Axial Flux Motor Manufacturing Feasibility

1. Decision summary: what makes an axial-flux design manufacturable?

Axial flux motor manufacturing feasibility means demonstrating a controllable assembly load path, repeatable magnet location and a measurable clearance budget before prototype release. The worksheet above identifies gaps in that review; it does not qualify a design.

Four decisions should leave the review with an owner and a requested piece of evidence:

The pressure model follows MIT's magnetic pressure explanation. Joint qualification and sensing limits are linked in sections 6 and 10. The decision workflow and review triggers here are editorial engineering screening guidance, not standard-mandated acceptance limits.

This page owns tooling, magnet placement and dimensional inspection decisions. Use Special Motor Topologies for the broader service scope, axial-flux cooling for heat-extraction decisions and air-gap control for a separate deflection budget.

2. Required inputs: separate known values from placeholders

Every numeric field is required to run this bounded model. Defaults are illustrative placeholders, not conservative design allowables. If an actual value is unknown, retain a clearly marked example for exploration or submit drawings with TBD values; do not treat the result as project evidence.

The tool accepts an active annulus with OD 80–500 mm, ID 40–350 mm and at least 20 mm diametral separation. These are worksheet operating limits, not proven manufacturing capabilities. Pole count changes the inspection checklist; core and retention choices record process intent. They do not simulate electromagnetic performance or material strength.

3. Topology and assembly load paths

These are layout implications to investigate, not comparative torque-density ratings. No universal torque-per-mass ranking or numerical bearing load is claimed.

4. Magnetic pressure: an estimate at a field condition

For a predominantly normal, uniform air-gap field, the simplified pressure and one-interface force are:

p = B_normal² / (2 × mu_0)
A = pi × (OD² − ID²) / 4     [diameters in metres]
F_est = p × A × 0.78         [newtons]
mu_0 ≈ 4 × pi × 10^-7 H/m

The field-energy pressure basis is described by MIT OpenCourseWare. Applying it to a uniformly covered annulus is this worksheet's approximation. Fringing, tangential field, local saturation, skew, current state and the complete assembly sequence are omitted.

Changing the gap does not change the input B in this calculator. It does not calculate a force-gap curve, transient snap peak or an inverse-square force law. Request a measured or FEA curve across the intended assembly travel.

For the illustrative default OD 220 mm, ID 130 mm and B = 0.88 T:

A ratio below 2.0× prompts review in this worksheet. That trigger is an illustrative policy choice. Exceeding it does not establish safe tooling. The fixture designer still needs load combinations, moments, stiffness, physical stops, controlled motion, restraint and protection against back-driving. Verify tooling materials for both mechanical and magnetic suitability rather than assuming a material name guarantees non-magnetic behaviour.

5. Magnet placement: datums, polarity and cure restraint

Use a hub-referenced nest or indexing tool to control radial and angular location. Neighbouring magnet interactions can include attraction, repulsion and torque depending on geometry and magnetization. A universal statement that all alternating neighbours repel is not a sound assembly plan.

Record the intended pole map, inspect polarity before irreversible cure and hold each magnet against its locating features during the approved cure cycle. Confirm that restraint does not damage coatings or distort the rotor. Set pitch, height and bondline acceptance limits from the electromagnetic and mechanical error budgets; this page supplies no universal ±angle or clamping-pressure specification.

No universal annual-volume threshold, pulse current, saturation field or capital cost separates these options. Obtain supplier trials before making that production decision.

6. Adhesive and mechanical retention: demand is not capacity

The tool uses an outer-radius acceleration as a conservative demand bound for a uniform-thickness magnet bonded across its full face:

v_tip = pi × OD × rpm / 60
acceleration_outer = v_tip² / r_outer
shear_demand = density × magnet_thickness × acceleration_outer

For the default 220 mm rotor at 4,500 rpm, with assumed density 7,600 kg/m³ and thickness 5 mm, tip speed is 51.84 m/s and shear demand is 0.928 MPa. These are calculated illustration values. A smaller effective bond area increases average demand; local peel, load sharing, temperature gradients, coating adhesion, fatigue and overspeed are not evaluated.

Retention safety factor remains unknown for every option. A mechanical pocket, composite hoop or metal ring has a different load path from a full-face adhesive joint; selecting one cannot assign it a generic allowable MPa value. There is no universal tip-speed cutoff that qualifies adhesive-only retention. Carbon-fibre composites can be electrically conductive; do not assume zero sleeve eddy-current loss.

ASTM D1002 describes apparent lap-shear testing of bonded metal specimens; it does not provide a motor joint design allowable. ASTM D3166 addresses adhesive shear fatigue testing; it supplies no universal endurance limit. ASTM D2093 concerns plastic surface preparation, so it is not a metal/NdFeB plasma-treatment recipe. Agree surface preparation, spacers and cure with the adhesive and coating suppliers, then qualify representative joints. Public evidence is insufficient here to rank named adhesives by hot strength or fatigue life.

7. Stator manufacturing: compare the process, not just the material name

A specific material record illustrates why properties cannot be mixed: Höganäs' Somaloy 700HR 3P at 600 MPa compaction, 80°C die temperature and 530°C steam heat treatment lists the following typical specimen properties. See the manufacturer's 2025 material-data PDF, printed page 6.

Do not turn these specimen values into tooth release limits. Tool draft, fillets, winding tension and handling fixtures need part/process trials. Changing grade can also change magnetic, thermal and dimensional behaviour; no generic polymer-reinforcement mechanism or guaranteed strength multiplier is asserted.

Copy or view the worksheet above to record open evidence, then Submit a Project with your drawing and TBD inputs.

8. Runout, bond height and the assembly datum

Record face TIR at a stated radius and against the bearing/hub datum used in the assembly. Sweep more than one radius to distinguish tilt, dish and local magnet-height variation. A magnet-face TIR measurement may already include bondline variation; adding both values again would overstate closure.

The tool deliberately uses full TIR as a one-sided conservative allowance. It does not assume TIR is a ± tolerance or halve it without a datum analysis. The same rule applies to stator flatness and bearing play: translate each drawing/measurement into a unique closing contribution first.

Rotor balancing is a separate task. ISO 21940-11:2016, with its listed amendment, covers procedures and tolerances for rotors with rigid behaviour and supersedes ISO 1940-1. Select the applicable rotor model, grade and correction planes from the actual machine requirements; topology alone does not mandate G1.0 or two-plane correction.

9. Cold gap budget and RSS limitations

The default calculation uses five closure allowances. No thermal term is silently added:

These calculated values match the default worksheet. They describe a cold dimensional budget, not a measured motor. The entered reserve is a budget to be justified against thermal growth, magnetic bending, bearing compliance, shock and measurement uncertainty. It does not calculate those effects or guarantee operating clearance.

Arithmetic RSS of maximum drawing bounds is not a statistical clearance prediction. The tool shows it only to explain sensitivity. A production statistical model needs measured distributions, consistent sigma/tolerance conventions, correlations, mean shifts and an agreed acceptance risk; a Cpk number by itself is insufficient.

10. Enclosed metrology: access, orientation and calibration

A radial access hole does not automatically let a straight probe measure an axial face. Show the probe tip orientation, viewing direction, reference offsets and reachable radii in the section drawing. Where direct access is impossible, investigate a calibrated transfer measurement or temporary assembly instrumentation and document its uncertainty.

These sensing dependencies follow Micro-Epsilon's eddyNCDT technical guidance and capacitive measurement explanation. The actual sensor and installation determine suitability. No universal percentage error, conductivity table, four-port arrangement or harmonic rejection limit is claimed here.

11. Three reproducible scenarios, not customer test results

The following are hypothetical worksheet exercises. No client, test report, achieved life or production yield is implied. Load the corresponding preset and use the stated inputs; no other fields need editing.

The numerical differences come from entered dimensions and allowances, not from a claim that one application or topology is inherently approved. Changing a retention option leaves load demand unchanged because qualified capacity is unknown.

Reproduce the default worksheet

Original software capture, 3 November 2026. The unmodified default uses OD/ID 220/130 mm, per-side gap 0.850 mm, B 0.88 T, speed 4,500 rpm and closure sum 0.165 mm: cold gap 0.685 mm, headroom 0.385 mm, pull 5.946 kN. This is reproducible software output, not measured motor performance or customer proof. The capture isolates the worksheet with site navigation hidden. Restore the example above to reproduce it.

12. Evidence-led prototype review and release

The responsible engineering team defines the release criteria. Refer to the applicable machine/application standards and their current editions under the project agreement. This guide is not a substitute for those documents.

13. Evidence and model boundaries

Sources checked: 3 November 2026. Only public catalog scopes, manufacturer data and open teaching material were used. Proprietary test data, detailed paid-standard clauses, joint allowables, prices, cycle times and supplier capability are not established here.

The 0.78 pole coverage, 7,600 kg/m³ density, worksheet ranges, presets and 2.0× fixture review trigger are explicit illustrative assumptions. The closure sum and kinematic equations are reproducible calculations. All process recommendations are items for project review rather than statements of proven factory capability.

14. Engineering questions by decision

15. Turn the worksheet into a project review

Copy or view the engineering worksheet, then paste it into the project summary with a section drawing, pole map, operating/overspeed conditions and available material/test data. List unknown values as TBD and identify who will resolve them. The contact link opens the existing structured inquiry form; it does not automatically transfer tool inputs.

Submit a Project for feasibility review →