Axial flux motor continuous vs peak power

An OEM specification should identify what must happen at the shaft, for how long and from what thermal state. Continuous output describes a sustained condition under stated boundaries. A timed overload describes a different condition; it must include an allowed duration and what happens before and after it.

As a document example, EMRAX 228 V1.7 separates an S2 30-second peak from S1 continuous columns and ties the columns to cooling conditions. That is evidence for this particular product document, not a universal 30-second definition of peak or a YokelessMotor rating.

ABB’s duty explainer describes S1 as constant load long enough to reach thermal equilibrium, and S2 as a limited run followed by de-energized rest sufficient to restore thermal equilibrium with the coolant. That distinction matters: an S2 label alone does not establish repeated hot-start operation. Ask the supplier to confirm the actual conditions and applicable duty designation; this guide does not certify compliance with an edition of a standard.

Use shaft mechanical power P(kW) = T(Nm) × rpm / 9550, where 9550 is a rounded conversion. At zero speed, useful shaft power is zero even if the winding carries substantial current. A maximum torque and a maximum power listed on the same page need not occur at the same speed.

Hypothetical example: a request for 100 Nm at 3,000 rpm means approximately 31.4 kW at the shaft. A separate 200 Nm request at 1,000 rpm means about 20.9 kW. The second point has more torque but less mechanical power; neither arithmetic result defines a permissible duration or temperature.

maxon explains that winding temperature and the ability to reject losses bound continuous operation, with speed-related losses influencing the available torque. This supports a general review question: does the supplier provide the relevant operating region, rather than one torque value for every speed?

For acceptance, agree the temperature measurement locations, sensor calibration, inferred hot-spot method, allowed temperatures and a quantitative stabilization criterion. Do not invent a universal soak time. Include winding, rotor and other limiting parts, not only the most convenient external sensor.

The EMRAX manual provides a useful installation counterexample: even its liquid-cooled arrangement needs rotor air renewal, and hotter coolant can reduce continuous output. This manufacturer-specific detail illustrates why a cooling label alone cannot transfer a bench rating into an enclosed installation.

For piecewise constant phase RMS current magnitudes Ij over durations tj, the cycle-equivalent value is Ieq = sqrt(sum(Ij² × tj) / sum(tj)). With the same phase resistance, average three-phase copper loss is 3 × Rphase × Ieq². This is an algebraic copper-loss screen for balanced phases; it is not a complete thermal model.

Our hypothetical 60-second cycle uses 100 A for 10 s, 40 A for 20 s and zero current for 30 s. Ieq is about 46.9 A. The time-average current magnitude is only 30 A; using 30 A for heating would understate the copper-loss screen. Resistance varies with temperature, and speed-dependent, harmonic and mechanical losses are absent here.

The order of the intervals and the initial temperatures can change the hot spot even if Ieq is identical. A zero-current rotating interval may still generate losses. A measured limit on peak current, torque or mechanical stress also remains binding irrespective of a low average.

Consider two hypothetical requests: one cold test with a long shutdown afterwards, and the same shaft overload repeated during a warm production shift. They can require different hardware even when peak kW matches. Ask for an event sequence rather than assuming the first successful pulse qualifies the second.

A timed overload test should state how the start condition is reached, what stops the event and how recovery is established. Loss of coolant flow, an already warm rotor or a drive current limit can invalidate a quoted duration. Define which conditions inhibit another event and how the controller reports that state.

Specific power is shaft output divided by the declared mass. A timed peak divided by active-part mass and a sustained point divided by a housed drive mass answer different questions. Record the included shaft, bearings, enclosure, inverter and cooling equipment before calculating a comparison.

For an aircraft concept, a heavier motor with a lighter cooling system may change the system result. For a marine installation, sustained heat rejection may govern instead. Those are engineering questions to evaluate with project evidence, not proof that one topology always wins.

Use the downloadable record to send one common brief to every candidate. The acceptance owner should approve the method, instrumentation and decision limits before the prototype run. A plot without a test configuration or raw time history is difficult to audit.

For a component-only purchase, separate the supplier component checks from the integrator motor rating. A dimensional or winding inspection cannot establish continuous system output. For a complete motor, clarify which installation and inverter conditions remain the buyer’s responsibility.

Reviewed on 21 October 2026. Manufacturer documents support the thermal and rating distinctions identified above. The matrices, hypothetical cycles and acceptance method are our editorial engineering synthesis. They do not establish any YokelessMotor model rating or supplier qualification.

The next decision needs your load sequence, initial and limiting temperatures, installation and agreed verification plan. This page addresses rating semantics and acceptance; it does not replace a thermal model, a detailed duty-cycle design or an application-specific approval.