Overcurrent protection field guide: breakers and fuses

Overcurrent protection, and the two things it guards against

Overcurrent protection is the breaker or fuse that opens a circuit when the current climbs past what the conductor can carry without overheating. The device that does it is called an overcurrent protective device, an OCPD. Every branch circuit, feeder, and service has one, and its whole job is to take a circuit offline before the wire gets hot enough to damage its insulation or start a fire.

There are two kinds of overcurrent, and they are not the same problem. An overload is too much current for too long, a circuit asked to carry more than it was built for, climbing maybe to two or three times its rating. A short circuit, or fault, is a sudden, enormous current that appears in a fraction of a cycle when a hot conductor touches a neutral, a ground, or another phase. One is a slow cook. The other is an explosion looking for a place to happen. The OCPD has to handle both, with different parts of itself, and that dual job is the core of everything that follows.

Get the device right and a fault is a tripped breaker and a service call. Get it wrong, undersized for the fault it can see, and the same fault can blow the gear apart. The size of the available fault current at each point is its own study, covered in the available fault current guide, and it feeds directly into the interrupting rating you read about below.

What is the difference between an overload and a short circuit?

An overload is current above the circuit rating flowing through the normal path, and a short circuit is a fault that creates a new, unintended path with almost no resistance. Both are overcurrents, but the magnitude and the timing are worlds apart, and a good OCPD treats them differently on purpose.

An overload is moderate and it builds. Plug too many tools into one circuit, stall a motor, or load a feeder past its design, and the current rises to maybe 150 to 600 percent of rating. Nothing is broken yet. The conductor just heats up over seconds or minutes, and the device gives it a deliberate time delay so a momentary inrush, like a motor starting, does not trip the circuit for no reason. The protection here is inverse-time: the more the overload, the faster it opens.

A short circuit is violent and instant. When insulation fails or a wire crosses, the only thing limiting the current is the impedance of the wire back to the source, so the current can jump to thousands or tens of thousands of amps in the first half-cycle. There is no time to wait. The device has to open as fast as it can, instantaneously, and it has to be built to interrupt that much current without failing. That last part, the interrupting rating, is where undersized gear gets people hurt.

What does a breaker or fuse actually protect?

The breaker protects the conductor, not the equipment plugged into it. This is the single most misunderstood idea in overcurrent protection, and it changes how you size everything. The amp rating on a breaker is not chosen to match the load. It is chosen so the device opens before the wire it feeds overheats.

The number that ties them together is ampacity, the current a conductor can carry continuously without exceeding its temperature rating. A 20 A breaker goes on a 12 AWG copper conductor because 12 AWG copper is good for roughly that current under common conditions, and the breaker opens before the wire cooks. Put a 30 A breaker on that same 12 AWG wire and you have removed the protection: the wire can be overheating badly and the breaker, watching for 30 A, never trips. The conductor becomes the fuse, and it fails by burning.

So the workflow runs in this order. Size the conductor for the load and its ampacity from the NEC ampacity tables, then pick the overcurrent device to protect that conductor. The load tells you the conductor. The conductor tells you the breaker. Equipment with its own internal protection, like a motor or an appliance, is a separate set of rules, but the default rule, the one that catches people, is that the device guards the wire.

How a circuit breaker trips

A standard molded-case circuit breaker, an MCCB, trips two different ways inside one case, one for each kind of overcurrent. The design is called thermal-magnetic, and the name is literal: a thermal element handles overloads and a magnetic element handles short circuits.

The thermal element is a bimetal strip, two metals bonded together that expand at different rates when warmed. Load current passes through it, and on a sustained overload the strip heats and bends, and once it bends far enough it pushes the trip bar and opens the breaker. Because the strip has to heat up first, this gives the inverse-time delay an overload needs: small overloads take longer, big ones trip sooner. That delay is a feature, not a flaw, so a motor inrush or a momentary surge rides through.

The magnetic element is an electromagnet in series with the load. Normal current makes a weak field that does nothing. A short circuit slams thousands of amps through that coil, the magnetic field spikes, and it yanks an armature that trips the breaker right now, with no intentional delay. That is the instantaneous trip. Two mechanisms, one handle: the bimetal watches for the slow cook, the magnet watches for the explosion, and either one drops the circuit.

Circuit breaker types

Breakers split into a few families by how they trip and how they are built, and the family decides what you can adjust and what you cannot. The everyday workhorse is the thermal-magnetic molded-case breaker, fixed trip, sealed case, the kind in nearly every panelboard. The miniature circuit breaker, the MCB, is the residential and light-commercial version of the same idea in a smaller body.

Move up in size and the trip mechanism changes. Larger molded-case and insulated-case breakers, and the low-voltage power breakers that live in switchgear, often carry an electronic trip unit instead of a bimetal and magnet. These read current electronically and let you dial in the trip behavior, which is what makes coordination possible on a big system. Power breakers, sometimes called air-frame breakers, are drawout units built to be maintained, with higher withstand ratings and short-time capability that lets them hold in briefly during a fault so a downstream device clears first.

One distinction worth keeping straight: a UL 489 listed breaker is a branch-circuit overcurrent device and can protect a circuit. A supplementary protector, listed to a different standard, looks similar but is not branch-circuit rated and cannot stand in for one. People swap them without realizing the listing is different.

Electronic trip units and adjustable settings

An electronic trip unit replaces the bimetal and magnet with current sensors and electronics, and the payoff is adjustability. The common feature set goes by the initials LSIG, for the four functions it can carry: Long-time, Short-time, Instantaneous, and Ground-fault. Each is a separately adjustable trip with its own pickup and, where it applies, its own time delay.

Long-time is the overload function, the electronic version of the bimetal. You set a current pickup, often adjustable from roughly 20 to 100 percent of the breaker's sensor rating, plus a time band. Short-time is a delayed response to a fault current: instead of tripping instantly, the breaker holds in for a set fraction of a second so a downstream device gets the chance to clear the fault first. Instantaneous is the no-delay trip for high fault current, with an adjustable pickup commonly in the range of several to fifteen times the rating. Ground-fault adds a trip on current returning through the ground path, set low compared to the phase functions.

Those settings are the tool that makes selective coordination possible, which is its own discipline covered in the selective coordination guide. The catch is that an adjustable breaker ships at some default and stays there until someone sets it. The settings have to be commissioned to match the study, and on too many jobs nobody ever does. A breaker set wrong is a breaker that trips the wrong device or does not trip in time.

How a fuse works

A fuse is a metal element that carries the load current and melts when the current gets too high, opening the circuit. There are no moving parts and nothing to reset. When a fuse operates, it is consumed, and you replace it. That simplicity is both its strength and the thing people hate about it.

The element is shaped and sized so that normal current passes without heating it past its melting point, but an overload heats it until it melts and clears, and a short circuit melts it almost instantly. A good fuse handles both kinds of overcurrent with the same piece of metal, with a time characteristic built into the element's shape. Many fuses run the element through a filler like quartz sand, which quenches the arc as the metal vaporizes and helps the fuse clear cleanly at high fault current.

Because there are no mechanics to wear or contacts to pit, a fuse holds its characteristics over time and clears a high fault fast. The flip side is real. A blown fuse has to be replaced with the exact same type, class, and rating, and on a three-phase load a single blown fuse can leave the load single-phased, running on two legs, which cooks a motor quietly. You also have to stock the right spares, because the wrong fuse grabbed off the shelf in a hurry is a common and dangerous shortcut.

Fuse classes

Fuses are sorted into classes, and the class is not a detail. It sets the physical dimensions, the voltage and interrupting rating, and how current-limiting the fuse is. A class is a package of guaranteed performance, listed under the UL 248 fuse standards, and you do not mix classes by eye because two fuses that look alike can behave nothing alike on a fault.

The common low-voltage classes you meet in the field include Class R, split into RK1 and RK5, Class J, Class L for the large amperages, Class CC for control and small branch circuits, and Class T in a compact body. Within Class R, RK1 and RK5 share the same dimensions but RK1 is more current-limiting than RK5, so it lets through less fault energy, which matters when you are protecting downstream equipment. Class R holders have a rejection feature meant to keep a less-protective older fuse out of a socket built for a current-limiting one.

The specific ratings differ by class and manufacturer. Many of these current-limiting classes carry interrupting ratings well up into the 200 kA range, but the exact interrupting rating, voltage rating, and let-through values are stamped on the fuse and published in the manufacturer's data, and that is what you size to. Confirm the class, the rating, and the current-limiting characteristic from the listing for the fuse you actually install, not from the one that used to be in the holder.

Current-limiting devices and let-through energy

A current-limiting device opens so fast on a high fault that it cuts the fault off before it reaches its full potential peak. Instead of letting the available fault current build through a full half-cycle, a current-limiting fuse or breaker clears it in the first few milliseconds, so the downstream gear and conductors never see the full energy the system could have delivered. The thing it limits is let-through, the peak current and the heating energy that actually gets past the device.

This matters because of what is downstream. A panel or a piece of equipment has its own withstand rating, and if the available fault current exceeds it, a current-limiting device ahead of it can knock the let-through energy down to something the downstream gear can survive. That is the basis of a series rating, covered further down, and it is one of the few legitimate ways to apply equipment in a system with more fault current than the equipment alone is rated for.

Current-limiting is most useful exactly where fault current is highest, near the service, where the numbers from the available fault current study are largest. The let-through values are published as curves and tables for each device. You read the available fault current off the study, then read the device's let-through at that current, then check it against what the downstream equipment can take.

What is an interrupting rating?

An interrupting rating, often called the AIC, is the maximum fault current a breaker or fuse can safely open without failing. It is the most important number on the device and the one that hurts people when it is ignored. The rule is not negotiable: the interrupting rating of the device must equal or exceed the available fault current at its terminals. Always.

Here is why. When a fault hits, the device has to break a current that is trying very hard not to be broken, drawing an arc across the opening contacts or the melting fuse element. A device rated for that current quenches the arc and clears. A device with an interrupting rating below the available fault current can fail to clear, sustaining the arc, and the gear can rupture, throwing molten metal and pressure. The breaker does not just fail to protect. It becomes the hazard.

The available fault current is found by the short-circuit study, covered in the available fault current guide, and it is highest near the source and lower as you move downstream through impedance. A breaker might see 42 kA at the main and 14 kA three panels down. A 10 kA breaker is fine at the panel and a bomb at the main. Check the rating against the calculated fault current at that specific point, every device, and never assume a standard breaker covers it. Common breaker interrupting ratings run from 10 kA up to 65 kA and beyond, and the available fault current decides which you need.

The time-current curve

Every OCPD trips according to a time-current curve, a plot of how long the device takes to open against how much current is flowing. Current runs along the bottom, time runs up the side, both on log scales, and the curve slopes down to the right: more current, less time to trip. Reading it is how you know what the device will actually do, instead of guessing.

The curve has regions that match the two overcurrents. The upper, sloping part is the overload region, the inverse-time behavior, where a small overload takes a long time and a larger one trips faster. Down at the bottom the curve goes nearly vertical, the instantaneous region, where any current above the magnetic pickup trips in a cycle or two regardless of how much. On a fuse the curve is a band between minimum melting time and total clearing time. On an electronic-trip breaker the L, S, and I settings literally reshape the curve, which is what makes those breakers coordinate.

Lay two devices' curves on the same plot and you can see whether they coordinate: if the downstream device's curve sits entirely below and to the left of the upstream device's, the downstream one clears first and the fault stays local. Where the curves overlap, both can trip, and you get a cascading outage. Keeping those curves apart across the full range of fault current is the whole subject of the selective coordination guide.

What is the difference between a breaker and a fuse?

A breaker is a resettable mechanical switch that trips and is switched back on, while a fuse is a one-time element that melts and gets replaced. That is the headline difference, and from it follows almost every reason to pick one over the other.

Breakers reset, so after a trip you find the problem and flip the handle, with no parts to stock. Larger breakers can be adjustable, which is what coordination on a big system needs. They cost more up front and they have moving parts that age, pit, and need exercising and testing to stay reliable. Fuses are cheap at the element, clear a high fault very fast, and current-limiting classes do it better than most breakers, which protects downstream gear. The price is that every operation consumes a fuse, you must stock the exact replacement, and a single opened fuse can single-phase a three-phase load.

In practice you see breakers in panelboards and most distribution because resetting beats replacing for everyday faults, and you see fuses where fast current-limiting clearing or a very high interrupting rating is wanted, often in switches feeding motors and in service equipment with high available fault current. Neither is simply better. Match the device to what the circuit needs and to what the available fault current demands.

GFCI: ground-fault circuit interrupter for people

A GFCI protects people, not equipment, and not the conductor. It watches the current going out on the hot and coming back on the neutral, and if those do not match, current is leaking somewhere it should not, possibly through a person, and the device opens. The Class A GFCI used for personnel protection trips at a ground-fault current in the range of about 4 to 6 milliamps, far below the level that would trip a normal breaker on overcurrent.

That is the key idea. A regular breaker cannot protect you from a shock. A few milliamps through a person is deadly but is nowhere near the amps a breaker is watching for. The GFCI fills that gap by responding to the imbalance, the leakage, not the overload. It comes as a receptacle, as a breaker that protects the whole branch circuit, and built into some cord ends.

GFCI protection is required in the wet and grounded locations where shock risk is highest, bathrooms, kitchens, garages, outdoors, and similar, with the list expanding each code cycle. The device has a test button for a reason. It uses electronics that can fail, and a GFCI that has quietly failed still passes current and still looks fine. Test it on the schedule, and confirm the required locations against the adopted code edition rather than memory.

AFCI: arc-fault circuit interrupter

An AFCI detects the electrical signature of a dangerous arc and opens the circuit before that arc can start a fire. A normal breaker cannot see an arc. A loose terminal or a nicked conductor can arc and throw sparks for a long time at a current well below the breaker's trip point, which is exactly how a lot of electrical fires begin. The AFCI listens for the chaotic current pattern an arc makes and trips on that.

There are two failure modes it cares about. A series arc is a break in a single conductor, like a loose screw at a device or a frayed cord, where the current jumps the gap. A parallel arc is arcing between conductors, hot to neutral or hot to ground. The combination-type AFCI required for dwelling branch circuits detects both, which is why the older branch-feeder type is no longer accepted for that use in current editions.

AFCI protection is required on most 120 V, 15 and 20 A branch circuits in dwelling units, and that scope has widened over several code cycles. The honest field note: AFCIs are sensitive, and nuisance trips on certain loads and on shared neutrals are a real source of callbacks. The fix is usually correct wiring and a good termination, not defeating the device. Confirm the required circuits against the adopted edition, because this is one of the faster-moving requirements in the code.

Ground-fault protection of equipment (GFP)

Ground-fault protection of equipment, GFP or GFPE, is a different animal from the GFCI. It protects the equipment and the building from the damage a sustained arcing ground fault does on a large service, not people from shock. A ground fault on a big solidly grounded system can draw a current too low to trip the main quickly but high enough to burn through gear, and GFP catches that middle ground.

The NEC requires this protection on large services, commonly cited at 230.95 for solidly grounded wye services of more than 150 volts to ground and rated 1000 A or more, the classic case being a 480Y/277 V service. The maximum trip setting for that equipment protection is limited, commonly to 1200 A, with a maximum time delay at a defined fault level. Those numbers, the threshold amperage and the setting limit, have shifted across code cycles, so confirm the current values against the adopted edition rather than treating these as fixed.

Two field cautions. The setting is up high, so GFP is not shock protection and is not a substitute for a GFCI. And GFP can defeat coordination if it is not studied: a ground fault on a downstream circuit can trip the main on its ground-fault function and black out everything, which is the opposite of what you want on critical power. Performance testing of GFP is required and it has to be set to match the study.

Sizing the overcurrent device

Size the overcurrent device to protect the conductor, then check it against the load and the standard ratings. Start from the conductor's ampacity, the current it can carry from the NEC ampacity tables under the actual installation conditions, after any temperature and bundling corrections. The device is then chosen so it opens before that ampacity is exceeded.

Two rules shape the number. Standard overcurrent ratings come from a fixed list in the NEC, the familiar 15, 20, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90, 100, 110, 125, 150, 175, 200 A and up, so you pick from that list. And where the conductor's ampacity does not land on a standard rating, the code generally lets you round up to the next standard size, within limits and with exceptions for the smaller conductors, commonly handled at 240.4 and its subsections. Those limits matter and they are not the same for 14, 12, and 10 AWG, which have their own fixed maximum protection.

Continuous load adds the other adjustment. A load that runs at its maximum for three hours or more is continuous, and the conductor and the overcurrent device are sized at 125 percent of that continuous current, so the device is loaded to no more than 80 percent of its rating. That is why a 100 A continuous load wants a 125 A device and conductor. Confirm the exact article and section numbers against the adopted edition, because they move, but the order does not: protect the conductor, round to a standard size within the limits, and add the 125 percent for continuous load.

Series rating vs fully rated

A fully rated system uses devices that each individually meet or beat the available fault current at their location. A series-rated system uses a tested combination, where an upstream current-limiting device protects a downstream device that has a lower interrupting rating than the fault current it would otherwise see. Both are legitimate. They are not interchangeable, and the difference trips people up at inspection.

The thing that makes a series rating valid is that it was tested as a combination by the manufacturer and is listed and marked as such. You cannot mix and match an upstream and downstream device on paper and call it series rated because the let-through math looks like it works. It has to be a tested, listed pairing, and the equipment has to be field-marked to show the series combination and the rating it achieves. Engineering a series rating from let-through curves is allowed only under narrow conditions and is not the everyday path.

There is one more limit that catches people: a series rating is generally not allowed where a motor is connected between the upstream and downstream devices, because the motor contributes its own current to a fault and can push the downstream device past what the test assumed. When in doubt, fully rate it. A series rating saves money on the downstream gear, but a fully rated system has no combination dependency to get wrong later when someone swaps a breaker.

Protection vs coordination, and critical power

Protection and coordination are two different bars, and clearing the first does not clear the second. Protection means the device opens on an overcurrent and the gear survives, which is the floor every system has to meet. Coordination, specifically selective coordination, means that when a fault happens only the device nearest the fault opens and everything upstream stays closed. A system can be fully protected and still coordinate terribly.

On ordinary commercial work, a cascading trip is an annoyance: a fault knocks out a panel or a floor, you reset, you move on. On critical power it is the event everyone is paid to prevent. A data center bus that drops because a single cabinet fault tripped the upstream main is not a reset. It is racks rebooting, storage rebuilding, and a service-level penalty. A hospital wing that goes dark because a branch fault opened a feeder is a patient-safety problem. That is why the code mandates selective coordination for emergency, legally required standby, and critical operations power systems.

Getting there is the work of the selective coordination guide, and it leans on everything here: the time-current curves, the adjustable LSIG settings on electronic-trip breakers, the short-time function that lets an upstream breaker hold in while a downstream device clears, and fuse selectivity ratios. The settings that achieve it have to be commissioned into the gear and verified before turnover, against the available fault current at each point. Coordination is not a property of the device. It is a property of how the whole set was selected and set.

Breaker maintenance and testing

A breaker is a mechanical device, and a mechanical device that sits closed for years can fail to open when it finally has to. The contacts pit, the lubricant hardens, the trip mechanism stiffens. A breaker that has never moved since startup is the one most likely to hang up on a fault, which is the worst possible time to find out.

The maintenance answer is to exercise and test on a schedule. Exercising means cycling the breaker open and closed to keep the mechanism free, and on larger gear it means primary or secondary injection testing to confirm the breaker actually trips at and near its settings and clears in the expected time. NETA acceptance and maintenance testing standards give the procedures and intervals, and on critical power the testing also confirms the trip settings still match the coordination study, since a setting can drift or get changed.

The field reality is that nobody budgets for this until a breaker fails to trip and takes out gear that should have survived. Insulated-case and power breakers are built to be maintained and racked out for service for exactly this reason. Molded-case breakers are largely sealed, but they still get tested at acceptance and exercised, and a breaker that has tripped on a hard fault should be evaluated rather than just reset, because clearing a high fault is hard on the contacts.

What to document

An overcurrent device that nobody recorded is a device nobody can verify later. When a setting is questioned, or a breaker is swapped, or the available fault current changes because the utility upgraded a transformer, the record is what tells you whether the protection still holds. Capture it at install and update it when the gear changes.

For each device, record what it is, what it trips on, and the ratings that make it safe at that location. The interrupting rating against the available fault current at that point is the line that has to be defensible, because that is the one that gets people hurt when it is wrong.

Common mistakes

• Installing a device with an interrupting rating below the available fault current at its terminals. This is the one that ruptures gear.
• Oversizing the breaker past the conductor ampacity so the wire becomes the fuse.
• Treating the breaker as protection for the load instead of the conductor.
• Swapping in the wrong fuse class or rating, or grabbing a less current-limiting fuse off the shelf.
• Leaving an adjustable electronic-trip breaker at its default settings instead of commissioning it to the study.
• Ignoring the continuous-load 125 percent adjustment when sizing the device and conductor.
• Calling a system series rated without a tested, listed, field-marked combination.
• Confusing GFP equipment protection with GFCI shock protection, or counting on a breaker to protect a person.

Field checklist

Standards and references

The NEC, NFPA 70, is where the installation rules live. Article 240 covers overcurrent protection: the standard ampere ratings, the rules for protecting conductors, and the next-standard-size allowance with its limits for the smaller conductors. Conductor ampacity comes from the ampacity tables in Article 310, with the correction and adjustment factors that change what the conductor can actually carry. Continuous-load sizing at 125 percent and the branch and feeder rules sit in Articles 210 and 215. Ground-fault protection of equipment on large services is commonly cited at 230.95, GFCI and AFCI requirements at 210.8 and 210.12, and series ratings at 240.86. Article and section numbers shift between code cycles, so confirm each against the edition the jurisdiction has adopted and any local amendments before you cite it on a submittal.

The devices themselves are listed under UL standards. Molded-case circuit breakers are listed to UL 489, which is what makes a breaker branch-circuit rated rather than a supplementary protector. Fuses fall under the UL 248 series, which defines the classes and their performance. NEMA standards cover device construction and ratings, and NETA gives the acceptance and maintenance testing procedures for the gear once it is installed.

The interrupting rating, the available fault current, and the trip settings are not generic. They come from the specific device listing, the short-circuit study, and the coordination study for the actual installation. Cite the standard that controls the point, size to the listing for the device you install, and let the engineered study govern the settings.

Units, terms, and abbreviations

Overcurrent protection carries a stack of acronyms that mean specific things, and they get used loosely on drawings and in conversation, so it helps to keep them straight.

The device is an OCPD, overcurrent protective device, whether it is a breaker or a fuse. Its safe interrupting capability is the interrupting rating, often written AIC, in amps or kiloamps. The amount of fault current the system can deliver at a point is the available fault current, also from the short-circuit study. Equipment that is not a breaker or fuse has a short-circuit current rating, the SCCR, which is the related withstand number on a piece of gear.

OCPD

Overcurrent protective device, the breaker or fuse that opens on overcurrent

Overload

Current above the circuit rating in the normal path, a slow rise over time

Short circuit / fault

A sudden high current through an unintended low-impedance path

Interrupting rating / AIC

Maximum fault current a device can safely open without failing

Available fault current

The fault current the system can deliver at a given point, from the study

LSIG

Long-time, short-time, instantaneous, ground-fault adjustable trip functions

Current-limiting

A device that clears a high fault fast enough to cut peak let-through energy

MCCB

Molded-case circuit breaker, the general-purpose thermal-magnetic standard

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What is selective coordination?
• What is available fault current?
• What is a tap conductor?
• What is cathodic protection?
• What is a motor overload relay?