The Manufacturing Engineer's Guide to Magnets in Assembly

The magnet as a process element

Three material facts drive every process decision downstream:

• It's a brittle ceramic-like part. Sintered NdFeB has essentially no ductility: press fits crack it, point loads chip it, and two parts snapping together can break both. Process forces must load magnets gently and in compression — the mechanical rules from the bonding & mounting guide apply to every station that touches one.
• It generates its own process forces. A magnetized part attracts the gripper, the fixture, the fasteners in the operator's tray, and the last assembly on the conveyor. Every station layout has to answer: what does this part want to fly toward, and what stops it?
• Its state is thermally fragile. One trip above its real temperature limit — which depends on geometry, not just the class letter, per the temperature guide — permanently changes the part, invisibly. The process flow either respects that or magnetizes after the hot steps.

Station design & line-side handling

• Non-magnetic everything in the working envelope: aluminum, brass, 300-series stainless, plastics for tools, trays, and fixtures at magnet stations. One forgotten carbon-steel bracket becomes a part trap and a chip generator.
• Controlled approach by design. Fixtures bring magnets in on a slide, pivot, or ramp — never free-handed above a steel target for anything beyond small parts. The acceleration distances surprise everyone the first time.
• Presentation matters: magnets arrive stacked or on spacers; the station needs a defined singulation method (slide off the stack sideways, never pry) and a defined orientation reference — pole marking on the part or a polarity-keyed nest, because a flipped magnet is a functional defect that looks identical.
• People protection is part of the layout: pinch-point guarding and gloves at stations handling palm-size and larger parts, spacing between staged magnets, and strong-field stations flagged for pacemaker wearers in the work instruction.
• Line-side storage discipline: original packaging or compartmented trays, away from heat sources and magnetic media — the shelf above the cure oven is the classic mistake.

Placement, bonding & cure fixturing

Most “magnet defects” at test are really attachment-process defects. The process-critical points from the bonding guide, in line order:

• Surface prep is a process step, not a note: degrease → abrade the bond area → re-clean, on both surfaces, with a defined time limit between prep and bond. Prepared surfaces re-contaminate; “bonded same shift” belongs in the work instruction.
• Control the bondline against magnetic clamping. Attraction squeezes joints to a starved film that passes inspection and fails in the field — spacer beads in the adhesive, shims, or a designed recess depth make the gap a controlled parameter.
• Fixture against cure-time walking: magnetic side-loads slide parts while the adhesive is green. Hard locating features (the pocket again), not friction; verify position after fixturing, before cure, and don't load green joints downstream.
• Sequence multi-magnet builds explicitly. Each placed magnet changes the field the next one sees — placement order, in-process spacing or shielding, and WIP handling rules are part of the process definition, not tribal knowledge.
• Keep the fillet. The small adhesive fillet around the perimeter is peel strength; a cosmetic wipe-down that removes it is a process regression that no inspection catches.

Magnetize before or after?

The single highest-leverage process decision. Assembling unmagnetized parts and pulsing the finished assembly eliminates most of sections 02–03 at a stroke — no attraction forces at placement, ordinary handling, ordinary freight (no UN 2807 classification), and no field constraints on hot process steps. The decision table:

Magnetizing direction and pattern define what a post-assembly fixture must do — axial discs are easy, clocked diametrics and multipole patterns are the supplier-fixture cases, per the directions guide. Raise the question at design freeze; it's an architecture decision wearing a process costume.

Process heat: the silent spec

The thermal profile that matters to a magnetized magnet is the process profile, not just the service profile — and the losses are invisible at the line:

• Reflow (~250 °C peak) exceeds every NdFeB class. Boards with magnetized parts must sequence the magnet after reflow, use SmCo, or magnetize after — there is no fourth option that yields.
• Cure ovens sit right in the danger band: heat-cured epoxies at 80–150 °C can approach or exceed the magnet's real limit — especially for thin parts whose geometry derates them below their class rating (the load-line effect in the temperature guide). Cure below the limit, verify flux after cure during process qualification, or bond unmagnetized.
• Welding is a double threat: heat locally, and the magnetic field around the cable/arc — both demagnetize. Route welding operations and cables away from magnetized WIP.
• Washing, marking, and pack-out are usually fine — but ultrasonic cleaning wipes marker-dot pole marks (specify permanent marking) and laser marking is a local heat event worth qualifying on the actual part.
• When line yield drifts down and dimensions check good, suspect a thermal event: a distorted pole pattern or uniformly low moment on a pole scan points at process heat — the diagnostic signatures are in the demagnetization guide.

In-process verification

• Verify polarity at or before placement — a cheap Hall sensor or gaussmeter check in the nest catches flipped parts at the only station where the fix is free. Downstream, a flipped magnet is a teardown.
• Verify presence and position after cure where the magnet is buried: a field check at a defined point doubles as both, and the limits come from measured good assemblies — the same anchor-to-hardware logic as the tolerances guide.
• After any hot or pulsed step, verify state: a moment or field-at-point sample check confirms magnetization survived the oven or that the post-assembly pulse saturated — fixture methods in how magnets are tested.
• Correlate line fixtures to the FAI reference set the quality team holds, and re-correlate on a schedule — drifted fixtures generate phantom yield problems and real arguments.
• Don't over-verify: a functional field check at the point of use beats re-measuring the incoming spec at three stations. Every extra magnet-handling step is a chip opportunity.

Yield killers & the launch checklist

Launch checklist: stations non-magnetic and approach-controlled · singulation and polarity reference defined · prep–bond–cure parameters and fixtures locked · magnetize-when decision made at design freeze · process thermal profile checked against the magnet's real limit · polarity/presence/state checks placed where fixes are cheap · fixtures correlated to FAI hardware · work instructions carry the safety rules. Process-qualify with production magnets — the prototype parts your pilot ran may not chip, bond, or survive the oven the way production parts do.

Related: bonding & mounting magnets · magnets & temperature · how magnets are tested