DC Arc-Fault Protection in Solar: What NEC 690.11 Requires
By Seamless Home Team, Solar fulfillment operations · September 2, 2026
Quick answer
DC arc-fault protection is the NEC 690.11 requirement that PV solar DC circuits operating at 80 volts or more between any two conductors, where those circuits are on or penetrate a building, must be protected by a listed PV arc-fault circuit interrupter or an equivalent listed component that detects an arc and interrupts the circuit. It exists because a direct-current arc behaves differently from an alternating-current one: AC crosses zero volts twice a cycle and an arc tends to extinguish itself at the crossing, whereas DC does not cross zero, so a DC arc that starts in a PV string can keep burning as long as the sun keeps the array producing. On residential systems the function is almost always built into the equipment rather than added separately, integrated into a string inverter or provided through module-level electronics. It is a different requirement from ground-fault protection, which detects current leaking to ground rather than an arc between conductors, and a system needs both. Systems built entirely on microinverters often fall outside the 80-volt threshold, because each module's DC circuit stays below it.
An alternating-current arc tends to put itself out. Twice per cycle the voltage sustaining it passes through zero, and at that moment the arc has nothing holding it up. This is not a complete protection against arcing in building wiring, but it is a real and helpful property of the physics.
A PV solar array has no zero crossing. While the sun is on the modules, the voltage across a string is continuous, and an arc that establishes itself somewhere in that string has nothing interrupting it. It burns until something stops it, at temperatures far above the ignition point of roofing felt and conductor insulation, on a roof, in a circuit that cannot be switched off at the panelboard because a module in sunlight is a current source.
NEC 690.11 exists because of that asymmetry.
What the requirement actually says
PV system DC circuits operating at 80 volts or more between any two conductors, where those circuits are on or penetrate a building, must be protected by a listed PV arc-fault circuit interrupter, or by an equivalent listed component that detects and interrupts arcing.
Both conditions carry weight:
- The voltage condition is a threshold about when a DC arc becomes reliably self-sustaining and destructive, rather than a bright line in physics.
- The location condition limits the requirement to circuits on or through a structure, which is where an arc's heat has something combustible to find and people nearby.
On a conventional string-inverter rooftop system, both conditions are satisfied comfortably. A series string of residential modules runs well above 80 volts, and it runs across a roof and usually through an attic or a wall. In other words, on the great majority of residential PV solar installations this requirement is simply universal.
How the detection works
An arc is electrically noisy. It produces broadband high-frequency content on the DC circuit that does not occur in normal operation, and that noise pattern is what the protection looks for. In a string inverter the function samples the DC input, recognises the signature, and shuts the input down. Where a system uses module-level electronics, an equivalent listed function may be distributed across the modules instead of concentrated at the inverter.
Two consequences follow for anyone specifying equipment.
The first is that this is almost never a separate box on a residential system. The protection is a property of the equipment you selected, which means the compliance question at plan review is a listing and documentation question: does the submitted equipment provide a listed PV arc-fault function, and can you show it. That is a datasheet and listing reference, not an assertion, which is the same standard that applies to other equipment claims in the plan set.
The second is that mixing equipment can break the claim. Where the arc-fault function depends on a particular combination of optimiser and inverter being listed together, substituting a model mid-project can invalidate it even where both parts are individually certified. That is the same listing trap that catches rapid shutdown compliance, and it is a good reason for equipment substitutions to go through design review rather than through a purchasing decision.
Arc fault and ground fault are not the same requirement
These two get conflated constantly, including by people who should know better, so it is worth being exact. They detect different physical events and a compliant system needs both.
| Ground-fault protection | Arc-fault protection | |
|---|---|---|
| What it detects | Current leaving the circuit and returning via ground | The high-frequency signature of an arc |
| Typical cause | Insulation failure to a module frame, raceway or grounded surface | A loose, corroded or mismatched connection; damaged conductor |
| Current signature | Measurable current flowing to ground | Often no change in circuit current at all |
| Primary hazard addressed | Shock, and fault current through structure | Ignition |
The decisive case is a series arc at a loose connector in the middle of a string. The circuit is intact. No current is reaching ground. The magnitude is unchanged. A ground-fault detector sees a perfectly healthy system, and the only evidence of the fault is a gap getting hot and a noise pattern on the conductors. Nothing but arc detection will find it.
It runs the other way too: a clean insulation breakdown onto a grounded module frame may produce no arcing signature worth detecting, and it is the ground-fault function that catches it.
What microinverters change, and what they do not
Systems built entirely on microinverters often fall outside 690.11 altogether, and the reasoning matters more than the conclusion.
A microinverter converts at the module. The only DC circuit in such a system is the short run between a single module and its own microinverter, and a residential module's open-circuit voltage is typically somewhere around 40 to 50 volts, below the 80-volt threshold. There is no long, high-voltage DC string on the roof to protect, because the architecture removed it. This is the same property that makes module-level conversion a clean answer to rapid shutdown.
But treat that as a design finding, not a product-category exemption:
- A design that puts modules in series before conversion can exceed the threshold.
- A DC-coupled battery introduces DC circuits elsewhere in the system, which changes the analysis and also changes what an inspector examines.
- What governs is the actual voltage between conductors in the DC circuits as designed.
The honest version of the answer is: check the design. On a straightforward one-module-per-microinverter rooftop system the threshold is not reached, and that is a real and useful consequence of the architecture.
When an inverter keeps tripping
Repeated arc-fault trips get treated as a nuisance more often than they should be, and the base rate of real causes is high enough that the assumption is dangerous.
The field causes cluster tightly around connectors:
- A connector that was never fully seated. The most common single cause.
- Mated connectors from different manufacturers. Physically compatible, not designed to mate, and a well-documented source of high-resistance contact over time.
- Moisture ingress and corrosion, particularly where a connector sits in standing water on a roof or in an enclosure that was not sealed properly.
Others include a conductor damaged during installation or by rodents, a module junction box defect, and water in a combiner.
There is a genuine false-trip category, mostly electromagnetic interference or a detection algorithm reacting badly to a specific equipment combination, and inverter manufacturers have issued firmware addressing particular cases. But "false trip" should be the conclusion of an inspection of the DC side, not the assumption that lets you skip one. An arc fault that is real and repeatedly reset is a fire on a roof waiting for the right conditions.
That distinction has a commercial edge as well as a safety one. A system that passed inspection and starts tripping months later is a service call, and whether it is a warranty visit or a chargeable one depends on whether the cause was workmanship. Connector specification and workmanship standards are therefore worth naming in a subcontract rather than left to be argued about after the second call-out, and they sit alongside the wider question of who stands behind the installation.
The edition qualification
As with the rest of the DC-side safety requirements, both the numbering and the wording of this section have been revised across code editions, and the edition that governs your project is the one your jurisdiction has adopted rather than the most recently published one. Adopted editions lag publication by years and differ between neighbouring jurisdictions, so which edition applies is a question with a local answer decided by the authority having jurisdiction.
For an addition to an existing array, the assessment is generally against the edition in force for the new permit rather than the one the original system was built under, which is how a small extension can pull arc-fault protection into scope on a system that never had it.
The bottom line
DC arc-fault protection exists because a DC arc does not extinguish itself, sits on a roof, and cannot be switched off from the panelboard. The requirement applies to DC circuits at 80 volts or more on or penetrating a building, which on a string-inverter rooftop system means always. It is satisfied by equipment selection rather than by a separate device, which makes it a listing and documentation question at plan review and an equipment-substitution risk during procurement.
It is not ground-fault protection and does not replace it. And a repeated trip is a report about the DC side of your system, not a fault in the detection.
Before booking a final inspection, walk the DC-side requirements alongside labels and clearance — and note that the arc-fault item changes depending on whether the array uses a string inverter or module-level conversion, since that is what decides whether the threshold is reached at all.
Seamless Home carries equipment listing and arc-fault documentation as part of the plan set through design and permitting, and reviews equipment substitutions against the listing rather than the specification sheet. Coverage is confirmed per service area rather than promised as blanket availability. If arc-fault trips are generating repeat service calls across your installed base, get in touch and we can look at whether the pattern points at a connector specification.
Frequently asked questions
What is DC arc-fault protection in solar?
It is a protective function that monitors a PV solar DC circuit for the electrical signature of an arc and interrupts the circuit when it finds one. An arc fault is not a short circuit and not a ground fault: it is current jumping a gap it should not be crossing, typically at a loose or corroded connection, a damaged conductor, or a compromised connector. The distinguishing feature is that an arc can draw far less current than a conventional fault while producing intense localised heat, which means ordinary overcurrent protection may never operate. NEC 690.11 addresses that by requiring detection of the arc itself rather than of the current magnitude. The protection is usually implemented inside the inverter, which samples the DC input for the characteristic high-frequency noise an arc produces and shuts the input down when the pattern is recognised. On systems using module-level electronics the equivalent function may be distributed across the modules instead.
Why do DC arcs matter more than AC arcs?
Because of the zero crossing, or rather the absence of one. Alternating current reverses direction and passes through zero volts twice per cycle, and an arc struck in an AC circuit tends to extinguish at each crossing because the voltage sustaining it momentarily disappears. That self-extinguishing behaviour is a large part of why arcing in ordinary building wiring, while dangerous, is often self-limiting. Direct current has no such crossing. The voltage across a PV string is continuous while the array is illuminated, so an arc that establishes itself has nothing interrupting it and can burn continuously, sustaining temperatures well above the ignition point of roofing materials and conductor insulation. The array cannot be switched off to stop it either, because a module in sunlight is a current source rather than something a panelboard controls. So the hazard is not merely that a DC arc is possible; it is that a DC arc is persistent, sits on a roof, and cannot be de-energised from inside the building.
What is the 80-volt threshold in NEC 690.11?
The requirement applies to PV system DC circuits operating at 80 volts or more between any two conductors, where those circuits are on or penetrate a building. Two conditions are doing work in that sentence. The voltage condition is a threshold judgement about when an arc becomes self-sustaining and dangerous in practice rather than merely theoretically. The location condition limits the requirement to circuits that are on or pass through a structure, which is where an arc's heat threatens something combustible and people. The practical consequence is that the requirement is universal on the great majority of rooftop residential systems, because a series string of modules is well above 80 volts. It also means the threshold is worth reading carefully on unusual designs, and worth confirming against the code edition your jurisdiction has adopted, because both the numbering and the wording of this section have been revised across editions.
Do microinverters need DC arc-fault protection?
Often they fall outside the requirement, and understanding why is more useful than memorising the answer. A microinverter converts at the module, so the only DC circuit in the system is the short run between one module and its own microinverter. A single residential module's open-circuit voltage is typically in the region of 40 to 50 volts, which is below the 80-volt threshold the requirement turns on. There is no long high-voltage DC string on the roof to protect, because the architecture eliminated it. This is the same architectural property that makes microinverters a straightforward answer to the rapid shutdown requirement, and it is worth being precise about the reasoning rather than treating it as a blanket exemption: what matters is the actual voltage between conductors in the DC circuits as designed, not the label on the equipment. A design combining modules in series before conversion, or a system with a DC-coupled battery, can put DC voltage above the threshold somewhere in the system even where the array itself is module-level. Confirm against the design rather than the product category.
How is arc-fault protection different from ground-fault protection?
They detect different faults and neither substitutes for the other. Ground-fault protection looks for current that has left the intended circuit and is returning through ground, which is the signature of insulation failure to a grounded surface such as a module frame or a raceway. It protects against shock and against fault current flowing through structure. Arc-fault protection looks for the electrical signature of an arc, which is characteristic high-frequency noise, and it does so regardless of whether any current is reaching ground. A series arc at a loose connector in the middle of a string produces no ground-fault current at all: the circuit is intact, the current is unchanged, and the only symptom is a hot gap and a noise pattern. A ground-fault detector will not see it. Conversely a clean insulation breakdown to a grounded frame may produce no arcing signature. Both requirements exist in the code for PV systems and a compliant system satisfies both, usually through separate functions in the same inverter.
Why does my solar inverter keep tripping on arc fault?
Treat a repeated arc-fault trip as a fault report rather than as a nuisance, because the base rate of real causes is high. The most common genuine cause is a connector problem: a DC connector that was not fully seated, a pair of connectors from different manufacturers mated together despite being physically compatible, or a connector that has admitted moisture and corroded. All three produce exactly the intermittent high-resistance contact the detection is designed to catch. Other real causes include a conductor damaged during installation or by rodents, a module junction box defect, and water ingress into a combiner or enclosure. There is a genuine category of false trips, most often from electromagnetic interference or from a detection algorithm reacting to a particular equipment combination, and firmware updates from inverter manufacturers have addressed specific cases. But diagnosing a trip as false should be the conclusion of an inspection of the DC side, not the assumption that avoids one. An arc fault that is real and dismissed as nuisance is a fire risk on a roof, and repeated resets without investigation are the wrong response.
Who is responsible for arc-fault protection on a subcontracted install?
The licensed contractor whose permit the work sits under carries the compliance obligation, whoever selected the equipment and whoever was on the roof. That party is the one the jurisdiction corresponds with and the one a correction notice names. The exposure that matters in practice, though, is not usually the code obligation but the workmanship one, because the field causes of arc faults are almost all installation-quality issues: connector seating, mixed connector brands, moisture management at enclosures, conductor protection. A system that passes inspection and then trips repeatedly three months later is a warranty and service problem, and the question of who returns to site and at whose cost is settled by the subcontract rather than by the code. That makes connector specification and workmanship standards worth naming explicitly among the things a subcontract should cover, alongside the responsibility for post-handover service calls.