Compare frameless motor rotor magnet retention with CFRP sleeves, metal cans and adhesive. Use the shock, thermal-cycle and RFQ checklist to scope your design.
Frameless motor rotor magnet retention is a coupled mechanical, electromagnetic and thermal selection problem. Adhesive-only retention can preserve radial packaging space when the bonded joint is qualified. A metallic can adds radial containment but can create harmonic losses. A carbon fiber sleeve offers low mass and high directional strength, with laminate, preload and heat-flow constraints.
Scope and review date: November 1, 2026; global OEM engineering and procurement guidance for inner-rotor, surface-mounted permanent-magnet (SPM) frameless torque motors. The kit image is illustrative. The diagram and calculations below are screening aids, not measured product results or retention certification. Outer-rotor and buried-magnet designs need different load paths.
For selection, start with rotor radius and speed, shock direction and pulse, rotor temperature history, and the available radial clearance. Then qualify the assembled rotor rather than selecting a sleeve by material name alone. Use the RFQ contact form when you have these inputs ready.
For a magnet segment of mass m_mag at center-of-mass radius r_cg, centrifugal force is:
omega = 2 * pi * N / 60
F_c = m_mag * r_cg * omega^2
p_c ≈ rho_mag * h_mag * r_cg * omega^2Here N is rpm, omega is rad/s, rho_mag is magnet density in kg/m³, and magnet thickness h_mag and radius are in meters. The pressure expression approximates a thin, uniformly covered magnet ring; segmentation, contact area and edge loading require a more detailed model.
Shock adds directional inertia (F_shock = m_mag * a); tangential acceleration also loads the bond in shear. A radial sleeve does not automatically restrain axial magnet motion or transfer motor torque to the hub.
Conceptual cross-section, not to scale. Adhesive below the magnets and any seating layer below the sleeve are separate interfaces; their dimensions must be identified in the rotor drawing.
For a thin sleeve carrying the assumed outward magnet pressure:
sigma_hoop ≈ p_c * r_sleeve / t_sleeve
+ rho_sleeve * omega^2 * r_sleeve^2r_sleeve is mean sleeve radius, t_sleeve is radial wall thickness, and rho_sleeve is density. This omits interference preload, cure residual stress, thermal mismatch, magnet gaps, local stress concentrations and fatigue. For CFRP, use laminate directional allowables and a suitable failure criterion rather than a metal yield check. Sleeve thickness and preload must be solved together with cold/hot contact pressure.
These are design tendencies, not guaranteed material speed limits, shock ratings or supplier quotations. No architecture has a universal safe rpm or temperature window.
| Selection criterion | Structural adhesive only | Metallic can: Inconel / titanium | CFRP sleeve |
|---|---|---|---|
| Radial retention path | Bond and magnet coating carry separation load | Sleeve contains magnets; bond still transfers other loads | Hoop laminate contains magnets; interfaces remain critical |
| Speed screening | Check centrifugal loading against aged joint strength | Check pressure, self-stress, preload and alloy condition | Check pressure, layup, preload and composite allowables |
| Radial packaging | No sleeve outside magnets | Sleeve and any seating layer consume clearance | Sleeve and any seating layer consume clearance |
| Harmonic loss | No added sleeve loss; magnets can still heat | Conductivity enables eddy-current loss and shielding | Often lower sleeve loss; directional conductivity still matters |
| Rotor cooling | Bond thickness and shaft path matter | Through-wall conduction can help; added loss can offset it | Radial heat flow may be limited by resin and layup |
| Shock and vibration | Peel, coating adhesion and axial restraint need testing | Ductility can help, but interfaces and axial capture need testing | Impact damage and delamination need testing |
| Thermal cycling | Test creep, fatigue and coating compatibility | Check interference at both hot and cold limits | Check differential expansion and resin behavior |
| Manufacturing controls | Cleaning, gap, dispensing and cure | Thin-wall tolerances, insertion and residual stress | Winding tension, fiber angle, cure and inspection |
| Buying decision | Request aged joint and rotor qualification records | Request material condition, fit model and total rotor loss | Request laminate properties, process traceability and thermal model |
Adhesive-only retention can be a candidate when radial space is limited and the assembled joint is validated. Selection needs the actual magnet coating, substrate, bond gap, cure and temperature-dependent properties. Room-temperature cylindrical-joint shear strength cannot establish magnet peel strength or fatigue life.
Do not approve a rotor using a generic 30 m/s or 25g cutoff. Rotor radius, magnet mass, acceleration direction, pulse duration and aged bond performance determine the load case. Secondary radial containment can reduce the consequence of bond failure, while axial restraint needs its own design.
Inconel and titanium sleeves offer different density, alloy strength, electrical resistivity and thermal expansion. Obtain properties for the supplied condition and operating temperature. A thin metallic sleeve can provide radial containment, but does not by itself establish shock survival or cooling capacity.
Conductive sleeves respond to slotting and inverter harmonics as seen in the rotor frame; the synchronous fundamental alone does not describe these losses. The IET sleeve-thickness study examines coupled loss and temperature effects for a specific PMSM. Its results support comparing sleeve and magnet losses together; they do not give a transferable wattage for another motor.
A simple cubic thickness rule is insufficient for a slotted motor with shielding, end effects and changing harmonic fields. Use the actual geometry and inverter spectrum in a validated electromagnetic model.
A hoop-wound CFRP sleeve can reduce rotating mass and sleeve loss relative to some metallic designs. Carbon fibers conduct electricity, however, so CFRP is not loss-free. Fiber type, orientation, resin and contact between fibers affect its electrical and thermal behavior.
Through-thickness heat flow may be much weaker than along the fibers. Check the radial and shaft heat paths together, including contact resistance. Compare calculated magnet temperature against the selected magnet grade's demagnetization limits at the actual magnetic operating point; maximum service temperature is not the Curie temperature.
Thermal expansion mismatch can increase or decrease interface pressure depending on the full hub–magnet–sleeve stack. Check both temperature extremes for contact loss, magnet damage, laminate stress and resin creep. Do not assume that low hoop expansion guarantees thermal-cycle durability.
Hypothetical inputs, not a tested motor: a 68 mm stator bore, 67 mm magnet OD, 3.5 mm magnet radial thickness, 31.75 mm magnet center-of-mass radius, and assumed magnet density of 7,600 kg/m³. The example compares 6,000 rpm operation with a chosen 9,000 rpm calculation point. The 1.5x speed ratio is not a burst-speed claim, a stress safety factor or a prescribed proof test.
omega = 2 * pi * 9000 / 60 = 942.48 rad/s
p_c = 7600 * 0.0035 * 0.03175 * 942.48^2
p_c ≈ 750,184 Pa = 0.750 MPaAt 6,000 rpm the same approximation gives 0.333 MPa. Increasing speed by 1.5x increases this load by 2.25x. Neither value includes shock or thermal preload.
For a 0.35 mm sleeve seated directly on the original magnet OD, mean sleeve radius is 33.675 mm. Using assumed screening densities, not certified material properties:
| Screening result at 9,000 rpm | Metallic sleeve, density 8,190 kg/m³ | CFRP sleeve, density 1,550 kg/m³ |
|---|---|---|
| Magnet-pressure contribution | 72.18 MPa | 72.18 MPa |
| Sleeve self-stress contribution | 8.25 MPa | 1.56 MPa |
| Sum of these contributions | 80.43 MPa | 73.74 MPa |
This calculation isolates density effects; it does not establish allowable stress, safe speed or shock resistance. Strength comparisons need temperature-dependent alloy or laminate allowables, preload, local stress and fatigue analysis. No measured loss, torque constant or rotor temperature is available for this example.
The table assumes no seating layer outside the magnets. The effective non-magnetic separation is approximated as g_eff = running_clearance + sleeve_thickness + outer_seating_layer. A hub-to-magnet adhesive layer is a different interface and is not added to this outer separation.
| Radial dimension | Original unsleeved rotor | Add sleeve to original magnets | Redesign to preserve running clearance |
|---|---|---|---|
| Stator bore radius | 34.00 mm | 34.00 mm | 34.00 mm |
| Magnet outer radius | 33.50 mm | 33.50 mm | 33.15 mm |
| Sleeve thickness | 0 mm | 0.35 mm | 0.35 mm |
| Finished rotor outer radius | 33.50 mm | 33.85 mm | 33.50 mm |
| Running mechanical clearance | 0.50 mm | 0.15 mm | 0.50 mm |
| Effective outer non-magnetic separation | 0.50 mm | 0.50 mm | 0.85 mm |
| Decision required | Validate original joint | Check tolerance, runout and hot growth | Recalculate magnet geometry, flux and retention |
Engineering consequence: adding a sleeve can consume running clearance; preserving clearance can require a smaller magnet envelope. The redesigned rotor needs a new stress calculation. Neither option supports a fixed percentage torque penalty without a magnetic circuit or field model. Do not grind finished magnets to fit a sleeve without an approved manufacturing process.
Use Henkel's LOCTITE 648 technical datasheet, May 2021 as an example of interpreting evidence, not as a blanket magnet-bond recommendation.
| Evidence to request | What this datasheet reports | Rotor-selection implication |
|---|---|---|
| Chemistry and cure | Anaerobic urethane methacrylate | Confirm cure with the actual substrates |
| Joint geometry | Cylindrical fitting parts; gaps approaching 0.15 mm | Magnet geometry needs separate validation |
| Strength test | ISO 10123 steel pins and collars | Not a magnet peel test |
| Cure condition | 72 hours at 23°C for the cited steel-joint result | Record production cure conditions |
| Typical compressive shear strength | 31 N/mm² for that steel-joint condition | Not a rotor design allowable |
| Glass transition | 100°C after 24 hours at 23°C, ISO 11359-2 | Not a universal service-temperature cutoff |
| Hot strength versus aging | Separate curves and test conditions | Separate hot-load performance from strength after cooling |
The complete rotor still needs qualification of its coating, bond, sleeve and interfaces under combined service loads.
Use the following sequence to make the design review and supplier handoff auditable:
For proof-spin testing, document the applicable standard edition, speed, duration, temperature, containment and sampling plan. Do not turn a generic 1.2x speed / two-minute example into a requirement for every frameless rotor. Production screening and design qualification serve different purposes.
Before requesting a quote, compile:
Use this checklist to request a rotor retention review and ask the supplier to identify which requirements are demonstrated, modelled or still open.
Match the candidate motor envelope to the required load case before selecting a sleeve material:
Send your rotor drawing, speed profile, shock specification and thermal envelope to [email protected], or use the online RFQ form. Ask which retention options and supporting calculations or test records are available for your program; confirm deliverables and timing during the review.
Reviewed November 1, 2026. These sources support specific methods and material/test boundaries; none certifies the hypothetical rotor above.
Only when the actual magnet coating, bond geometry and aged adhesive joint are qualified for the specified speed, acceleration, shock and rotor temperature. There is no universal 30 m/s or 25g limit. A sleeve can provide radial containment, but axial retention and torque transfer still need separate checks.
Conductive metallic sleeves can generate eddy-current heat under rotor-frame slotting and inverter harmonics. CFRP may reduce that loss, but its conductivity is anisotropic and not zero. Compare total sleeve-plus-magnet loss for the actual winding, inverter and operating point.
Check cold and hot contact pressure, laminate stress, resin temperature limits and radial heat flow. A hoop-wound sleeve can expand differently from the hub and magnets; thermal cycling can change preload or damage interfaces. Validate the actual layup after the required environmental exposure.
There is no universal percentage. For a non-magnetic sleeve, the radial magnetic separation includes running clearance, sleeve thickness and any layer outside the magnets. Keeping the stator bore and running clearance fixed usually requires a smaller magnet envelope. Recalculate flux and torque constant for the redesigned geometry.
Agree the applicable standard edition, test speed, duration, rotor temperature, sampling plan and acceptance criteria with the supplier. A 1.5x speed screening case is not a stress safety factor or a production proof-test requirement. Balance acceptance does not certify retention strength.
Application engineers and manufacturing specialists supporting OEM motion programs.
Frameless torque motor air gap eccentricity and UMP: compare bearing loads, tolerance budgets, assembly methods, and inspection limits. Prepare your OEM RFQ.
Compare axial vs radial flux frameless torque motors: bearing loads, air gaps, inertia, cooling and total cost. Use the RFQ checklist to size a robot joint kit.
Review 800V frameless motor requests with terminal waveform, insulation scope, PDIV evidence and configuration checks before agreeing on sample acceptance.