Lightning protection system design
IEC 62305 does not leave the design order to preference. The risk assessment decides whether protection is needed and to what class, and every dimension after that follows from the class. This guide walks the sequence in order, with the clause behind each step.
Lightning protection design has a fixed order, and the first step is not on the roof. It is the risk assessment, because that is what sets every dimension that follows.
IEC 62305 is a risk-based standard. It does not begin by telling you what to install. It begins by asking whether protection is needed at all, and if it is, how much. The answer to that question fixes the class of the lightning protection system, and the class then determines the rolling sphere radius you work to, the mesh size, the spacing between down conductors and the coefficient in the separation distance calculation.
That is why designs go wrong in a particular way. Someone lays out air terminations that look right, or copies a scheme from a previous building, and the risk assessment is produced afterwards to accompany the drawings. Done in that order the assessment is not deciding anything, and if it turns out to call for a higher class than the drawings assumed, every dimension on them is wrong at once. This guide follows the sequence in the order the standard intends, naming the clause behind each step.
One note on what this page does and does not contain. IEC 62305 is a copyrighted document sold by the IEC and its national adoption bodies. We name the clauses, tables and equations you need by number and explain what each one governs, so you can go to the right place, but we do not reproduce their contents. You need the standard itself to design to it.
Seven steps, and why the order is not optional
Each step consumes the output of the one before it. Work them out of order and you will redo the ones you skipped ahead of.
Risk assessment
IEC 62305-2. Establishes whether protection is required, and which measures bring the risk within the tolerable value. Output: a need, and a protection level.
Class of LPS
IEC 62305-3 clause 4.1. The protection level maps to the class. Output: the single parameter that every later dimension is read against.
Air termination
Clause 5.2, with the class parameters from Table 2. Output: the positions that leave no part of the protected structure exposed.
Down conductors
Clause 5.3, spacing from Table 5. Output: the routes that carry the current to earth, as short and as direct as the building allows.
Earth termination
Clause 5.4. Output: the arrangement that disperses the current into the soil, bonded into one earthing system for the whole structure.
Internal LPS
Clauses 6.2 and 6.3. Output: bonding, a verified separation distance, and coordinated surge protection so the electronics survive too.
Step 7 is the deliverable itself: the drawings, calculations and the assessment that justifies them, assembled so the system can be built, inspected under clause 7 and defended if it is questioned. It is treated as a step rather than as paperwork because a design that cannot be traced back to its assessment is difficult to defend later.
The risk assessment decides everything after it
IEC 62305-2 computes the risk of injury to people R and the frequency of damage F from the structure, its surroundings and the services connected to it, and compares each against the tolerable value that applies. Where the computed risk is already below that value, no lightning protection system is required, and a documented assessment saying so is a legitimate design outcome rather than a failure to produce one.
Where it is above, the assessment does more than say yes. It identifies which protection measures bring the risk down, and it is the protection level implied by those measures that the physical design has to deliver. This is the step that makes the rest of the design deterministic instead of a matter of judgement.
One caution that belongs here rather than further down. The tolerable risk is not a fixed constant. Clause 7.3 NOTE 1 of IEC 62305-2:2024 gives RT = 1×10-5 per year as a representative value and allows another value once the case has been investigated in detail, and printed p.12 lets national or local regulation fix RT, the tolerable frequency of damage FT, and the Annex A, B, C and E calculation rules and parameter values. So the value your design is judged against is a property of the jurisdiction and the project, and the assessment should state which value it used.
For the method itself see the IEC 62305-2 risk method, and for the arithmetic worked through on an example building see how an IEC 62305 risk is calculated.
Fixing the class of LPS
IEC 62305-3 opens at clause 4.1 with the class of LPS, before any design clause, and that placement is the point. The class is an input to everything in clause 5 and clause 6, not something that emerges from them. Once it is fixed, the parameters you work to are read from the tables rather than chosen: the protection angle, the mesh size and the rolling sphere radius from Table 2, the preferred spacing between down conductors from Table 5, and the coefficient ki used in the separation distance from Table 11.
A higher class is more demanding on every one of those, which is why treating the class as a safety margin to be inflated is a mistake. It is an output of the assessment, and raising it without the assessment calling for it multiplies cost across the whole design for no defensible reason. Lowering it without the assessment supporting it is worse. For what the levels mean and how they map to the class, see lightning protection levels.
Air-termination design
Clause 5.2 governs the air-termination system. The positioning rules in clause 5.2.2 permit three methods, the rolling sphere, the mesh and the protection angle, with the parameters for each set by the class in Table 2. The design requirement is a volume, not a component count: no part of the structure being protected may fall outside the protected volume that the chosen method defines.
The methods can be combined on one building, and on anything other than a simple rectangle they usually are. A flat roof suits the mesh, a tall mast or a lift overrun suits the protection angle, and the rolling sphere is the general case that resolves the awkward geometry between them. Two things are easy to miss. Roof plant, which arrives late in a project and is often outside the protected volume that was drawn without it. And flashes to the side of tall structures, which clause 5.2.3 addresses specifically and which a plan-view-only design does not consider at all.
Clause 5.2.5 covers natural components of the air-termination system, and clause 5.2.4 the construction requirements. Where a metal roof or metal pipework qualifies as a natural air termination, the standard sets minimum thicknesses that depend on the material and on whether puncture is acceptable, so this is a check to make against the clause rather than an assumption to carry.
The mechanics of the three methods are covered in full in air-termination methods, and you can estimate the protected radius of a single vertical mast with the free rolling sphere calculator. A note on devices marketed as early streamer emission air terminals: IEC 62305 specifies conventional air terminations, and the ESE method belongs to a separate national standard, as covered in ESE air terminals and IEC 62305.
Down-conductor layout
Clause 5.3 governs the down-conductor system, and the design intent is to divide the current among several paths and get it to earth quickly. More parallel paths means less current in each, which reduces both the magnetic field inside the structure and the separation distance you will have to achieve at step 6. Table 5 gives the preferred distance between down conductors for each class of LPS, so the spacing is read from the class rather than judged.
Routing matters as much as spacing. Down conductors should take the most direct route available, distributed around the perimeter, avoiding loops where the geometry permits, because a longer path and a tighter loop both raise the potential difference that the separation distance has to hold off. Clause 5.3.2 covers positioning for an isolated LPS and clause 5.3.3 for an attached one, which are different problems and worth reading separately.
Two practical points. Clause 5.3.5 covers natural down conductors, and interconnected steel framework or reinforcement often qualifies, which is usually cheaper and far less visually intrusive than dedicated conductors. That check is worth making early, while the structural design can still accommodate durable connections; retrofitting continuity into a frame that is already cast is a different exercise. Clause 5.3.6 covers test joints and test points, which exist so the installation can be tested at each inspection: a design that omits them is a design that cannot be verified later.
The earth-termination system
Clause 5.4 governs the earth termination. Its job is to pass the current into the soil without a large rise in potential, and the arrangement that achieves that depends on the soil and the structure rather than on a single preferred answer. Clause 5.4.2 covers the earthing arrangement in general conditions, clause 5.4.3 the installation of electrodes, and clause 5.4.4 natural earth electrodes, of which foundation reinforcement is the most common and frequently the best.
The design rule that catches people out is not about the electrodes at all. The lightning earth is bonded into a single earthing system shared with the power and telecommunications earths. Separate earths at different potentials during a strike are a hazard rather than a precaution, and the desire to keep the lightning earth apart from the electrical one is a durable and dangerous instinct. See earthing and the earth termination for the arrangements, why soil resistivity shapes them, and why earth resistance is measured at every inspection.
Internal LPS: bonding, separation distance, surge protection
The external system stops the structure burning down. It does nothing on its own for the electronics inside, and on most modern buildings that is where the loss actually falls. Clause 6 is the internal lightning protection system, and it has two halves.
Lightning equipotential bonding, clause 6.2, ties metal installations, external conductive parts, internal systems and incoming lines to the same potential so a strike cannot spark between them inside the structure. Clause 6.2.5 deals with lines connected to the structure, which is the route by which a strike some distance away reaches the equipment indoors.
Separation distance, clause 6.3, is the clearance needed between the LPS and conductive parts so that no dangerous sparking occurs. Clause 6.3.1 gives the general approach in equation (5), summing the contributions of successive lengths along the current path. Clause 6.3.2 gives the simplified approach in equation (6), where s is ki divided by km, multiplied by kc and by the length l measured from the first equipotential bonding point or earth termination back along the down conductor and the air termination. The three coefficients come from tables: ki from Table 11, set by the class of LPS; km from Table 12, set by the insulating material in the gap; and kc from Table 13, set by how the current divides between down conductors. Annex B is the informative treatment of evaluating s.
Read that formula and the design lever is visible: because kc depends on the current division, adding down conductors at step 4 reduces the separation distance you need at step 6. This is the clearest example of why the sequence is worth respecting, and of why a late change to the down-conductor layout is never local. Where the required distance genuinely cannot be achieved, the answer is to bond the part in question rather than to hope the gap holds.
Surge protection completes the internal system, coordinated across the zone boundaries so each stage passes a smaller remnant to the next. That is the subject of SPD Types 1, 2 and 3, and the zone concept it depends on is covered in lightning protection zones. Both belong to this step of the design rather than to a separate electrical exercise carried out later by someone else.
The sequence on one building
A four-storey office with a flat roof, rooftop plant and a small server room. Illustrative, to show how each step feeds the next; the values on a real project come from its own assessment.
Assess, then classify
The Part 2 assessment models the structure, its surroundings and its incoming power and telecom services, and returns R above the tolerable value, driven by injury to people and by failure of the internal systems in the server room. It shows which measures bring R into line, and the protection level those imply fixes the class of LPS under clause 4.1. Nothing has been drawn yet, and that is the point.
Air termination on a flat roof
The flat roof suits a mesh at the size Table 2 gives for the class, with the rolling sphere applied at the parapet and around the rooftop plant, where a mesh alone would leave the units exposed. The plant is the item most often missed, because it is specified after the lightning design is drawn. Building height is checked against clause 5.2.3 for flashes to the side.
Down conductors and earth
Down conductors are distributed around the perimeter at the preferred spacing in Table 5 for the class. The frame is checked against clause 5.3.5 to see whether it qualifies as a natural down conductor, which would remove most of the dedicated conductors. Test points go in per clause 5.3.6. The earth termination uses foundation reinforcement as a natural electrode under clause 5.4.4, bonded into the single site earthing system.
The server room drives the internal design
Separation distance is checked with equation (6) on the path passing nearest the server room, with ki from the class, km for the material in the gap and kc from the number of down conductors. If it fails, the two available moves are adding a down conductor to lower kc or bonding the part. Bonding follows clause 6.2, and coordinated SPDs sit at the zone boundaries protecting the server room.
Notice what happened at step 6: the fix for a separation distance failure was a change at step 4. That back-pressure is normal and is the reason the steps are worked in order and then checked as a whole, rather than signed off one at a time. Notice too that the server room, which the risk assessment flagged at step 1 through the failure of internal systems, is what drives the internal design at step 6. The assessment was not a formality that preceded the design; it was the thing that told you where the design would be tested.
Step 7: the design deliverable
Clause 5.6 covers materials and dimensions, with Table 7 giving the material, configuration and minimum cross-section for down conductors, earth lead-in conductors, rods and air-termination conductors, and Table 8 doing the same for earth electrodes. These are compatibility and corrosion decisions as much as current-carrying ones, and they are read from the clause rather than carried over from a previous job on a different site.
The deliverable itself should let someone who was not there build the system, inspect it and challenge it. In practice that means the risk assessment justifying the class, stating the tolerable value it was judged against; drawings showing air-termination positions with the method used and the resulting protected volume, down-conductor routes and spacing, test point locations, and the earth-termination arrangement; the separation distance calculation with its coefficients and lengths shown, not just its result; the material and dimension schedule; the bonding and surge protection scheme; and the basis for the inspection regime that clause 7 sets out.
That last item is the one most often left out, and clause 7 is explicit that maintenance and inspection are part of the system rather than an afterthought: clause 7.2 covers maintenance, clause 7.3 the objective of inspections and clause 7.4 the need for them. A design that does not say what will be inspected, or that omits the test points to inspect it with, has handed a problem to whoever owns the building. See IEC 62305-3 inspection and testing, and audit-ready risk reports for what a reviewer looks for in the assessment half of the pack.
Five failures worth designing against
Each of these is common, and each traces back to a step taken out of order or skipped.
Geometry before assessment
Drawing the system first and producing the assessment to accompany it. If the assessment then calls for a different class, every dimension is wrong at once, and the reviewer can see which came first.
Copying a previous building
Reusing a scheme from a similar structure. The inputs that drove that class were its location, surroundings, services and contents, and none of those transfer, even when the buildings look identical.
Roof plant added later
Air-termination geometry drawn before the mechanical layout is fixed, then plant installed outside the protected volume with nobody re-checking. Worth an explicit hold point rather than good intentions.
Separation distance left to site
Not calculating s at design stage and discovering on site that a service runs too close to a down conductor. At that point the cheap remedies from step 4 are gone and bonding is the only move left.
A separate lightning earth
Keeping the lightning earth apart from the electrical earth, which feels prudent and creates the potential difference the bonding exists to prevent. One bonded earthing system, always.
External system only
Treating the design as finished once the structure is protected, leaving bonding and surge protection to a later electrical package. On a building with a server room or control systems, that is where most of the actual loss sits.
Begin where the standard begins
If you are picking up a structure and do not yet know whether it needs protection, step 1 is the whole job for now. Run the IEC 62305-2 assessment, record the tolerable value you judged it against, and let the result tell you whether there is a design to do and what class it has to be. Everything on this page after step 2 is downstream of that answer.
Lumex runs that assessment in the browser on IEC 62305-2:2024, computing the risk of injury to people R and the frequency of damage F from the nine risk components, comparing each against the tolerable value in force for the project rather than a hardcoded constant, and producing a report with every coefficient traceable to the clause, equation or table behind it. It covers step 1 and the documentation around it. It does not do the physical design of steps 3 to 6, so the drawings remain yours or your LPS contractor's. It is sold per seat per month with a 14 day free trial.
New to the standard? Start with what is IEC 62305, or what a lightning protection system is for the anatomy this guide assumes. Working to a US code instead? See NFPA 780 vs IEC 62305. Choosing a tool for step 1? See lightning risk assessment software compared.
Questions answered
How do you design a lightning protection system?
Where does lightning protection design start?
What determines the class of LPS?
What are the steps of an IEC 62305 lightning protection design?
What is the separation distance and how is it calculated?
Do all buildings need a lightning protection system?
Can existing structural steel be used instead of dedicated down conductors?
What does the lightning protection design deliverable contain?
Lumex computes the IEC 62305-2 method and shows the working. It does not certify a structure. You may not issue or submit a Lumex output until a competent person, qualified where the structure is located, has reviewed the inputs and the result and signed it.
The tolerable risk in IEC 62305-2 is not a fixed constant. Clause 7.3 NOTE 1 gives RT = 1×10-5 per year as a representative value of tolerable risk and adds that another value may be set once the case has been investigated in detail. Printed p.12 then lets national or local regulations fix RT, the tolerable frequency of damage FT, and the Annex A, B, C and E calculation rules and parameter values. Every Lumex assessment states the jurisdiction it was computed under and the values that applied.
Get started today