What decides whether an SPD actually protects anything

1. The number that matters is not on the datasheet

An SPD's voltage protection level is measured at its own terminals. The equipment is not at the SPD's terminals — it is at the end of some wiring, behind some connecting leads. What has to be compared against the equipment's rated impulse withstand voltage is the effective protection level: the device's level, plus everything the installation adds on top of it.

That distinction is the whole of this article. A device with an excellent protection level, installed with long leads at the wrong end of a long circuit, can deliver a worse result to the equipment than a mediocre device installed properly. The datasheet is necessary and nowhere near sufficient.

2. The half-metre rule, and what breaking it costs

The total length of the SPD's connecting leads is limited to half a metre. Total is the word that gets missed: it is the sum of the path from the live conductor to the SPD and from the SPD back to the PE conductor, not just whichever leg is longer or whichever one you can see in the photograph. Worth knowing where the rule actually binds, because the two books word it differently: IEC 62305-4 carries it in Annex C, which is informative and says consideration should be given. IEC 60364-5-53 — which Part 4's normative text says the installation shall comply with — says consideration shall be given, and that where the total exceeds 0,5 m one of the remedies below shall be chosen.

The reason is inductance. Surge current changes fast, and a fast current change through even a short conductor develops a real voltage across it. That voltage is in series with the SPD, so the equipment sees the sum. As a scale: at ten kiloamps through the device, one metre of lead develops roughly a kilovolt on the standard's own approximation — comparable with the entire protection level of a decent low-voltage device, thrown away by routing.

Two honest qualifications, because this is where a tool can quietly flatter you. The first is the standard's own: ΔU = 0,1 × I_SPD kV/m is a value the leads “can be approximated by”, and the note under it says ΔU depends on the current waveshapes, on which source of damage is being considered, and on the conductor type — and that for sources S1 and S3 the real voltage drop can be greater than the formula gives. The formula is a floor for those cases, not an answer. The second is that published manufacturer figures for the same case run higher too — one major vendor's catalogue gives about 1.2 kV where the formula gives 1.0 kV for that one-metre, ten-kiloamp case. We compute the standard's figure and say on the report that the real one may be up to about a fifth higher, because the alternative is presenting a number we know to be optimistic without saying so.

That is the rule as our sample design is checked against it, and the reason to look at it is the shape rather than the verdict: three legs are being added up, and one of them is the branch down to the back-up overcurrent device. That third leg is the one that gets left out of the sum on site, because it does not look like part of the SPD's connection — and leaving it out is how an installation that measures compliant on two legs is not.

That gap is exactly why the half metre binds as a shall in the installation standard rather than as advice: kept short, the disagreement between the formula and reality does not matter. Let the leads grow and it does. Where the half metre genuinely cannot be achieved, there are three recognised ways out — choose a device with a lower protection level, add a second coordinated SPD close to the equipment, or adopt the connection arrangement of the standard's Figure C.2 — captioned there as an example of installing an SPD to reduce the effect of lead length, and known in the trade as the V-shaped connection — so the lead voltage is not in series with the protection. Accepting the long leads is not on the list.

3. Ten metres, and why distance is not just attenuation

Intuition says a longer run between the SPD and the equipment can only help — more impedance, more damping. Intuition is wrong here, and the error is in the direction that hurts.

A surge travelling down a circuit reflects at the far end, where the impedance changes. The reflected wave adds to the incident one, and at an open or high-impedance termination the voltage at the equipment can approach twice the value the SPD is holding at its own terminals. The standard handles this with a reflection coefficient k_r whose value lies between one and two, and directs that where it is unknown the worst case of two be used.

So the rule splits at ten metres:

• Equipment within 10 m of the entrance SPD. The effective protection level should not exceed 80 % of the equipment's rated impulse withstand voltage. The 20 % is the margin absorbing what the calculation does not model.
• Equipment beyond 10 m, protected only by a one-port SPD at the origin. The available withstand is effectively halved, which is the reflection worst case made explicit. Many installations that look compliant on the datasheet fail here. Note also the condition the clause attaches: taking this route is meant to be accompanied by other measures, such as shielded wiring throughout the protected circuit — the halved allowance is not a licence to leave the installation as it is.
• Equipment beyond 10 m, done properly. Either a second coordinated SPD close to the equipment, or a two-port SPD at the origin. Both restore the full withstand — but the origin-mounted route carries a condition that is easy to skip: it requires shielded wiring throughout the protected circuit. Claim the two-port allowance on unshielded wiring and the calculation no longer applies.

The practical consequence for a real building: a single entrance SPD does not protect a distribution board two floors up, however good it is. That is a coordination problem, and it is solved with a second device, not a better first one.

Which is what the second device in each of these chains is doing. Every incoming service gets its own, because a telecom line entering the same building is a separate path to the same equipment — and each SPD is placed at a zone boundary rather than at a convenient distance, which is section 4:

4. Zone boundaries decide how many devices you need

The lightning protection zone concept is the part of Part 4 most often reduced to a diagram in a tender document and then ignored. It is doing real work: it is what turns "put in a surge protector" into a number of devices in specific places.

The rule is that an SPD belongs at the entrance of a line into each inner zone — and where only one protected zone exists, at least one at its entrance. The zones themselves are defined by what a location is exposed to: whether a direct strike can reach it, whether the full lightning electromagnetic field is present, and how much surge current can arrive. A boundary is where one of those things changes, and every crossing is a place where the conducted surge has to be dealt with.

There are two recognised ways to reduce the count, and both are design decisions rather than omissions: interconnecting zones of the same order so they behave as one, and extending a zone into the next using shielded cable bonded at both ends, so the run itself carries the protection. Both are legitimate; neither is "we only fitted one".

This is also why an SPD schedule is a design output rather than a shopping list. The count, the locations, the class at each location and the coordination between them all follow from the zone layout — which is why our tool produces a manufacturer-neutral schedule you can issue to a supplier, rather than a single recommended part number.

5. How to check any of this, including ours

Three questions to put to any surge protection design — one you have been handed, one you have produced, or one a tool has produced for you:

1. Does the comparison use the effective level or the datasheet level? If nothing in the calculation accounts for the connecting leads, it is comparing the wrong number against the equipment withstand.
2. Is the distance to the protected equipment an input at all? A design that never asked cannot have applied the 10 m rule, which means it silently assumed the favourable case.
3. Does the device count follow from the zone boundaries? If the answer is one SPD because there is one incoming service, the zone concept was skipped.

Our own working is public, including where it disagrees with other published sources — the lead-voltage figure above is one such case, and there is an unresolved question about how one manufacturer's handbook reads a table in the series that we set out rather than quietly pick a side on. It is all on the validation page, with the sources named. If you run these checks against our output and something looks wrong, that is the message we most want to receive.