Nameplate Ratings Are a Starting Point, Not a Comparison Method

A technical comparison by performance under load should answer a narrower and more consequential question than “Which unit has the higher rating?”: which solution remains stable, efficient, controllable, and compliant across the load profile it will actually see?

Two devices with the same nominal power can behave very differently at 25%, 75%, 100%, and transient overload. A drive may deliver an attractive peak efficiency while producing unacceptable harmonic current at part load. A transformer may meet its rated temperature-rise limit in a controlled test but have limited thermal margin when installed in a high-ambient enclosure. A motor may achieve its nominal output at rated frequency yet lose usable torque, cooling capability, or insulation life when paired with a variable-speed drive.

The comparison therefore needs to move from a single rated point to a verified operating envelope. That envelope includes steady-state loading, cyclic loading, overload duration, supply-quality sensitivity, ambient conditions, cooling arrangement, control behavior, and interaction with the connected network or machine.

Define the Load Before Comparing Equipment

The most common weakness in technical evaluations is comparing supplier data obtained under different load assumptions. A declared efficiency at full load and unity power factor cannot be treated as equivalent to field performance in an application dominated by partial load, variable torque, frequent starts, regenerative operation, or distorted supply voltage.

A useful load definition contains more than minimum, normal, and maximum power. It should specify:

• the time spent at each load band, including idle, partial-load, rated-load, and peak-load periods;
• the shape of the load: constant torque, variable torque, constant power, intermittent duty, cyclic duty, or pulsating duty;
• the rate of load change and the expected number of starts, reversals, or step changes;
• the permitted overload magnitude and duration;
• source conditions, including voltage tolerance, frequency variation, available short-circuit level, and background harmonic distortion;
• site conditions such as altitude, ambient temperature, enclosure ventilation, contamination, and cooling-water temperature where relevant.

For motion systems, a pump and a conveyor may have the same rated motor power but impose fundamentally different demands. Pump loads generally follow a variable-torque pattern, so part-load efficiency and low-speed control behavior may matter more than short-term overload. Conveyors, hoists, crushers, extruders, and rolling equipment can require sustained torque at low speed, repeated acceleration, or high breakaway torque. A comparison based only on motor kW or inverter current rating will not expose these differences.

For grid-connected equipment, the relevant load is also electrical. A feeder with a high share of nonlinear loads, distributed generation, capacitor banks, or rapid demand changes should not be assessed as if it were a balanced, sinusoidal, constant-load system.

Compare Curves, Not Isolated Numbers

Performance curves reveal whether a quoted rating is representative or merely a favorable point. The most useful comparison is built around identical load points and identical boundary conditions. For each candidate, record values at a defined sequence of load levels—for example, no-load where meaningful, 25%, 50%, 75%, 100%, and the specified overload point. The chosen points should reflect the duty profile rather than an arbitrary template.

Efficiency is central, but it must be interpreted with the associated losses. The basic relationship is:

η = P_out / P_in

At low load, fixed losses can dominate. Magnetizing losses in transformers, control power consumption in converters, fan power, and core losses in rotating machines do not fall in proportion to useful output. A unit with a modestly better full-load efficiency can therefore consume more energy over its annual duty cycle if it performs poorly at the load levels where it operates most often.

Loss separation helps explain the result. In a transformer, no-load loss and load loss respond differently to loading. In a motor, stator and rotor copper losses, iron losses, mechanical losses, and stray-load losses change differently with speed and torque. In a variable-frequency drive, semiconductor conduction loss, switching loss, DC-link loss, cooling consumption, and filter losses are affected by current, switching frequency, output frequency, and control mode.

Efficiency figures should also state whether auxiliary consumption is included. A power conversion system evaluated without its external cooling unit, harmonic filter, control transformer, or enclosure ventilation may appear materially better than the installed system actually is.

Thermal Performance Determines the Usable Rating

Electrical equipment fails under load less often because of an immediate power shortfall than because heat accumulates beyond what the insulation, semiconductors, bearings, capacitors, or contacts can tolerate. Thermal comparison should therefore examine temperature rise, thermal time constants, hot-spot behavior, cooling dependence, and derating rules.

A steady-state temperature test at rated load is necessary but incomplete. Equipment with significant thermal mass can survive a short overload even if it cannot sustain the same load continuously. Conversely, repeated peaks may prevent meaningful cooling between cycles. The thermal question is not simply whether an overload is permitted; it is whether the stated overload profile is compatible with the equipment’s thermal model and protection settings.

For motors, the applicable duty classification and temperature-rise limits should be considered in conjunction with IEC 60034-series requirements and the insulation system used. Where a motor is inverter-fed, the evaluation should also address low-speed cooling, voltage stress associated with fast switching edges, bearing-current risk, and the suitability of the insulation system for converter supply. A self-cooled motor that performs acceptably at base speed may not provide the same continuous torque at low speed without forced ventilation or a revised derating curve.

For switchgear and busbar systems, temperature rise depends not only on nominal current but also on conductor arrangement, enclosure design, joint resistance, ventilation, proximity effects, and installation spacing. For transformers, ambient temperature, altitude, cooling mode, harmonic current, and load-cycle profile all affect hot-spot temperature. A technically valid comparison should identify the reference ambient and the assumed cooling conditions behind every thermal claim.

Overload Capability Must Be Read With Its Time Basis

“150% overload” is not a complete specification. It may mean 150% current for 60 seconds, 150% torque for a limited interval at a specified output frequency, or a capability available only when the unit begins from a defined thermal state. Some drive ratings distinguish between normal-duty and heavy-duty operation; this distinction can alter the required frame size even when the connected motor has the same nominal power.

Comparison should establish four linked values:

• the overload magnitude available at the equipment output;
• the permitted duration and repetition interval;
• the initial thermal condition assumed by the manufacturer;
• the limiting mechanism: current limit, semiconductor junction temperature, torque limit, voltage limit, mechanical constraint, or protection setting.

For a drive system, torque capability should be checked across the relevant speed range rather than at one nominal speed. In the field-weakening region, available torque decreases as voltage becomes the constraining factor. At very low speed, sensorless control may have lower torque accuracy or dynamic response than encoder-based control, depending on the control design and application conditions. These are application constraints, not minor control features.

For generators, UPS systems, converters, and energy-storage interfaces, overload behavior must be evaluated together with voltage recovery and fault contribution. A unit can support a temporary current increase while still failing to maintain acceptable voltage for downstream loads or protection coordination.

Power Quality Is Part of Loaded Performance

A system that delivers its rated kW but injects excessive harmonic current, causes voltage distortion, or has poor reactive-power behavior is not performing satisfactorily at the point where it connects to the network. Harmonic comparison must distinguish between device emissions and system-level compliance.

IEEE 519 is commonly used as a framework for evaluating harmonic distortion at the point of common coupling, with limits dependent on system conditions such as the ratio of available short-circuit current to maximum demand load current. It should not be reduced to a generic claim that a single drive or converter is “IEEE 519 compliant.” Compliance is determined by the installation, source impedance, connected loads, filters, and operating state at the point of evaluation.

IEC 61000 standards address electromagnetic compatibility and harmonic-related requirements across different equipment categories, while IEC 61800-3 addresses EMC requirements and test methods for adjustable-speed electrical power drive systems. The relevant standard edition, equipment category, installation environment, cable arrangement, filter configuration, and test setup must be aligned before comparing declarations.

Under load, assess at least total harmonic distortion, individual harmonic components where they affect the network, displacement power factor, true power factor, unbalance tolerance, and sensitivity to voltage dips. For active-front-end drives and grid-forming or grid-following converters, reactive-power control range, current limiting behavior, and response during voltage disturbance can be more important than a low harmonic figure under ideal supply conditions.

Control Response Should Be Tested as a System Property

Control performance is often presented through attractive terms such as “fast response,” “high precision,” or “stable regulation.” These descriptions are insufficient unless the command, measurement point, load disturbance, and acceptance band are defined.

A meaningful dynamic test uses a repeatable disturbance: a load step, speed command step, torque reversal, grid-voltage variation, or transition between operating modes. The comparison records overshoot, settling time, steady-state error, oscillation, current excursion, and protective trips. The test should also state whether the system includes the actual motor, gearbox, inertia, transformer, filter, and cable length expected in service.

In converter-fed motor systems, a short cable and a laboratory motor can conceal issues that emerge with long motor leads, output filters, high-inertia loads, or weak supply systems. In grid equipment, a stiff test source can conceal control interactions that become visible when the short-circuit ratio is lower. The objective is not to reproduce every site condition in a factory test, but to identify which assumptions the published response depends on.

Normalise the Test Boundary

Technical comparisons lose validity when one proposal includes the equipment needed for reliable operation and another excludes it from the stated performance result. The comparison boundary should be explicit.

This discipline is particularly important when comparing integrated solutions against component-based assemblies. An integrated unit may include cooling, filtering, protection, and coordinated controls within its published rating. A component proposal may require separate equipment to achieve the same operating result. Neither approach is automatically superior, but their loaded performance cannot be compared fairly until both are evaluated at the same functional boundary.

Use Standards as a Common Language, Not a Substitute for Duty Definition

Standards make comparisons more reliable by defining test methods, rating conventions, safety expectations, and terminology. They do not eliminate the need to verify application-specific conditions. IEC 60034-1, for example, provides a foundation for rotating electrical machine ratings and performance, but it does not itself determine whether a particular motor-drive combination will meet a low-speed, high-torque process duty. IEC 60076 provides a framework for power transformer requirements, yet harmonic loading and site cooling can still require additional assessment. IEC 61800-series documents support the evaluation of adjustable-speed drive systems, but installation-dependent EMC and motor compatibility remain essential.

The appropriate question is: which standard governs the declared value, under what test condition, and does that condition correspond to the intended duty? A certificate or declaration is valuable evidence only when its scope matches the supplied configuration.

Turn Supplier Claims Into Comparable Evidence

A robust technical comparison by performance under load should request raw or traceable evidence rather than relying only on brochure tables. Useful documentation includes guaranteed efficiency maps, loss breakdowns, thermal derating curves, overload-time curves, harmonic spectra under stated load and source conditions, protection coordination data, and test reports identifying instruments and test arrangements.

Where guarantees are required, acceptance criteria should use measurable language. “Stable operation” can be converted into defined voltage deviation, speed error, temperature limit, recovery time, or harmonic threshold at stated conditions. “High overload capacity” can be converted into current or torque magnitude, duration, repetition, and allowable thermal state. “Grid-ready” can be converted into identified ride-through, reactive-power, fault-current, EMC, and interoperability requirements applicable to the connection point.

The strongest result is not the equipment with the best isolated number. It is the option whose performance evidence remains valid when load profile, thermal environment, network condition, auxiliaries, and control interactions are placed inside the same evaluation boundary. That is where nominal equivalence gives way to a technically defensible selection.