When a servo motor designer specifies a rotor magnet, the conversation almost always starts with magnet grade — N38, N42SH, N48UH. Grade gets the attention because it’s the number on the datasheet, the variable that maps cleanly to remanence and coercivity, and the easiest thing to compare across suppliers. But for servo motor magnets specifically — where the entire value proposition is smooth, precise, repeatable motion — grade is rarely the parameter that determines whether the motor meets spec. Dimensional tolerance is.
This isn’t a claim that grade doesn’t matter. It does, particularly for thermal margin and demagnetization resistance in duty-cycle-heavy applications. But once a grade is selected that clears the motor’s thermal and load requirements with reasonable margin, further grade upgrades produce diminishing returns on the metrics servo applications actually care about: cogging torque, torque ripple, and positioning accuracy. Tolerance, on the other hand, has a direct and largely unavoidable relationship to all three.
What Makes Servo Motors Different
Induction motors and many BLDC applications tolerate a certain amount of torque irregularity — it averages out under load, or the application doesn’t need fine control. Servo motors don’t get that luxury. They’re specified for closed-loop position and velocity control, often driving axes where torque ripple translates directly into position error, vibration, or acoustic noise that shows up in the finished product: a CNC surface finish defect, a robotic arm’s repeatability spec, a camera gimbal’s jitter.
Cogging torque and ripple are generated by the interaction between the rotor magnet field and the stator slot geometry. Magnet grade sets the strength of that field. Magnet geometry and placement — arc angle, radial thickness, chamfer consistency, pole-to-pole spacing, and outer/inner radius accuracy — set the shape and symmetry of that field. A servo rotor with magnets that vary by even a few hundredths of a millimeter pole-to-pole introduces a periodic asymmetry that shows up directly in the torque ripple spectrum, regardless of how strong or expensive the magnet grade is.
The Tolerance Parameters That Actually Control Servo Performance
| Parameter | Typical Servo-Grade Tolerance | Effect if Out of Spec |
|---|---|---|
| Arc angle (span) | ±0.1°–0.2° | Uneven pole width → asymmetric flux distribution → harmonic torque ripple |
| Radial thickness | ±0.02–0.03 mm | Airgap flux density variance pole-to-pole → cogging torque increase |
| Outer radius (OD) | ±0.02–0.05 mm | Airgap non-uniformity → localized flux concentration, potential rotor rub risk at high speed |
| Chamfer/skew edge consistency | ±0.05 mm | Inconsistent flux edge transition → higher-order harmonics in back-EMF |
| Pole-to-pole spacing (on assembled rotor) | ±0.05–0.1 mm | Mechanical unbalance at speed; periodic torque ripple at pole-pass frequency |
| Magnetization axis alignment | ±1°–2° | Back-EMF waveform distortion, reduced torque constant accuracy |
These are the numbers that show up in a servo motor OEM’s incoming inspection report, and they’re the numbers that separate a magnet supplier capable of servo-grade work from one that isn’t — independent of what grade is stamped on the packing slip.
Why Grade Upgrades Don’t Fix Ripple
A common but costly mistake is responding to a torque ripple problem by moving up in grade — say, from N42 to N48 — on the assumption that a “better” magnet will smooth out performance. In most cases this doesn’t help, because ripple driven by geometric asymmetry scales with the variation in field strength between poles, not the absolute field strength. A stronger magnet with the same dimensional inconsistency produces a stronger ripple signal, not a weaker one. The fix has to happen at the dimensional level: tighter arc tolerance, tighter thickness control, and consistent pole-to-pole matching during assembly.
Where grade does matter is thermal margin and demagnetization resistance — an H, SH, or UH suffix buys headroom against the temperature rise a servo motor sees under sustained load or in a compact frame with limited cooling. That’s a real engineering requirement, but it’s a separate one from ripple and positioning accuracy, and it’s worth specifying independently rather than treating grade as a proxy for overall magnet quality.
Grinding, Not Just Pressing, Is What Gets You There
Sintered NdFeB blanks come out of the pressing and sintering process with tolerances too loose for servo work — typically several tenths of a millimeter. Reaching the tolerances in the table above requires precision grinding on the arc surface, OD, ID, and axial faces, along with dimensional sorting before magnetization and assembly. This is a process capability question as much as a materials question: two suppliers using an identical grade of NdFeB powder can produce magnets with meaningfully different servo performance purely based on grinding equipment, in-process measurement, and sorting discipline.
When evaluating a magnet supplier for a servo application, it’s worth asking directly about grinding tolerance capability and whether pole-to-pole matching is performed on finished pieces — not just relying on the grade specification in the RFQ.
Comparison: Tolerance Tightening vs. Grade Upgrade
| Design Response to Ripple Problem | Effect on Cogging/Ripple | Effect on Thermal Margin | Typical Cost Impact |
|---|---|---|---|
| Tighten arc angle & thickness tolerance | Significant reduction | None | Moderate (grinding precision) |
| Improve pole-to-pole matching/sorting | Significant reduction | None | Low–moderate (added QC step) |
| Upgrade magnetic grade (e.g., N42 → N48) | Minimal to none | Improves | High |
| Upgrade to H/SH/UH temperature suffix | None | Significant | Moderate–high |
FAQ
Does a higher NdFeB grade reduce torque ripple in a servo motor? Not directly. Torque ripple is driven primarily by dimensional and magnetization consistency between poles. A higher grade increases field strength but doesn’t correct geometric asymmetry, and can amplify ripple caused by uneven magnets rather than reduce it.
What dimensional tolerance should I specify for servo rotor magnets? It depends on pole count, airgap, and the ripple spec of the application, but arc angle tolerances of ±0.1°–0.2° and radial thickness tolerances of ±0.02–0.03 mm are typical starting points for precision servo work.
When does magnetic grade matter more than tolerance? When the motor’s limiting factor is thermal — sustained high-load operation, compact frames with limited cooling, or high ambient temperature environments — where demagnetization resistance from an H/SH/UH grade provides necessary margin.
Can pole-to-pole magnet matching be done after magnets are magnetized? Sorting by flux output is typically done before final assembly, using flux meters to group magnets into matched sets per rotor, which reduces pole-to-pole variance beyond what dimensional tolerance alone controls.
For servo rotor magnets requiring dimensional tolerances beyond standard catalog specification, our engineering team can review your ripple and positioning requirements alongside grade selection. Reach out to tony@xh-magnet.com to discuss tolerance capability for your application.
