AFPM Motor Concentrated Winding: Tool & Guide

1. What the screening result lets you decide

• Shortlist a connection, then validate it. The 12S/10P and 12S/14P double-layer reference layouts both have kw = 0.9330. This describes fundamental EMF utilization, not efficiency. Winding synthesis method.
• Account for frequency before increasing pole count. At 3,600 rpm, 10 poles give 300 Hz and 14 poles give 420 Hz, using fe = (2p)n/120. The higher frequency changes the inverter and loss-analysis requirements.
• Separate winding-factor quality from rotor heating. Nonworking spatial harmonics need their own loss analysis; a good fundamental factor cannot establish magnet temperature. Rotor harmonic-loss study.
• Use topology for packaging decisions. A yokeless segmented armature enables individually wound teeth, but retaining structure and cooling remain design tasks. YASA primary record.

2. Method: coil pitch, phasors and reference layouts

Q denotes slots, 2p rotor poles and m phases. This tool fixes m = 3 and assumes identical turns, one-slot coil pitch and an iron-tooth stator. Double layer winds every tooth; single layer winds alternate teeth on an even-tooth ring. Changing polarity, pitch or parallel branches requires a new winding calculation. Cros & Viarouge.

For the 12S/10P double-layer reference, coil phasors assigned to phase A resolve to ±15° about its resultant: kp = sin(75°), kd = cos(15°), so kw = 0.9330127. The alternate-tooth reference has aligned phase-A EMFs, kd = 1, and kw = 0.9659258. Each other phase must have equal magnitude and be displaced by 120°. These are ideal connection calculations, not measured motor data.

3. Slot/pole catalog: compare like-for-like layouts

Calculated reference factors below use the method above. The 9-slot family uses phase-coil angles 0°, +20° and −20° after polarity assignment. Repeated spatial layouts retain their factor. A dash means the single-layer connection is not cataloged here; it does not certify impossibility. See Salminen’s thesis for published winding-factor context, matching layer and connection before comparison.

Selection rule: compare electrical frequency at the same required mechanical speed, then evaluate harmonic fields, inverter voltage margin, manufacturability and ripple. The highest kw or LCM is not automatically the best machine.

4. Concentrated versus distributed: packaging and loss trade-offs

Nonoverlapping tooth coils reduce winding crossings and can simplify segmented-coil production. A distributed arrangement offers different spatial filtering. A useful comparison fixes active loading and operating duty before comparing drawn mean turn length and losses. Neither architecture has a universal copper-saving percentage. AFPM construction context.

5. Stator topology: what transfers to an AFPM disc

The original YASA work describes a yokeless segmented armature. Its individually wound teeth fit the tooth-coil concept, while retention and thermal paths must be designed around the dual-rotor structure. Current commercial slot counts, conductor fill, segmentation and cooling ratings are not established by this source. Woolmer & McCulloch (2007).

6. Spatial harmonics: use the rotor frame correctly

A working wave has p pole pairs; other signed spatial waves can move relative to the rotor. For positive rotor rotation n and a harmonic of ν pole pairs with direction s = ±1, its rotor-frame frequency is frotor = |s fe − ν n/60|. For a forward working wave ν = p, this is zero. Direction and order both matter. The tool does not infer a spectrum from slots alone. Pyrhönen et al. (2012).

As an illustrative kinematic check, at 3,600 rpm and fe = 300 Hz, a backward ν = 1 wave gives |−300 − 60| = 360 Hz. It is not a 1,500 Hz field merely because the working wave has five pole pairs. This example specifies direction; it makes no claim about the wave amplitude in a particular layout.

7. Magnet eddy currents: test segmentation rather than prescribing it

Magnet resistivity, dimensions, harmonic flux, electrical insulation between pieces and inverter spectrum all affect loss. A rotor-loss model needs these inputs; the calculator supplies none of them. Compare monolithic and segmented candidates under the same excitation and cooling boundary. Validated loss-model study.

Skin depth δ = √(ρ/(π f μ)) compares diffusion length with conductor dimensions. Segmentation can reduce circulating paths, but frequency and geometry can also produce a segmentation anomaly. Do not transfer an inverse-square trend to every rotor or promise a fixed percentage reduction. Plait & Dubas (2022).

8. Disc geometry and copper accounting

The active radial span is (OD − ID)/2, not OD − ID. At radius r, slot pitch is 2πr/Q. If tooth width is constant, available slot width varies with radius. Check the inner-radius bend and clearance before committing a rectangular bobbin; report net copper area after insulation and potting allowances.

For the baseline OD/ID 280/180 mm, radial span is 50 mm and mean diameter is 230 mm. The 12-slot mean tooth pitch is 60.21 mm and 10-pole pitch is 72.26 mm. Neither pitch defines the actual tooth width. To compare copper, use mass = density × total conductor area × actual mean turn length, including joints and leads. The checker does not calculate savings from assumed tooth-width multipliers.

9. Isolation, field weakening and fault limits

Separate bobbins can support physical separation, yet magnetic coupling, interturn faults and thermal propagation still require assessment. A layer choice cannot establish galvanic independence of two inverters or a remaining torque percentage. Review the actual neutral connections, protection and phase isolation architecture.

The characteristic current ψm/Ld is useful in a defined machine model, but neither quantity is known here. It cannot establish a constant-power speed range, a bounded transient short-circuit current, or freedom from irreversible demagnetization. Require temperature-dependent material data and fault analysis before making those decisions.

10. Conductor choice and manufacturing cost

Screen manufacturability alongside losses. A large wire cross-section may improve DC resistance while increasing AC losses or bend difficulty. Cooling and duty determine acceptable current density; insulation temperature class alone is not a permissible coil hotspot specification.

11. Validation plan and standards boundary

Standards checked 2026-10-24: IEC 60034-15:2025 replaces the 2009 edition and excludes converter-fed windings. For Type I insulation in PWM-fed machines, review IEC 60034-18-41 and its applicable amendments. Agree classification, terminal waveform, sample type and test limits with the responsible engineer. The following is a project planning checklist, not a normative acceptance specification. IEC 60034-15; IEC 60034-18-41.

Before serial tooling, resolve the inner-radius coil clearance, tooth retention, magnet bonding and potting process. Quote setup, scrap, inspection and rework separately from recurring part cost. Public evidence here does not establish a supplier MOQ, lead time or commercial teardown specification.

12. Five illustrative screening scenarios

These are input-driven exercises, not measured product case studies. Use the matching tool preset to reproduce each result. Torque, efficiency, temperature, flight capability and manufacturing savings are unknown.

13. Frequently asked decision questions

14. Primary sources, scope and review date

Reviewed 2026-10-24. Source links identify the underlying work or publisher scope statement. Some full texts require access; the sources do not verify proprietary product details or certify this checker.

1. Gieras, Wang & Kamper — Axial Flux Permanent Magnet Brushless Machines

   Springer, 2nd edition, 2008

   Supports: AFPM topology, fabrication and cooling context.

   Boundary: Book overview is public; detailed chapters may require access. No supplier performance is inferred.

2. Cros & Viarouge — Synthesis of high performance PM motors with concentrated windings

   IEEE Transactions on Energy Conversion 17(2), 248–253, 2002

   Supports: Reference method for three-phase concentrated winding synthesis.

   Boundary: A winding method is not a validated AFPM product. Catalog values here are analytical reference layouts.

3. Salminen — Fractional slot permanent magnet synchronous motors for low speed applications

   LUT doctoral thesis, 2004

   Supports: Fractional-slot winding factors and harmonic context.

   Boundary: Layer, pitch and phase connection must match before transferring a published factor. Repository may challenge automated access.

4. Woolmer & McCulloch — Analysis of the yokeless and segmented armature machine

   IEMDC 2007, pp. 704–708; DOI 10.1109/IEMDC.2007.382753

   Supports: Primary YASA topology record: independent wound segments without a stator yoke.

   Boundary: Does not disclose current commercial models, coil fills or fault certification.

5. Bianchi & Bolognani — Design techniques for reducing the cogging torque in surface-mounted PM motors

   IEEE Transactions on Industry Applications 38(5), 1259–1265, 2002

   Supports: Cogging mitigation depends on geometry and can affect back-EMF.

   Boundary: Repetition order alone cannot predict amplitude. University record provides abstract and DOI 10.1109/TIA.2002.802989.

6. Pyrhönen et al. — Harmonic Loss Calculation in Rotor Surface Permanent Magnets—New Analytic Approach

   IEEE Transactions on Magnetics 48(8), 2012; DOI 10.1109/TMAG.2012.2190518

   Supports: Harmonic-dependent rotor loss model compared with 3D FEA and measurements, including an axial-flux machine.

   Boundary: Loss calculation needs harmonic flux and magnet geometry/material inputs absent from this screening tool.

7. Plait & Dubas — Volumic Eddy-Current Losses in Conductive Massive Parts with Experimental Validations

   Energies 15(24), 9413, 2022

   Supports: Skin-depth effects and conditions under which segmentation can increase losses.

   Boundary: Experimental conductive-part geometry is not a universal NdFeB AFPM rotor benchmark.

8. IEC 60034-15:2025 — Impulse voltage withstand levels of form-wound stator coils

   Edition 4.0, published 2025-06-06; checked 2026-10-24

   Supports: Supersedes the 2009 edition; covers grid-connected form-wound coils and excludes converter-fed coils.

   Boundary: Cannot justify a universal surge voltage for inverter-fed AFPM traction windings.

9. IEC 60034-18-41:2014+AMD1:2019 CSV — Type I converter-fed insulation systems

   Edition 1.1; includes December 2020 corrigendum; checked 2026-10-24

   Supports: Qualification and quality-control scope for Type I insulation under voltage-source PWM converters.

   Boundary: Applicability and test levels require the actual insulation system and terminal waveform; not a general motor certification.

15. Turn the shortlist into a reviewable specification

Use the tool brief as a starting point. Add a signed coil diagram, turn count, conductor cross-section, duty, inverter envelope, cooling and acceptance criteria before requesting an engineering or manufacturing quote.

• AFPM design and validation
• Back-EMF and voltage margin
• Bearing and rotor-load checks
• Assembly and process planning
• Axial-flux topology overview
• Power-envelope screening