Circuit Breaker

What a circuit breaker actually protects

A circuit breaker does not protect the appliance plugged into it — it protects the wiring. A cable has a current it can carry indefinitely without its insulation cooking; push past that and the conductor heats up by I²R, the insulation degrades, and eventually you get a fire inside a wall. The breaker is sized to trip before the cable reaches that point. A 16 A breaker on a 2.5 mm² copper cable is a promise that the current will never sit long enough above 16 A to overheat the copper.

There are two completely different threats. A slow overload is a modest, sustained excess — plug a heater into a circuit already running near its limit and you might draw 20 A on a 16 A circuit. Nothing is on fire yet; the cable can carry that for a while. A short circuit is a catastrophic one — live touches neutral or earth directly and the only thing limiting current is the impedance of the supply network. That can be thousands of amps, and the cable would melt in a fraction of a second. The breaker must answer both, and it uses two separate sensing elements to do so.

The thermal element: a bimetal strip

The overload sensor is a bimetal strip — two metals with different thermal expansion coefficients bonded face to face, typically a high-expansion alloy like FeNiMn against a low-expansion Invar (FeNi36). When current flows through (or near) the strip, I²R heating warms it, the high-expansion side grows more than the low-expansion side, and the strip curls toward the low-expansion side. The tip deflection of a cantilevered bimetal strip is:

δ = K · (α₁ − α₂) · ΔT · L² / t

where:
  α₁, α₂ = expansion coefficients of the two metals (1/K)
  ΔT     = temperature rise (K)
  L      = free length of the strip (m)
  t      = total strip thickness (m)
  K      ≈ 0.75  (geometry constant, ~3/4 for equal-thickness layers)

The deflection grows with the square of length and inversely with thickness, so makers tune sensitivity with geometry. The strip is heated either directly (current passes through it) or indirectly (a heater coil wrapped around it), and once the tip moves far enough it nudges a trip latch and the breaker opens. Because heating takes time, the response is deliberately slow: a small overload trips after minutes, a larger one after seconds. This is the inverse-time behaviour — the bigger the overload, the faster the trip — and it mirrors the thermal time constant of the cable being protected, so the breaker rides through harmless transient surges but acts on real sustained overloads.

The magnetic element: a solenoid for fast faults

A bimetal is far too slow for a short circuit; by the time it warmed up the cable would be gone. So the breaker carries a second element: a solenoid coil (the magnetic trip) carrying the load current. The magnetic force on the solenoid's plunger scales with the square of current:

F = (N·I)² · μ₀ · A / (2 · g²)

where:
  N  = turns on the coil
  I  = instantaneous current (A)
  μ₀ = 4π × 10⁻⁷  (T·m/A)
  A  = pole-face area (m²)
  g  = air gap (m)

At normal current the force is tiny and a spring holds the plunger out. When current jumps to many times rated — a multiple set by the trip curve — the F ∝ I² force overwhelms the spring almost instantly, the plunger flies in, strikes the trip bar, and the contacts blow open. There is no thermal delay: a Type C 16 A breaker fed a 200 A fault trips magnetically in well under 10 milliseconds. The two mechanisms share one trip bar, giving the composite inverse-time trip curve: a sloped thermal region for moderate overcurrent and a vertical magnetic cliff for short circuits.

Trip curves and breaker types

	Type B	Type C	Type D	Fuse (gG)	RCD / GFCI
Magnetic trip threshold	3–5 × Iₙ	5–10 × Iₙ	10–20 × Iₙ	melting curve	n/a (residual)
Senses	Overcurrent	Overcurrent	Overcurrent	Overcurrent	Earth leakage
Trip quantity	Phase current	Phase current	Phase current	Phase current	L − N imbalance
Typical setting	—	—	—	—	30 mA / 100 mA / 300 mA
Resettable	Yes	Yes	Yes	No (replace)	Yes
Speed on hard fault	< 10 ms	< 10 ms	< 10 ms	< 5 ms	< 40 ms (shock)
Best for	Lighting, sockets, long cables	Mixed loads, small motors, fluorescent banks	Transformers, large motors, high inrush	Industrial mains, high-fault points	Personnel shock protection

The thermal portion of the curve is essentially the same across Types B, C and D — only the vertical magnetic threshold shifts. Choosing a higher type avoids nuisance tripping on inrush but, crucially, it raises the fault current needed to trip instantaneously, so you must verify the cable's actual prospective fault current is high enough to reach that threshold. A Type D breaker at the end of a long, high-impedance cable run may never see enough fault current to trip magnetically — a real and dangerous design trap.

Worked example: how long does an overload take to trip?

The thermal trip approximates a fixed-energy (I²t) limit on the bimetal heater. Consider a breaker whose bimetal needs a temperature rise that corresponds to about 40 A²·s of heating energy above the rated holding current, on a 16 A breaker carrying a 1.45× overload (23.2 A). The excess heating relative to rated scales as the square of the multiplier:

Overload multiple m = 23.2 / 16 = 1.45
Excess heating ∝ (m² − 1) = (2.10 − 1) = 1.10  (relative units)

Trip time t ≈ thermal-energy constant / excess heating
            ≈ (large) → order of tens to a few hundred seconds

By standard (IEC 60898) the "conventional tripping current"
is 1.45 × Iₙ and the breaker MUST trip within 1 hour;
the "conventional non-tripping current" is 1.13 × Iₙ
and it must NOT trip within 1 hour.

The exact time depends on the breaker's thermal model, but the takeaway is the structure: a 13% overload is tolerated indefinitely, a 45% overload clears within an hour, and a 200% overload clears in seconds — the inverse-time slope. This is why the standard specifies a band, not a single point: every bimetal has a slightly different cold-start temperature and thermal mass, so manufacturers guarantee a tripping region.

Worked example: short-circuit force on the solenoid

Suppose a short circuit drives a peak instantaneous current of 200 A through a magnetic coil of N = 50 turns, pole area A = 1 × 10⁻⁴ m², air gap g = 1.5 mm:

F = (N·I)² · μ₀ · A / (2 · g²)

  N·I = 50 × 200 = 10,000 A-turns
  (N·I)² = 1.0 × 10⁸
  μ₀·A = 4π×10⁻⁷ × 1×10⁻⁴ = 1.26 × 10⁻¹⁰
  2·g² = 2 × (1.5×10⁻³)² = 4.5 × 10⁻⁶

F = 1.0×10⁸ × 1.26×10⁻¹⁰ / 4.5×10⁻⁶
  = 1.26×10⁻² / 4.5×10⁻⁶
  ≈ 2,800 N

Roughly 2,800 newtons — the weight of a small car — slams the plunger inward, easily overcoming a hold-off spring of a few newtons. Because force scales as I², the same coil at normal 16 A current produces only about (16/200)² ≈ 0.6% of that, around 18 N reduced further by the larger working gap — comfortably held by the spring. The squared dependence is what makes the magnetic trip sharply selective: nothing at rated current, a sledgehammer on a fault.

Killing the arc

Opening the contacts does not stop the current — it converts it into an arc, a column of ionized plasma hotter than 10,000 K that conducts almost as well as the metal it replaced. Interrupting the arc is the hardest part of the breaker's job, and it is what sets the interrupting (breaking) rating in kiloamps. Several mechanisms cooperate:

• Contact separation speed. A fast-acting spring blows the contacts apart quickly so the arc is stretched and starved before it can stabilize.
• Magnetic blowout. The arc is itself a current-carrying conductor, so a field (from the fault current looping through "arc runners") pushes it upward and lengthens it — the same F = BIL force used in a motor.
• Arc chute / splitter plates. A stack of steel plates divides the long arc into many short series arcs. Each short arc has a fixed cathode/anode voltage drop, so N plates raise total arc voltage to N times that drop, pushing arc voltage above the supply voltage.
• Current zero. In AC, current naturally passes through zero 100 or 120 times per second. Once the arc voltage exceeds the source and the gap deionizes, the arc cannot reignite after a zero crossing — and the current is interrupted.

This is why AC breakers are far easier to design than DC: AC hands you a free current zero every half cycle, while a DC arc has none and must be force-extinguished, which is why DC breakers (and EV/solar disconnects) need much larger gaps, blowout magnets, or solid-state switching.

Interruption technologies by voltage class

• Air (MCB / MCCB / ACB): arc chute in air; the workhorse from household 6 A up to substation 6300 A low-voltage.
• Vacuum: contacts open in a sealed vacuum bottle; the arc cannot sustain without gas to ionize, so it dies almost instantly. Dominant for medium voltage (1–38 kV) — switchgear, distribution.
• SF₆ gas: sulfur hexafluoride is an excellent electron scavenger that quenches arcs aggressively; used at high and extra-high voltage, though its global-warming potential is driving a shift to clean-air and vacuum alternatives.
• Oil: older high-voltage breakers used the arc to crack oil into hydrogen, cooling and de-ionizing the gap; largely retired for environmental reasons.
• Solid-state / hybrid: IGBTs or thyristors switch off without mechanical contacts, giving microsecond interruption — essential for DC grids and fast fault isolation, at the cost of conduction losses.

Failure modes and trade-offs

• Contact welding. On a fault near the breaker's interrupting limit, the inrush can momentarily weld the contacts shut before they fully part — the breaker fails to open and the fault persists. This is precisely what the kA interrupting rating guards against; never install a breaker rated below the prospective fault current.
• Nuisance tripping. Motor and transformer inrush, or LED-driver capacitor charging, can briefly exceed the magnetic threshold and trip a B-curve breaker on switch-on. The fix is selecting a higher trip type (C or D), not increasing the current rating, which would under-protect the cable.
• Failure to trip on a remote fault. A high-type breaker behind a long, thin cable may never see enough fault current to reach its magnetic threshold; only the slow thermal element responds, which can be too slow to protect people from shock — hence the cable-length and earth-loop-impedance limits in wiring codes.
• Thermal drift. A bimetal's trip point depends on ambient temperature; a breaker in a hot enclosure trips earlier (already pre-heated), one in the cold trips later. Datasheets give a temperature-derating curve, and adjacent breakers heat each other.
• Arc-flash energy. Slower clearing means more energy dumped into an arc-flash event; this is why protective coordination and fast breakers matter for worker safety, quantified as incident energy in cal/cm².
• Contact erosion. Every fault interruption burns a little contact metal away. Breakers have a rated number of operations at full fault current (often just a handful) versus thousands at rated load; after a major fault, inspection or replacement may be required.