Motor current signature analysis (MCSA) diagnoses faults inside a running induction motor by measuring its current and studying the frequency spectrum: specific faults print telltale sidebands around the line frequency, so the current itself becomes the sensor. The signature technique is best known for catching broken rotor bars, a fault you cannot see, hear, or usually feel from the outside until the motor is in real trouble.
MCSA belongs in the condition-based maintenance toolbox alongside vibration, thermography, and insulation resistance testing. Its two standout traits: it reads current from the supply without touching the machine, and it looks straight through the motor casing at the rotor, which most other methods struggle to inspect while the machine runs. This guide explains how the sidebands form, what other faults current reveals, how to judge severity, and how to run a test.
What is motor current signature analysis?
It is a spectral technique: record the stator current of an operating motor, transform it to the frequency domain, and read the fault frequencies. A perfectly healthy motor running on a clean supply would show almost all of its current energy at one place, the line frequency, 50 or 60 Hz. Real motors never look that clean, and the interesting part is where the extra energy shows up. A broken rotor bar, an eccentric air gap, and a load problem each add current components at their own predictable frequencies. Knowing the formulas for those frequencies turns a messy spectrum into a diagnosis.
The appeal for maintenance teams is practical. You clamp a current transducer on the motor's supply leads, often at the motor control center, safely away from the machine, capture a few seconds of data under load, and process it. No shutdown, no uncoupling, no opening the motor. For large, hard-to-reach, or continuously running motors, that non-intrusive quality is the whole game.
How does MCSA detect a broken rotor bar?
A broken or cracked rotor bar disturbs the rotor's magnetic field as it spins. That disturbance modulates the current the stator draws, and the modulation shows up as a pair of sidebands at the line frequency plus and minus twice the slip frequency the classic f ± 2sf signature, where f is line frequency and s is the slip.
Slip is the small gap between the motor's synchronous speed and its actual running speed. Take a 60 Hz motor running at 2% slip: twice the slip frequency is 2 × 0.02 × 60 = 2.4 Hz, so the rotor-bar sidebands sit at about 57.6 Hz and 62.4 Hz, very close to the towering 60 Hz line peak. Pulling those small sidebands out from beside a peak that may be tens of thousands of times taller is why MCSA needs fine frequency resolution and several seconds of clean data.
How do you tell how bad the rotor is?
Read the sideband height in decibels below the main line peak. The bigger the difference, the healthier the rotor; the closer the sidebands climb toward the peak, the worse the damage. These are working guidelines, not a hard standard, the exact thresholds vary with motor design and analyst, but they orient the reading:
| Sideband level vs. line peak | Typical interpretation |
|---|---|
| 50–60 dB down | Healthy rotor; no significant bar breakage |
| 45–50 dB down | Early degradation; establish a trend and recheck |
| 40–45 dB down | One or two broken rotor bars likely |
| 25–35 dB down | Multiple broken bars or a cracked end ring; plan repair |
Two cautions. First, this only works with the motor properly loaded, at light load, slip is tiny, the sidebands crowd against the line peak, and the reading gets unreliable. Second, the sidebands are meaningful only when you trust the slip value, because that is what tells you where to look. A speed reading or a nameplate-plus-load estimate anchors the whole analysis.
What faults can MCSA find besides broken rotor bars?
Rotor bars are the headline, but current carries more than one story. Air-gap eccentricity, a rotor not centered in the stator, from a bent shaft, worn bearing, or bad assembly, adds its own frequency components tied to rotor speed. Load and coupling problems, and issues in the driven equipment, modulate current at the rotational and vane- or tooth-passing frequencies of the machine being driven. Some bearing faults also leave a current fingerprint, though vibration usually reads bearings more clearly.
Where MCSA is weak: winding insulation condition and turn-to-turn shorts are better assessed with insulation resistance and PI testing and surge testing, and many mechanical and bearing faults read more cleanly on vibration. Treat current analysis as the specialist for rotor and electrical-load faults, not a universal instrument. That is why serious electric-motor maintenance programs layer several techniques.
Why does MCSA beat opening the motor to inspect the rotor?
Because the fault you most want to find is the one you cannot reach. Broken rotor bars are buried in the rotor cage, sealed inside the machine, spinning at full speed. To inspect them directly you would uncouple the motor, pull it, and disassemble it, hours of skilled labor and a production outage, often to confirm the rotor was fine. MCSA answers the same question from the motor control center in a few minutes while the machine keeps running.
That inverts the usual condition-monitoring bargain. Most inspection methods trade access for downtime; current analysis gives you a look inside a running, loaded motor with neither. The tradeoff is interpretation: a current spectrum is not a photograph, so it takes a trained analyst, a trustworthy slip value, and ideally a trend of past readings to separate a real rotor fault from supply harmonics or load noise. Get those three things right and MCSA turns an invisible, buried failure mode into a number you can trend and plan around.
How do you run an MCSA test?
- Confirm the motor is loaded. Run it at a representative, steady load, ideally above roughly half load. Low load shrinks slip and pushes the fault sidebands into the line peak where you cannot resolve them.
- Clamp a current transducer on the supply. A current clamp or CT on one phase, typically at the motor control center, feeds a data collector. Follow your plant's electrical-safety and lockout rules for working around energized conductors.
- Capture enough data at fine resolution. You need several seconds of steady-state current so the analyzer can resolve sidebands sitting within a couple of hertz of the line frequency. Note the running speed or slip while you capture.
- Transform and locate the fault frequencies. Compute the spectrum, mark the line peak, then look for sidebands at f ± 2sf for rotor bars and at the eccentricity and load frequencies for the other faults.
- Measure severity and trend it. Record the sideband level in decibels below the peak. One reading is a snapshot; the value comes from watching that number climb over successive tests on the same motor.
- Confirm and act. For a serious finding, repeat the test and cross-check with vibration or a physical inspection before you pull a motor. Then turn the diagnosis into a planned maintenance work order rather than a surprise failure.
What the numbers say
- Motor faults caught early are cheap; caught late they are outages. The U.S. Department of Energy's FEMP O&M guidance, maintained by PNNL, documents that condition-based programs, the family MCSA belongs to, save 8–12% over preventive-only maintenance, and the opportunity versus heavily reactive operation can exceed 30–40% (PNNL, O&M Best Practices: Maintenance Approaches). A broken rotor bar found at 45 dB and repaired on schedule is a work order; found at failure it is a line-down event.
- The technicians who run these tests are in short supply. The U.S. Bureau of Labor Statistics projects 13% employment growth (2024–2034) for industrial machinery mechanics, maintenance workers, and millwrights, with roughly 54,200 openings a year (BLS Occupational Outlook Handbook). Non-intrusive tests that need no shutdown stretch that scarce labor further.
Where does MCSA fit in a reliability program?
Think of current signature analysis as one instrument in a condition-monitoring program not a standalone answer. It shines on rotor-bar and electrical-load faults and it earns its keep on big, critical, hard-to-reach motors precisely because it needs no shutdown. Pair it with vibration for mechanical faults, insulation testing for winding condition, and thermography for connections, and you cover the failure modes that actually take motors down. The broader case for catching degradation before failure is the subject of our predictive maintenance and equipment reliability guides, and the common failure modes are cataloged in electric motor failure causes.
The catch, as always, is data discipline. A single spectrum is a snapshot; the diagnosis lives in the trend, which means capturing readings on a schedule, storing them against the right asset, and turning findings into planned work. Plants that feed condition data into one operational layer, the way Harmony pulls floor data together for machine monitoring get to act on a rising sideband weeks before it becomes a 2 a.m. failure. For how one plant built that trustworthy data foundation, see the CLS case study.