A motor around your finger.

1. Record finger circumference and comfortable fit separately from knuckle passage. Keep the requested 15 mm bore as a target; do not force a smaller ring onto a finger.
2. Decide removal method. A one-piece structural sleeve must slip off easily; a split design needs a new bearing/structure solution.
3. Declare initial testing on a dummy finger/fixture. Define what useful output is wanted: a rotating mass is not yet an accessible shaft output.

Record: Drawing version, measured fit and test-fixture plan.

1. Model liner, sleeve, bearing seats, hub, rotors, cores, insulation, covers and wire routing as separate bodies.
2. Start from the proposed 38 mm OD / 24 mm axial budget, not the image’s earlier 32 × 12 mm visual proportions.
3. Run full 360° interference checks, including wire loops, screw tips and rotating magnet pockets.

Record: Section CAD with clearance map; any collision stops part manufacture.

1. Use common bearing-seat axis as datum A, bearing locating face as datum B and sensor reference as datum C.
2. Specify coaxiality, rotor face runout, flatness and balance from an actual gap/stiffness budget.
3. Nominal 0.5 mm face gap is not a tolerance: subtract manufacturing error, axial play, magnetic deflection and thermal growth.

Record: Toleranced drawing; no guessed universal machining tolerance.

1. Check candidate 61803 17 × 26 × 5 mm envelope against actual variant, seals, friction, life and axial load.
2. Choose housing/sleeve fits from manufacturer guidance; rotating outer race load matters.
3. Design positive axial retention and assembly tools that act on the fitted race; do not transmit press force through the balls.

Record: Bearing datasheet, fit callouts and axial location stack.

1. Use magnet grade and gap geometry in magnetic analysis or measured fixture tests.
2. Opposed rotor attractions can nominally balance, but unequal gaps create net axial force.
3. Check backplate bending, hub stiffness and residual bearing thrust; do not equate an axial-force estimate to rotational torque.

Record: Flux/force map and deflection allowance.

1. Choose qualified ferromagnetic plate material and corrosion protection.
2. Machine matching annular plates and magnet pockets concentric to hub datum.
3. Remove burrs and sharp edges; inspect thickness, pocket depth and face runout before adding magnets.

Record: Two serialized plates with inspection record.

1. Keep central sleeve fixed and hollow hub free to rotate around it.
2. Provide outer-race seats and rotor joints; preserve rotor-to-stator radial clearance.
3. Include positive retention, access for assembly and a balancing feature; inspect seats before pressing bearings.

Record: Hub drawing and measured concentricity.

1. Join sleeve to end covers, then covers to outer stator casing; this is the stator torque reaction path.
2. Protect any copper-to-metal contact with the specified electrical barrier.
3. Keep cover screws and fixed ribs outside the swept rotor volume.

Record: Fixed frame that leaves clear bore and coil channel.

1. Choose SMC or correctly oriented lamination design suitable for axial flux; validate loss and saturation.
2. Six cores are a study topology, not a guarantee of smooth torque or buildability.
3. Insulate core edges and provide a repeatable winding former.

Record: Core material, insulation class and dimensional samples.

1. Label coils C1–C6; mark start S and finish F and maintain identical winding sense when viewed from one defined axial side.
2. Count illustrative 40 turns per coil using a counter; actual turns depend on slot area and voltage design.
3. Use controlled tension, smooth guides and bend radius from wire data; enamel must not scrape on core edges.
4. Measure bare and insulated wire diameter; record actual mean turn length and all lead lengths.

Record: Turn counts, lead labels, per-coil resistance and photo.

1. Dry-fit bobbin and wound coil before potting. Include enamel, leads, bobbin, impregnation and growth.
2. Bare copper fill estimate does not verify a real winding envelope.
3. Reject pinched insulation or any coil projecting into a rotor gap.

Record: Measured coil pack and insulation barrier check.

1. Preliminary coil mechanical centers: 0°, 60°, 120°, 180°, 240°, 300°. With four pole pairs, electrical centers are 0°, 240°, 120°, 0°, 240°, 120°.
2. Candidate pairs are C1+C4 (A), C2+C5 (C), C3+C6 (B); same defined winding sense, series aiding.
3. Prove series aiding by rotating a pole test fixture and comparing induced voltage before permanent joining. A misplaced or reversed coil can cancel flux.
4. Join to a star neutral if using the assumed wye model; insulate and secure every joint. Final phase sequence requires calibration.

Record: Verified phase pair polarity, wye schematic and terminal map.

1. Use a polarity indicator; do not rely on painted colors. Mark the face that will point toward the stator.
2. Lay eight magnets on each rotor at 45° mechanical intervals with alternating N/S facing poles.
3. At matching angular positions, upper inward face N opposes lower inward face S, then upper S opposes lower N.
4. The second rotor doubles active gap surfaces; it does not double the number of pole pairs.

Record: Two face maps with angular zero and 4 pole pairs.

1. Use nonmagnetic assembly tooling and a fixture to prevent sudden attraction.
2. Clean and prepare pockets according to selected adhesive data; use specified bondline and cure, not an invented time.
3. Add mechanical capture compatible with the axial gap. Confirm no magnet stands proud beyond drawing allowance.

Record: Cure record, retention inspection and matched magnet mass.

1. First compare rotor pair and magnet masses, then balance completed linked assembly using a suitable fixture.
2. Remove/add balance mass only at designed locations; never loosen magnet retention to balance.
3. Record residual unbalance and recheck after rotor joints are locked.

Record: Balance report; acceptable grade remains an engineering selection.

1. Fix cores to outer carrier using the specified locating method; no support bridges to the rotating hub.
2. Bond temperature sensor to insulated coil carrier and reserve lead path.
3. Mount Hall sensors to sense rotor poles; locate and calibrate from electrical angle, not arbitrary mechanical spacing.
4. Keep sensor leads and phase leads restrained clear of both rotating disks.

Record: Sensor mounting record and coil-to-frame insulation test.

1. Install bearing races using controlled race-specific tooling and clean work surfaces.
2. Locate fixed inner rings on sleeve; rotating outer rings in hub. Preserve planned endplay/preload.
3. Verify finger bore remains open and retain bearings positively. Lubricant, seals and preload must match selected bearing data.

Record: Smooth hand rotation, measured axial play and assembly record.

1. Use an axial jig and temporary nonmagnetic spacers to control attraction.
2. Install lower rotor on hub, stationary stator onto casing, then upper rotor onto same hub. Both rotors must be mechanically coupled.
3. Set both nominal face gaps with shims and measure at multiple angular positions.
4. Remove temporary spacers only when joints and bearing retention are secured; turn by hand through 360°.

Record: Gap map, runout map and no-contact rotation.

1. Provide fixed strain relief on casing. No conductor rides a rotating surface or passes between bearing rings.
2. Document exact controller pin names; phase outputs U/V/W are not sensor supply terminals.
3. Use keyed connectors, insulation sleeves and separate sensor return paths as controller documentation requires.

Record: Continuity table and marked connector diagram.

1. Close all finger-accessible moving surfaces; a clear display shell is not automatically a containment guard.
2. Check smooth bore, removable liner, cleaning path and absence of burrs or exposed metal joints.
3. Choose guard material/thickness for tested retention, fragments, heat and impact; do not use a human finger as the fixture.

Record: Guard inspection and dummy-finger clearance check.

1. Measure phase-to-phase resistances; for balanced wye, each equals 2 × phase resistance (allow lead resistance).
2. Verify phase-to-frame insulation with a test voltage suitable for components; disconnect sensitive electronics as needed.
3. Check sensor supply polarity, ground continuity and connector pin map. Check no solder whisker or trapped wire.

Record: Resistance, insulation and wiring inspection record.

1. On a guarded fixture, rotate externally at known speed with controller disconnected; measure phase-neutral peak back EMF using a safe accessible neutral.
2. For this dq convention, λ = E_phase_peak / (p × ω). If measuring line-line RMS, E_phase_peak = √(2/3) × V_LL_RMS for sinusoidal balanced wye.
3. Measure coil temperature, inductance, breakaway torque and coastdown; fit losses across speed rather than guessing them from a photo.

Record: Measured constants with instruments, units, uncertainty and date.

1. Use the chosen drive board’s rated supply and a current-limited source; confirm output-enable initially off.
2. Verify current polarity/offset, rotor sensor sequence and phase mapping. Start with a limited low-current alignment trial on the guarded bench.
3. Log current, voltage, RPM, temperature, vibration and gap contact. Stop on rubbing, abnormal noise, sensor faults, loose parts or excessive heat.
4. No general safe RPM follows from bearing catalog speed alone. Retention, resonance, imbalance and enclosure limits govern the assembly.

Record: Bring-up report; device remains bench-mounted.

1. Characterize loaded and unloaded performance and thermal rise over defined duty cycles.
2. Check fault disable, disconnect, stalled rotor heating, restraint reaction torque and coast behavior.
3. Determine skin-contact temperature, electrical insulation, snag/removal behavior and mechanical containment before any energized wearable demonstration.

Record: Qualified operating envelope; absent today.

1. Replace all assumed inputs with measured values or approved drawings; freeze BOM and versions.
2. Record fits, torque values from joint analysis, cure schedules from suppliers, balance tolerance and acceptance limits.
3. Attach test curves and uncertainties. The guide’s checkboxes track your work; they do not certify compliance or mark fictional construction complete.

Record: Complete revision package and evidence-linked sign-off.