DC Fuse and Breaker Sizing for Small Solar Systems

A fuse does not protect your equipment. It protects the wire. Equipment mostly protects itself, or fails in ways a fuse cannot prevent. The conductor is the thing that will happily absorb hundreds of amps, heat until its insulation ignites, and set fire to whatever it is stapled to.

Once you internalise that, the sizing rule becomes obvious: the overcurrent device must open before the conductor it feeds is damaged. It is sized to the wire, and it must be small enough to protect that wire while being large enough not to nuisance-trip on the load.

Where devices belong

A small off-grid system has four places that need protection, and one that is frequently missed.

Between battery and everything else. This is the non-negotiable one. The battery is the only component in the system capable of delivering thousands of amps into a fault. Every conductor leaving the battery positive terminal needs a fuse, and it needs to be as close to the terminal as physically practical. ABYC E-11 specifies within 7 inches (18 cm) for marine installations, and that is a sound target on land too. The unprotected stub between the battery post and the fuse is the one piece of wire in your system that nothing can save.

Between charge controller and battery. Protects the controller’s output conductors.

Between PV array and charge controller. Needed when three or more strings are in parallel, since the other strings can back-feed a fault in one string. With one or two strings there is no back-feed path capable of exceeding the wire rating, so a fuse is often omitted — though a disconnect is still valuable for servicing.

Between battery and inverter. Usually the highest-current circuit in the entire system, and the one people most often leave unfused because the cable is short and “obviously fine.”

The one that gets missed: the DC load distribution feed. If a busbar feeds six branch circuits through individual small fuses, the conductor feeding that busbar still needs its own protection sized to its ampacity, not to the sum of the branches.

Sizing PV source circuits: NEC 690.8

PV modules are current-limited, which makes them unusual. A dead short across a panel produces only slightly more than its short-circuit current, so the fusing logic is not “what happens in a fault” but “how much can this thing produce on a very good day.”

NEC 690.8 applies 125 percent twice, for two different reasons:

690.8(A) — maximum circuit current. Take the module’s rated short-circuit current, multiply by 1.25. This covers irradiance above standard test conditions: edge-of-cloud effects and snow reflection can briefly push a panel above its nameplate Isc.

I_max = I_sc × 1.25 × (number of parallel strings)

690.8(B) — conductor and overcurrent device sizing. Take that result and multiply by 1.25 again, this time because PV output is treated as a continuous load.

I_conductor = I_max × 1.25

Combined, that is 1.25 × 1.25 = 1.5625, which is why you will hear people talk about “the 156 percent rule.”

Worked example

Two parallel strings of a module rated Isc = 11.2 A:

690.8(A):  11.2 A × 1.25 × 2 strings = 28 A
690.8(B):  28 A × 1.25              = 35 A

The conductors must have an ampacity of at least 35 A after all derating, and the overcurrent device is selected at the next standard size that protects them — 35 A here, since it is a standard rating.

Note the order of operations. Derating for temperature and bundling reduces your conductor’s ampacity; the 156 percent calculation increases the current it must carry. Both push you toward bigger wire, and both apply.

Sizing the continuous-load side: the 80 percent rule

For circuits that are not PV source circuits, NEC 210.20(A) and 215.3 give the general rule:

OCPD rating ≥ non-continuous load + (1.25 × continuous load)

A load is “continuous” if it is expected to run for three hours or more. The 1.25 multiplier is mathematically the same thing as saying a standard breaker should not be loaded past 80 percent of its rating continuously — unless it is specifically listed for 100 percent continuous duty, which residential-scale devices generally are not.

For a charge controller output, manufacturer practice follows the same shape. Renogy, for example, recommends fusing at 1.25× the controller’s rated charge current, so a 40 A controller gets a 50 A fuse. That is vendor guidance rather than a code requirement, but it agrees with the continuous-load logic and is a sensible default.

Interrupting rating: the specification nobody reads

Here is the part that catches lithium builds.

Every fuse has two ratings. The one on the label is the current rating — the current at which it opens. The one in the datasheet is the interrupting rating, or AIC — the maximum fault current it can actually break safely.

Exceed the AIC and the fuse does not politely fail. The arc struck across the opening element is not extinguished, the fuse body can rupture, and in the worst case the contacts weld together — leaving a permanently closed circuit feeding a dead short, which is precisely the scenario the fuse existed to prevent.

Fuse type	Typical interrupting rating	Appropriate for
ANL	~5,000–6,000 A	Flooded and AGM lead-acid banks
MRBF (terminal fuse)	~10,000 A	Smaller lead-acid and AGM banks
Class T	~20,000 A (to 125 VDC)	LiFePO4 and other lithium chemistries

The reason lithium needs more is internal resistance. A LiFePO4 cell has dramatically lower internal resistance than a comparable lead-acid cell, so the prospective short-circuit current it can deliver is far higher — high enough to exceed an ANL fuse’s interrupting rating on a bank of quite ordinary size. The battery that is safer in almost every other respect is the one that demands a better fuse.

Practical rule: if the bank is lithium, specify Class T unless you have the manufacturer’s short-circuit current figure in hand and can demonstrate a lower-AIC device is adequate. The price difference is trivial relative to the battery.

Check your battery’s datasheet for a prospective short-circuit current or maximum fault current figure. Reputable LiFePO4 manufacturers publish it. If yours does not, that is information about the manufacturer.

Fuses versus breakers

Both interrupt. They differ in what else they do.

Fuses are cheaper, more compact, generally available in higher interrupting ratings, and more predictable in their time-current behaviour. They are single-use, and you need a spare on hand.

Breakers double as a disconnect switch, which is genuinely useful for servicing, and they reset. But DC-rated breakers are notably more expensive than AC ones, and this is a place where substitution is dangerous.

An AC breaker is not a DC breaker. AC current crosses zero volts 100 or 120 times a second, and that zero crossing extinguishes the arc when contacts open. DC never crosses zero, so the arc must be forced out by the breaker’s own arc chutes and contact geometry. An AC-rated breaker used on DC may fail to interrupt at all, sustaining an arc across open contacts. Use devices explicitly rated for DC service at or above your system voltage — and note that a device rated 250 VAC may carry a DC rating of only 60 V or less.

Many builds sensibly use both: a Class T fuse at the battery for fault protection, and a DC-rated breaker or a heavy disconnect switch downstream for isolation during maintenance.

A worked system

A 24 V system with a 2000 W inverter and a LiFePO4 bank.

Inverter DC input current at full load, assuming 88 percent efficiency:

2000 W / 0.88 / 24 V = 94.7 A

Design the conductor for continuous duty at 1.25×, so about 118 A. That indicates 2 AWG at minimum by NEC 75 °C ampacity (115 A), and 1 AWG or 1/0 is the more comfortable choice once voltage drop and inverter surge are considered.

Fuse selection. Somewhere between the conductor’s ampacity and the load current — commonly 125 A or 150 A here, chosen so it protects the chosen conductor while riding through inverter surge. Because the bank is lithium, it should be a Class T.

Placement. Within a few inches of the battery positive terminal, in an enclosed holder, with the stub between post and fuse as short as it can physically be made.

Things that are not overcurrent protection

A battery management system is not a fuse. A BMS protects cells and will disconnect under fault conditions, but it is an electronic device with its own failure modes, and it is protecting the battery, not your cables. Fuse the system regardless of what the BMS claims to do.

A charge controller’s internal protection is not a fuse. It protects the controller.

“The wire is short” is not overcurrent protection. A twelve-inch unfused cable across a battery terminal will vaporise just as enthusiastically as a twelve-foot one.

Before you build

Overcurrent protection is the area where the gap between “works” and “safe” is invisible until the day it is not. Every number on this page is a planning figure derived from published rules and manufacturer guidance; none of it is a substitute for the installation instructions that came with your equipment, your local code, or an electrician’s review. If the system is permanent, connected to a structure, or carrying more than a few hundred watts, have someone qualified look at it before you energise it.