Troubleshooting Broken Rotor Bars Under Motor Load

Integral-horsepower induction motors with a developing rotor defect may show elevated current-spectrum components, pole-pass sidebands in vibration, heating marks, or unstable behavior only while loaded. Start here: reproduce the operating load and collect current and vibration together. A single spectrum or visual mark is not a diagnosis.

Stop Trying the Wrong Fixes First

Several common checks can point you in the wrong direction when a rotor bar is cracked, intermittently connected, or merely suspected.

• Do not condemn the rotor from vibration sidebands alone. Rotor eccentricity can produce similar magnetic asymmetry. Pole-pass-frequency sidebands around running-speed harmonics identify modulation, not its cause.
• Do not accept one current spectrum as proof. Load variation, supply conditions, mechanical torque variation, and eccentricity can create or change spectral components. Compare repeated measurements at comparable loads.
• Do not treat burn marks as confirmed broken bars. Discoloration away from the bar-to-end-ring joint needs another test. The common fracture area is the joint between a bar and an end ring, but location alone still does not prove electrical discontinuity.
• Do not rely on an unloaded shop run to clear the rotor. Temperature, centrifugal force, and operating load can open an intermittent connection that appears intact when the rotor is cold or lightly loaded.
• Do not keep restarting the motor to reproduce the symptom. Starting produces high rotor current. Once one bar opens, adjacent bars carry more current and face greater stress.

Current analysis is usually the stronger operating confirmation, while vibration adds corroboration and helps expose competing mechanical causes. Final confirmation may require an offline electrical, thermal, or surface inspection.

Work Back from the Symptom

Observed symptom	Probable mechanism or competing cause	Next check
Repeatable current-spectrum sidebands under load	Rotor-field asymmetry or load-related modulation	Repeat at the same load and compare sideband spacing with calculated rotor slip and pole-pass frequency.
Pole-pass sidebands around 1×, 2×, or 3× running speed	Pulsating torque, speed modulation, eccentricity, or another magnetic-force modulation	Correlate with current data and inspect for eccentricity before assigning the pattern to broken bars.
Indication appears in service but disappears in the shop	An intermittent bar or joint changes with temperature, centrifugal force, or load	Reproduce the loaded condition or use an offline current-and-infrared test.
Localized heating or burn marks	High-resistance joint, discontinuity, or unrelated surface heating	Map the location and confirm it with infrared, penetrant, magnetic-particle, or direct electrical testing.
Sidebands without visible shaft-speed oscillation	Magnetic-force or eccentricity effects may be producing the vibration pattern	Do not use the absence of visible oscillation to clear the rotor; compare electrical and mechanical evidence.

The key distinction is modulation versus failure identity. A nonuniform rotor field can produce nonuniform torque, but a stator measurement does not automatically distinguish a broken bar from an eccentric rotor. Use independent measurements to separate those causes.

Calculate the Frequencies Before Collecting Data

Record supply frequency, motor pole count, actual shaft speed, load, and measurement time. Nameplate speed is not a substitute for measured speed when you need accurate sideband locations.

Calculate synchronous speed and slip from the same operating point:

n_sync = 120 × f_supply / p

F_slip = (n_sync - n_running) / 60

For the vibration relationship used here, calculate pole-pass frequency as:

F_p = p × F_slip

where p is the number of poles, F_slip is slip frequency in hertz, and speeds are in revolutions per minute. Search for components separated by F_p around 1× running speed and its harmonics. Do not move the cursors until measured speed and pole count are correct; a small speed error moves every predicted sideband.

The proposed mechanism is torque pulsation at F_p. If rotor angle is represented by theta(t) = cos(Wr×t + m×sin(Wp×t)) and the modulation index m is small, the expression contains components at Wr-Wp and Wr+Wp. That produces sidebands spaced by pole-pass frequency around running speed. The math explains the spacing but does not identify whether the modulation came from a broken bar or eccentricity.

Capture Current and Vibration Under Load

1. Stabilize the process. Run the motor at a normal, steady load. Record the operating state so the test can be repeated at a comparable point.
2. Measure actual running speed. Use that value to calculate slip and F_p. Recalculate whenever speed or load changes.
3. Collect clamp-current data. Use a current sensor and acquisition setup with enough bandwidth, record length, and spectral resolution to separate the expected sidebands. Prevent clipping and keep the sensor orientation fixed between tests.
4. Collect vibration at the same operating point. Acquire data at stable bearing-housing locations and retain the spectrum around 1×, 2×, and 3× running speed.
5. Compare spacing, not just peaks. Look for repeatable sidebands at the calculated spacing in both measurement domains. A peak close to a cursor during one run is weak evidence.
6. Repeat the measurement. Trend amplitude and pattern at comparable load, temperature, and speed. Separate a persistent rotor-related pattern from a transient process disturbance.

Current and vibration must describe the same operating interval. Comparing a loaded current spectrum with vibration taken after the process changed destroys the correlation needed to separate rotor modulation from load modulation.

Separate Broken Bars from Eccentricity

Start with repeatability. A suspected rotor-bar pattern should track rotor slip as the operating point changes. Recalculate the expected spacing for each measurement rather than forcing every spectrum to match one stored cursor set.

Then check for eccentricity. Rotor eccentricity also creates an asymmetric air gap and nonuniform magnetic pull. That can place pole-pass-related components in vibration even when no bar is broken. Inspect the mechanical air-gap condition, bearing condition, shaft position, and evidence of rotor-to-stator geometry problems using the maintenance methods available for the motor.

Use the strobe only as a supporting observation. Frequency modulation from torque pulsation can theoretically create sidebands around running speed, but the absence of visible shaft oscillation does not disprove the spectral components. A case with strong pole-pass sidebands and no broken bars demonstrates why vibration alone produces false alarms.

Give more weight to a pattern that meets all three conditions: it repeats under comparable load, its spacing follows calculated slip, and an independent electrical or thermal test points to a bar or end-ring discontinuity. If only the first two conditions are present, keep eccentricity and process torque variation in the decision tree.

Confirm the Rotor Offline

Remove the motor from service when operating risk or trend growth justifies intrusive testing. A cold visual inspection is the first offline check, not the last word.

1. Clean and inspect the rotor. Concentrate on bar-to-end-ring joints. Map every crack indication, hot spot, and discolored area rather than treating all marks as equivalent.
2. Perform a loop test with infrared observation. Watch for localized heating that identifies unequal current paths or a high-resistance connection.
3. Apply current directly between the rotor end rings when the approved test method permits it. Support the rotor on nonconducting wood blocks and monitor it with infrared imaging. Use controlled test equipment and the motor repair procedure applicable to the rotor.
4. Apply a surface-discontinuity method. Dye penetrant can expose surface-breaking cracks. Magnetic-particle inspection may be applicable to suitable ferromagnetic regions and must follow the inspection procedure for the material and geometry.
5. Reconcile the results. Match the offline indication to the loaded current and vibration pattern. A thermal or crack indication at a conductive joint carries more diagnostic weight than discoloration by itself.

An intermittent defect may close after shutdown. If the offline tests are negative but the loaded signature repeats, test at the temperature and mechanical condition that produces the symptom or continue controlled trending until the decision becomes clear.

Verify the Diagnosis Before Repair

Build a verification record that another engineer can reproduce. Save measured speed, load, supply condition, current-sensor placement, vibration measurement points, spectral resolution, calculated F_slip, calculated F_p, and the marked sideband frequencies.

• Confirm that the sideband spacing matches the calculation within the resolution of the acquired spectrum.
• Confirm that the pattern repeats during another stable loaded run.
• Check whether the pattern moves when slip changes.
• Compare electrical and vibration trends from the same operating states.
• Require an independent offline indication before calling ambiguous sidebands a broken bar.

After repair or rotor replacement, repeat the loaded test with the same acquisition settings and comparable process load. The suspect components should fall materially relative to the stored baseline, while running speed and load remain comparable. If they do not, reopen the diagnosis and examine eccentricity, load modulation, and measurement setup.

Control the Risk While the Motor Remains in Service

For a critical machine, use zero confirmed broken bars as the acceptance target. A broken bar transfers current and thermal stress to adjacent bars, so continued operation can expand the damage even when the motor still carries load.

If production requires continued operation, reduce avoidable starts and shorten the monitoring interval according to trend growth and machine consequence. Do not choose an interval from one amplitude reading. Base it on repeat measurements taken at comparable speed, load, and temperature.

Plan removal when the signature grows, the process becomes unstable, offline inspection confirms a discontinuity, or the consequence of an in-service failure exceeds the value of continued operation. Trending estimates stability; it does not make a cracked rotor safe.

FAQ