrected in the angular domain:

x_{corr}(t)=x_{meas}(t)-x_{sr}(\theta(t))

This compensation is only defensible when the angular reference, probe chain, target surface, shaft direction, installed gap, filtering and scaling remain compatible. A slow-roll record taken after polishing, demagnetizing, changing a probe, changing a cable or moving the shaft axially may no longer represent the operating measurement.

Worked Example: Correct a 1x Vector for Slow-Roll Runout

A turbine bearing has an eddy-current proximity probe with calibrated sensitivity:

S=7.87\ \text{mV}/\mu\text{m}

During slow roll, the once-per-revolution voltage component is:

\Delta V_{sr,pp}=0.118\ \text{V}=118\ \text{mV}

The equivalent peak-to-peak slow-roll displacement is:

\displaystyle X_{sr,pp}=\frac{118}{7.87}=15.0\ \mu\text{m}

At operating speed, the measured 1x voltage component is:

\Delta V_{op,pp}=0.472\ \text{V}=472\ \text{mV}

The measured peak-to-peak operating displacement is:

\displaystyle X_{op,pp}=\frac{472}{7.87}=60.0\ \mu\text{m}

The engineer also has a phase reference. The operating 1x vector is:

X_{op}=60.0\angle 35^{\circ}\ \mu\text{m}_{pp}

and the slow-roll vector at the same angular reference is:

X_{sr}=15.0\angle 20^{\circ}\ \mu\text{m}_{pp}

Convert both vectors into rectangular form:

X_{op}=60.0(\cos 35^{\circ}+j\sin 35^{\circ})=49.1+j34.4

X_{sr}=15.0(\cos 20^{\circ}+j\sin 20^{\circ})=14.1+j5.13

Subtract the slow-roll vector, not just the scalar magnitude:

X_{corr}=X_{op}-X_{sr}=(49.1-14.1)+j(34.4-5.13)

X_{corr}=35.0+j29.3

The corrected magnitude is:

|X_{corr}|=\sqrt{35.0^2+29.3^2}=45.7\ \mu\text{m}_{pp}

and the corrected phase is:

\displaystyle \phi_{corr}=\tan^{-1}\left(\frac{29.3}{35.0}\right)=39.9^{\circ}

Engineering comment: scalar subtraction would give $60.0-15.0=45.0\ \mu\text{m}_{pp}$ , which is close here only because the phase angles are similar. With different phase, scalar subtraction could be very wrong. Runout compensation for balancing, orbit interpretation or acceptance testing must preserve amplitude, phase and the convention used for peak, peak-to-peak or RMS values.

Electrical runout is not mechanical runout. Mechanical runout is geometric or datum-related surface variation during rotation. Electrical runout is an apparent probe signal caused by material or electromagnetic variation, even when geometry is acceptable.

Electrical runout is not true dynamic vibration. Dynamic vibration changes with speed, load, damping, stiffness, unbalance, instability and operating condition. Electrical runout is usually tied to shaft angular position and may appear at slow roll.

Electrical runout is not random sensor noise. Noise is usually broadband or stochastic. Electrical runout is often repeatable once per revolution or at harmonics of rotation.

Electrical runout is not an orbit plot. An orbit plot may contain electrical runout, but the orbit is the X-Y display. Electrical runout is one possible contaminating contribution inside the displacement channels.

Electrical runout is not proximity-probe sensitivity. Sensitivity converts voltage to displacement for a calibrated target and geometry. Electrical runout is an error mechanism that violates the assumption that all voltage change is caused by distance change.

Validation and Common Mistakes

A defensible electrical-runout assessment states probe model, driver, extension cable, sensitivity, shaft material, target area, probe gap, shaft speed, angular reference, filtering, temperature, surface condition, magnetic condition, compensation method and residual runout after correction.

Common mistakes include:

assuming a slow-roll waveform is entirely electrical when mechanical runout or shaft bow may also be present;
subtracting slow-roll amplitude without phase;
applying compensation after changing probe gap, cable, driver, filter settings, reference mark or shaft surface condition;
measuring slow roll at a speed where bearing friction, rubs, oil-film effects or support motion are no longer negligible;
ignoring shaft magnetization, plating transitions, keyways, scratches or repaired regions under the probe track;
using runout compensation to hide a poor target surface instead of correcting the measurement surface when acceptance limits are tight;
comparing compensated and uncompensated orbit plots without labeling them clearly.

REF

Disciplines

Electrical Runout

Definition

Engineering Role

Residual Runout and Acceptance Use

Slow-Roll Compensation

Worked Example: Correct a 1x Vector for Slow-Roll Runout

Validation and Common Mistakes

See also

Engineering Role

Residual Runout and Acceptance Use

Slow-Roll Compensation

Worked Example: Correct a 1x Vector for Slow-Roll Runout

Distinction from Related Terms

Validation and Common Mistakes