
Radial vs. Diametric vs. Segmented: Choosing a Ring Magnet Magnetization
“Ring magnet, N42, 20 mm OD” tells a supplier almost nothing about how the part will work. The same ring can be magnetized through its thickness, across its diameter, or radially through its wall — and it can be built as one sintered piece or glued up from arc segments. Each choice produces a different field, a different set of manufacturing constraints and a different price, and the wrong one on a drawing is one of the most common reasons a first article fails.
This guide covers the three magnetization patterns that matter for rings and cylinders, how each is manufactured, where each is the right answer, and what to write on the drawing so the supplier builds what you meant.
Orientation vs. Magnetization
Two things happen to a sintered NdFeB magnet that are easy to confuse:
- Orientation is set during pressing. The green powder is compacted in a strong magnetic field that aligns the easy axis of every grain. This is permanent — the magnet’s crystal structure is now pointed in that direction and cannot be changed after sintering.
- Magnetization is done after sintering and machining, in a magnetizing fixture. It sets the actual poles. A magnet can be magnetized along its orientation axis with full strength, or at an angle to it with reduced and unpredictable strength.
The practical consequence: a ring pressed with an axial orientation cannot be turned into a good radial ring in the magnetizer, no matter how many poles you ask for. The orientation has to be designed in from the start. Most magnetization mistakes trace back to a drawing that specified the poles but not the orientation, or assumed they were the same thing.
Axial: The Default
Axial magnetization runs parallel to the ring’s axis — north on one flat face, south on the other. It’s the simplest to press, the cheapest to buy and what most stock rings are. The field concentrates on the two faces, with a hole in the middle that has little useful flux.
Use it for holding, latching, magnetic couplings where the parts mate face-to-face, and anywhere the ring is really just a disc with a mounting hole. It’s mentioned here mostly to rule it out: for anything that rotates and needs a field at the outer surface, axial is the wrong answer.
Diametric: One Pole Pair Across the Diameter
A diametric ring is oriented and magnetized across its diameter — north on one side of the outside surface, south on the opposite side, with the field passing straight through the bore. It’s a single dipole, two poles, 180° apart.
How it’s made
Diametric orientation is straightforward: the ring is pressed in a transverse field, the same way a block is. That makes diametric rings almost as economical as axial ones, available in the full range of grades up to N52, and manufacturable at nearly any size.
Where it fits
- Angle and position sensing. A diametric magnet on a rotating shaft produces a clean sinusoidal field at a Hall or TMR sensor placed on-axis or at the periphery — the standard configuration for absolute rotary encoders, throttle and pedal position sensors, and steering angle sensors.
- Two-pole rotors. Small brushless and stepper motors, magnetic stirrers and small pumps where a single pole pair is sufficient.
- Magnetic couplings and gears in the two-pole configuration.
- Reed and Hall switch actuation where the switching angle matters.
Limitations
The field at the outer surface is strongly non-uniform — maximum at the two poles and zero at the two points 90° away. That’s exactly what a sensor wants and exactly what a multi-pole motor doesn’t. Diametric is also inherently two-pole; it can’t be “magnetized with four poles” and still be diametric. A ring pressed with diametric orientation and then magnetized with four poles produces two weak, distorted pole pairs — a frequent and expensive mistake.
A diametric ring’s operating point is set by its diameter relative to wall thickness and length, so a large-bore, thin-wall diametric ring has a low permeance coefficient and derates harder in temperature than its grade suggests. The Demagnetization Calculator handles this geometry.
Radial: The Field Points Out of the Wall
In a radially magnetized ring, the magnetization vector at every point in the wall points along the radius — straight out from the center (or straight in). The ring can be:
- Monopolar radial: the entire outside surface is one pole, the entire bore is the other. Field is uniform all the way around the circumference.
- Multipolar radial: the outside surface is divided into alternating N and S poles (4, 6, 8, up to several dozen), but within each pole the magnetization is still radial. This is the rotor configuration for most brushless motors.
The distinction matters for the drawing: both are “radial,” and a supplier that assumes monopolar when you meant 8-pole has built you a paperweight.
How it’s made — and why it’s hard
A true radial ring is pressed in a die with a radial orienting field, so every grain’s easy axis points along its own radius. Producing that field uniformly around a ring is much harder than producing a straight field across a block. The die tooling is complex, the achievable ring length is limited by the field the tooling can generate, and sintering shrinkage on a radially oriented ring is anisotropic — the wall shrinks differently than the circumference — which can crack the part or leave it out of round.
For those reasons, true radial rings were historically restricted to small sizes and mid grades. Tooling has improved considerably; Radial Magnets produces them across a broad range of ODs and grades, but they remain a specialized product, and the size, length-to-diameter and grade envelope should be confirmed early in a design rather than assumed.
Where it fits
- Brushless DC and PMSM rotors where a uniform, sinusoidal air-gap flux matters — spindle motors, drone and e-bike hub motors, servo motors, fans, pumps.
- Precision rotary sensors that need a multi-pole target with consistent pole-to-pole amplitude.
- Magnetic couplings and torque limiters with more than one pole pair.
- Anywhere the alternative is a segmented assembly whose glue lines, balance or assembly cost are causing problems.
Advantages over a segmented assembly
- Smooth flux. No gaps between segments, so no dips in the air-gap field at the joints. Lower cogging torque, lower torque ripple, lower acoustic noise.
- Single part. No adhesive, no fixture, no assembly labor, no segment-to-segment tolerance stack. One press-fit or bonding operation onto the shaft or hub.
- Inherent balance. A sintered ring is concentric by manufacture; a segmented rotor has to be balanced after assembly.
- Retention. A ring cannot throw a segment at speed. High-RPM segmented rotors often need a retaining sleeve; a ring frequently doesn’t.
Limitations
- Size envelope. Very large diameters and long rings are difficult or impossible as one piece.
- Grade ceiling. Radial orientation is less perfect than parallel orientation, so the maximum achievable grade in a radial ring is typically somewhat lower than in a block of the same class.
- Cost per part is higher than an equivalent axial or diametric ring. Whether it’s higher than an equivalent segmented assembly depends on volume and how much the assembly labor and balancing cost — often it isn’t.
- Thermal expansion mismatch with a steel hub can crack a ring on a press fit. Bonding with a compliant adhesive, or a slightly interference-free fit with a retaining feature, is the usual solution.
Segmented: Arcs Assembled on a Hub
A segmented rotor is built from individual arc-shaped magnets — typically 4 to 48 of them — bonded to a steel or aluminum hub. Each segment is pressed and magnetized as a separate part.
Each arc can be oriented and magnetized in one of two ways:
- Parallel (straight-through): the orientation is a single straight line across the arc, like a block that’s been curved. Cheapest and highest-grade, but the field at the arc’s edges is off-radial, which shows up as flux harmonics in the motor.
- Radial arc: the orientation curves with the segment, so the field is truly radial across its whole width. Closer to a true radial ring in performance, at a higher tooling cost per segment.
Where it fits
- Large rotors. Traction motors, wind generators, industrial servos and anything beyond the diameter a one-piece ring can be made in. Above roughly 100 mm OD, segmented is usually the only option.
- Maximum grade. Segments are pressed with a parallel field and can reach the top of the grade table; a radial ring can’t.
- Halbach arrays and skewed poles. Any arrangement where the magnetization direction varies segment-to-segment in a way a single ring can’t reproduce.
- Interior-magnet rotors (IPM) that use flat or arc segments buried in laminations.
- Field-replaceable or rework-friendly designs.
Trade-offs
- Flux gaps at every joint. Even with tight assembly, the gaps between segments produce dips in the air-gap field that raise cogging and torque ripple. This can be designed around (skewing, shaped segments) at a cost.
- Adhesive as a structural member. Bond strength, cure schedule, thermal cycling, and the magnet’s temperature exposure during cure all become part of the design — a 150 °C epoxy cure on an N-grade segment demagnetizes it (see Temperature Classes).
- Tolerance stack. Segment width and thickness tolerance, hub tolerance and adhesive thickness all stack around the circumference. The last segment may not fit, or the assembly may be out of round.
- Balancing and retention. Assembled rotors are balanced after assembly, and high-speed designs typically need a carbon fiber or stainless sleeve.
- Handling. Magnetized segments are individually strong and want to slam together. Assembly fixtures and safety procedures add cost.
Side-by-Side
| Diametric | True radial ring | Segmented assembly | |
|---|---|---|---|
| Poles | 2 only | 1 pair or many | Any |
| Field at OD | Sinusoidal, 2-pole | Uniform per pole, smooth | Per-segment, dips at joints |
| Max practical OD | Very large | Limited (confirm with supplier) | Effectively unlimited |
| Max grade | Full range (N52) | Somewhat below block max | Full range |
| Part count | 1 | 1 | 4–48+ plus hub |
| Assembly | None | None | Bond, cure, balance, sleeve |
| Cogging / ripple | N/A (2-pole) | Lowest | Highest without mitigation |
| Piece price | Low | Medium–high | Low per segment, high assembled |
| Typical use | Rotary sensors, small 2-pole motors | BLDC rotors, precision sensors, couplings | Large motors, generators, Halbach, IPM |
How to Decide
- How many poles? Two, and the field shape at the sensor is what matters: diametric. More than two: radial ring or segmented.
- How big? Under about 60 mm OD with a moderate length, a true radial ring is very likely available. 60–100 mm, ask. Above that, plan on segments.
- What does the motor care about? If cogging, ripple, noise or balance are on the requirements list, a radial ring earns its price. If peak torque density is the only metric and the rotor is big, segments with the highest available grade win.
- What’s the volume? At low volumes, segmented assembly labor is cheap and ring tooling isn’t. At high volumes the per-unit assembly, adhesive, balancing and scrap on a segmented rotor usually exceed the ring premium.
- What happens in the assembly process? If the rotor goes through a hot cure, potting or reflow step, a pre-magnetized segmented assembly may not survive it. A ring can be pressed on and magnetized after assembly, in a multi-pole fixture, which sidesteps the problem entirely.
Not sure? Send the rotor drawing and we’ll tell you whether a one-piece ring is manufacturable in the size and grade you need, and what it costs against the segmented alternative.
What to Put on the Drawing
The magnetization callout is the single most under-specified item on magnet drawings. A complete one includes:
- Orientation direction — axial, diametric, radial — stated explicitly, not implied by the pole count.
- Pole count and pole arrangement for radial parts: “radially oriented, magnetized 8 poles on OD, alternating N/S, equal pole pitch.” For monopolar: “radially oriented, N on OD, S on ID.”
- Pole position reference — a keyway, flat, notch or paint mark that locates pole 1, if the sensor or stator needs to align to it.
- Pole boundary tolerance — the allowable angular error on each pole transition, and the allowable pole-to-pole amplitude variation. ±2° and ±5% are common starting points for sensor targets; motors are often looser.
- Acceptance test — how the supplier proves it. A flux measurement on a Helmholtz coil for total moment, a surface gauss scan at a specified radius and axial position for pole uniformity, or both. State the fixture position; a reading 1 mm from the surface and a reading 3 mm away differ by half or more.
- Magnetized or unmagnetized at delivery — and if unmagnetized, whether the supplier or the customer owns the magnetizing fixture.
Our guide to reading a magnet datasheet covers the grade side of the callout; the Magnet Grade Selector will recommend a grade and class once you know the geometry.
Frequently Asked Questions
Can a diametric ring be magnetized with four poles?
Not usefully. Diametric orientation aligns every grain in one direction across the diameter; a four-pole magnetizing pattern fights that alignment over half the ring, producing weak, distorted poles. Four or more poles need radial orientation or a segmented assembly.
What is the difference between a radial ring and radial arc segments?
Both have radial orientation. A radial ring is one sintered piece with no joints; radial arc segments are individual pieces assembled on a hub. The ring gives smoother flux and no assembly; the segments allow larger diameters and higher grades.
Are radially magnetized rings more expensive?
Per magnet, yes — the tooling and process are more demanding than axial or diametric pressing. Per finished rotor, often no, once the segment count, adhesive, assembly labor, balancing and scrap of a segmented alternative are included.
Why is my sensor reading uneven around a radial ring?
Usually pole-to-pole amplitude variation or pole boundary position error in the magnet, or the sensor’s radial and axial position relative to the ring not being where the acceptance test assumed. Check the drawing’s pole tolerance and test position; if neither was specified, that’s the cause.
Can a radial ring be magnetized after it’s pressed onto the shaft?
Yes, and it’s frequently the best approach. An unmagnetized ring is easy to handle and survives hot assembly steps; magnetizing in a multi-pole fixture afterward sets the poles with the ring already concentric to the shaft. The fixture is application-specific tooling, so plan for it.
What’s the largest true radial ring available?
It depends on wall thickness, length and grade as much as diameter, and the envelope keeps growing as tooling improves. Send the dimensions rather than assuming — parts that were impossible a few years ago are now routine.
Related Magnetization & Rotor Resources
- Magnet Demagnetization & Permeance Coefficient Calculator
- NdFeB Temperature Classes Explained
- How to Read a Magnet Datasheet
- Magnets 201: Circuits, Load Lines and Permeance
- Magnet Grade Selector
- Submit an Engineered Magnet RFQ
Radial Magnets manufactures true radially magnetized NdFeB rings — monopolar and multipole — alongside diametric rings and arc segments, with over 10 million magnets stocked in the U.S. Request a quote or contact us with your rotor or sensor drawing.

