How coded magnets work

Three ideas explain everything on this site: magnetic fields from patterned surfaces die exponentially with distance; a pattern designed like a signal-processing code makes force behave like a correlation; and therefore the force curve itself — not just its strength — becomes something you engineer.

1. A dipole versus a coded surface

A conventional magnet is a dipole: one N face, one S face, flux looping far out into the room. Its behavior is fixed by geometry and grade — you can pick stronger or weaker, and that is roughly the whole menu.

A coded magnet (the category CMR trademarked as Polymagnet®) is a single piece of ordinary sintered NdFeB whose face has been re-magnetized into a pattern of many small N and S regions — maxels, magnetic pixels roughly 1–4 mm across. Each maxel’s flux no longer has to travel around the whole magnet; it closes its loop into the opposite-polarity neighbor a millimeter away. The field piles up at the surface and vanishes fast beyond it.

No new energy is created. Coding rides on the material’s fixed energy product (NdFeB tops out around 52 MGOe). The pattern doesn’t add force — it redistributes the field, choosing where it lives. Concentrated at the surface means more force at contact and less everywhere else.

2. Near field, far field — the exponential law

For a periodically patterned surface the field decays as

F ∝ e^−2πz/λ — z = gap, λ = pole pitch (maxel spacing)

Small pitch → intense but short-throw. Large pitch → gentler at contact but longer reach. Reach itself is a design knob. CMR recommends working gaps under 2 mm for fine patterns; in Fullerton’s own 2009 demonstrations, one coded magnet ignored a steel plate entirely until it came within inches. Play with the law yourself:

Interactive · The reach designer

Pole pitch λ 3.0 mm Useful reach (force > 5%): 2.3 mm

Drag the pitch. Fine maxels (1–2 mm) give a wall of force that is gone within a couple of millimeters; coarse patterns trade contact intensity for reach. A conventional dipole’s power-law tail (dashed) is why it grabs your screwdriver from across the bench — and why it also wipes cards and collects swarf.

3. The correlation trick — magnets as matched filters

Larry Fullerton spent his career in ultra-wideband radar, where a receiver finds a coded pulse train by correlation: signal and template line up exactly → huge spike; any misalignment → near-cancellation. His insight, patented as “field emission structures” (US 7,800,471), was that magnet faces can do the same trick with force.

Print a code — the patents literally use Barker codes, radar’s favorite low-sidelobe sequences — on one face and its complement on the mate. At exact registration every N sits over an S: every maxel pair attracts, forces add, peak attraction. Slide or rotate a little and attracting pairs start landing on repelling pairs: the sum collapses. The patent puts the off-peak force at “typically about an order of magnitude” below the peak.

A Barker-7 code (+ + + − − + −) and its complement at registration: seven attracting pairs, zero fighting each other. The engineered result — force as a function of position — is what the patents call the spatial force function.

Interactive · Slide the code

Lateral offset 0 maxels Net force: +7 (peak)

The top strip is the moving magnet; green ticks are attracting pairs, red are repelling. The bar chart underneath is the autocorrelation — the force you’d feel at each offset. One position wins by ~7×; a Barker code’s sidelobes never exceed 1. This is a self-aligning connector, a lock-and-key, and a “snaps only where it should” fastener — all from one principle.

4. What the code gives you

A designed force curve. A conventional magnet’s force-vs-distance curve has one fixed shape. A coded magnet’s can be sculpted — it can cross zero, change sign, hold a plateau, or click through detents. That single fact generates all eight behaviors.
Identity. Differently-coded pairs simply ignore each other — mechanical lock-and-key, anti-counterfeit seating, connectors that refuse the wrong socket.
Field containment. Flux that closes to a neighbor a millimeter away is flux that never reaches your credit card, compass, or Hall sensor.

5. Where it stops — the honest physics

The defining constraint is short field reach. The same local flux-closing that makes coded magnets ferocious at contact makes them nearly dead an inch away. They are contact and near-contact devices — brilliant fasteners, latches and couplings; wrong tool for long-range attraction. Any application that needs pull across a real gap belongs to conventional dipoles.

No levitation. Earnshaw’s theorem still applies: no static arrangement of permanent magnets can float free. Every magnetic “spring” demo rides a pin or shaft. (CMR’s own FAQ says this plainly — credit where due.)
No free energy, no new strength record. “Up to 4–5× stronger” and “first advance in magnets in 180 years” are vendor claim — vendor statements about specific contact-force comparisons, not new physics. Directionally supported at contact (see the K&J test), but always attributed here.
Temperature and materials limits are the substrate’s: standard demo parts are rated to about 60 °C; SmCo versions exist for heat and corrosion. Smallest engineered parts run about 4 mm diameter.

Next: what the sculpted force curve actually does — the eight behaviors, with datasheet numbers →