Radial Magnets, Inc.
Radial Magnets, Inc.we know magnets
Radial Magnets, Inc.
Radial Magnets, Inc.we know magnets
Radial Magnets · Technical Resource
The arc segment is the motor builder's magnet, and it is the geometry where a specification most often arrives incomplete. Six numbers and one direction fully define an arc; leaving any of them out means the supplier guesses, and the guess is usually the cheap one. This guide covers how to dimension an arc, what pole arc ratio does to the back-EMF waveform, and why parallel and radial magnetization are not interchangeable.
An arc segment is fully defined by six dimensions and one magnetization direction. Anything less and the drawing is ambiguous.
Where the outer surface uses a different radius from the inner — thicker at the pole centre, thinner at the edges — the air gap varies across the pole and the flux distribution moves toward sinusoidal. It reduces cogging and torque ripple without changing the winding. It also requires profile grinding rather than plain cylindrical grinding, so state both radii and the offset explicitly, and expect a cost step.
Pole arc ratio is the magnet's subtended angle divided by the pole pitch, where pole pitch is 360° divided by the pole count. For an eight-pole rotor the pitch is 45°, so a magnet subtending 40° gives a ratio of 0.89.
This single number does more to shape motor behaviour than the grade does. It controls the back-EMF waveform, and the back-EMF waveform has to match the drive.
| Pole arc ratio | Back-EMF | Cogging / ripple | Matches |
|---|---|---|---|
| 0.65–0.70 | Trapezoidal | Higher ripple | Six-step BLDC commutation |
| 0.75–0.85 | Between the two | Moderate | General purpose SPM |
| 0.85–0.95 | Approaching sinusoidal | Lower ripple | FOC / PMSM sinusoidal drive |
| >0.95 | Sinusoidal but leaky | Low ripple, lower torque constant | Rarely optimal — inter-pole leakage rises |
The instinct is to fill as much of the pole pitch as possible on the theory that more magnet means more torque. Past roughly 0.95 it stops being true: adjacent poles begin to leak flux directly into each other across the gap rather than crossing the air gap into the stator. The leakage is wasted magnet. The inter-pole gap is a design parameter, not scrap space.
Where cogging torque is the dominant concern, pole arc ratio is normally tuned together with skew — either skewing the magnet itself or stepping a stack of segments angularly. Skewing a sintered arc is expensive; stepping a stack of straight segments achieves most of the benefit at a fraction of the cost, and is the usual production answer.
Two arcs of identical geometry magnetised differently are different parts with different performance and different prices. This is the specification line most often left blank.
On a narrow arc — a high pole count rotor, where each magnet spans a small angle — parallel and radial magnetisation produce very similar fields, because across a small angle “parallel” and “radial” barely diverge. On a wide arc in a four- or six-pole machine they diverge substantially. Before paying for a radial fixture, check whether the arc is wide enough for it to buy anything.
Multipole magnetisation of a single continuous ring is an alternative to discrete segments and removes assembly positioning error entirely — covered in ring and annular magnets. The broader treatment of orientation options is in magnetization directions explained.
On a rotor, magnet tolerance becomes air gap tolerance and air gap tolerance becomes torque ripple, acoustic noise and unbalance. The magnets are usually the loosest components in the stack, so they dominate.
| Feature | Standard | Precision | Why it matters |
|---|---|---|---|
| Outer radius | ±0.10 mm | ±0.05 mm | Direct air gap variation pole to pole |
| Inner radius | ±0.10 mm | ±0.05 mm | Bond line thickness and seating |
| Wall thickness | ±0.10 mm | ±0.05 mm | Flux per pole consistency |
| Arc angle | ±0.5° | ±0.2° | Pole pitch error; drives ripple harmonics |
| Axial length | ±0.15 mm | ±0.08 mm | Stack height in segmented rotors |
| Flux per pole | ±5 % | ±2–3 % | Pole-to-pole imbalance and unbalanced pull |
A ±5 % flux tolerance on individual magnets permits a 10 % spread across a single rotor if you happen to draw one magnet from each end of the distribution. That spread produces unbalanced magnetic pull, bearing load and a torque ripple harmonic at rotational frequency. Where it matters, specify matched sets: magnets supplied grouped so that all poles on one rotor fall within a stated band of each other. It costs more than a wider individual tolerance and less than a tighter one.
An arc segment RFQ that contains all of the following can be quoted accurately first time. One missing and the quote is a guess.
If the arc dimensions came out of a finite element model, the underlying requirement is a flux density in the air gap at a temperature. Send that alongside the geometry. It lets the supplier flag the case where the specified grade cannot deliver the modelled field at the modelled temperature — which is a problem far cheaper to find at quotation than at first article. The general approach is in how to prepare a magnet RFQ.
We supply arc segments from N38SH through N42UH and in SmCo for high-temperature service, with parallel or radial magnetization and PPAP documentation where the programme requires it. Send the rotor geometry and pole count and we will work from there.