A broken rotor bar is one of the few faults that MCSA detects more reliably than vibration analysis. The sideband it produces in the current spectrum at f1(1 - 2s) and f1(1 + 2s) appears weeks before the fault generates measurable torque pulsation or vibration.
What Are Rotor Bars?
The squirrel-cage rotor of an AC induction motor consists of aluminum or copper bars embedded in slots cut into the rotor core. Short-circuit rings at each end connect all bars in parallel. Current induced in the bars by the stator's rotating magnetic field generates the electromagnetic torque that drives the motor.
The cage gets its name from the visual resemblance of bars and end rings to a squirrel exercise wheel. In smaller motors, the cage is typically die-cast aluminum poured directly into the rotor slots. Larger motors use individually inserted copper bars brazed or welded to the end rings.
Each bar carries a fraction of the total rotor current. If one bar is cracked or broken, its share of current redistributes to adjacent bars. That redistribution creates rotor asymmetry, which has measurable consequences in the stator current waveform.
Why Rotor Bars Fail
Several distinct failure mechanisms can initiate a rotor bar crack. Most real failures involve more than one:
- Thermal fatigue: Each motor start drives 600 to 700% of full-load current through the bars for several seconds, creating severe resistive heating. Thousands of starts accumulate thermal stress at bar-to-end-ring joints, the highest-stress location in the cage structure.
- Mechanical fatigue: Torsional shock from abrupt load changes or direct-on-line starting into high-inertia loads such as fans and compressors creates cyclic mechanical stress that propagates existing micro-cracks.
- Casting defects: Die-cast aluminum rotors contain porosity and inclusions from the manufacturing process. These create pre-existing stress concentrations that develop into cracks under cyclic loading, sometimes within the first few years of service.
- Electrical fatigue: Unbalanced supply voltage drives additional circulating currents in the rotor cage, increasing local heating at connection points and accelerating crack development at joints already stressed by starts.
- End ring fracture: Cracking at the bar-to-end-ring connection has the same electrical effect as a broken bar. Rotor asymmetry develops regardless of whether the break is in the bar body or at the joint.
High-cycle applications, such as compressors with frequent starts or motors subjected to regular plug-reversing, are the most common candidates for early rotor bar failure.
The MCSA Detection Mechanism
A balanced rotor produces a symmetric rotating magnetic field. The stator current at the line frequency (f1) reflects this symmetry. When a bar breaks, the rotor field becomes asymmetric: the damaged section generates less flux than the intact bars, creating a rotating imbalance.
This asymmetry modulates the stator current at twice the slip frequency. The modulation creates sidebands in the current spectrum at predictable frequencies on either side of the fundamental. Those sidebands are the diagnostic signature.
For a 60 Hz motor running at 1,746 RPM on a 4-pole supply:
- Synchronous speed: 1,800 RPM
- Per-unit slip: (1,800 - 1,746) / 1,800 = 0.03
- Lower sideband: 60 x (1 - 0.06) = 56.4 Hz
- Upper sideband: 60 x (1 + 0.06) = 63.6 Hz
The lower sideband at f1(1 - 2s) is typically larger in amplitude and is the primary diagnostic indicator. The upper sideband at f1(1 + 2s) serves as confirmation. Both must be present and consistent with the calculated slip frequency before a fault call is made.
Severity Thresholds
Sideband amplitude is measured in decibels relative to the supply frequency component. The table below reflects widely cited starting-point guidance, not universal fixed limits. Motor geometry, rotor bar count, and slip all affect the baseline amplitude for a healthy motor.
| Sideband Amplitude (dB relative to f1) | Interpretation | Recommended Action |
|---|---|---|
| Better than -54 dB | No significant fault indicated | Continue routine monitoring interval |
| -54 dB to -40 dB | Developing fault | Increase measurement frequency; plan inspection at next opportunity |
| Worse than -40 dB | Significant bar damage | Prioritize repair; assess remaining run time with engineering |
A motor with a healthy baseline of -60 dB that drifts to -48 dB has changed by 12 dB, which is a meaningful progression. A motor whose healthy baseline was -45 dB reading -48 dB has changed very little. The threshold table provides orientation; the trend against each motor's own history provides the diagnosis.
Minimum Load Requirement
At low load, the motor runs close to synchronous speed. Slip is small, and the sidebands fall very close to the fundamental frequency. At 1% slip on a 60 Hz supply, sidebands are only 1.2 Hz away from the 60 Hz component. Resolving them requires very high frequency resolution and long data acquisition windows.
Reliable detection typically requires 40% load or higher. At full load, slip is at its maximum, sidebands are furthest from the fundamental, and frequency resolution requirements are most easily met. When possible, schedule MCSA measurements at representative operating load rather than at light load or during transitions.
If a motor rarely reaches 40% load in normal operation, document that constraint in the measurement record. A single measurement at higher-than-normal load during a planned test condition is more diagnostic than multiple measurements at 15% load.
Rotor Bar Faults on VFD-Driven Motors
Variable speed drives create two complications for rotor bar MCSA. The fundamental output frequency changes with drive speed commands, shifting the sidebands with it. And the drive's switching waveform introduces harmonic noise that can mask low-amplitude sideband components.
Standard MCSA algorithms assume a fixed supply frequency. Applied to a VFD-driven motor, they produce unreliable results. Specialized VFD-aware platforms process the variable-frequency current signal, tracking the instantaneous fundamental and reconstructing the sideband pattern relative to the changing carrier frequency.
An alternative approach uses flux monitoring: a search coil placed near the motor frame measures stray magnetic flux rather than stator current. The stray flux signal carries the same rotor asymmetry signature but is less contaminated by drive switching harmonics. Some platforms use both signals in parallel, using agreement between the two as additional confidence in a fault call.
For VFD applications, confirm that the monitoring platform is specifically validated for drive-driven motors before relying on its rotor bar results. The measurement method matters as much as the sensor placement.
Trending vs Single-Point Measurement
A single MCSA measurement establishes a data point, not a diagnosis. Its value depends entirely on context. The same sideband amplitude can mean a severe developing fault on one motor and normal operation on another, depending on rotor geometry and the motor's own history.
The correct reference is the earliest available measurement on that specific motor, taken at comparable load and speed. Trending sideband amplitude over months and years provides the signal; week-to-week variation is usually within measurement uncertainty.
When setting up a monitoring program on motors without historical data, the first two or three measurements taken over the first six months serve as the baseline. Flag any motor where the first measurement already shows a sideband above -50 dB: the motor may have a pre-existing fault, or a low healthy baseline, and the distinction matters for setting alert levels.
Confirming a Rotor Bar Fault
MCSA provides evidence; it does not provide certainty. Sidebands at the predicted frequency are a strong indicator, but other sources of rotor asymmetry, including air gap eccentricity and stator winding faults, can produce current spectrum components in similar frequency ranges.
Before scheduling a rewind or rotor replacement, consider the following confirmation steps:
- Repeat at different load levels. A genuine rotor bar fault produces sidebands that track with slip. As load increases, slip increases, and the sideband frequencies shift predictably outward from the fundamental. A consistent pattern across two or three load points strengthens the diagnosis.
- Second measurement or second analyst. For high-consequence machines, obtain a second measurement with a different instrument or have an independent analyst review the spectrum. Interpretation of borderline spectra benefits from independent review.
- Coast-down or high-current test. During a planned outage on a critical machine, a current measurement taken as the motor decelerates through slip frequencies can confirm rotor bar asymmetry. High-current testing with a motor circuit analyzer measures rotor bar resistance directly and confirms bar integrity without running the motor.
The repair decision should combine the MCSA evidence with operational context: motor criticality, availability of a spare, production impact of an outage, and the cost of a false positive removal versus a run-to-failure event.
Frequently Asked Questions
What causes rotor bar failures in induction motors?
The main causes are thermal fatigue from repeated starting cycles, mechanical fatigue from torsional shock on high-inertia loads, casting defects in aluminum rotors, unbalanced supply voltage driving extra circulating currents, and cracking at bar-to-end-ring connection points. High-start-frequency applications combine thermal and mechanical fatigue, making them the highest-risk category.
What load is required for reliable rotor bar MCSA detection?
Reliable detection typically requires 40% load or higher. At low load, the slip is small and sidebands fall very close to the fundamental frequency, limiting frequency resolution. At full load, sidebands are furthest from the fundamental and easiest to resolve.
What are the MCSA sideband severity thresholds for rotor bar faults?
General guidance: above -54 dB indicates a potential developing fault requiring increased measurement frequency; -54 to -40 dB indicates a developing fault to plan for at the next inspection opportunity; above -40 dB indicates significant bar damage requiring prioritized repair. Trending over time against the motor's own baseline is more reliable than applying fixed thresholds alone.
Can MCSA detect rotor bar faults on VFD-driven motors?
Yes, but it requires specialized algorithms. Variable speed drives shift the fundamental frequency and move the sidebands with it. VFD-aware MCSA platforms process the variable-frequency current signal to reconstruct the sideband pattern. Flux monitoring using a search coil at the motor frame is an alternative that is less affected by drive switching noise.
How do you calculate rotor bar fault frequencies?
The broken rotor bar sidebands appear at f1 x (1 +/- 2s), where f1 is the supply frequency in Hz and s is the per-unit slip: (synchronous speed minus actual speed) divided by synchronous speed. Use the MCSA Fault Frequency Calculator to compute sideband frequencies for any motor nameplate data.
Continuous Rotor Bar Sideband Trending
Tractian's MCSA-based monitoring continuously trends rotor bar sideband amplitude on connected motors, flagging progression before the fault reaches critical severity.
See Electrical Monitoring