Solar Grounding and Bonding: What an Inspector Actually Checks
By Seamless Home Team, Solar fulfillment operations · September 3, 2026
Quick answer
Grounding on a residential PV solar system has two separate jobs, and inspectors check them separately. Equipment grounding connects every exposed metal part — module frames, rails, enclosures, metallic raceway, the inverter and the racking itself — back to the service equipment grounding system, so that a fault has a low-impedance path and clears the overcurrent device. Bonding is what makes those metal parts electrically continuous in the first place, and on a modern array it is almost always achieved through hardware listed for that purpose rather than through a separate wire to every module. The equipment grounding conductor is sized from the overcurrent device protecting the circuit, using the general sizing table in Article 250, and the array is connected to the building's existing grounding electrode system rather than to a separate rod of its own. The most common failures are not calculation errors: they are unlisted or incorrectly installed bonding hardware, a bonding path broken between roof planes or across a rail splice, a conductor terminated on painted metal or under a sheet-metal screw, and a plan set that shows a grounding detail the crew did not actually build.
Grounding is the part of a PV solar installation where the code gives the least ground and the workmanship shows the least. A missing label is visible from the driveway. A rapid shutdown failure announces itself when the device is tested. A bonding path broken at a rail splice under a module looks exactly like a bonding path that works, and the only thing that distinguishes them is whether someone checked.
It is also the area where the most common failures are not electrical mistakes. They are hardware and documentation mistakes: the wrong clamp, the right clamp fitted wrongly, a conductor landed on paint, or a plan set that describes a grounding scheme nobody built. An electrician who understands fault current perfectly can still fail this inspection, because compliance on a modern array rests on a product listing rather than on first principles.
Two jobs, checked separately
The words get used interchangeably on job sites and they are not interchangeable.
Bonding makes separate pieces of metal into one electrically continuous body. On an array that means every module frame connected to every rail, every rail section connected across its splices, the racking connected to the attachments, and each enclosure connected to the conduit or cable entering it.
Grounding connects that continuous body back to the grounding system of the building, so that a fault between an energised conductor and any metal part has a low-impedance path back to the source and opens the overcurrent device protecting the circuit. The purpose is not to send fault current into the earth. It is to make the breaker trip.
Reading them as one requirement is what produces the most expensive version of this failure: a single equipment grounding conductor, correctly sized and correctly terminated, serving an array whose two roof planes were never bonded to each other. The grounding is right. Half the metal on the roof is not connected to it.
The equipment grounding conductor, and the sizing question people get backwards
The conductor running from the array back to the equipment grounding system is sized from the rating of the overcurrent device protecting the circuit, using the general sizing table in Article 250. It is the same basis as any other branch circuit in the building.
The array's short-circuit current and the inverter's output current matter, but indirectly: they determine the overcurrent device, and the overcurrent device determines the grounding conductor. Sizing the grounding conductor directly from the array current is a common error in both directions.
The point that causes the most avoidable argument at inspection is this: an equipment grounding conductor is permitted to be smaller than the circuit conductors it accompanies. A drawing showing a visibly smaller grounding conductor is usually correct. It is worth knowing which question is being asked before defending the drawing.
Two things can still make a table-correct conductor wrong. There are minimum sizes below which you cannot go regardless of the calculation. And there are physical protection requirements, so a conductor that is correct on paper can be non-compliant in a location where it is exposed to damage. Earlier code editions also carried PV-specific sizing language that has since been simplified, which is one of several reasons the adopted code edition is the first thing to establish on any project rather than the last.
Bonding through listed hardware, which is where compliance actually lives
Almost no residential array built in the last decade has a grounding wire looping from module to module. That is not a shortcut. Module mounting systems are evaluated and listed as bonding and grounding systems, and where a system carries that evaluation, the rails and the specified clamps are permitted to provide the bonding path between the module frames and the rail.
This is efficient and it moves the basis of compliance. The listing becomes the governing document, and three consequences follow that account for a large share of grounding corrections.
The specified components are requirements, not preferences. A mid-clamp from a different manufacturer that fits the rail and appears to make the same metal-to-metal contact breaks the listed assembly. So does a washer left in the box, a bonding clip fitted without the tooth engagement the instructions describe, or a substituted fastener.
Module frames are anodised, and anodising is an insulator. Listed bonding devices work by penetrating that coating in a defined way. Clamping two anodised surfaces together firmly is not bonding, however tight it is, which is why torque values in the installation instructions are compliance requirements and not workmanship advice.
The listing has limits written into it — permitted module frame types, and often a maximum number of modules in a single bonded run. Exceeding those is a defect that no amount of care in the installation corrects.
Where bonding paths actually break
| Location | Why it breaks | What it looks like |
|---|---|---|
| Rail splice | Splice hardware not listed for bonding, or a slip joint deliberately allowing thermal movement | Two rail sections mechanically joined, electrically separate |
| Between roof planes | Sub-arrays wired as one circuit but bonded as two islands | One conductor at one plane, nothing crossing to the other |
| Attachment to racking | Coated or gasketed attachment breaking continuity to the rail | Structurally sound, electrically open |
| Dissimilar metals | Copper conductor onto aluminium rail without a connector listed for the pair | Works on day one, corrodes into an open circuit |
| Anodised frame | Clamp bearing on the coating rather than a device penetrating it | Tight, clean, and not a bond |
The rail splice and the second roof plane are the two worth checking on every job, because both produce an installation that passes a visual scan and fails a continuity test.
The grounding electrode question, and a rod you probably do not need
This is the most persistent piece of outdated practice in residential PV solar.
What the code requires is that the array is connected to the grounding electrode system that already exists for the premises. Earlier editions required an auxiliary electrode at the array; that requirement was removed, and later editions describe additional electrodes as permitted rather than required.
What survives is a large population of plan-set templates showing a ground rod at the array, and a smaller population of local amendments that genuinely still require one. So:
- A rod at the array with no connection back to the premises grounding electrode system is not a compliant substitute for that connection, and never was.
- An additional rod properly bonded to the existing system is generally harmless, and adds cost, labour and a second thing to corrode.
- Because local amendments exist, this is a plan-review question for the specific jurisdiction rather than one the national code settles on its own. The authority having jurisdiction is the only source that answers it for your permit.
Terminations, which is where good installations fail
The conductor can be correctly sized and the array correctly bonded, and the system still fails at the two ends of the wire.
- Landing on coated metal. Paint, powder coat and anodising are insulators. A lug bolted to a painted enclosure surface without the coating removed, or without a device designed to penetrate it, is a mechanical connection and not an electrical one.
- Sheet-metal screws. A grounding conductor secured under a sheet-metal screw instead of a listed lug or a screw with the required thread engagement is a defect, and a common one in retrofit work inside existing enclosures.
- Two conductors under one lug. Permitted only where the lug is listed for more than one conductor. Most are not.
- Wrong lug for the metal. A copper conductor terminated directly to aluminium needs a connector listed for the combination. Direct contact is a galvanic cell with a roof over it.
None of these require test equipment to find. All of them are visible in a photograph taken before the inspector arrives, which is the cheapest control available on the whole subject.
Ground-fault protection is a separate requirement
Equipment grounding gives fault current a path. Ground-fault protection is the device that detects a fault in the array's DC circuits and responds to it, and it is required for PV arrays independently of the grounding and bonding requirements above.
On modern residential equipment this function is integral to the inverter, which means it is usually satisfied by the equipment selection rather than by anything the crew installs. It is worth naming separately for two reasons. It is a distinct line item an inspector can ask about, and it is distinct again from the DC arc-fault and rapid shutdown requirements that apply to the same conductors — three different requirements addressing three different hazards on the same wires, frequently discussed as though they were one.
Where a battery is part of the project, the energy storage equipment brings its own bonding and grounding requirements and its own inspection scope, which is not simply the array's requirements repeated.
The documentation failure that fails a compliant installation
A meaningful share of grounding corrections are issued against installations that are physically correct.
The mechanism is straightforward. The plan set carries a grounding detail — a bonding method, a mounting system, a conductor size, sometimes a ground rod inherited from a template. The crew installs a system that is compliant and different: a different mounting system with its own listing, no rod because none is required, a conductor sized from the actual overcurrent device.
The inspector's job is to verify the installation against the approved documents. A mismatch is a finding whether or not the built version is compliant, and resolving it means either a revision or a re-inspection. This is the same failure mode that makes design changes more expensive than they look: the drafting is trivial and the queue is not.
The specific items worth reconciling before submission are the mounting system and its listing, the bonding method, the equipment grounding conductor size and its basis, the grounding electrode connection, and whether any rod shown on the drawing is actually being installed.
Who carries it
The licensed contractor holding the permit carries the code obligation, whoever was on the roof. That is the party a correction notice names and the party that has to stand for the re-inspection. Where a sales organisation sold the project and a separate entity holds the permit, the selling party owns the homeowner relationship and the delay; the contractor of record owns the compliance duty.
What is worth settling in advance is narrower: who pays for corrective labour and hardware, who pays the re-inspection fee, and what happens to the funding milestone while the correction is open. That belongs in the subcontract rather than in a conversation after the notice arrives — and it is worth naming grounding specifically, because the corrective cost can exceed the cost of the original work.
The bottom line
Grounding failures are rarely failures of electrical understanding. They are a clamp that was not the listed clamp, a splice that moved, a lug on paint, a rod that was drawn but not needed, or a drawing that describes a different system from the one on the roof. Every one of those is identifiable before an inspector is booked, and most are identifiable from photographs.
The two structural controls are worth more than any amount of care on the day: establish the adopted code edition and any local amendments at plan review, and make the mounting system's listing — with its specified components and torque values — the document the crew works from rather than the drawing's generic detail.
Seamless Home carries design, permitting and inspection coordination as part of design and permitting, which includes reconciling the grounding detail on the plan set against the mounting system actually being installed before a crew is scheduled. Coverage is confirmed per service area rather than promised as blanket availability. If grounding corrections are recurring across your jurisdictions, get in touch and we can look at where they are originating.
Frequently asked questions
Do solar panels need to be grounded?
The metal parts do, and on almost every residential installation that is the module frames, the mounting rails, the racking attachments, the inverter enclosure, any metallic raceway or junction box, and a disconnect enclosure. These are exposed non-current-carrying metal parts, and the requirement is that they are connected to an equipment grounding conductor so that a fault between an energised conductor and the metal has a low-impedance path back to the source and trips the protecting device rather than leaving the frame energised. The module laminate itself is not grounded — there is nothing to ground — so the requirement lands on the frame and the structure holding it. What has changed over successive code editions is not whether this is required but how it is permitted to be achieved. Running an individual grounding conductor to every module frame is still compliant and is now rare, because module mounting systems evaluated and listed for bonding do the same job through the clamps and the rail. Where that listing is relied on, the listing is the compliance document, and the installation has to match it.
What is the difference between grounding and bonding on a solar array?
Bonding makes separate metal parts into one electrically continuous body. Grounding connects that body to the grounding system of the premises. On a PV array the two are easy to conflate because one piece of hardware often does both, but inspectors look at them as different questions. The bonding question is whether every module frame, every rail section, every attachment and every enclosure is actually connected to its neighbours — which is where rail splices, separate roof planes and dissimilar metals cause trouble. The grounding question is whether that continuous metal body has a properly sized conductor running back to the equipment grounding system at the service, and whether that conductor is terminated correctly at both ends. A system can be perfectly bonded and not grounded, if the conductor back to the service was never pulled or was terminated on a surface that does not conduct. It can also be grounded and not fully bonded, which is the more common defect, because a single conductor at one end of a roof does nothing for a second sub-array whose bonding path to it is broken.
How is the equipment grounding conductor sized for solar?
From the rating of the overcurrent device protecting the circuit, using the general equipment grounding conductor sizing table in Article 250, which is the same basis used for any other branch circuit or feeder. It is not sized from the conductor ampacity, from the inverter output, or from the array short-circuit current directly — those figures matter because they determine the overcurrent device, and the overcurrent device then determines the grounding conductor. Two practical points cause most of the errors. First, an equipment grounding conductor is permitted to be smaller than the circuit conductors, so a plan set showing a conductor that looks undersized next to the current-carrying conductors is often correct, and an inspector reading it as an error is asking the wrong question. Second, there are minimum sizes and there are physical-protection requirements, so a conductor that is correct by the table can still be wrong if it is exposed to damage in the location it occupies. Earlier code editions carried additional PV-specific sizing language that has since been simplified, so the applicable rule depends on the edition your jurisdiction has adopted.
Does a solar array need its own ground rod?
In current code editions, no — and this is one of the most persistent pieces of outdated practice in residential PV solar. What the code requires is that the array is connected to the grounding electrode system that already exists for the premises. Earlier editions required an auxiliary grounding electrode at the array, that requirement was removed, and later editions describe additional electrodes as permitted rather than required. The result is a large stock of plan-set templates, and a number of local amendments, that still show a rod at the array. Three things follow. A driven rod at the array with no connection back to the premises grounding electrode system is not a compliant substitute for that connection and never was. An extra rod that is properly bonded to the existing system is generally harmless but adds cost and a second point to corrode. And because some jurisdictions have adopted local amendments that do still require one, the honest answer is that this is a question to settle at plan review for the specific jurisdiction rather than from the national code alone.
Why did my solar installation fail inspection on grounding?
Usually for one of a small number of reasons, and almost none of them are calculation errors. The bonding hardware was not the hardware the mounting system's listing specifies, or it was the right hardware installed the wrong way — a bonding clip fitted without the required tooth engagement, a washer omitted, a clamp under-torqued, or module frames clamped on an anodised surface where the listing requires the coating to be penetrated by a specified device. The bonding path was broken at a rail splice, at a slip joint intended to allow thermal movement, or between two roof planes wired as one array but bonded as two. The equipment grounding conductor was terminated on painted or coated metal, under a sheet-metal screw rather than a listed lug, or two conductors were placed under a lug listed for one. A dissimilar-metal connection was made directly, copper onto aluminium, without a connector listed for the combination. The ground-fault protection device required for the array was absent or not functional. Or, frequently, the installation is fine and the plan set shows a different grounding detail from the one that was built, which is a documentation failure that fails an inspection just as effectively.
Can the mounting rails be used as the grounding path?
Yes, where the mounting system is listed and identified for that purpose, and only in the way the listing describes. Module mounting systems are evaluated as bonding and grounding systems in their own right, and where a system carries that evaluation the rails and the specified clamps are permitted to provide the bonding path between module frames and the rail. This is why almost no modern residential array has a grounding wire looping between panels. The consequence is that compliance rests on the listing rather than on judgement, and three things follow from that. The specified components have to be the ones used, so substituting a mid-clamp from another manufacturer because it fits breaks the listed assembly even where the metal contact looks identical. The installation instructions carry the torque values, the permitted module frame types and often a maximum number of modules per bonded run, and those are requirements rather than guidance. And the documentation matters at inspection: an inspector who cannot tell which system is installed cannot verify that the bonding method is listed, which is why the mounting system and its listing belong on the plan set rather than being left to be identified from the roof.
Who is responsible for grounding defects on a subcontracted solar install?
The licensed contractor holding the permit carries the code obligation, whoever was on the roof. That is the party the jurisdiction issues a correction notice to, the party that has to produce a compliant installation and stand for the re-inspection, and the party whose licence is exposed if the defect is not corrected. Where a sales organisation sold the project and a separate installing entity holds the permit, the selling party owns the homeowner relationship and the schedule consequence but not the compliance duty. The exposure worth settling in the subcontract before it arises is narrower and more practical: which party pays for the corrective labour and hardware, which pays a re-inspection fee, and what happens to the funding milestone while the correction is open. Grounding corrections are particularly worth naming, because unlike a missing label they can require partial disassembly of the array, which makes them one of the few inspection failures where the corrective cost is materially more than the original work.