HRC Fuse: The Silver Strip That Kills a Fault Before It Peaks

What an HRC Fuse Is, and What “HRC” Really Means

An HRC fuse (high rupturing capacity, also called HBC, high breaking capacity) is a sealed cartridge built to interrupt the largest fault current an installation can deliver without bursting or venting flame. The rating that earns the name is its breaking capacity: typically 80 kA for a BS 88 industrial link at 400 V, 120 kA for an NH knife-blade link at 500 V, and 200 kA for a UL Class J fuse. “HRC” says nothing about load current; the same construction is sold from a few amps to well over 1,000 A.

• Body: a ceramic barrel that survives internal pressure and thermal shock.
• End caps: brass or plated copper, bolted, clipped or blade-mounted into the base.
• Element: thin silver strips punched with rows of holes that leave narrow necks. Silver conducts best of all metals and builds up no insulating oxide, so its behaviour stays stable for decades.
• Filler: high-purity quartz sand, compacted around the element.
• Indicator: a fine parallel wire that melts after the element and releases a flag or striker pin.

A gG link (“g” full-range, “G” general purpose) clears any current from its minimum fusing current up to its breaking capacity, protecting cables against overload and short circuit. An aM link breaks only large overcurrents and relies on a motor overload relay for the rest.

Step by Step: From Dead Short to Glass

1. Current crowds into the necks. Where a neck removes three-quarters of the strip's width, current density is 4× higher and Joule heating per unit volume, proportional to J², is 16× higher.

2. Heating is adiabatic. On a short circuit the current climbs by tens of kiloamps per millisecond, far too fast for heat to escape. A neck's temperature then depends only on the integral ∫J²dt. For silver, reaching the 961.8 °C melting point takes roughly 6×10¹⁶ A²s/m⁴. When the necks get there, the wide strip has received one-sixteenth of that and has warmed by only ~25 °C. Because the necks are cut to the same size and carry the same current, they all melt at almost the same moment, within the first quarter-cycle: under 5 ms at 50 Hz.

3. Each neck becomes an arc. The molten silver breaks up, the gap ionises and an arc strikes. A strip with eight rows of necks now carries eight arcs in series, and their voltages add. Higher voltage ratings use more rows.

4. The arcs burn back. Each arc erodes silver on both sides of its gap, so the gaps lengthen and total arc voltage keeps climbing. Squeezed between sand grains, the arc is cooled hard and runs at a far higher voltage gradient than an arc in open air.

5. The sand swallows the energy. Silver vapour condenses onto the grains, and silica (melting ~1,700 °C) around each arc channel fuses into a glassy fulgurite, the name also given to glass formed where lightning strikes sand. Nothing is vented.

6. Current reaches zero and stays there. The fulgurite and silver-laced sand form an insulating plug that withstands the returning supply voltage. The fault is cleared within half a cycle, under 10 ms at 50 Hz, before the current's natural zero.

The Governing Relation: Arc Voltage Must Beat the Supply

Around the fault loop, v_supply = L·di/dt + R·i + v_arc. Neglecting resistance gives L·di/dt = v_supply − v_arc. While the necks are solid, v_arc ≈ 0 and the current rises. Once total arc voltage exceeds the instantaneous supply voltage, di/dt turns negative and the current is driven down. A 400 V line peaks at ~565 V (400 × √2), so a 400 V fuse must reliably generate more than that.

Worked example. Take a 400 V supply able to deliver a 50 kA rms fault, modelled as a pure inductance: L = 400 / (2π × 50 × 50,000) ≈ 25 µH. Unchecked, the first peak would be 50 × √2 ≈ 70.7 kA, and more if the fault struck at a point on the wave that adds a DC offset. The steepest rise is 565 V ÷ 25 µH ≈ 22 kA per millisecond.

• Melting: a linearly rising current accumulates I²t = (di/dt)²·t³/3. If a 100 A gG element needs ~15,000 A²s to melt its necks (a sub-millisecond, near-adiabatic figure, below the 0.01 s values used for selectivity), that arrives at t ≈ 0.45 ms, when the current is only ~10 kA.
• Arcing: suppose the series arcs build to ~800 V while the supply sits near its 565 V peak (as it does when a symmetrical inductive fault current starts from zero). Then di/dt ≈ (565 − 800) ÷ 25 µH ≈ −9 kA per millisecond, and the current reaches zero in about a millisecond. Real arc voltage takes time to build, so arcing lasts a few milliseconds, still well inside half a cycle.
• Result: a 50 kA fault that would peak above 70 kA lets through only ~10 kA. Electromagnetic forces scale with i², so busbars feel about (70.7 ÷ 10)² ≈ 50× less peak force. The prospective I²t of one half-cycle is 50,000² × 0.01 = 2.5×10⁷ A²s; the fuse lets through of order 10⁵ A²s, hundreds of times less.

The arc energy, ∫v_arc·i dt, is several kilojoules, including ~1.3 kJ stored in the loop inductance (½LI² at 10 kA). Heating and fusing quartz takes roughly 2 kJ per gram, so a few grams of sand turn to glass. Arc voltage cannot simply be made enormous, because the overshoot stresses the rest of the installation; standards cap it, and designers shape the necks so it rises in a controlled way.

Overloads, the M-Effect and I²t Discrimination

Adiabatic necks handle short circuits, but a sustained overload needs a fuse that operates without first heating a strip to 961.8 °C while the fuse body and holder cook. In 1939 A. W. Metcalf described what became known as the M-effect: a small spot of tin on the silver element. Under overload the element runs hot enough to melt the tin (~232 °C), and the liquid tin dissolves the silver beneath it, forming lower-melting alloys that part the strip at a few hundred °C. In a short circuit there is no time for this diffusion, so the spot plays no part.

IEC 60269 sets the overload window for gG links of 16 A and above. A conventional non-fusing current of 1.25 × rating must not blow the fuse, and a conventional fusing current of 1.6 × rating must, within a conventional time of 1 h for ratings up to 63 A, 2 h up to 160 A, 3 h up to 400 A and 4 h above.

Discrimination (selectivity) means only the fuse nearest the fault operates. For fuses it is decided by I²t, not by current: the downstream link's total operating I²t (pre-arcing plus arcing) must be smaller than the upstream link's pre-arcing I²t, so the upstream necks never start to melt. IEC 60269 builds this in with staggered I²t limits at 0.01 s: the upper limit for a 63 A gG link, about 27×10³ A²s, is also the lower limit for a 100 A link. The ratings differ by 100 ÷ 63 ≈ 1.59, which is why gG fuses are selective at a 1.6:1 rating ratio.

History, Real Hardware and How Fuses Are Tested

Fuses are as old as electric lighting; in the 1880s William Preece measured the currents at which wires of different diameters melt. As networks grew and fault levels reached tens of kiloamps, the enclosed, sand-filled cartridge with a notched element replaced bare wire. It was well established by the 1930s, when Britain formed ASTA (1938) to certify equipment by short-circuit testing. Today's families include:

• BS 88 (IEC 60269-2 and -3): British bolted and clip-in industrial links plus domestic cut-out fuses, with 80 kA at 400 V typical for industrial links.
• NH knife-blade (IEC 60269-2, DIN VDE 0636): sizes 000 to 4a covering up to 1,250 A, withdrawn with an insulated puller handle; 120 kA at 500 V.
• UL Class J (UL 248-8): 600 V, up to 600 A, 200 kA interrupting rating, with unique dimensions so the holder rejects lower-performance fuse classes.
• Semiconductor aR links (IEC 60269-4): many very fine necks for the lowest possible I²t, because a thyristor or diode junction can be destroyed in milliseconds.

Breaking capacity is proven, not calculated. Links are type-tested in high-power laboratories at low power factor and at or above rated voltage, through a sequence of test duties from the full rated breaking capacity down to small overcurrents. It includes the current that produces the most arc energy, because a fuse that breaks 80 kA cleanly can still fail at a much lower current where the arc lingers. Oscillograms yield the cut-off current, pre-arcing and operating I²t and peak arc voltage, published as time-current curves, cut-off charts (let-through against prospective current) and I²t tables.

How HRC Fuses Fail or Get Misused

• Wrong voltage, or AC links on DC. Arc voltage must beat the supply, so a link rated below the circuit voltage may never clear. DC has no natural current zero, so solar strings and battery banks need DC-rated links, such as gPV fuses to IEC 60269-6.
• Fault level above the breaking capacity. The arc outlasts what the sand can absorb, and the ceramic can crack or burst.
• Too little fault current. A long cable or a generator supply may keep the fuse out of its current-limiting region, so it clears slowly and arc-flash energy can exceed that of a larger fault.
• aM links without an overload relay. A partial-range link facing a sustained overcurrent below its minimum breaking current can overheat instead of clearing.
• Single-phasing. When one fuse of three blows, a three-phase motor keeps running on two phases and overheats. Strikers linked to a switch disconnect all three.
• Fatigue and hidden damage. Heavy cyclic loads fatigue the necks and can cause nuisance operation years later; after a severe fault, the unblown fuses in a set may be damaged, so good practice replaces them all.
• Bridging. A bolt, bar or wire in place of a link has no breaking capacity or current limitation, so the next short circuit becomes an explosion inside the switchboard.

HRC Fuse vs Rewirable Fuse vs MCB

Rewirable fuses (BS 3036) hold a bare tinned-copper wire in a porcelain carrier. The wire melts and arcs in air in a semi-enclosed space, with no sand, no necks and no meaningful current limitation. Breaking capacity is only 1–4 kA, anyone with a reel can fit the wrong gauge, and the fusing factor can approach 2. That is why BS 7671 requires the cable they protect to be rated so that the fuse rating is at most 0.725 × the cable's current capacity.

Miniature circuit breakers (BS EN 60898-1) use a bimetal for overloads and a magnetic trip that throws the contacts apart on a short circuit, driving the arc into steel splitter plates. They are resettable, but their rated short-circuit capacity is typically 6 or 10 kA. Where the prospective fault is higher, an upstream HRC fuse provides back-up protection, limiting the let-through so the breaker survives.

The HRC cartridge is single-use, but no other device its size combines 80–200 kA breaking capacity, current limitation within half a cycle, no moving parts and selectivity at a simple 1.6:1 ratio. That is why it still guards service heads and industrial switchgear, while MCBs handle the final circuits.

HRC fuse versus the devices it is confused with

Device	Typical breaking capacity	How the arc is quenched	Limits the first peak?
HRC cartridge (BS 88, NH, UL Class J)	80–200 kA	Series arcs at silver necks; energy absorbed by quartz sand, which fuses into fulgurite	Yes: cut-off within the first quarter-cycle, cleared within half a cycle
Rewirable fuse (BS 3036)	1–4 kA (S1A, S2A, S4A)	Single copper wire melts and arcs in air in a semi-enclosed carrier	No meaningful limitation; fusing factor up to ~2
MCB (BS EN 60898-1)	Typically 6 or 10 kA (up to 25 kA)	Contacts part; arc driven into steel splitter plates	Partly (energy-limiting class 3); resettable
Miniature 5 × 20 mm fuse (IEC 60127)	35 A (glass, low-breaking) to 1,500 A (sand-filled ceramic, high-breaking) at 250 V	Wire vaporises in air or sand	Only at small fault levels
Semiconductor aR fuse (IEC 60269-4)	~100–200 kA	Many very fine silver necks for the fastest arc build-up	Yes, with the lowest I²t of any fuse type