Why a Faraday Cage Shields You

Equivalent forms

E_{leak} \sim E_{ext}\,(d/\lambda)\ (d\ll\lambda)

\oint \vec{E}\cdot d\vec{l}=0\ \text{(conductor)}

Perfect interior shielding falls straight out of one fact — E = 0 inside a conductor in equilibrium — with no need to know anything about the external sources.

Symbol	Name	SI unit	Dimension	Range
$E_ext$	External field Field applied outside the cage	kV/m	M·L·T⁻³·I⁻¹	0 – 200
$d$	Mesh hole size Aperture size of the conducting mesh	mm	L	0.1 – 5
$λ$	Wavelength Wavelength of the incident radiation	mm	L	1 – 100
$E_in$	Leaked interior fieldoutput Residual field inside (fraction of E_ext)	kV/m	M·L·T⁻³·I⁻¹	0 – 200

Unit systems

Symbol	SI	CGS	Imperial
$E_ext$	kV/m	statV/cm	kV/m
$d$	mm	mm	in
$λ$	mm	mm	in
$E_in$	kV/m	statV/cm	kV/m

Where it holds

Exact for static fields inside any closed conductor. For time-varying fields, shielding degrades when holes approach

\lambda

, at very low frequencies for thin/poorly-conducting shells (skin depth), and against static magnetic fields (which need high-permeability mu-metal, not a conductor).

Dimensional analysis

$E_in \propto E_ext\cdot (d/\lambda )$ : $(kV/m)\cdot (mm/mm) = kV/m \checkmark (d/\lambda$ dimensionless)

Discovery

Michael Faraday · 1836

Faraday lined a room with metal foil and sat inside with sensitive electroscopes while the exterior was charged to high voltage. No field penetrated — his 'ice-pail' and cage experiments proved charge resides on the outside of a conductor and the interior is shielded.

Try this

Lightning hits a car and the people inside are fine; your phone loses signal in an elevator. Same physics — what's happening to the field?

Explain why the electric field inside a hollow conductor is zero, and estimate how much a wire mesh with hole size d leaks a wave of wavelength λ.

Research status: stable

Real-world applications

Microwave-oven door screens
Shielded rooms for sensitive measurements (MRI, EEG)
Aircraft and cars protecting occupants from lightning
Coax cable shields and EMI enclosures

Common misconceptions

The cage must be grounded to work — grounding sets the potential but interior shielding happens regardless
A cage blocks magnetic fields too — static B passes through; you need mu-metal
Any metal box is a perfect RF shield — seams and holes near $\lambda$ leak badly

Experimental verification

Faraday's foil-lined room, the classic 'man in a cage' lightning demos, and microwave-oven door screens all confirm it. Place a phone in a metal mesh box and the signal drops by tens of dB as predicted by aperture theory.

Derivation

Inside a conductor in equilibrium

E = 0

(else charges would keep moving).

For a cavity with no enclosed charge, Gauss's law plus the fact that the cavity wall is an equipotential forces

E = 0

everywhere inside: any field line would have to start and end on the wall,

but \oint E\cdot dl = 0

around a loop through the cavity forbids it.

Induced charge lives entirely on the outer surface.

Limiting cases

Solid conductor

(d \to 0)

⟶

E_in = 0

A continuous shell perfectly cancels the interior field.

d ≪

\lambda

⟶

E_in \approx 0

Holes much smaller than the wavelength barely leak — a microwave-oven door screen blocks 12 cm waves.

d ≳

\lambda

⟶ wave passesApertures comparable to the wavelength let radiation through — visible light sails through the same door screen.

What if…

What if you stand inside during a lightning strike?

The current flows over the outer skin of the shell; the interior field stays $\sim 0$ and you're safe — the car/cage effect.

What if the field is a static magnet?

A plain conductor won't shield it — magnetic field lines pass through. You need high-permeability mu-metal to reroute them.

What if the mesh holes grow to the size of the wavelength?

Shielding collapses; the apertures act like antennas and the wave leaks straight in.

1

Microwave door screen

Given ·

E ext:: 100
d:: 1
λ:: 122

Find · E_in (approx)

Steps

Microwaves at 2.45 GHz have $\lambda \approx 122\,\mathrm{mm}$
Holes $of \sim 1\,\mathrm{mm}$ give $d/\lambda \approx 0.008$
Leakage is sub-percent — yet visible light $(\lambda \approx 0.0005\,\mathrm{mm}$ ≪ d) passes freely, so you can watch your food

Result ·

d/\lambda = 1/122 \approx 0.008

, so

E_in \approx 0.8\,\mathrm{kV/m}

— under 1% leaks for

2.45\,\mathrm{GHz} (\lambda \approx 122\,\mathrm{mm})

Gauss's Law→Electric Field of a Point Charge→Speed of Electromagnetic Waves→