A surge protective device catches the brief, high-voltage spikes — from lightning, or from switching on the network — that destroy electronics in a fraction of a second. Whether you have to fit one is a short test in Regulation 443.4.1, and the answer is usually yes.

The harder question is whether the one you fitted is doing anything. An SPD's protection level Up is measured at its own terminals. Its connecting conductors sit in series with the surge path, and their inductance adds a volt drop on top. BS 7671 puts a number on it, and the number is large.

The short version: Reg 534.4.8(c) states that a 1 m length of rectilinear conductor carrying a discharge current of 10 kA (8/20) adds a voltage drop of about 1000 V. That drop adds directly to Up. A perfectly compliant SPD on a metre of lead can therefore deliver more voltage to your equipment than the equipment is rated to withstand — which is why the standard constrains the lead length, and caps the installed assembly, rather than just the device.

This article is written against BS 7671:2018+A4:2026, and describes the rules as they read now rather than as any one amendment changed them — with the chronology in its own section, because a good deal of what is written about SPDs online is describing the 2018 edition.

1. Do you need one?

Regulation 443.4.1 is short enough to quote in full. Protection against transient overvoltages shall be provided where the consequence caused by the overvoltage could result in:

  • (a) serious injury to, or loss of, human life;
  • (b) — this item was deleted by the Corrigendum of May 2023;
  • (c) significant financial or data loss.

And then the sentence that catches almost everything else: "For all other cases, protection against transient overvoltages shall be provided unless the owner of the installation declares it is not required due to any loss or damage being tolerable and they accept the risk of damage to equipment and any consequential loss."

Decision tree for Regulation 443.4.1. An SPD is required where the consequence of a transient overvoltage could be serious injury or loss of human life, or significant financial or data loss. A third trigger, failure of a safety service, was deleted in May 2023 and is shown struck through. In all other cases an SPD is still required by default unless the owner of the installation declares the loss tolerable.Decision tree for Regulation 443.4.1. An SPD is required where the consequence of a transient overvoltage could be serious injury or loss of human life, or significant financial or data loss. A third trigger, failure of a safety service, was deleted in May 2023 and is shown struck through. In all other cases an SPD is still required by default unless the owner of the installation declares the loss tolerable.
Two live triggers, not three — and the only way out of "all other cases" is a declaration by the owner of the installation.

Two things are worth pinning down, because both are commonly got wrong.

Item (b) used to read "failure of a safety service, as defined in Part 2", and it was removed by the Corrigendum of May 2023. The lettering was left with a hole in it, which is probably why it is still quoted as live in a great deal of published guidance — including, until this update, the earlier version of this article. Chapter 56 adds no SPD requirement to compensate. In practice a fire alarm panel will still very often bring an installation into scope, but through route (a), or on the default in the last paragraph, not because it is a safety service.

The declaration route belongs to the owner of the installation, not to a householder. The 2018 edition did have a special dispensation for single dwelling units; that is long gone, and it was never the same thing as this. Any installation's owner can decline, and the decision should be recorded — it is the one place in this regulation where somebody takes on a liability by writing something down.

Note also what the standard does not do: it never defines "significant financial or data loss". There is no threshold, no table, no monetary test. That judgement is the designer's, and it is worth having a defensible reason for it in the file.

2. What changed, and when

Surge protection has moved more than most parts of BS 7671, which is why so much online material is out of date. The short chronology:

  • BS 7671:2018 — protection required on a four-item consequence list, with the AQ external-influence criteria removed, a risk assessment for everything else, and an exception for single dwelling units in certain situations.
  • Amendment 2, 2022 — the calculated risk level method was deleted outright: Regulation 443.5, Table 443.1, Figures 44.2 and 44.3 and Annex A443 all went, and the single-dwelling exception went with them. The consequence list was cut down and the owner-declaration default introduced.
  • Corrigendum, May 2023 — item (b), failure of a safety service, was deleted from 443.4.1.
  • Amendment 4, 2026 — two deletions in this area: Annex B443, and Regulation 712.443 for solar PV. Regulation 443.4.1 and Section 534 themselves are unchanged from the A2 text plus that Corrigendum.
The risk assessment is gone, not relocated. Search the current standard for the letters CRL and they appear exactly once in over six hundred pages — in the note recording that the calculated risk level method was removed. There is no environmental factor, no ground flash density and no threshold value anywhere in the book. If a guide is showing you that arithmetic, it is describing the 2018 edition.

3. Overvoltage categories, and the number that matters

Everything downstream turns on one value: the equipment's rated impulse voltage Uw, from Table 443.2. It is the highest peak impulse the insulation will take without breaking down, and it falls as you move further into the installation.

BS 7671 Table 443.2 for a 230/400 volt system. Category IV equipment such as an energy meter withstands 6 kV, Category III distribution boards 4 kV, Category II appliances 2.5 kV and Category I sensitive electronics 1.5 kV. The further into the installation, the less the equipment can survive.BS 7671 Table 443.2 for a 230/400 volt system. Category IV equipment such as an energy meter withstands 6 kV, Category III distribution boards 4 kV, Category II appliances 2.5 kV and Category I sensitive electronics 1.5 kV. The further into the installation, the less the equipment can survive.
Table 443.2. Up is selected against Category II and must never exceed the withstand of the equipment being protected.
Table 443.2 — required rated impulse voltage of equipment (Uw), kV
Nominal voltage (V) Line to neutral, up to (V) Cat IV Cat III Cat II Cat I
120/208 150 4 2.5 1.5 0.8
230/400 300 6 4 2.5 1.5
400/690 600 8 6 4 2.5
1000 1000 12 8 6 4

Category IV is equipment at or near the origin, such as the meter. Category III is the fixed installation from the main distribution board down. Category II is current-using equipment. Category I is equipment that is only suitable for use where SPDs limit the overvoltage for it — Regulation 443.6.2 says so explicitly, which means a Category I device connected without surge protection is outside the terms it was built to.

Regulation 534.4.4.2 then sets the rule: Up "shall be selected in accordance with impulse withstand voltage Category II of Table 443.2 and in no case exceed the required rated impulse voltage of the equipment". On a 230/400 V system that is 2.5 kV.

4. Type 1, Type 2, Type 3

The type is a test class, not a quality grade. It tells you which waveform the device was proved against, and therefore where in the installation it belongs.

TypeTest waveformDeals withWhere BS 7671 puts it
Type 110/350 µs (Class I)Partial direct-strike lightning currentAt the origin, where the structure has an external lightning protection system — Reg 534.4.1.3
Type 28/20 µs (Class II)Switching and indirect lightning surgesAt the origin where there is no external LPS (Reg 534.4.1.4), and at sub-distribution boards
Type 3Combination wave (Class III)Residual surges near equipmentClose to sensitive equipment, coordinated with an upstream device — Reg 534.4.1.5

Note what decides between Type 1 and Type 2 at the origin: it is the presence of an external lightning protection system, not the size or importance of the building. Combined devices classified Type 1+2 do both jobs in one unit.

A Type 1 alone is not a complete answer. Regulation 534.4.1.1's note is unusually direct about it: a lightning protection system which only employs equipotential bonding SPDs "provides no effective protection against failure of sensitive electrical and electronic systems". Type 1 devices exist to stop dangerous sparking that could cause fire or shock. Keeping equipment alive is a separate job for the Type 2 and Type 3 devices downstream.

5. The number on the SPD is not the voltage your equipment sees

Here is the part that catches people. Up is a measurement taken across the SPD's own terminals. It says nothing about the wires you use to connect it.

Those wires are in series with the surge path. Under an impulse, an inductance of roughly a microhenry per metre and a rate of rise measured in kiloamps per microsecond produce a volt drop that is anything but negligible, and it adds to Up. The standard states the result directly in Regulation 534.4.8(c):

a 1 m length of rectilinear conductor carrying a discharge current of 10 kA (8/20)
adds a voltage drop of about 1000 V

Stacked bar chart. A 600 volt surge protective device is shown with connecting leads of zero, 0.25, 0.5, 1 and 2 metres. Each metre of lead adds about 1000 volts, so the voltage reaching the equipment rises from 0.6 kV to 2.6 kV, crossing the 2.5 kV Category II withstand threshold at about 2 metres.Stacked bar chart. A 600 volt surge protective device is shown with connecting leads of zero, 0.25, 0.5, 1 and 2 metres. Each metre of lead adds about 1000 volts, so the voltage reaching the equipment rises from 0.6 kV to 2.6 kV, crossing the 2.5 kV Category II withstand threshold at about 2 metres.
The device is compliant at every point on this chart. The installation is not — the leads alone carry it past the withstand of the equipment it is supposed to protect.

This is not an aside in a guidance note. It is the reason Regulation 534.4.4.2 carries a hard ceiling: "In installations operating at 230/400 V, the voltage protection level of the installed SPD assembly shall not exceed 2.5 kV, as the SPD's connecting leads have additional inductive voltage drop across them (see Regulation 534.4.8). It may, therefore, be necessary to select an SPD with a lower voltage protection level."

Read that last sentence as an instruction. The leads consume part of your budget, so the device has to be better than the equipment's withstand alone would suggest. A note to the same regulation recommends going further and keeping Up within 80 % of the Category II value — 2.0 kV at 230/400 V — while listing four cases where that margin is unnecessary, including where the equipment is connected directly to the SPD terminals.

6. The connecting leads (Reg 534.4.8)

Regulation 534.4.8 requires all conductors and interconnections to be "as short and as straight as possible" with "any unnecessary cable loops avoided", and then gives the limit:

The total wiring length of conductors between the connection points of the SPD assembly should preferably not exceed 0.5 m and shall in no case exceed 1.0 m.

Two subtleties, both of which decide whether a real installation complies.

0.5 m is a recommendation; 1.0 m is absolute. "Should preferably" and "shall in no case" are doing different work, and it is worth knowing which side of the line you are on when you are arguing about a board that cannot physically be wired shorter.

The length is measured A to B, through the overcurrent device. The standard defines it as the sum of the path length from the line conductor to the PE between connection points A and B — in the figure's own terms, a + b + c. That includes the conductor from the busbar to the backup device and from the device to the SPD, not just the SPD tails. Where there is no overcurrent device in the branch, length b is zero. Conductors run to an intermediate earthing terminal beside the SPD are excluded, which is the whole point of that arrangement.

An SPD tapped off a line conductor through an overcurrent protective device and down to earth, with the three lead segments labelled a, b and c between connection points A and B, whose total should preferably not exceed 0.5 metres. On the right, the same line running more than 10 metres to a load, where oscillation can double the voltage at the equipment terminals.An SPD tapped off a line conductor through an overcurrent protective device and down to earth, with the three lead segments labelled a, b and c between connection points A and B, whose total should preferably not exceed 0.5 metres. On the right, the same line running more than 10 metres to a load, where oscillation can double the voltage at the equipment terminals.
Two different lengths, two different effects. The leads ADD a fixed volt drop at the SPD; distance MULTIPLIES whatever arrives at the far end.

If a + b + c exceeds 0.5 m, the regulation does not simply disapprove — it requires you to choose at least one of three remedies:

  • (c) select an SPD with a lower voltage protection level Up;
  • (d) install a second coordinated SPD close to the equipment to be protected, adapting Up to the equipment's rated impulse voltage;
  • (e) use the Figure 534.9 installation method — an intermediate earthing terminal placed as near as possible to the SPD, whose connections back to the main earthing terminal are then excluded from the length.

Binding the leads matters as much as shortening them

A parallel-connected SPD has current flowing into it and out of it in opposite directions, so its two leads carry opposing magnetic fields. Bring them close together and the fields — and therefore much of the inductance — cancel. The BEAMA Guide to Surge Protection Devices measured it, applying a 6 kV/3 kA combination wave to a 230 V parallel SPD with Up = 600 V, protecting equipment with a 700 V safe level:

Measured voltage at the equipment — BEAMA Guide to SPDs, Figure 33
Connecting leadsAt the equipmentAgainst a 700 V limit
25 cm, tightly bound 630 V passes
25 cm, unbound 810 V fails
2 m, tightly bound 1,200 V fails
2 m, unbound 2,300 V fails

Three of the four installations fail, with the same compliant device in each. Note this is a Class III combination wave with a much lower rate of rise than the 10 kA 8/20 impulse behind the 1000 V per metre figure, so the volts per metre it implies are lower — the two should not be added together or averaged. What it shows is the shape of the problem: length and routing both count, and cable ties are free.

7. Protective distance — a different length entirely

There is a second distance in Section 534, and it is constantly confused with the first. Regulation 534.4.4.2 continues: "If the distance between the SPD and equipment to be protected (protective distance) is greater than 10 m, oscillations could lead to a voltage at the equipment terminals of up to twice the SPD's voltage protection level."

The leads ADD; the distance MULTIPLIES. Connecting leads add a fixed volt drop at the SPD itself. Protective distance can double whatever arrives at the far end. They compound: a 1 kV device on 0.5 m of lead delivers 1.5 kV, and 30 m away that can become 3 kV.

The remedy is the one you would expect — additional coordinated SPDs closer to the equipment, or a device with a lower Up.

8. Selecting Uc (Table 534.2)

Uc is the maximum continuous operating voltage the SPD can sit at indefinitely. Set it too low and the device conducts during normal service or a routine fault and destroys itself. Table 534.2 gives the minimum, and there is a trap in the header: "U is the line-to-line voltage of the low voltage system" — not the line-to-neutral voltage.

Table 534.2 — minimum required Uc by supply system
SPD connected betweenTN systemTT systemIT system
Line conductor and neutral conductor1.1U/√31.1U/√31.1U/√3
Line conductor and PE conductor1.1U/√31.1U/√31.1 U
Neutral conductor and PE conductorU/√3U/√31.1U/√3
Line conductors1.1 U1.1 U1.1 U

On a 400 V system, 1.1U/√3 is 254 V — which is why 275 V devices are the norm. The neutral-to-PE cells deliberately drop the 10 % tolerance: a note to the table explains that those values relate to worst-case fault conditions. And on an IT system the line-to-PE requirement is the full 1.1U, 440 V on a 400 V supply — a 275 V device there is badly under-selected.

9. Discharge current, and the neutral-to-PE trap

A Type 2 device is rated by its nominal discharge current In (8/20); a Type 1 by its impulse discharge current Iimp (10/350). The minima depend on the connection type — CT1, the 4+0 arrangement with each live conductor to PE, or CT2, the 3+1 arrangement with the lives to neutral and the neutral to PE.

Tables 534.3 and 534.4 — minimum discharge current at or near the origin (kA)
Mode Type 2 — In (8/20) Type 1 — Iimp (10/350)
1ph CT11ph CT23ph CT13ph CT2 1ph CT11ph CT23ph CT13ph CT2
L – N 5512.512.5
L – PE 5512.512.5
N – PE 51052012.52512.550

Table 534.4 applies where the building is protected against direct lightning strike and no BS EN 62305-2 risk analysis has been done; its note records that the values refer to lightning protection levels III and IV. Where a risk analysis has been done, Iimp is determined from BS EN 62305 instead.

On a 3+1 arrangement the neutral-to-PE device is the one that matters. The lightning current returning toward the protective conductor passes through that single device rather than sharing between the lines, so its rating jumps to 20 kA for a three-phase Type 2 and 50 kA for a Type 1 — four times the line figure. CT2 is the normal arrangement on a TT system, so this is not a corner case.

10. Conductor sizes (Reg 534.4.10)

Regulation 534.4.10 gives two pairs of minimum cross-sectional areas for SPDs at or near the origin, and most write-ups quote only the first:

  • Earth side, between the SPD and the main earthing terminal or protective conductor — 16 mm² copper for a Type 1, 6 mm² for a Type 2.
  • Live side, connecting the SPD and its overcurrent device to the live conductors, per Reg 433.3.1(b) — 6 mm² for a Type 1, 2.5 mm² for a Type 2.

These are sized for the surge current, not the load current. A parallel SPD draws essentially nothing in normal service, so the same device and the same leads serve a 100 A board and a 1000 A board alike.

11. SPDs, RCDs and connection type

Fault protection has to survive the SPD failing. Regulation 534.4.6 requires that in as many words, and Table 534.5 sets out which connection type each system permits:

  • TN — CT1 and CT2 both apply; the overcurrent device on the supply side of the SPD generally satisfies fault protection.
  • TT — with CT1 the table's cell is not a plain yes: the SPD may only be installed downstream of the RCD. CT2, the 3+1 arrangement, can be installed upstream of the main RCD instead.
  • IT — CT1 and CT2 with a neutral; CT2 is not applicable without one.

Where an SPD does sit on the load side of an RCD, Regulation 534.4.7 requires an RCD having an immunity to surge currents of at least 3 kA 8/20, and notes that Type S devices to BS EN 61008-1 and BS EN 61009-1 satisfy it. Installing a Type 1 SPD downstream of an RCD is not recommended. Above 3 kA the RCD may simply trip and interrupt the supply.

12. Backup protection, end of life and testing

Regulation 534.4.5.1 requires the SPD installation to be protected against short-circuit currents, internally or externally to the manufacturer's instructions. The external device should be "the highest permissible rating to provide a high surge current capability for the complete assembly" without exceeding what the manufacturer allows — a bigger fuse is better here, up to the stated limit, because that device carries the surge too.

Regulation 534.4.5.2 makes an important point about the arrangement: where the overcurrent device operates because the SPD has failed, supply continuity is unaffected, but the installation is no longer protected against further overvoltages — and, while it is in circuit, that device's own volt drop increases the effective protection level, because it is in series with the SPD.

An SPD normally reaches end of life through a slow thermal overload, which its internal disconnector handles. The external device is there for the fast short-circuit case. Since a parallel SPD's disconnection is invisible from outside, the status indicator is what tells you: Appendix 6's guidance for recipients asks that it is checked and expert advice sought if it shows the device is not operational, and the periodic inspection schedule carries an item confirming the SPD is functional. The model forms now include SPD type and functionality fields.

Disconnect SPDs before an insulation resistance test. They will treat the 500 V test voltage as a transient overvoltage and conduct, which both defeats the test and risks the device. Reconnect them afterwards — a shunt-connected SPD left disconnected is invisible.

Finally, Regulation 514.16.1 requires an information notice at or near the relevant distribution board indicating that SPDs are present. It need not be applied at domestic premises where the information is recorded on the certificate or condition report and issued to the person ordering the work.

13. Where BS EN 62305 comes in

BS 7671 deals with transient overvoltages arriving via the supply. It says plainly that it does not cover protection against direct or nearby lightning strokes to the structure — that is BS EN 62305's territory, and Section 534 points at it directly.

The organising idea there is the lightning protection zone. Equipment should sit in a zone whose electromagnetic environment matches what it can withstand, and every metal part and service crossing a zone boundary must be bonded. A live conductor obviously cannot be bonded to earth, so an SPD does the bonding instead — which is why the type of device follows the boundary it sits on.

Nested lightning protection zones from BS EN 62305. LPZ 0A is subject to direct strokes, LPZ 0B to the full electromagnetic field but no direct stroke, LPZ 1 is the service entrance and main board, LPZ 2 the sub-board. A Type 1 SPD sits at the LPZ 0/1 boundary, a Type 2 at LPZ 1/2 and a Type 3 at LPZ 2/3.Nested lightning protection zones from BS EN 62305. LPZ 0A is subject to direct strokes, LPZ 0B to the full electromagnetic field but no direct stroke, LPZ 1 is the service entrance and main board, LPZ 2 the sub-board. A Type 1 SPD sits at the LPZ 0/1 boundary, a Type 2 at LPZ 1/2 and a Type 3 at LPZ 2/3.
BS EN 62305 guidance, not a BS 7671 requirement. A live conductor cannot simply be bonded where it crosses a zone boundary, so an SPD does the bonding instead.

The risk assessment in BS EN 62305-2 is a different animal from Regulation 443.4.1. It computes four risks — R1 loss of human life, R2 loss of service to the public, R3 loss of cultural heritage, R4 loss of economic value — and compares each against a tolerable value: 10−5 per year for loss of life, 10−4 for the other two. There is deliberately no tolerable risk for R4; economic loss is a commercial judgement, not a safety one. You add protection measures until the calculated risk falls below the tolerable value.

It also offers a design margin that is stricter than BS 7671's, and useful to know even when you are not doing a 62305 assessment. Manufacturers' guidance derives a target of Up ≤ (Uw × 0.8) ÷ 2 — the 0.8 budgeting 20 % for the additive inductive volt drop on the connecting leads, and the divide-by-two covering the worst-case doubling over the protective distance. That gives 1600 V at a 4 kV board and 600 V for 1.5 kV terminal equipment.

Withstand is not the same as susceptibility. Uw is the voltage at which insulation breaks down — the destruction threshold. Equipment stops working long before that. Where the maker states nothing, BS EN 62305 takes susceptibility as twice the peak operating voltage: for a 230 V supply, 230 × √2 × 1.1 gives 358 V peak, so roughly 715 V — against a 1.5 kV rated withstand. Meeting BS 7671 keeps equipment alive. It does not keep it running.

Everything in this section is guidance rather than a BS 7671 requirement, and the manufacturers' guides it draws on predate the current editions of both standards. Verify against BS EN 62305 itself before relying on it.

14. Four things that are widely believed and are not in the standard

  • "EV charge points need an SPD." Section 722 contains no SPD, surge or overvoltage requirement at all. Neither do the sections for agricultural premises, caravan parks or medical locations. The only special location with its own SPD rules is Section 712 for solar PV — and Amendment 4 deleted 712.443 from that. EV installations are covered by the general Reg 443.4.1 test like everything else.
  • "You do a risk assessment to decide." Not under BS 7671. That method was deleted in 2022 and there is nothing to calculate — only the consequence test and the owner's declaration. A BS EN 62305-2 risk assessment is a real and separate thing, but it is not what Regulation 443 asks for.
  • "The withstand table is Table 44.3." That is 17th Edition numbering, still repeated in a lot of otherwise good material. It is Table 443.2 now. Tables 44.1 and 44.2 exist but are about power-frequency stress voltage from HV earth faults and have nothing to do with SPDs.
  • "Buy the biggest kA rating you can." Appendix 16 is blunt about this: the most important aspect in selecting an SPD is its protection level Up during the surge, "and not the energy withstand (for example, Iimp) which it can handle". A high energy withstand may only mean a longer operating life; it is a low Up that actually protects the equipment.

Glossary

Up
Voltage protection level. Characterises the SPD's ability to limit the voltage across its own terminals. Does not include the connecting leads.
Uw
Rated impulse withstand voltage of the equipment — the highest peak impulse its insulation takes without breaking down. From Table 443.2.
Uc
Maximum continuous operating voltage of the SPD; the voltage it can sit at indefinitely without conducting. Minimum set by Table 534.2.
In
Nominal discharge current, 8/20 µs. The Type 2 rating (Table 534.3).
Iimp
Impulse discharge current, 10/350 µs. The Type 1 rating (Table 534.4).
CT1 / CT2
Connection types. CT1 is 4+0, each live conductor to PE. CT2 is 3+1, lives to neutral and neutral to PE.
Protective distance
The run from the SPD to the equipment. Beyond about 10 m, oscillation can double the voltage at the equipment terminals.
LPZ
Lightning protection zone (BS EN 62305). A region with a defined electromagnetic environment; SPDs sit at the boundaries.

Frequently Asked Questions

Do I need an SPD under BS 7671?

In most cases yes. Regulation 443.4.1 requires protection against transient overvoltages wherever the consequence could be serious injury to, or loss of, human life, or significant financial or data loss. For every other case an SPD shall still be provided unless the owner of the installation declares it is not required because the loss or damage would be tolerable, and accepts the risk. So the default is that one is fitted, and not fitting one is a decision somebody has to own and record.

Is failure of a safety service still a reason to fit an SPD?

Not as a trigger in its own right. Regulation 443.4.1 originally listed three consequences, and item (b), failure of a safety service, was deleted by the Corrigendum of May 2023. The lettering still runs (a), (b), (c) with a gap where it used to be, which is why it is so often quoted as live. Chapter 56, which covers safety services, adds no SPD requirement of its own. A fire alarm panel will very often still bring an installation into scope, but through the injury-or-loss-of-life route, or on the owner-declaration default, rather than because it is a safety service.

What happened to the SPD risk assessment?

It no longer exists. The calculated risk level method, Regulation 443.5 with its Table 443.1, Figures 44.2 and 44.3 and Annex A443, was deleted by Amendment 2 in 2022 and has not returned. Search the current standard and the letters CRL appear once in more than six hundred pages, in the note explaining that it was removed. Any article that still shows you the arithmetic with the environmental factor and the ground flash density is describing the 2018 edition.

What is the difference between Type 1, 2 and 3 SPDs?

It is a test class, not a quality grade. A Type 1 is tested with a 10/350 microsecond impulse, the waveform of a partial direct lightning current, and Regulation 534.4.1.3 puts one at the origin where the structure has an external lightning protection system. A Type 2 is tested with an 8/20 wave and goes at the origin where there is no such system, and at sub-distribution boards. A Type 3 is tested with a combination wave and belongs close to sensitive equipment, never on its own. Combined devices such as Type 1+2 do more than one job in one unit.

How long can the connecting leads to an SPD be?

Regulation 534.4.8 says the total wiring length between the connection points of the SPD assembly should preferably not exceed 0.5 m and shall in no case exceed 1.0 m. The length being measured is the whole path from the line conductor to the PE between points A and B, through any overcurrent device in the branch, not just the SPD tails. If it exceeds 0.5 m you must choose at least one of three remedies: a device with a lower Up, a second coordinated SPD near the equipment, or the Figure 534.9 method using an intermediate earthing terminal beside the SPD.

Why does lead length matter so much?

Because the leads are in series with the surge path and their inductance produces a volt drop that adds to the SPD protection level as the equipment sees it. The standard quantifies it in Regulation 534.4.8(c): a 1 m length of rectilinear conductor carrying a discharge current of 10 kA (8/20) adds a voltage drop of about 1000 V. That is why Regulation 534.4.4.2 caps the installed assembly at 2.5 kV on a 230/400 V system and says in as many words that it may therefore be necessary to select an SPD with a lower voltage protection level.

Can an SPD go on the load side of an RCD?

Yes, but Regulation 534.4.7 then requires an RCD with an immunity to surge currents of at least 3 kA 8/20, and notes that Type S devices to BS EN 61008-1 and BS EN 61009-1 satisfy that. Installing a Type 1 SPD downstream of an RCD is not recommended. On a TT system Table 534.5 makes the choice sharper: with connection type CT1 the SPD may only be installed downstream of the RCD, while the 3+1 CT2 arrangement can sit on the supply side.

Do EV charge points need surge protection?

Not by any regulation specific to them. Section 722 contains no SPD, surge or overvoltage requirement, and neither do the sections for agricultural premises, caravan parks or medical locations. The only special location with its own SPD rules is Section 712 for solar PV, and even there Amendment 4 deleted the old 712.443. An EV installation still falls under the general Regulation 443.4.1 test like everything else, and usually lands on the default that an SPD is provided.

Does an SPD need its own overcurrent protection, and how do I know when it has failed?

Regulation 534.4.5.1 requires the SPD installation to be protected against short-circuit currents, internally or externally per the maker instructions, and says the external device shall be the highest permissible rating so the assembly keeps its surge capability. The device is normally at end of life through a slow thermal overload, which its internal disconnector handles, so the visible status indicator is what tells you. Appendix 6 asks that it is checked, and the periodic inspection schedule carries an item for confirming the SPD is functional. Note also that SPDs must be disconnected before an insulation resistance test, or they will treat the test voltage as a surge.

Where this fits

SPD provision is part of the protection and coordination design for a board, alongside RCD selection, discrimination and the cable checks — and the earthing arrangement decides which connection type you can use, which is covered in TN-S, TN-C-S and TT explained. Deciding where SPDs are required across a distribution system and coordinating the types is something we fold into a full design package, and lightning risk work sits with our earthing and lightning protection studies.

Indicative, for learning — not a design. Every value here is transcribed from BS 7671:2018+A4:2026 and is reproduced as an aid to understanding it, not as a substitute. Check each figure against your own copy and against the SPD manufacturer's declared ratings before you specify anything. For the exact wording see Regulations 443 and 534 and Appendix 16, and the IET's Wiring Matters guidance on surge protection.

Compliant Protection, Board by Board

We design the protection — SPDs, RCDs and discrimination — as part of a full cable calculation package.

Discuss a project See the design service