Industrial Fuse Selection Guide: Classes, Ratings and How to Specify the Right Fuse

Fuse class describes a body style and a performance envelope. It does not replace the three ratings that determine whether the fuse is legal and safe in the circuit.

• Voltage rating. A fuse may be applied at or below its voltage rating, never above. A 250 V fuse does not belong in a 480 V circuit, regardless of ampere rating.
• Ampere rating. Set by the conductor, the load, and the applicable NEC article — not by what was in the holder when you opened the panel.
• Interrupting rating. The fault current the fuse can safely clear. NEC 110.9 requires overcurrent devices intended to interrupt fault current to have an interrupting rating at least equal to the fault current available at their line terminals.

Interrupting rating is where legacy equipment causes problems. Class H fuses carry a 10,000 A interrupting rating. Available fault current at a modern service can be several times that. A Class H fuse left in place through a service upgrade is a code violation that nothing on the panel will announce.

All figures below are the class definitions under UL 248. Individual manufacturers publish higher interrupting ratings on some lines — 300 kA is common on Class J, L and R product — so confirm against the datasheet for the part you are actually buying.

Classes H and K share dimensions and are interchangeable in the same holders. That is precisely the problem with them: nothing physically stops a 10 kA Class H fuse going into a holder that was carrying a 200 kA Class K fuse. Classes R, J, T, CC, CD, CF and G all carry rejection features that prevent a lower-rated fuse being installed in their place.

Class L is the exception at the top end. It is bolt-in construction and its dimensions vary with ampere rating, so it has no defined rejection element in the way Class R or J does — the physical mounting is what constrains the fit.

A current-limiting fuse clears a fault in less than half a cycle, so the prospective peak current is never reached. The NEC definition is a device that, when interrupting currents in its current-limiting range, reduces the current in the faulted circuit to a magnitude substantially less than would flow if the device were replaced by a solid conductor of comparable impedance.

Manufacturer-published figures show the size of the effect. Mersen gives an example where 35,000 A rms available fault current — which would otherwise produce an 80,500 A peak — is limited by a 200 A Class J fuse to 12,000 A peak let-through. Eaton's cable-whip demonstration shows a 200 A fuse limiting peak current to 10.2 kA against 48.1 kA unlimited.

Here is the fact that most fuse guides get wrong. Class K fuses are current-limiting in performance — K1 to a high degree, K5 to a moderate degree. But they are not permitted to carry the "Current Limiting" marking, because NEC 240.60(B) requires that fuseholders for current-limiting fuses reject fuses that are not current-limiting. A Class K fuse fits an unmodified Class H holder, so allowing the marking would defeat the rule the marking exists to support.

Class H is the class that genuinely is not current-limiting. If you read a page that says "Class K is not current-limiting," it has confused the marking rule with the physics.

NEC 240.60(C) requires cartridge fuses and fuseholders to be plainly marked with the ampere rating, the voltage rating, the interrupting rating where it is other than 10,000 A, the words "current limiting" where applicable, and the manufacturer's name or trademark. Manufacturers additionally mark the UL/CSA class.

That class marking is how you know the fuse is branch-circuit rated at all. A fuse with no UL class designation is a supplemental fuse under UL 248-14, and a supplemental fuse cannot be used as branch-circuit protection. It is intended to protect a component inside a listed assembly, downstream of something that already provides branch-circuit protection.

RK1 and RK5 are both Class R. Both are 200 kA. Both fit the same rejection holder. The difference is how much energy they let through while clearing a fault.

RK1 fuses are more current-limiting than RK5. They clear faster under high fault current, producing lower peak let-through current and lower let-through energy. RK5 fuses let more energy through before the fault clears.

That matters in three places:

• Equipment short-circuit current ratings. A series-rated or component-protected assembly may have been evaluated with a specific fuse class. Substituting a less current-limiting fuse can invalidate the rating the assembly was labelled with.
• Arc-flash studies. Incident energy calculations depend on the clearing characteristics of the device. Change the device, and the study no longer describes the installation the labels are attached to. We cover this in more detail in our guide to the 2026 NEC arc flash labeling requirements.
• Selective coordination. Where the design required upstream and downstream devices to coordinate, the time-current curves were chosen deliberately. A substitution changes the curve.

If the equipment label or the drawing specifies RK1, order RK1. If nobody knows why the original was RK1, that is a question for the engineer of record before it is a purchasing decision.

Class R dimensions are set by the standard, so they are the same across manufacturers for a given voltage and ampere band. The table below is from Littelfuse FLSR (600 V) and FLNR (250 V) datasheets and applies to Class H, K and R bodies alike.

Note what that table shows: a 30 A Class R fuse is 5.00 inches long at 600 V and 2.00 inches at 250 V. The voltage rating changes the physical size, so 250 V and 600 V Class R fuses cannot be interchanged even before the rejection groove is considered. If you are measuring a fuse to identify it, measure length and diameter together and note the voltage marking.

Time-delay fuses tolerate a temporary overload before opening. Fast-acting fuses do not. The choice follows the load, not preference.

• Motors, transformers and other inductive loads draw high inrush at start. A fast-acting fuse sized for running current will open on every start. Time-delay fuses ride through inrush and still protect against sustained overload and short circuit.
• Resistive and non-inductive loads — heaters, lighting, ovens — have no meaningful inrush, so fast-acting protection is appropriate.
• Semiconductor and drive applications need fuses specifically rated for that duty. See the section below.

These get used interchangeably and they should not be. Dual-element describes construction — a fuse with two current-responsive elements in series, one for overload and one for short circuit. Time-delay describes the curve — the fuse tolerates temporary overload before opening.

Every dual-element fuse is time-delay. Not every time-delay fuse is dual-element: Eaton's Limitron KLU, for example, is a single-element fuse with a time-delay characteristic. When both apply, manufacturers write "dual-element, time-delay," and that is the phrase to use on a purchase order.

NEC Table 430.52 sets the maximum sizing for branch-circuit short-circuit and ground-fault protection on AC polyphase squirrel-cage motors: 300% of full-load current for non-time-delay fuses and 175% for dual-element time-delay fuses, against 250% for an inverse time breaker. Where the calculated value does not land on a standard fuse size, the next higher standard rating is permitted. Where the motor will not start on the calculated size, 430.52(C)(1) Exception No. 2 permits up to 400% for non-time-delay fuses at 600 A or less, and up to 225% for time-delay fuses.

Those are the code maximums. For optimal protection rather than maximum permitted protection, manufacturers publish much tighter numbers. Eaton's Bussmann motor circuit selection tables give 125% of motor full-load amps for RK5 Fusetron, 130% for RK1 Low-Peak, and 150% for Class J and Class CF Low-Peak. Mersen's general guidance is that sizing time-delay fuses between 125% and 150% of motor full-load current provides advantages.

Read those figures carefully, because there is a persistent claim online that Class J can be sized closer to full-load current than RK5 for high-inrush loads. Eaton's own tables show the reverse — RK5 at 125%, Class J at 150%. The real comparison worth making is not between classes but between constructions: a dual-element time-delay fuse sizes at 125–150% of motor FLA, where a single-element fuse may need 300% simply to hold the starting current. That is where the sizing advantage lives.

Note what all those percentages do: they size the device for short-circuit and ground-fault protection. Motor overload protection is a separate requirement and a separate device — normally an overload relay in the starter. Our guide to Sprecher + Schuh motor control components covers overload relay selection and starter coordination.

Class J, T, CC, CD and CF exist largely because Class R is physically large.

• Class J delivers Class R performance in a smaller body, with its own rejection dimensions. Mersen states its AJT Class J fuses require only half the mounting space of 600 V Class RK fuses. Common on feeders and branch circuits in newer equipment where panel space was designed around it.
• Class T is the most compact of the group and reaches 1200 A. Used where a high interrupting rating is required and enclosure depth is limited — meter stacks, compact switchboards, control assemblies.
• Class CC covers up to 30 A in the smallest body of any 200 kA class. Standard for control circuits, transformer primaries and short-circuit protection on IEC-style motor starters. Class CD extends the same concept from 35 A to 60 A.
• Class L starts where the others stop, at 601 A, and runs to 6000 A. Bolt-in construction for service entrance equipment and large feeders.

These are not interchangeable with each other. Each rejection profile exists to keep the wrong fuse out.

Class CF is a finger-safe, rectangular, current-limiting dual-element fuse rated up to 400 A at 600 V ac, with a 300 kA interrupting rating at 100 A and below. It delivers Class J electrical performance in a package Eaton describes as requiring up to 70% less space than comparable Class CC, J or RK solutions, and its holders and switches carry an ampere-rating rejection feature that helps prevent overfusing.

Two details worth knowing, because almost nothing published online gets them right:

• Class CF is defined by UL 248-17, which is an Outline of Investigation rather than a full ANSI/UL standard — Edition 5, published September 2018. That is a real distinction. It is a recognised evaluation basis, not a consensus standard in the way UL 248-8 or 248-12 are.
• "CUBEFuse" is Eaton's trademark, not the class name. Class CF is an open class — Littelfuse builds to it too and publishes a Class CF to Class J cross-reference. If a specification says CUBEFuse, it is naming a product line; if it says Class CF, it is naming the class.

Power semiconductors — diodes, SCRs, IGBTs, GTOs — cannot withstand heavy short-circuit current. Their silicon has very low transient thermal capacity, and damage occurs faster than a general-purpose branch-circuit fuse can clear. Protecting them requires ultra current-limiting fuses selected on low I²t and low peak let-through, covered by UL 248-13 (Semiconductor Fuses, Edition 3, 2022) and internationally by IEC 60269-4 with its aR (short-circuit only) and gR (full-range) classes.

There is one point here that is regularly missed and that will fail an inspection. Traditional high-speed fuses are UL and CSA Recognized to 248-13 but are not considered branch-circuit protection under the NEC. They protect the semiconductor; something else has to protect the branch circuit.

The exception is dual-listed product. Eaton's Bussmann DFJ drive fuse meets UL listing requirements for Class J fuses as well, so it provides both semiconductor and branch-circuit protection in one device. Mersen's HSJ is the equivalent approach. If you are specifying fuses for a VFD or a rectifier, that dual listing is the thing to check on the datasheet — not the catalogue description.

Legacy equipment and discontinued lines are where fuse orders stall. The label is unreadable, the manufacturer has been absorbed by another, or the catalogue number returns nothing.

Before you order, gather:

• A photograph of the fuse body, including every marking on the label, and of the holder or block it sits in
• Overall length and body diameter, measured — the dimension table above will narrow the class and voltage immediately
• The equipment nameplate, and the panel schedule if there is one
• The circuit voltage and the load being protected
• Any equipment short-circuit current rating shown on the assembly label
• The quantity, and whether this is a one-off replacement or a stocking item

That set is usually enough to identify the original and confirm a current equivalent. Manufacturer consolidation is common in this category — Mersen, whose fuses ATEK distributes, carries the former Ferraz Shawmut line, so a discontinued catalogue number frequently has a direct successor rather than no answer at all.

1. Leaving Class H in place after a service upgrade. 10 kA interrupting rating against a fault current that has grown. Nothing on the panel warns you.
2. Swapping RK1 for RK5 on availability. It fits, it runs, and it invalidates the arc-flash study and possibly the equipment SCCR.
3. Treating a supplemental fuse as branch-circuit protection. No UL class marking means it is not branch-circuit rated, whatever the ampere rating says.
4. Using a high-speed semiconductor fuse as the only protection on a drive circuit. Unless it is dual-listed to a branch-circuit class, it is not branch-circuit protection.
5. Sizing a single-element fuse for motor running current. It will open on the first start. Dual-element time-delay is the answer, not a bigger fast-acting fuse.