Troubleshooting
Why motors burn up with no overload trips on 480 V
Short answer
Crystallized insulation several feet back from motor terminals is the signature of high-resistance terminations cooking the leads from I²R heat, while unbalanced voltage adds negative-sequence heating that current-magnitude overload heaters never see. Record all three line-to-line voltages and all three phase currents under load, verify the transformer grounding, and measure insulation resistance on every failed branch before any restart. Welded contacts and ruptured starters are consequences of closing onto failed motors, not a separate mystery.
On this page
- Two failure mechanisms the heaters never see
- What the reported readings actually mean
- Worked example: turning readings into unbalance numbers
- Why the contactors welded and the starters ruptured
- The test sequence before any restart
- Fix decisions by finding
- Protection that catches it next time
- Questions asked
Motors burning up with the overload heaters never tripping, insulation crystallized 4 to 5 ft back from the motor terminals, contactors welded shut and two wall-mounted starters ruptured — that is not one problem. The crystallized leads point to high-resistance terminations heating the conductors locally, at I²R losses the heaters never see. The reported 62 A on one phase against 78 A on another points to voltage unbalance driving negative-sequence heating that current-magnitude protection also misses. The welded contacts and ruptured starters are the consequence: someone resetting and restarting motors that had already failed to ground. Separate the three, in that order.
Two failure mechanisms the heaters never see #
Termination heating at the lug #
Insulation does not crystallize 4 ft from a termination on its own. A loose lug, a cold crimp, corroded strands or an aluminum conductor on a connector not prepared for aluminum adds series resistance where the loss is P = I²R. At the reported 78 A, an illustrative joint resistance of 0.10 Ω dissipates 78 × 78 × 0.10 ≈ 608 W at one lug — a small space heater clamped to the connection. That heat soaks into the terminal box and conducts along the copper, cooking the jacket well away from the joint while the phase current, and therefore the heater, stays exactly where it was.
Winding-failure investigations trace unequal terminal voltages to unbalanced source loading, poor motor-terminal connections and high-resistance contacts, which puts termination quality first on the list when every symptom is heat.
Voltage unbalance and negative-sequence heating #
Unbalanced line voltages in a three-phase induction motor set up a stator field component that rotates opposite to the rotor. The rotor sees that field at roughly twice line frequency, so a few percent of voltage unbalance produces large negative-sequence currents concentrated in the rotor bars and stator winding, raising losses and temperature well before the average current looks abnormal. NEMA MG 1-2021 puts motor current unbalance at roughly six to ten times the applied voltage unbalance, requires derating once the line-to-line unbalance measured at the motor terminals exceeds 1%, and its derating curve runs only to 5% — continuous operation beyond that is outside the standard.
A bimetal or melting-alloy heater models winding heat from current magnitude, so rotor-concentrated negative-sequence heat at modest average current simply never reaches it. If the spread appears on every motor fed from the transformer rather than on one machine, work upstream: diagnosing current imbalance across every motor on a service walks that trace.
What the reported readings actually mean #
On a 480 V wye system, nominal phase-to-ground voltage is 480 / √3 ≈ 277 V. The reported 265 to 280 V spread is therefore almost certainly phase-to-ground data — a motor actually fed 280 V line-to-line would stall long before drawing 78 A. Useful as a flag, but not an unbalance number: unbalance must be computed from three simultaneous line-to-line readings. Verify the transformer's neutral-to-ground connection at the same time, because a wye system that has lost its ground reference shifts phase-to-ground voltages and can let transient overvoltages build through repetitive restriking.
The currents are a reported symptom too: 62 A on phase A and 78 A on phase C, with phase B unrecorded, is at least a 16 A spread, and "around 100 A" means nothing until it is pinned to the feeder or to one motor. Neither set of numbers closes the case; both tell you where to put the meters.
Worked example: turning readings into unbalance numbers #
Illustrative, using the method you will run on site. Three line-to-line voltages read simultaneously at the motor terminals under steady load: 476 V, 482 V and 464 V.
- Average = (476 + 482 + 464) / 3 = 1422 / 3 = 474 V.
- Deviations from average: 2 V, 8 V and 10 V; the largest is 10 V.
- Unbalance = 100 × 10 / 474 = 2.1%, the NEMA MG 1 definition.
At six to ten times that in current, expect roughly 13 to 21% current unbalance — on a 70 A average, about ±9 to ±15 A per phase. The reported 62/78 A pair averages 70 A with a ±8 A spread if phase B sits at 70 A: an 11.4% current unbalance, implying a supply unbalance of roughly 1.1 to 1.9%, already past the 1% derating threshold. The decision at 2% voltage unbalance: find and fix the source or derate the motor per the MG 1 curve; do not plan continuous running above 5%.
Why the contactors welded and the starters ruptured #
Welded contacts in motor-starting contactors are, most often, the result of closing onto a short circuit: a motor fails, the breaker or overload is reset — or never functioned — and the next start attempt makes the contacts onto fault current far beyond starting duty. The contacts weld; enough fault energy ruptures the enclosure, which is what a starter coming off the wall looks like. Chatter, rapid cycling, contamination, coil undervoltage and incorrect duty ratings can also weld contacts, but with six months of motor failures and zero recorded trips, restart onto an existing fault is the working hypothesis until teardown says otherwise.
"No overload ever tripped" is itself a finding: bench-test the overload on every affected machine before trusting it again. With enclosures rupturing, also have the available fault current at that panel calculated and compare it against the short-circuit current rating marked on each starter — UL 508A requires that marking on industrial control panels, and an installation whose available fault current exceeds the marked SCCR is wrong regardless of which failure came first. Confirm the requirement and the marked value with your authority having jurisdiction and the panel documentation.
The test sequence before any restart #
One safety note first. NFPA 70E, 2024 edition, treats energized work as the exception that requires documented justification and protective equipment; confirm its application with your authority having jurisdiction and the equipment documentation. Every step below that opens an enclosure happens de-energized and verified for absence of voltage, and steps 6 and 7 are measurements on operating equipment under your site's safety procedure.
- Stop restarting failed motors — every reset-and-restart cycle risks welding another contactor or worse.
- Lock out each branch and prove it de-energized, then document motor nameplate data (hp, FLA, voltage, service factor), conductor material and size, overload type and setting, heater size, contactor size, upstream protective device and interrupting rating, and transformer kVA and impedance.
- Inspect every termination from disconnect to motor: discoloration, hardened or missing insulation, broken strands, cold crimps, oxidation, aluminum on connectors not rated for it, and water ingress, blocked ventilation or contamination while you are in there.
- Measure insulation resistance with a megohmmeter on the isolated circuit, phase to phase and phase to ground. IEEE Std 43 gives 500 V DC as the test voltage for windings rated below 1 kV and 5 MΩ as a minimum for modern windings — treat any phase below that, or far from its neighbors, as the finding.
- Measure each phase conductor and each winding phase to phase with a low-reading ohmmeter; a high-resistance joint shows as an unequal reading.
- Re-energize only circuits that passed steps 4 and 5, then measure under steady load all three line-to-line and all three phase-to-ground voltages at the transformer secondary, the starter line side, the starter load side and the motor terminals, and compute unbalance at the motor terminals.
- At the same load, record all three phase currents and compute current unbalance the same way.
- Verify the transformer neutral and system grounding connections, including the bonding path back to the service.
- Bench-test each overload relay against its manufacturer's procedure; confirm it senses all three phases and that the setting matches the motor nameplate FLA.
- Inspect the control circuit for coil undervoltage and chatter, and check how many times each starter has been cycled.
Fix decisions by finding #
| Finding | Decision |
|---|---|
| Crystallized insulation on motor leads | Replace the cable back to sound, supple insulation; sleeving over crystallized jacket hides nothing. |
| Loose, oxidized or misapplied termination | Rebuild with the correct connector, proper conductor preparation and the torque printed on the lug or device; use connectors rated for aluminum on aluminum. |
| Line-to-line unbalance above 1% at the motor terminals | Find the source — transformer, unbalanced single-phase loading, an upstream high-resistance joint — and derate per NEMA MG 1 until fixed; above 5%, do not run continuously. |
| Available fault current above the starter's marked SCCR | Correct the installation so the marked rating covers the available current, typically properly rated fusing in the feeder per the panel documentation. |
| Overload dead, single-phase sensing, or heaters upsized to stop trips | Reset to nameplate-based selection and install three-phase-sensing protection. |
| Welded contactor | Replace it after the initiating fault is found and cleared, never before. |
Protection that catches it next time #
The heaters did their design job and the motors burned anyway, so the protection changes. Electronic overload relays add phase-loss detection — one documented design, the CEP7 relay, trips on phase loss once current in any leg drops below 30% of its FLA dial setting — and separate phase-loss relays are available with an adjustable response delay of 0.2 s to 20 s so momentary dips do not cause nuisance trips while a real loss still gets caught. Those numbers are model-specific: verify them against the relay series' own documentation before setting anything.
On a service with a proven unbalance problem, the decision is simple: three-phase-sensing electronic overloads with phase-loss and current-imbalance elements on every motor, set from the nameplate FLA. The setting trade-offs for an individual machine are covered in selecting overload and jam protection for a reversible motor.
Questions asked
Why did the overload heaters never trip while the motors cooked?
Thermal overload heaters model motor heating from current magnitude, phase by phase. A high-resistance termination dissipates its heat at the lug and along the conductor, outside both the winding and the heater element, so the phase current never rises toward pickup. Negative-sequence heating from unbalanced voltage raises winding temperature without a matching rise in average current either. If the motors were also running below full load, neither mechanism ever reaches the trip point.
What voltage unbalance should I hold this service to?
NEMA MG 1 defines unbalance as the largest deviation of the three line-to-line voltages from their average, divided by that average, measured at the motor terminals under load. Hold it under 1%: above that the motor must be derated on the standard's curve, which ends at 5% unbalance. Because current unbalance typically runs six to ten times the voltage unbalance, a 2% voltage problem already shows up as a double-digit current spread.
Why did the contactors weld shut and the starters come off the wall?
Welded contacts in motor-starting contactors are most often the result of closing onto a short circuit — a motor fails, someone resets and restarts, and the contactor makes onto fault current far beyond starting duty. Ruptured enclosures are the same event with more energy. That is why the sequence stops all restarts and tests insulation before anything closes again, and why available fault current should be checked against each starter's marked short-circuit current rating.
Can I sleeve over crystallized insulation on motor leads and keep the cable?
No. Crystallized, embrittled insulation has been overheated past recovery; tape and heat-shrink only hide it. Replace the damaged section back to clean, supple insulation, rebuild the termination with the correct lug, preparation and torque, then re-measure under load to confirm the heating is actually gone rather than moved.
Are 265 to 280 V readings a problem on a 480 V wye service?
Nominal phase-to-ground on a 480 V wye system is 480 divided by the square root of three, about 277 V, so those values are plausible phase-to-ground readings. The 15 V spread only matters if all three phases were read at the same moment under the same load; readings taken at different times prove nothing. Wide phase-to-ground shifts also warrant checking the transformer's neutral-to-ground connection before trusting any of them.