Why Magnet Assembly Is Where Motor Projects Fail
Every motor project eventually arrives at the assembly bench. The magnets are specified, graded, and coated correctly. The rotor laminations are machined to drawing. And then something goes wrong: a magnet snaps onto a steel fixture during handling and chips, or the measured air gap varies by 0.3mm around the circumference and the motor vibrates at full load, or three out of twenty rotors have back-EMF waveforms that don't match the simulation because magnets shifted during adhesive cure.
These failures are not caused by bad magnets. They're caused by assembly processes designed without accounting for what NdFeB magnets actually do when they're near steel and each other. The material properties that make NdFeB useful - high remanence, high coercivity, very high BHmax - also make it genuinely dangerous to handle and dimensionally unforgiving to assemble. This article covers the specific tolerance and process decisions that separate first-pass assembly success from chronic rework.
1. Magnetic Force During Assembly: The Numbers Most Engineers Ignore
The attractive force between a sintered NdFeB magnet and a ferromagnetic surface scales with the square of the gap distance closing - meaning the force increases rapidly as the gap closes, not linearly. A 40mm × 20mm × 5mm N42 magnet has a pull force of approximately 14kg against a flat steel surface when in full contact. At 5mm separation, that force is already 6–8kg depending on surface area. At 1mm, it's close to the full contact value.
This matters because assembly fixtures, tooling, and nearby components are almost always ferromagnetic. A technician holding a magnet 100mm away from the rotor hub feels almost nothing. At 20mm, the attraction becomes noticeable. At 5mm, the magnet will pull out of an ungloved hand and impact the rotor surface at velocity. The chip failure mode on sintered NdFeB is well-documented: the material has tensile strength of roughly 80 MPa but compressive strength of only about 1100 MPa in the magnetization direction and is much weaker in other orientations. Impact at speed produces corner chips and edge spalls that are invisible under nickel coating but reduce effective magnetic volume by 2–8% and create stress concentration points for long-term fatigue.
The practical rule: any assembly operation where a magnet is within 50mm of a ferromagnetic surface requires either a non-magnetic fixture, a controlled-approach jig that limits the closing velocity, or both. Aluminum and austenitic stainless steel (304, 316) are the standard fixture materials. Fixtures should be designed so that the magnet approaches the rotor pocket along its intended final axis - side-loading during assembly is the second most common cause of chip damage after impact.
2. Tolerance Stack-Up in Bonded Magnet Assemblies
Most permanent magnet motors bond magnets into rotor pockets using anaerobic adhesive, epoxy, or structural acrylic. The bonded joint serves two purposes: it holds the magnet in position under centrifugal load, and it provides the dimensional reference that sets the air gap. When the bonded joint is the air gap reference, tolerance stack-up becomes critical.
Consider a typical radial flux rotor pocket design: the pocket depth tolerance is ±0.05mm (achievable with CNC turning), the magnet thickness tolerance is ±0.05mm (standard for sintered NdFeB), and the adhesive bondline thickness adds another ±0.03mm if controlled. The worst-case stack is ±0.13mm radial variation in magnet surface position. For a motor with a 0.5mm air gap, that's ±26% gap variation. For a motor with a 0.8mm air gap, it's ±16%. Both figures are too large for a precision servo drive - they produce back-EMF harmonic content that translates directly into torque ripple at electrical frequency multiples.
There are three ways to manage this. First, tighten the magnet thickness tolerance to ±0.02mm through grinding - this adds cost but is standard for precision servo motor magnets. Second, specify the adhesive bondline thickness explicitly (typically 0.05–0.10mm) and use a fixture that controls it during cure. Third, use a post-assembly grinding operation on the rotor OD to bring all magnets to the same radial position - this is common for high-pole-count servo rotors where the magnet surface is the air gap reference surface. Each approach has a different cost and process complexity; the right choice depends on the motor's air gap specification and the production volume.
3. Pre- vs. Post-Magnetization Assembly
Sintered NdFeB can be assembled either as a fully magnetized part or as an unmagnetized blank. The choice determines what tooling, handling precautions, and quality checks are required at each stage. This is not an arbitrary decision - it has direct consequences for yield and process risk.
Assembling pre-magnetized magnets (the most common approach for small production runs) means every handling step involves live magnetic force. The handling risks described above apply at every stage. The advantage is that magnetization uniformity can be verified on the incoming magnet before it goes into the motor - this is straightforward with a flux meter or Hall probe array. The disadvantage is that assembly fixtures must be entirely non-magnetic, and adjacent magnets on the rotor must be loaded one by one with controlled polarity sequencing to avoid the pole-to-pole repulsion force snapping magnets sideways during installation.
Assembling unmagnetized blanks and magnetizing after assembly eliminates most handling hazards. The magnets are inert during bonding and fixture operations. Magnetization is done in a single fixture pass using a pulsed magnetizer, which simultaneously magnetizes all poles to the correct orientation. The disadvantage is that post-assembly magnetization requires a magnetizing fixture designed specifically for that rotor geometry - this is a non-trivial tooling investment, and the fixture must be validated to confirm saturation of all poles evenly. For high-pole-count rotors (>16 poles), inter-pole flux leakage during magnetization can leave some poles under-magnetized if the magnetizing fixture design is not carefully optimized.
The crossover point for most motor manufacturers is around 500–1000 units per year. Below that volume, pre-magnetized assembly with careful handling is generally lower total cost. Above that volume, the quality consistency and handling efficiency of post-assembly magnetization typically justifies the fixture investment.
4. Adhesive Selection and Cure Process for Magnetic Joints
The magnetic force between a bonded magnet and the steel rotor pocket does not disappear during adhesive cure - it's present and active throughout. For most adhesive systems, the bond forms while the magnetic pre-load is holding the magnet in contact with the pocket surface. This is actually useful: the magnet self-locates against the pocket bottom, giving consistent depth positioning without a dedicated fixture for that axis.
The dimension that requires control is the lateral position - tangential alignment within the pocket determines pole-to-pole spacing uniformity, which affects back-EMF waveform symmetry. A standard approach is to use pocket geometry that constrains lateral position mechanically (tight side clearance on the tangential faces of the pocket), with the adhesive gap controlled on the radial face. Pocket side clearance of 0.05–0.10mm per side allows manual loading without binding while providing lateral positioning accuracy better than 0.1mm total.
Adhesive selection matters more than most motor designers realize. Anaerobic adhesives (Loctite 638, 648) work well for steel-to-magnet bonds at temperatures up to 150°C, cure quickly under controlled conditions, and have good resistance to the centrifugal stress mode (shear along the bond face). Epoxies (Araldite 2011, 3M DP420) offer higher temperature resistance (up to 180°C) and fill larger bondline gaps more reliably, but require longer cure time and fixture holding. For motors operating above 120°C continuously, epoxy is the correct choice - anaerobics soften at elevated temperature and can allow magnet movement under centrifugal load at high-speed operation.
5. Post-Assembly Inspection: What to Measure and Why
The three measurements that predict whether a bonded magnet rotor will perform to specification are: back-EMF waveform at a reference speed, magnet surface runout (TIR on the rotor OD measured across magnet faces), and flux balance between opposing poles.
Back-EMF measured at 500–1000 RPM against a reference winding gives the aggregate result of all assembly tolerances - pole-to-pole flux variation, angular spacing errors, and magnetization non-uniformity all appear as harmonic content in the waveform. For a well-assembled rotor, the 5th harmonic amplitude should be below 1% of the fundamental. Values above 2–3% indicate a systematic assembly problem worth investigating before the motor enters the winding and housing stage.
Magnet surface runout (TIR) measured with a dial indicator across the magnet faces directly measures radial position consistency. For air gaps of 0.5mm or more, TIR below 0.05mm is achievable and adequate. For air gaps below 0.4mm (precision servo applications), TIR below 0.03mm is the target - this typically requires the post-assembly grinding operation mentioned in Section 2.
HIMAGNET supplies motor magnets with inspection reports including remanence, coercivity, and dimensional verification. We also work with customers on assembly fixture design and post-assembly magnetization tooling validation. Contact us with your rotor geometry and production volume for a process recommendation.
