BLDC Motor Troubleshooting: 12 Common Faults, Causes and Fixes

Here is the statistic that shapes this whole guide: roughly half the motors returned to our factory as "faulty" pass every bench test we run on them. The winding is balanced, the Hall sensors switch cleanly, the bearings are quiet, and the back-EMF is textbook. The fault was somewhere else — a crimp that backed out of a connector, a controller current limit set below what the application actually demands, a sagging battery, or a phase-and-Hall sequence that was never correct in the first place. This guide is the diagnostic sequence we use, in the order we use it, with the numbers we measure against.

Before You Start: Safety and Tools

Everything below assumes a 12–72V DC system, which is below the 60V DC SELV threshold in IEC 61140 for 12V through 48V buses. That does not make it harmless — a 48V lithium pack can deliver hundreds of amps into a shorted spanner. Disconnect the battery or open the main contactor before touching the power wiring, and discharge the controller's DC-link capacitors (typically 10–30 seconds) before probing the output stage.

You need five instruments, and four of them are cheap:

• Digital multimeter with a milliohm-capable resistance range — phase resistances are often below 1 Ω, so a meter that only resolves to 0.1 Ω will mislead you.
• Clamp meter with DC current capability. The single most useful tool for this work. An AC-only clamp is useless on a DC bus.
• Infrared thermometer for housing and controller heatsink temperatures.
• Insulation tester (megger), 500 V DC range — for suspected moisture ingress or winding-to-frame breakdown.
• Oscilloscope (optional but decisive for Hall timing and PWM problems).

Warning: never apply a 500 V megger test with the Hall sensors or controller connected. Hall ICs are rated for a few volts and will be destroyed instantly. Disconnect the Hall harness and the controller, then test the three phase leads to the motor housing only.

Baseline Measurements: What "Healthy" Looks Like

You cannot diagnose a motor without knowing what a good one reads. These are the ranges from our production test bench, measured phase-to-phase at 20°C on 48V windings.

Measured at 20°C. Copper resistance rises about 0.39% per °C, so a hot motor reads roughly 20% higher at 70°C — always compare like with like. The 24V column is approximately one quarter of the 48V value because the winding uses half the turns at twice the cross-section; see our 24V vs 48V bus voltage guide for why.

The three pass/fail rules that matter more than any absolute number:

• Phase balance: the three phase-to-phase readings must agree within 5%. Our factory QC limit is 3%. Above 10% means a shorted turn or a bad internal joint — the motor is not repairable in the field.
• Phase to housing: should read open circuit on a multimeter, and above 5 MΩ at 500 V DC per IEEE 43-2013 for low-voltage windings. A dry, healthy 48V winding normally reads well above 100 MΩ. Anything under 1 MΩ means moisture or insulation damage.
• Back-EMF balance: spin the shaft by hand and measure AC volts across each phase pair in turn. The three readings must be within 10% of each other. Unbalanced back-EMF with balanced resistance points to partially demagnetised rotor magnets — usually from a prior overheating event.

Hall sensor baseline

With 5 V applied to the Hall supply pin and the shaft turned slowly by hand, each of the three Hall outputs should switch cleanly between roughly 0 V and 5 V. Across one electrical revolution you should observe exactly six states, and you must never see 000 or 111 at any rotor position. Both of those are invalid and indicate a dead sensor, a broken signal wire, or a missing pull-up resistor (open-collector Hall ICs need a 1–10 kΩ pull-up to 5 V; most controllers provide it internally, but bench test rigs often do not). Our Hall sensor guide has the full commutation table and pin-by-pin wiring.

Symptom Triage Table

Start here. Find your symptom, then jump to the detailed fault section below.

Fault 1: Motor Will Not Start (Silent, No Vibration)

A completely silent motor is almost never a motor fault. Work the supply chain from the battery outwards:

1. Bus voltage at the controller terminals, measured under load. Not at the battery. A 48V pack that reads 51 V open-circuit and collapses to 38 V when the run command is given has either an undersized cable, a corroded connector, a tired cell group or a BMS current limit cutting in. Most controllers assert undervoltage lockout at 80–85% of nominal (about 40 V on a 48V bus).
2. Enable / brake / start input. Many industrial BLDC controllers use an active-low enable and will remain silent with the pin floating. Check the wiring diagram for your controller — this accounts for a surprising share of "dead on arrival" reports on new installations.
3. Speed reference. A 0–5 V or potentiometer input sitting at 0 V is a valid "stop" command. Measure it.
4. Fuse and contactor. Obvious, and worth 10 seconds.
5. Motor winding continuity. Only now do you suspect the motor. An open phase reads infinite resistance on one of the three pairs.

Fault 2: Motor Hums or Jitters But Will Not Rotate

The controller is energising the windings, so power and enable are fine. Two candidates:

Mechanically jammed. Remove the run command, disconnect the load, and turn the output shaft by hand. A geared motor should turn with resistance but turn. If it will not, the fault is in the gearbox, the driven machine, or an engaged brake — not the motor.

Commutation cannot start. The controller does not know the rotor position, so it energises the wrong pair and the rotor oscillates in place. This is usually a Hall sensor with no supply (check 5 V at the Hall connector), a broken Hall signal wire (reading stuck at 0 V or 5 V through a full shaft rotation), or a sensorless controller attempting to start into a high-inertia or high-friction load, which is a known limitation of sensorless control at zero speed.

Fault 3: Motor Runs Rough, Cogs and Draws High Current

This is the classic symptom of a phase/Hall mismatch, and it is the fault we most often talk customers through by phone.

There are three phase wires (U, V, W) and three Hall signals (Ha, Hb, Hc). That gives 3! × 3! = 36 possible wiring permutations, of which only 6 commutate correctly. Several of the wrong 30 will still spin the motor — badly. The signature is unmistakable once you know it: the motor turns, but noisily and with visible torque ripple, it draws two to four times normal no-load current, it gets hot within a minute, and it produces perhaps 30–50% of rated torque.

Test each combination at no load and record the DC current draw with a clamp meter. The correct combination is the one with the lowest no-load current and the smoothest sound — typically 5–12% of rated current at no load.

Many modern controllers have an auto-learn or self-identification routine that walks these combinations for you in about 30 seconds. All of our BLDC controllers supplied with a matched motor are pre-configured at the factory, which is the main practical argument for buying the motor and controller as a set rather than sourcing them separately.

If the roughness survives all six combinations, the cause is elsewhere: partially demagnetised rotor magnets (check back-EMF balance), a Hall sensor that has drifted out of alignment mechanically, or damaged bearings.

Fault 4: Motor Runs in One Direction Only

Check the direction input first — a floating DIR pin defaults one way on most controllers. If the input toggles correctly and the motor still refuses to reverse, or reverses roughly while running smoothly forward, the Hall sequence is one of the "half-correct" permutations that happens to work in one rotation direction. Return to the six-combination table above.

Fault 5: Overheating

The field alarm threshold is a housing surface temperature above 80°C. The real limit is the winding hot-spot, governed by insulation class under IEC 60034-1:

Hot-spot limits per IEC 60034-1 at 40°C reference ambient. Every 10°C above the class limit roughly halves winding insulation life — the Arrhenius relationship that governs all thermal ageing in motors.

Work through the four causes in this order:

1. Overload. Clamp-meter the DC bus current and compare it to the nameplate. If the motor is drawing 1.3× rated current continuously, it is being asked for 1.3× rated torque and it will overheat — no amount of cooling fixes an undersized motor. Recheck the sizing with our torque and power calculation guide.
2. Duty-cycle mismatch. A motor rated for S1 continuous duty used in S3 or S5 intermittent service with frequent starts sees very high RMS current from the acceleration peaks even when average load looks modest. Every start of a high-inertia load dumps energy into the windings. Count the starts per hour, not just the average torque.
3. Wrong commutation. Covered in Fault 3 — a mis-sequenced motor is inefficient and heats fast. Confirm no-load current is in the normal 5–12% band before blaming the load.
4. No heat path. A BLDC motor sheds most of its heat through the mounting flange into the machine frame. Mounting on a plastic bracket, wrapping it in insulation, or enclosing it in a sealed box without airflow can add 20–30°C. Bolt the flange to metal.

Fault 6: Controller Overcurrent Trips

The timing of the trip tells you the cause:

Fault 7: Speed Instability and Hunting

If the speed oscillates around setpoint, run the drive in open loop. If it is stable open-loop, the fault is in the closed-loop control, not the motor: PI gains set too aggressive for the load inertia, or a speed reference signal polluted by noise. If it is unstable open-loop as well, look at bus voltage sag under load and at the Hall signal quality — a marginal Hall sensor produces intermittent position errors that show up as speed jitter.

A practical starting point for PI tuning on a high-inertia traction load: reduce proportional gain until the oscillation stops, then increase integral gain slowly until steady-state error is acceptable. High-inertia loads need lower proportional gain than the factory default, which is usually tuned for a low-inertia bench setup.

Fault 8: Noise and Vibration

Motor vibration is measurable, and IEC 60034-14 gives you a limit to measure against: for Grade A machines with shaft height up to 132 mm, free suspension, the limit is 1.6 mm/s RMS in the 600–3,600 rpm range. Beyond that number, the character of the noise tells you the source:

Bearings are the only genuinely wear-limited part of a BLDC motor. We specify sealed 6000/6200-series bearings with an L10 life of 20,000 hours at rated radial load; in AGV and conveyor service, geared units typically run 20,000–40,000 hours before noise appears. If a bearing fails in under 5,000 hours, look for a cause: overhung load beyond the rated radial limit, shaft misalignment, or a coupling installed by hammering the shaft, which brinells the races immediately.

Fault 9: Low Torque Under Load

The motor runs, but stalls or slows badly when the machine is loaded. Three checks:

• Is the current reaching rated? If the motor stalls while drawing well under rated current, the controller is limiting it — check the current limit setting.
• Is the bus voltage holding? Torque falls with available voltage. A pack sagging from 48 V to 40 V under load has lost 17% of its speed capability, and the controller will limit current to protect against undervoltage.
• Is commutation correct? A mis-sequenced motor produces 30–50% of rated torque. This is the most commonly missed cause.

If all three pass, the motor is genuinely undersized for the application. Re-run the sizing calculation — for mobile equipment, our AGV motor sizing guide covers payload, gradient, rolling resistance and wheel diameter, and the usual fix is a higher gearbox ratio rather than a bigger motor.

Fault 10: Electromagnetic Brake Will Not Release

Motors fitted with a spring-applied electromagnetic brake are fail-safe: the spring clamps the rotor and the coil must be energised to release it. If the brake stays engaged, the motor will hum, draw heavy current and trip.

• Measure the brake coil voltage during a run command. Our brakes are 24 V DC; anything below about 21 V may not overcome the spring, and a common cause is powering the brake through a long, thin wire that drops several volts.
• Check the brake release timing. The controller should energise the brake 100–300 ms before commanding torque. Reversed sequencing makes the motor fight an engaged brake on every start.
• Check the air gap. Brake pads wear; as the gap grows, the coil can no longer pull the armature. Most brakes have an adjustable gap of 0.2–0.5 mm.
• Check the flyback diode. An open flyback across the brake coil lets the collapsing field arc the switching contacts, which eventually welds or destroys them.

Fault 11: Moisture and Contamination Failures

Motors on outdoor equipment — PV cleaning robots, agricultural machinery, gate operators — fail differently from indoor motors. IEC 60034-5 defines the IP protection ratings, and it is worth knowing what your rating actually promises:

The tell for moisture ingress is a phase-to-housing insulation resistance that has fallen below 1 MΩ. Sometimes a motor that has taken in condensation can be recovered by baking it at 80–90°C for 4–6 hours and re-testing; if the reading returns above 100 MΩ it is usually serviceable. If water has reached the Hall sensor PCB, the sensors will need replacing regardless. Note also that repeated heating and cooling cycles pump humid air in and out through the shaft seal — a motor that is switched on and off daily outdoors accumulates moisture faster than one that runs continuously.

Fault 12: EMI and Erratic Behaviour

Symptoms that appear only when a nearby contactor, VFD or welder operates are electromagnetic interference, not motor faults. IEC 61000-6-2 sets the immunity requirements for industrial environments; the practical countermeasures are:

• Route Hall sensor and speed reference cables at least 200 mm away from motor phase cables, and cross them at 90° where they must meet.
• Use shielded cable for Hall and encoder signals, with the shield grounded at the controller end only — grounding both ends creates a ground loop that makes things worse.
• Keep motor phase cables as short as practical and twist them together; long parallel runs act as antennas.
• Fit an RC snubber or flyback diode across every inductive load switched near the drive (contactors, solenoids, the brake coil).
• Bond the motor housing to the machine frame with a short, wide strap, not a thin wire.

What Actually Fails: Our Warranty Return Data

This is the distribution we see across motors returned to our factory for analysis. It is worth studying before you order a replacement motor, because the numbers say the motor usually is not the problem.

Distribution from our own warranty and field-return analysis across AGV, conveyor, gate and cleaning-robot applications. Roughly half of all returned motors pass full bench testing with no fault found.

The pattern is consistent: the two largest categories — wiring and controller — account for over half of all failures, and neither requires a new motor. Before you order a replacement, spend 15 minutes on the baseline measurements at the top of this guide. If the phase resistances are balanced and the back-EMF is balanced, the motor is fine.

Preventive Maintenance Schedule

Frequently Asked Questions