AFDDs Under BS 7671 Amendment 3 — The Most Controversial Requirement in Recent UK Wiring History

An AFDD monitors the current waveform on a circuit and identifies the characteristic high-frequency signature of electrical arcs. The physics is straightforward: when current flows across a gap (a loose terminal, damaged insulation, a broken conductor), it forms a plasma arc with a distinctive broadband noise signature in the 1 kHz to 100 MHz range, concentrated around specific frequency bands.

AFDDs conforming to IEC 62606:2013+A1:2017 must detect two fault types:

Series arc faults

A break in a conductor where current continues to flow through the arc. This occurs at loose screw terminals, damaged cable insulation where the conductor is partially severed, or deteriorated connections. The critical characteristic of a series arc fault is that current does not increase — it may actually decrease slightly because the arc introduces additional impedance. This means:

• An MCB will not trip (current is at or below normal load)
• An RCD will not trip (no earth leakage is present)
• The arc generates localised heating at temperatures exceeding 3,000°C, sufficient to ignite surrounding materials

Series arc faults are the fire risk that no other protective device addresses. This is the core justification for AFDDs.

Parallel arc faults

A fault between live and neutral (or live and earth) through an arc rather than a solid connection. These generate higher currents than series faults and may eventually trip an MCB or RCD — but the arc may ignite material before the conventional device operates, particularly at lower fault currents where MCB tripping times are longer.

AFDD manufacturers use proprietary algorithms, but the fundamental approach is consistent across devices:

1. High-frequency current sensing. A current transformer with a bandwidth of several MHz monitors the circuit current for broadband noise characteristic of arcing.
2. Pattern recognition. The algorithm distinguishes arc signatures from normal high-frequency noise sources (SMPS power supplies, LED drivers, motor commutators). This is the hardest engineering problem — nuisance tripping from legitimate loads has been the primary objection to AFDDs.
3. Current waveform analysis. Arcs produce characteristic flat shoulders on the current waveform where the current re-ignites each half-cycle. The algorithm detects these shoulder patterns.
4. Persistence check. The device confirms the arc signature persists for a defined duration (typically 0.5–2 seconds depending on severity) before tripping, reducing false positives from transient events.
5. Trip action. The AFDD disconnects the circuit. Most devices incorporate MCB and optionally RCD functionality in the same unit, providing combined protection.

Modern AFDDs (2024–2026 generation) have significantly improved discrimination compared to early devices. Nuisance tripping rates have reduced from approximately 1 in 50 installations to approximately 1 in 500, based on field data from major UK manufacturers. This improvement is real but the nuisance tripping concern has not been fully eliminated.

BS 7671:2018 Amendment 3:2024 introduces AFDD requirements through Regulation 421.1.7, which specifies that AFDDs conforming to BS EN 62606 shall be provided for final circuits in the following locations:

• Single-phase AC final circuits not exceeding 32 A supplying socket-outlets in dwellings
• Final circuits in higher-risk residential buildings (HMOs, purpose-built flats, care homes)
• Final circuits in locations with sleeping accommodation not covered above
• Final circuits supplying socket-outlets in buildings with combustible construction (timber frame, etc.)

Critically, Regulation 421.1.7 includes a risk assessment route: the designer may omit AFDDs where a documented risk assessment demonstrates that the risk of fire from arc faults is acceptably low. This provision is both a pragmatic compromise and a source of ongoing dispute — what constitutes an acceptable risk assessment is not defined, leaving considerable room for interpretation.

The requirement applies to new installations and rewires from the effective date. Existing installations are not required to be retrofitted, though the IET Guidance Note 3 recommends considering AFDDs when circuits are modified.

No recent change to BS 7671 has generated as much debate as the AFDD requirement. The arguments on each side are substantive:

In favour of mandatory AFDDs

• Fire statistics. The UK Home Office reports approximately 13,000 dwelling fires per year with an electrical cause (2023/24 data). Even if only 10–20% involve series arc faults (estimates vary widely), that is 1,300–2,600 fires per year that no existing protective device can prevent.
• NEC precedent. The US National Electrical Code has required AFCIs (Arc Fault Circuit Interrupters, the US equivalent) in bedroom circuits since 2002, extended to all habitable rooms by 2014. US fire statistics show a measurable reduction in electrical fires in jurisdictions that adopted AFCI requirements early.
• Technology maturity. The current generation of AFDDs has 20+ years of development behind it (counting US AFCI experience). Nuisance tripping has been substantially reduced. The devices work.
• Life safety. A single fire prevented justifies the cost of thousands of installations. The value-of-statistical-life calculation strongly favours mandatory AFDDs.

Against mandatory AFDDs (or for risk-assessment-only)

• Cost. AFDDs add £40–£80 per circuit to installation cost. A typical 10-circuit domestic consumer unit increases from approximately £300 to £700–£1,100. For social housing providers installing thousands of units, this is a significant budget impact.
• Nuisance tripping. While improved, AFDDs can still trip on some LED dimmer combinations, vacuum cleaners with brush motors, and certain power tool types. Each nuisance trip risks the occupant bypassing the device entirely.
• Data quality. UK fire statistics do not separately identify series arc fault fires from other electrical causes. Without this granularity, the cost-per-prevented-fire calculation is based on estimates, not measurement.
• Existing protection. RCDs already prevent many electrical fires (earth fault fires). MCBs prevent sustained overcurrent fires. The incremental risk addressed by AFDDs — series arc faults only — may be smaller than advocates suggest.

The fundamental challenge in the AFDD debate is the absence of UK fire investigation data that specifically identifies series arc faults as the ignition mechanism.

UK fire investigation reports typically classify fires as “electrical” without distinguishing between overload, earth fault, poor connection (series arc), insulation failure (parallel arc), or external heat source affecting wiring. This means the number of fires that AFDDs would actually prevent is unknown — we have only estimates.

The Electrical Safety First charity commissioned research in 2021 estimating that 38% of domestic electrical fires involved “faulty or deteriorated connections” — a category that includes series arc faults but also includes connections that fail through overheating without arcing. If even half of these involved series arcs, AFDDs could prevent approximately 2,470 fires per year in the UK.

Opponents argue that this figure is inflated because many “faulty connection” fires involve high-resistance joints that overheat without producing the broadband frequency signature that an AFDD detects. The true preventable number may be lower — but by how much is genuinely unknown.

Until the UK adopts fire investigation protocols that specifically identify arc fault ignition (as the US NFPA 921 methodology does), this debate will continue to be driven by estimates rather than measurement.

For engineers designing installations under Amendment 3, the practical implications are:

1. Consumer unit sizing. AFDD modules are typically 1.5–2 module widths per circuit (when combined with MCB). A domestic consumer unit that previously required 12 modules may now require 18–20. Specify enclosure sizes early.
2. Load compatibility testing. Before handover, test each AFDD-protected circuit with the actual connected loads. Vacuum cleaners, hair dryers, dimmed LED lighting, and power tools should all be operated to verify no nuisance tripping.
3. Client communication. Explain to the client what AFDDs do, why they trip, and what to do when they trip. Nuisance trips that the client does not understand lead to unauthorised bypassing — eliminating the protection entirely.
4. Risk assessment documentation. If omitting AFDDs based on Regulation 421.1.7’s risk assessment provision, document the assessment thoroughly. Include the basis for concluding that series arc fault risk is acceptably low. This document will be scrutinised if a fire occurs.
5. Cable sizing unchanged. AFDDs do not affect cable sizing calculations. The cable must still be sized for the MCB rating, and voltage drop and earth fault loop impedance requirements remain unchanged.

Standards referenced: BS 7671:2018 Amendment 3:2024, Regulation 421.1.7. AFDD product standard: BS EN 62606:2013+A1:2017 (IEC 62606). US equivalent: UL 1699:2017 (AFCIs). Fire data: UK Home Office Fire Statistics 2023/24, NFPA Fire Loss Reports 2023.