Axial Flux Motor Test Bench: Testing Pancake Motors for EV Drivetrains and Aerospace

Axial Flux Motors Are Entering Production — Testing Must Follow

Axial flux motors have moved from aerospace curiosities to mainstream EV components faster than most engineers anticipated. Ferrari’s SF90 Stradale uses a YASA-derived axial flux motor for its hybrid rear unit. Mercedes-Benz acquired YASA outright in 2021 to secure the technology for its AMG line-up. Emrax and Magnax supply units to aircraft, racing, and naval propulsion OEMs. Turntide and Saietta address the commercial vehicle and two-wheeler markets respectively.

The common characteristic is a pancake geometry: the rotor and stator discs are arranged face-to-face, separated by a small axial air gap. Power density values of 5–10 kW/kg at peak are routinely demonstrated, roughly double what a comparable radial flux machine achieves. But the test bench engineer faces several challenges that have no equivalent in conventional motor testing.

Why the Form Factor Changes Everything on the Test Bench

A radial flux motor has a shaft protruding from both ends of a cylindrical housing — the test bench coupling is straightforward. An axial flux motor’s output shaft may protrude from only one side of a very flat housing, or in twin-rotor / single-stator (TRSM) configurations the shaft may be of unusually short bearing span. The test bench must accommodate:

• Custom adaptor flanges: Axial flux housings rarely use standard IEC or NEMA foot mounting. A precision-machined bell-housing adaptor matching the motor’s bolt circle diameter and centring spigot is mandatory. Misalignment of more than 0.05 mm TIR at the coupling face will introduce parasitic bending moments that distort torque measurement.
• Axial load management: In dual-rotor configurations, the attractive force between rotor discs and stator can exceed the motor’s own rated thrust by a factor of three or more. The bench’s test table must be capable of resisting this axial pre-load without allowing the motor to shift, which would alter the air gap and corrupt efficiency readings.
• Thermal access: The stator in a TRSM design is enclosed on both faces by rotating disc rotors, making stator slot temperature measurement by thermocouple extremely difficult. Fibre-optic probes routed through the stator windings are preferred; alternatively, winding resistance measurement at thermal equilibrium provides an average conductor temperature.

Core Test Items

1. Back-EMF Waveform and Harmonic Analysis

Axial flux motors with concentrated windings often exhibit a back-EMF waveform that is more trapezoidal than sinusoidal. This affects how the motor interacts with a sine-wave inverter and must be characterised before control tuning. The test bench spins the motor at rated speed with no load applied; the open-circuit phase voltages are recorded at a sample rate of at least 50 kS/s. Harmonic content up to the 50th order is reported.

Key metrics: THD of open-circuit back-EMF voltage; fundamental frequency; peak line-to-line back-EMF at rated speed; cogging torque magnitude and angular period from the same spin-down measurement.

2. Efficiency Mapping (Speed-Torque Grid)

Axial flux motors commonly achieve peak efficiency above 96%, but the narrow high-efficiency island can collapse rapidly at off-design operating points. A minimum grid of 10 × 10 speed-torque points is recommended, with finer resolution (20 × 20) in the region of expected WLTP / NEDC weighted operating points. Input power must be measured on the DC bus; phase current and voltage measurement alone misses inverter switching losses.

Key metrics: Peak system efficiency (motor + inverter); efficiency at 50% speed / 50% torque; weighted efficiency per WLTP cycle.

3. Thermal Mapping Under Continuous Load

Heat dissipation is the axial flux motor’s primary constraint. The stator, sandwiched between two spinning discs, has limited convective cooling unless active liquid cooling channels are embedded in the stator yoke. Continuous load testing at 100% rated power for 30 minutes minimum is required to identify any thermal hot-spots. If the motor uses a water jacket, the coolant inlet temperature, flow rate, and pressure must be held constant throughout the test run.

4. NVH and Structural Mode Testing

The thin disc structure of an axial flux rotor is susceptible to acoustic resonances at specific speed-torque combinations. The disc can act as a drum skin, radiating structure-borne sound at twice the electrical fundamental frequency plus rotor slot harmonics. Accelerometers on the motor housing flanges and a sound power measurement in a semi-anechoic environment are standard. Identify any operating points above 80 dB(A) at 1 m early so the drive control team can apply harmonic injection to mitigate.

5. Demagnetisation Risk Assessment (PM Axial Flux)

Permanent magnet axial flux rotors (e.g., TRSM) are vulnerable to partial demagnetisation under short-circuit fault conditions because the magnet depth in the axial direction is limited by the compact disc design. Testing involves applying a controlled peak negative d-axis current pulse equal to 2× rated peak current for 100 ms and immediately re-measuring peak flux linkage. Any reduction greater than 0.5% indicates irreversible demagnetisation.

Adapting an Existing Test Bench for Axial Flux Work

Most motor test benches can be retrofitted for axial flux work without replacing the load dynamometer or power electronics. The primary modifications involve the mechanical fixturing (custom bell-housing adaptor and axial restraint frame), the thermal instrumentation (fibre-optic probes or contactless IR), and the data acquisition configuration. Our engineering team has supported axial flux programmes ranging from 10 kW e-bike mid-drives to 250 kW motorsport traction motors. Reach out to discuss your application.