How do you find the root cause of a motor failure?

Split the investigation into mechanical and electrical first, because bearings and stator windings account for most failures and leave different evidence. Read the physical evidence, the bearing damage mode or the winding burn pattern, then trace it back to a controllable cause with 5 whys: burn pattern to supply fault to loose connection to missing PM check, or bearing mode to lubrication, alignment, or stray current. Verify against the is/is-not boundary and a measurement, then correct the condition and confirm non-recurrence.

How can you tell why a motor winding burned out?

Read the burn pattern. Two of three phase groups burned with one left clean indicates single-phasing from a lost supply phase. A single localized hot spot indicates a turn-to-turn short from insulation weakness or a voltage spike. A burn against the core or slot edge indicates a ground fault from moisture, damage, or age. Even discoloration across all three phases indicates general overload or lost cooling. Each pattern points to a different root cause.

Why does the same motor keep failing?

Because the real condition was never fixed, the investigation stopped at 'bad motor' or 'bad rewind.' A motor that single-phases again has an unfixed supply or connection fault; one that flutes its bearings again is missing a grounding path; one that cooks its insulation again is starved of cooling or running over its load. The motor is the victim. Stop rewinding, investigate the circuit, load, or environment, and engineer the condition out.

What tests help find the cause of a motor failure?

On the electrical path, insulation-resistance and polarization-index testing (IEEE Std 43) catch moisture and aging, and surge testing finds turn-to-turn weakness; motor current signature analysis catches rotor-bar and winding faults on a running motor. On the mechanical path, vibration spectra and temperature trends catch bearing defects early. Pull the protection-relay and drive fault logs too, they often record the failure sequence.

Is it better to rewind or replace a failed motor?

That is a cost-and-efficiency decision, and it is separate from the RCA, neither rewinding nor replacing fixes the cause if the cause is external. Before spending on either, find why the motor failed: a rewound or new motor bolted back into the same misalignment, single-phasing supply, blocked cooling, or ungrounded VFD will fail the same way. Do the RCA first, correct the condition, then decide rewind versus replace.

How do you verify you found the real cause of a motor failure?

The cause must explain the evidence and the is/is-not boundary, why this motor failed and its identical neighbor on the same feeder did not, and be confirmable with a measurement such as a phase-voltage check, an insulation-resistance reading, or a shaft-voltage test. After correcting the condition, keep temperature, vibration, and insulation trends on watch through a defined window. If the signature does not return, the cause was root; if it does, reopen the analysis.

Root Cause Analysis of Motor Failure

Q: What is the most common cause of electric motor failure?

Bearings are the single largest cause. Long-standing IEEE and EPRI motor reliability surveys put bearings behind roughly 40 to 50 percent of failures and stator-winding faults behind another 28 to 37 percent, with the rotor and other causes making up a smaller remainder. That is why a motor RCA forks into mechanical (mostly bearings) and electrical (mostly windings) paths before chasing details.

Split the investigation into mechanical and electrical, read the burned winding or the failed bearing as evidence, and work a specific motor from symptom back to the true cause.

Root cause analysis of a motor failure is a structured investigation that splits into two paths, mechanical, mostly bearings, and electrical, mostly windings and insulation, reads the physical evidence each leaves, and traces the specific symptom back to the true cause so the replacement motor does not fail the same way.

A failed motor gets one of two verdicts in most plants: send it out to be rewound, or scrap it and hang a new one. Both skip the only question that stops the next failure, why did this one die? An electric motor is unusual among plant assets because its two dominant failure paths are almost independent and leave completely different evidence. Bearings fail mechanically and leave marks on raceways; windings fail electrically and leave a burn pattern that names the fault. Reading that evidence is fast, cheap, and specific. This guide applies the general root cause analysis method to a specific motor, and it is the investigation companion to electric motor failure causes which surveys the five causes across the population; here we work one motor from symptom to cause.

What is root cause analysis of a motor failure?

It is the work of getting from "the motor tripped and won't restart" to a controllable cause you can remove. A root cause is a condition within your control, supported by evidence, whose removal prevents recurrence. For motors that condition is rarely "the motor", it is the blocked cooling that cooked the insulation, the misalignment that killed the bearing, the lost phase that burned two windings, or the VFD without a grounding path that flutted the bearing. The failed part is where the damage surfaced; the cause is the operating or supply condition upstream.

What makes motor RCA distinctive is the clean fork at the top. Before chasing details, you decide whether you are investigating a mechanical death or an electrical one, because that choice sends you down a different evidence trail, a different set of tests, and a different family of causes.

Why split the investigation into mechanical and electrical?

Because the two paths account for almost every failure and barely overlap. Long-standing reliability surveys of industrial motors, the classic IEEE and EPRI studies, put bearings behind roughly 40 to 50 percent of failures and stator-winding faults behind another 28 to 37 percent, with the rotor and everything else making up a modest remainder. So the first branch of the investigation is not a detail; it is most of the answer. Establish which path you are on and you have already halved the search.

Two paths carry most motor failures and leave different evidence. Deciding mechanical versus electrical first cuts the search in half and sends you to the right tests, vibration and raceway reading on one side, insulation and burn-pattern reading on the other.

How do you read the failed motor as evidence?

Each path has a signature. On the mechanical side, the bearing tells its own story exactly as it would in any bearing failure RCA: fluting means stray current (chase the VFD grounding, see electrical bearing damage), a frosted band means lube starvation, load-zone spalling means misalignment or overload. On the electrical side, the winding burn pattern is remarkably specific about the fault, and reading it is the fastest cause-narrowing move in the whole investigation.

A burned stator is not scrap until it is read. Two phase groups gone with one clean means single-phasing; a localized hot spot means a turn-to-turn short; a burn against the core means a ground fault; even discoloration across all three phases means general overheating. Each points to a different cause.

How do you trace a symptom back to the true cause?

The mode narrows the cause class; a short 5 whys drives it to something controllable. Single-phasing: why did a phase disappear? A loose lug in the starter that overheated and opened. Why loose? No torque check in the PM. That is a fixable process. Confirm with the is/is-not boundary and a test: single-phasing burns two groups and leaves one clean, so a symmetric three-phase burn rules it out and points at overload or cooling instead. The electrical trail uses insulation-resistance and polarization-index testing (see motor insulation resistance testing) and, for a motor still running, motor current signature analysis to catch rotor bar and winding faults before the burn. The mechanical trail uses vibration and temperature trends. The evidence, not the loudest theory, decides.

The smaller remainder still matters. Rotor faults, broken or cracked rotor bars, show up as a distinctive current-signature sideband and as torque pulsations, and they are easy to misread as a bearing or supply problem if you skip the electrical tests. Shaft, coupling, and external causes round out the balance. The point of naming the two big paths is not to pretend the others do not exist; it is to make sure the first hour of the investigation goes where the evidence is statistically most likely to be, so you confirm or rule out bearings and windings before spending time on the rare causes. And as with any RCA, preserve the evidence before you disturb it: a motor that gets reset, spun, and hosed down before anyone looks has lost the burn pattern, the fault log, and the bearing condition that would have named the cause.

A 6-step motor failure RCA

Capture the state before disturbing it. Record how it failed, tripped, stalled, smoked, ran hot, pull the protection-relay and drive fault logs, and note load, ambient, and run hours before anything is disassembled or reset.
Fork the investigation. Decide mechanical or electrical from the first evidence: bearing noise and vibration history versus a smell of burnt varnish and a relay trip. This sets the test path.
Read the physical evidence. On the mechanical side, inspect the bearing as removed for its damage mode. On the electrical side, read the winding burn pattern and run insulation and, where useful, surge tests.
Trace the mode to a controllable cause. Run 5 whys down the chain, burn pattern to supply fault to loose lug to missing PM check; bearing mode to lube, alignment, or stray current. Push past "the winding failed."
Verify against the evidence. The cause must explain the pattern and the is/is-not boundary, why this motor and not its identical neighbor on the same feeder. Confirm with a measurement.
Correct the condition and confirm non-recurrence. Fix the feeder, load, cooling, grounding, or lube, then keep the motor's temperature, vibration, and insulation trends on watch through a defined window, and log the cause with consistent failure coding.

Why does the same motor keep burning out?

A motor position that eats windings or bearings on a schedule is an RCA that reached the wrong conclusion, usually "bad motor" or "bad rewind." The recurrence is proof the real condition survived. A motor that single-phases again has a supply or connection fault nobody fixed; a motor that flutes its bearings again is missing a grounding path; a motor that cooks its insulation again is starved of cooling or running over its load. Stop rewinding and investigate the circuit, the load, or the environment, because the motor is the victim, not the culprit. Roll repeat offenders into defect elimination so the condition is engineered out, and let the fix protect every identical motor on the same duty rather than just this one.

How do you keep it from failing the same way?

Non-recurrence needs a monitored window, not a closed work order. After the fix, keep the motor's temperature, vibration, and insulation trends on watch with a trigger to reopen if the signature returns. Rank the countermeasure honestly: a corrected feeder, a torque check added to the PM, a grounding ring, or a cooling fix beats a note to "watch this motor," which beats nothing. Feed the closed case into equipment reliability data and your maintenance KPIs so the next person facing the same burn pattern inherits your answer. Prevention is the job of electric motor maintenance and predictive maintenance.

What do the reliability studies say?

The two paths and their proportions come from long-standing industrial-motor reliability surveys: the IEEE motor reliability working-group study and the EPRI study both put bearings first at roughly 40–50% of failures and stator windings second at roughly 28–37%, with the rotor near 9–10% and the balance in other causes (summarized in Reliability Connect, Large Electric Motor Reliability). The exact splits differ between studies; read them as directional.
Insulation condition, the electrical path's leading indicator, is assessed by insulation-resistance and polarization-index testing per IEEE Std 43 Recommended Practice for Testing Insulation Resistance of Rotating Machinery, the test that catches moisture and aging before a ground fault.
Feeding motor failures into reliability data uses the standard equipment taxonomy in ISO 14224 (ISO 14224), so a recurring burn pattern across a fleet becomes a visible pattern instead of scattered notes.

A motor RCA only compounds if its conclusions land where the whole team can see them. Harmony pulls maintenance history, failure codes, protection-relay and drive faults, and vibration and temperature signals into one operational data layer, so a motor that burns out on the same duty every year surfaces as a pattern instead of a string of rewinds, and it can draft the corrective work order for a human to approve. It layers onto the CMMS and machines you already run, with no rip-and-replace; see how it works or the CLS case study.