Current-Spectrum Triage: Using MCSA on Bosch-Driven Plant Without Overselling It

Motor current signature analysis, usually shortened to MCSA, has a persuasive appeal for maintenance teams: the motor is already connected, the drive is already measuring current, and no additional transducer needs to be mounted on the bearing housing. That appeal is justified, but only within limits. Used carefully, MCSA can provide an early indication of rotor-bar deterioration, air-gap eccentricity and some forms of coupling or alignment wear. Used as a substitute for vibration monitoring, it will disappoint.

The practical question for UK engineers is therefore not whether MCSA works, but where it fits in a condition-monitoring regime alongside Bosch drives and motors. The answer lies in treating the current spectrum as a triage tool: a way to rank suspicion, direct further investigation and establish a repeatable trend.

What the current spectrum actually contains

Stator current is rarely a clean sinusoid. Even on a healthy machine fed from a stiff 400 V 50 Hz supply, the spectrum contains supply harmonics, slot-passing components and load-induced modulation. Once a variable-speed drive is introduced, the picture gains inverter switching components, control-loop effects and DC-link ripple. MCSA does not remove this complexity; it asks the analyst to separate fault signatures from normal operating artefacts.

The signatures of interest are sidebands. Rotor-bar damage typically produces sidebands around the supply frequency at twice slip frequency. Dynamic eccentricity can produce components related to rotational frequency. Coupling wear, misalignment and torsional variation may appear as modulation around the fundamental or around slot harmonics. These are indirect measurements. The current spectrum shows that torque or flux is being modulated, not which mechanical component is responsible.

That distinction matters. A sideband is evidence of modulation, not a diagnosis. The same spectrum may be produced by a genuine fault, a pulsating load, a control-loop interaction or a supply anomaly. The technique earns its place when it is used comparatively: against a known baseline, across phases, and at stable operating points.

Where MCSA earns its keep

MCSA is most useful where access is difficult or where vibration measurement is impractical. Motors mounted in congested skid frames, submersible or enclosed machines, and repeat-duty assets in food or water plants can all benefit from a current-based screen. It is also useful for rotor integrity checks, where the current spectrum responds directly to rotor asymmetry, and for trend monitoring where a stable load profile exists.

On Bosch-fed machines, the drive itself may provide useful current data. This is not automatically an MCSA measurement. Many drives expose current values for control and protection, not raw phase-current waveforms at sufficient bandwidth and resolution for spectral analysis. Where the drive offers high-resolution current logging or a dedicated diagnostic function, it can form part of the evidence. Where it does not, a portable analyser with a suitable current probe is the more defensible route.

The technique should be positioned as a screening layer. It can flag a change, prompt a vibration check or an offline rotor test, and provide evidence that a suspected fault is developing. It should not be used to clear a machine for continued service on the basis of a single clean spectrum.

The inverter caveat

Inverter-fed motors present particular difficulty. The drive controls current continuously, so a fault signature may be suppressed, shifted or modulated by the control loop. Switching ripple can obscure low-amplitude sidebands. The current sensor bandwidth and the analyser sampling rate must be adequate for the frequencies of interest. A portable analyser connected to a motor phase conductor is generally preferable to relying on a drive display, provided the probe is rated for the environment and the measurement is taken consistently.

Phase comparison is valuable. A rotor fault may produce similar sidebands in all phases, while a supply imbalance may not. Recording more than one phase, and repeating at the same load and speed, helps separate machine effects from supply effects.

A UK plant trial checklist

A defensible trial begins with a defined question. Decide whether the aim is fault confirmation, baseline creation or trend monitoring. Do not set a blanket pass or fail threshold before the machine has a baseline.

• Capture a genuine baseline. Record at least three repeat spectra at the same speed, load and thermal condition. Note ambient temperature, drive parameters and the date. A baseline taken at a different load point is not a baseline.
• Use stable load conditions. Avoid no-load or very light load, where fault sidebands may be weak or masked. Capture at two or three stable points, ideally including normal duty and a representative high-load condition. Record whether the machine is inverter-fed and note the switching frequency.
• Fix sensor placement. Use the same phase conductor, probe orientation and analyser settings for every repeat. Clamp around a single conductor, not a multicore cable. Keep the probe away from adjacent conductors and noisy power cables. For permanent installations, label the measurement point.
• Record supply-side artefacts. Note 5th and 7th harmonic levels, capacitor-bank switching, neighbouring loads and any recent supply changes. Compare spectra taken at different times. A component that moves with the supply, rather than with motor load, is unlikely to be a rotor fault.
• Do not chase normal slot harmonics. Slot-passing frequencies and their multiples are normal. Mark them on the baseline and judge deviations from that baseline, not their absolute presence.
• Confirm before intervention. A current-spectrum anomaly should be supported by vibration data, thermography, offline resistance or rotor tests, or a repeat measurement under controlled load. Do not authorise a motor removal on a single spectrum.
• Keep the record. Store the raw data, analyser settings, load conditions and interpretation together. A future analyst needs the context, not just the image.

Where MCSA falls short

MCSA is weak at localising mechanical faults. A bearing defect that is obvious in a vibration spectrum may barely register in current, particularly in the early stages. Lubrication problems, localised gear damage and looseness are better detected by vibration or ultrasound. Supply unbalance and load variation can create components that look like fault signatures to an inexperienced analyst. Inverter control can mask or shift the very sidebands being sought.

Vibration monitoring also offers spatial information: which bearing, which direction, which component. MCSA offers a system-level view. That makes it a useful complement, not a replacement.

The sensible position for a UK plant is to use MCSA where it is technically credible, document its limits and keep it in its proper place: a current-spectrum triage method that directs attention, supports trending and helps avoid unnecessary intervention. It is not a vibration sensor disguised as a drive.