Maglev: How Trains Float on Magnetic Fields

Two separate jobs: hold it up, then push it along

Every maglev splits into two independent magnetic subsystems that people often blur together. Suspension (levitation) generates a vertical force that exactly balances the vehicle weight so a millimetre-to-centimetre air gap opens between vehicle and guideway. Propulsion generates a horizontal force to accelerate, cruise, and brake. Neither uses mechanical contact, so there is no rolling resistance and — critically — no wear surface that degrades with speed.

The key numbers to hold in mind: a Transrapid car weighs on the order of 60–100 t loaded, so the levitation system must produce roughly m·g ≈ 80,000 kg × 9.81 = 785 kN of steady lift, spread over the magnet faces along both sides of the vehicle. That works out to a magnetic pressure of tens of kN per square metre of pole face — comparable to the contact-patch pressure a steel wheel puts on a rail, but delivered across a gap of air rather than a point of steel.

Because the two functions are decoupled, engineers can optimize them separately: you can pick a tiny stiff gap for compactness (EMS) or a big forgiving gap for high speed (EDS), independently of how you choose to make thrust.

EMS: pulling upward under active control

The electromagnetic suspension (EMS) scheme — used by Transrapid and the Shanghai line — wraps iron-cored electromagnets under the vehicle around the underside of a steel guideway rail and energizes them so they are attracted upward toward the rail. The vehicle literally hangs from the track from below, wrapping around it like a hand cupping a ledge.

The force between an electromagnet and a ferromagnetic rail across an air gap g goes roughly as F ∝ (N·I)² / g² — proportional to the square of the ampere-turns and inversely to the square of the gap. That inverse-square law is the whole problem: attraction gets stronger as the gap shrinks. If the car drifts up, force rises and pulls it up further; if it sags, force falls and it sags more. EMS is therefore statically unstable — Earnshaw's theorem guarantees no arrangement of fixed currents can passively trap it.

The fix is fast feedback. A gap sensor measures the air gap (target ≈ 8–10 mm), and a controller modulates the magnet current at roughly 1–4 kHz to hold that gap within about ±1–2 mm. This is a hard mechatronics problem: the guideway must be built to millimetre straightness because the control loop has only a few millimetres of authority, and a power loss must be caught by backup batteries and skids in the fraction of a second before the car drops the gap.

EDS: repelling with induced currents and superconductors

The electrodynamic suspension (EDS) scheme — Japan's SCMaglev for the Chūō Shinkansen — flips the sign. Powerful superconducting magnets on the vehicle sweep past figure-8 coils set in the guideway walls. A moving magnet changes the flux through those coils, and by Faraday's and Lenz's laws it induces currents that oppose the change: emf = −dΦ/dt. Those induced currents create fields that repel the vehicle magnets, lifting the car.

Because lift comes from motion-induced currents, EDS produces almost no lift at standstill. The SCMaglev runs on rubber tyres until roughly 100–150 km/h, at which point induced currents become strong enough to levitate and the wheels retract. In exchange for that complexity, EDS gives a large, forgiving air gap of about 100 mm and is passively stable — displace the car and the induced currents naturally restore it, no fast feedback loop required.

The magnets are the marvel: SCMaglev coils are niobium–titanium wound and cooled to about 4 K (liquid helium), carrying currents of ~700 kA-turns with essentially zero resistive loss once persistent. A room-temperature variant, Inductrack (Post, LLNL, late 1990s), replaces superconductors with a Halbach array of permanent magnets over a passive ladder track — cheaper and fail-safe, though with lower field strength.

Propulsion: a magnetic wave that drags the train

With no wheels to turn, maglev cannot use a rotary motor. Instead the vehicle rides a linear motor — imagine taking an ordinary rotary AC motor, slicing it radially, and unrolling the stator flat along the track. Three-phase windings embedded in the guideway are driven so the current pattern shifts down the rail, producing a traveling magnetic wave. The vehicle's magnets lock onto that wave and are dragged along by it.

The wave's speed is set by the drive frequency and pole pitch: v = 2·τ·f, where τ is the pole pitch (magnet spacing) and f the supply frequency. To reach 500 km/h ≈ 139 m/s with a pole pitch near 1.35 m, the guideway inverters must feed roughly 50 Hz and sweep the frequency continuously from zero at start-up.

• Linear Synchronous Motor (LSM) — vehicle magnets are actively energized; the wave and the magnets stay in lockstep. Used by Transrapid and SCMaglev. Needs long, expensive powered guideway sections switched on only where a train is.
• Linear Induction Motor (LIM) — the vehicle carries only a passive conductor plate; the wave induces currents in it (slip drive). Simpler and cheaper, used on lower-speed urban lines like Incheon and the Linimo (Aichi).

Because the powered coils sit in the ground, most of the propulsion mass and heat stays off the vehicle — a big reason maglev can chase very high speeds.

Why bother: speed, wear, and the numbers that justify it

Conventional high-speed rail hits two hard ceilings. First, above roughly 350–400 km/h, wheel-rail adhesion and rolling contact fatigue become punishing, and the pantograph that slides against the overhead wire starts to lose contact and arc. Second, everything wears — wheels, rails, contact strips, brakes — driving relentless maintenance. Maglev removes the mechanical interface entirely, so its dominant resistance at speed is aerodynamic drag plus a small magnetic drag, not rolling friction.

The payoff is concrete: on 21 April 2015 an SCMaglev test set reached 603 km/h on the Yamanashi test line, the fastest a crewed train has ever gone. The Shanghai Maglev (Transrapid, opened 2004) covers the 30 km airport run at up to 431 km/h in about 7.5 minutes, cruising 431 for only a short window. Japan's Chūō Shinkansen is building a 286 km Tokyo–Nagoya SCMaglev line for ~505 km/h service — a project whose cost estimate has ballooned from an initial ¥5.5 trillion to roughly ¥11 trillion (on the order of US$70 billion), with much of it in deep tunnels, which is the honest counter-argument: maglev guideway is extraordinarily capital-intensive per kilometre.

Energy-wise, a maglev at 500 km/h is comparable to or slightly better than steel-wheel HSR at the same speed because magnetic drag is small — but at 500 vs 300 km/h aerodynamic drag (∝ v²) dominates, so faster inevitably means more energy per seat-km.

Failure modes, misconceptions, and design pitfalls

Misconception #1: maglev has no friction, so it needs no energy to move. Levitation and cruising still cost power. EMS burns continuous current in its electromagnets just to hover; EDS pays a magnetic drag that is worst at low-to-mid speed (where induced currents lag) and eases at high speed. And aerodynamic drag never goes away — it grows with the square of speed and is the main energy sink at cruise.

Misconception #2: the whole track is one giant magnet. In an LSM, only the guideway block a train currently occupies is energized; the inverter substations switch power block-to-block as the train passes, both to save energy and to control thrust precisely.

Key EMS failure mode: loss of levitation control. With an 8–10 mm gap and active-only stability, a control or power failure means the car settles onto emergency skids in milliseconds — survivable, but it hard-stops levitation, so redundancy (independent battery-backed control channels, multiple magnet segments) is mandatory. Key EDS pitfall: the vehicle cannot levitate below its transition speed, so it needs a wheeled landing gear and the associated weight and complexity, and a quench of the superconducting magnets (loss of cryogenic cooling) removes lift entirely. A subtler EDS issue is magnetic drag and ride-quality resonance — the induced-current suspension has low damping, so secondary suspension and coil geometry must be tuned to avoid bounce, much like a tuned mass damper problem. Finally, both schemes demand guideway tolerances measured in millimetres over kilometres, which is why the civil works, not the magnets, dominate the cost sheet.

EMS (attractive) vs EDS (repulsive) levitation — the two dominant maglev architectures

Property	EMS (attractive)	EDS (repulsive)
Force type	Attraction, iron-cored electromagnets pulled up under a steel rail	Repulsion, eddy/induced currents in guideway coils
Air gap	Small, ≈ 8–10 mm	Large, ≈ 100 mm
Stability	Inherently unstable — needs active feedback control (~kHz)	Passively stable at speed; unstable/no lift at low speed
Levitation at standstill	Yes, hovers on the spot	No — needs ~100–150 km/h before coils lift, runs on wheels below that
On-board magnets	Conventional copper electromagnets	Superconducting (SCMaglev) or permanent (Inductrack)
Example system	Transrapid / Shanghai, Incheon, Changsha	JR–Maglev SCMaglev (Chūō Shinkansen)