Lightning rarely strikes the array, and surges reach it anyway

The direct hit is the rare and uninteresting case

A rooftop array is not notably more likely to be struck than the building it sits on. Lightning is drawn by height and by geometry rather than by the presence of electrical equipment, so a house’s aerial matters as much as its array, and a metal frame on a house that already has an aerial, a flue and a satellite dish does not change the picture much. Where a structure does warrant protection, that protection is a building matter rather than a solar one, and it isn’t bolted on afterwards.

And in a genuine direct strike nothing survives. The energy is far beyond what any protective component can absorb, and the conversation afterwards is with an insurer rather than an electrician. The damage that actually occurs is subtler and much more common. Strikes several hundred metres away, or cloud-to-cloud discharges that never reach the ground, put substantial voltages into wiring nowhere near the event. This is the mechanism worth understanding, because it is the one that can be mitigated cheaply.

Induced surges travel in the loops you built

A lightning channel is an enormous current that appears and vanishes in microseconds, and around it is a magnetic field doing the same. Any conductive loop within reach of that field has a voltage induced in it, and the size of that voltage depends on the area the loop encloses and on how fast the field is changing. Both are the whole story.

A string of panels is a loop by construction. Current goes out along the positive conductor, passes through the modules and comes back along the negative, and whatever area is enclosed between those two paths is the loop the field sees. Run the two conductors side by side and the area is small. Route one down the left of the array and the other down the right, and it is large.

This is the cheapest mitigation in the whole subject, because it costs nothing except attention at the moment of installation. Keeping positive and negative conductors of each string together for their whole run, and avoiding large decorative loops of spare cable, reduces induced voltage without a single component being purchased. It is also invisible afterwards, which is why it is worth asking about beforehand.

What a surge protective device actually does

A surge protective device sits between the live conductors and earth and does nothing at all under normal conditions. Its components are effectively insulators at working voltage. When the voltage across them rises past a threshold they conduct heavily, diverting the surge to earth and clamping the voltage seen by whatever is downstream to something survivable.

Two things follow from that description. The first is that the device needs somewhere for the diverted energy to go, so a low-impedance earth connection and correct bonding are not optional extras; a protector fitted to a poor earth doesn’t do much beyond looking reassuring. The second is that the components degrade as they work, which is why many devices carry an indicator window that changes state once the device has spent itself.

That indicator is worth glancing at occasionally, because a protector reaching end of life gives no other sign. It carries on looking exactly like a working device and protects nothing. Replacement is a task for an electrician, since the enclosure is live from more than one direction on a solar installation.

There are three routes in, not one

Surges reach an inverter along the direct-current cables from the roof, along the alternating-current cables from the consumer unit, and along whatever data cabling connects it to a monitor or a router. Protecting one route and ignoring the others is a common and slightly pointless arrangement, and the data path is the one most often forgotten.

The alternating-current side is the more likely entry point in most locations, because the distribution network is long, largely overhead in rural areas, and shared with everything else on the feeder. A surge originating three streets away arrives through the meter like any other current. Protection here also covers the rest of the house rather than only the solar equipment.

The direct-current side matters most where cable runs are long, where the array is spread across separate roof planes, or where a ground-mounted array returns to the building by a buried cable. Communication links are cheap to protect and are a well-documented route for damage that then propagates into equipment nobody associated with the storm.

Whether it is worth fitting depends on where you live

This is a genuine cost-benefit question with a location-dependent answer, and the honest position is that surge protection is not always worth the money. In a dense urban area with a short underground supply, a compact array and a low local lightning frequency, the expected loss it prevents is small. Fitting it everywhere as a reflex is a habit, not an analysis.

The case strengthens sharply with exposure. Long overhead supply lines, elevated or rural sites, regions with frequent summer storms, long direct-current runs and ground-mounted arrays all push the calculation the other way. So does the simple observation that an inverter is the most expensive single component in the system and the one a surge is most likely to find.

Many wiring codes now require a documented risk assessment rather than leaving the decision to preference, and the parameters that feed it — local strike density, supply arrangement, the consequence of losing the equipment — are exactly the sort of thing a qualified designer is expected to evaluate. Ask what was decided and why. An installer who can’t answer that hasn’t done the assessment.