Motor Current Signature Analysis: Method, Findings and Evaluation

What it is

Motor current signature analysis, often shortened to MCSA, is a diagnostic technique in which the current drawn by an energized, running motor is captured with a clamp-on current sensor and converted from a time waveform into a frequency spectrum. The motor itself acts as the transducer: any mechanical or electrical abnormality that modulates the magnetic field in the air gap leaves a small but recognizable imprint on the supply current. The method is applied to three-phase induction motors and generators and to synchronous machines, and it is usually performed with the machine coupled to its normal load. Because the sensor is clamped around a conductor at the starter, motor control center or switchgear, nothing has to be opened up and the measurement can be taken from a safe location. In acceptance and maintenance specifications the test is listed as optional, so it is normally added when a machine is large, critical, expensive to repair or hard to reach, or when another test has raised a question that a running-condition check might answer. It is also commonly requested as a baseline so that later surveys have a trustworthy reference to be compared with.

Why it is done

Many rotating-machine defects develop slowly and give no outward sign until the machine trips or is destroyed. A cracked or broken rotor bar, a porous end ring or a loose rotor lamination stack increases heating and torque ripple, can shed fragments into the stator winding and eventually forces an unplanned outage. Static or dynamic air-gap eccentricity, usually from a bent shaft, worn bearings or a soft foot, drives uneven magnetic pull that accelerates bearing wear and can finally let the rotor rub on the stator. Shorted turns, supply unbalance and load-side problems such as misalignment, gear damage or a pulsating process also show up. Detecting these conditions on the current signal is attractive because it needs no shutdown, no access to the machine and no sensor installed on the bearings, and it can look at the driven load as well as the motor. A baseline taken at acceptance and repeated on a maintenance cycle lets a developing fault be spotted months before it becomes a failure, so repairs can be scheduled instead of forced.

The principle

In an ideal, perfectly symmetrical motor the supply current contains essentially only the line frequency. Real machines are never perfect, and when a fault is present the rotating magnetic field in the air gap is disturbed at frequencies tied to the rotor speed and the slip. Those disturbances induce extra components in the stator current that sit as sidebands on either side of the line-frequency peak. A broken rotor bar, for example, produces sidebands spaced from the supply peak by a multiple of the slip frequency, and the height of those sidebands relative to the main peak, expressed on a logarithmic scale, grows as the damage grows. Eccentricity adds components related to the rotor rotational frequency and to the slot passing frequency, while bearing defects and load faults add their own characteristic frequencies. Because slip depends on load, the machine should be carrying a steady and reasonably substantial load during the capture; a lightly loaded motor has a very small slip, which squeezes the sidebands against the line peak and makes them hard to resolve. Good frequency resolution, which means a long enough capture, is therefore as important as the load.

How it is performed

The test is performed with the motor running at a stable operating condition, preferably near its normal load and speed. A technician with the proper training and personal protective equipment places a split-core or flexible current sensor around one phase lead at a point that is safely accessible, such as the starter or control center, and connects it to a dedicated analyzer or a data collector with a suitable input. Before recording, the nameplate speed, rated voltage, rated current, number of poles and line frequency are entered so the instrument can compute the expected slip and mark the frequencies of interest. The record must be long enough to give fine frequency resolution, and the load should be steady throughout; a swinging load, a variable-frequency drive, or a nearby machine drawing a changing current can smear the spectrum or create false peaks. Where the drive is variable-speed the setup and the interpretation differ and need extra care. Several records are usually taken and compared for repeatability. Operating conditions are noted, including the load level and how it was determined, since a spectrum can only be judged fairly against another one taken under similar conditions. The test must never be carried out by opening a live circuit or reaching into energized equipment.

What is recorded

The record states the machine identification and rating, the supply frequency and the operating load condition during capture, along with how the load was estimated. It notes the type and model class of the current sensor and analyzer, the capture length and frequency resolution used, and the date and operator. The measured spectrum, or the key sideband and harmonic amplitudes taken from it, is saved together with the slip or speed used for the analysis. Any anomalies observed, such as load fluctuations, drive interference or visible peaks that could not be explained, are written down, as is the as-found mechanical state of the machine and any related results from vibration or other tests so the readings can be interpreted together.

How results are judged

Neither ANSI/NETA ATS-2025 nor ANSI/NETA MTS-2023 attaches a single universal pass-fail number to this optional test, so judgment rests on comparison and engineering interpretation. Sideband amplitudes are read relative to the main line-frequency peak and compared with the baseline recorded when the machine was new, with the same machine on earlier surveys, and with sister machines of the same design and duty. A steady spectrum that matches the earlier record is reassuring; a rising trend or a sideband that grows against the baseline calls for corroboration with vibration data, insulation and winding tests, and a closer inspection. Severity bands published by the instrument maker or by an in-house program can help rank findings, but they are guidance, not a substitute for the standard. The same machine should be tested at a comparable load each time, because readings taken at different loads are not equivalent. Consult the current edition of the standards named here, and any applicable manufacturer data, for the governing criteria and for how results are to be reported.

Variations

At acceptance the main purpose is to document a healthy baseline on a new or reconditioned machine while it is carrying real or simulated load, so that every later survey has a reference. In maintenance the same signature is repeated on a periodic cycle, on a schedule that reflects how critical the machine is, and the trend is more important than any single result. Large and medium-voltage induction motors are the most common subjects, while synchronous machines can also be examined, with the interpretation adjusted for their different rotor construction and excitation. A variable-frequency drive, a gear train or a coupling to a fluctuating load complicates the spectrum and may call for specialized analysis. When two machines share one supply or one bus, the capture should confirm which machine the signature belongs to. The test is optional in both documents and is frequently paired with a running vibration survey and a temperature check, so findings that agree across methods carry more weight.