The Question Behind Every Proposed Solution
When neutral current appears on an earthing conductor, gas pipe or shared building structure, the understandable response is to look for something that will stop it. Fit an electrode. Change the earthing arrangement. Install a protective device. Disconnect the electricity supply. Generate electricity independently. Each suggestion contains an idea worth examining. None can be judged properly until we identify the current path, the source driving it and the people who remain exposed after the proposed change.
This article examines seven proposals: an earth rod at every home; protective neutral bonding for every property; widespread conversion to TN-S; matt:e Guardian protection; local generation with island capability; an open-PEN detector combined with a neutral-to-earth relay and an electrode; and a three-pole switch intended to disconnect a property completely. The purpose is to explain their strengths, limitations and possible unintended consequences. It is not a wiring specification for modifying a service intake or building a protection assembly.
The central distinction is between repairing a defective network, limiting the consequences at one installation and maintaining electricity during a disruption. Those are three different objectives. A device can contribute to one without delivering the other two. Likewise, a technically sound earthing arrangement can still be unsuitable for a particular street because adjoining buildings, buried services or alternative supplies cross the boundary on which its safety depends.
Diverted neutral current is often abbreviated to DNC. The IET also uses neutral current diversion, or NCD. Its industry research identifies faults involving an open combined protective and neutral (PEN) conductor as a cause of current returning through unintended routes, including shared metallic services. The hazard is therefore wider than a faulty appliance within one home. The supply network and connections between premises matter. [1]
However, the explanation must not stop at a completely broken PEN conductor. Parallel return paths can carry current while the intended neutral remains continuous. A poor connection can progressively change the division of current. A neighbouring installation can drive current through local metalwork even when the local consumer unit is switched off. Treating all these conditions as one identical fault produces misleading promises about what a solution will achieve.
The conclusions here are deliberately specific. An electrode does not make current disappear. PNB is not another name for adding rods along a PME street. A separate-looking service cable does not establish end-to-end TN-S. A detector is not the same thing as a complete protective system. Island operation does not automatically remove imported voltages. A neutral-to-earth link does not automatically create TT. Finally, the number of poles on a switch says very little about whether all hazardous connections have been addressed.
The worked examples are original teaching calculations with stated assumptions. They are not measured network results, equipment settings or installation acceptance limits. References identify the published basis for factual claims. Where product performance depends on information not established by the public documentation, that uncertainty is retained rather than replaced with a reassuring assumption.
What PME and TN-C-S Actually Connect Together
In a TN-C-S arrangement, neutral and protective functions are combined in part of the supply system and separated elsewhere. The combined conductor is the PEN. At a typical domestic supply intake, the distributor provides separate neutral and earthing connections to the consumer installation. The consumer's circuit protective conductors and main protective bonding connect to the main earthing terminal, usually shortened to MET.
PME means protective multiple earthing. It describes a network arrangement in which the supply neutral is connected to Earth at multiple points. It is not a promise that every piece of bonded metal remains at exactly the potential of the ground beneath someone's feet. Real conductors and soil have impedance. Current through impedance produces a voltage difference. That remains true when the conductor is called neutral, earth or protective bonding.
Protective earthing provides a connection used in the protective arrangement. Protective bonding connects relevant conductive parts to reduce dangerous differences of potential between them. Neither function turns a gas pipe into an approved replacement neutral. Yet a bonded metallic service can become part of a parallel circuit if it connects premises whose METs are connected to different points on the neutral network.
Picture two houses joined by electrically continuous metallic water pipework. Each has appropriate bonding. Each also has a connection to the network earthing facility. There may now be a route from one neutral connection, through one property's earthing and bonding, along the shared pipe and back through the other property's bonding and earthing. The electrical circuit exists regardless of whether anyone intended the pipe to carry operating current.
Current does not select a single route because it is described as the easiest one. It divides among available paths according to their impedances and the driving voltages. If the intended PEN path has a much lower impedance, it carries most of the return current. If that path deteriorates, a greater share can use alternatives. If it opens, the alternatives may become the only available return routes for connected loads.
The IET's protective-earthing questions and answers explicitly recognise that diverted neutral current can occur without a broken PEN when parallel earthing paths have sufficiently low resistance. That is an important correction to the idea that any current on bonding proves a completely severed neutral. It also means a continuity indication alone cannot establish that a network has no parallel-current problem. [3]
Four Conditions That Should Not Be Confused
First is ordinary protective-conductor current from connected equipment. Filters and other components can produce such current. Its existence does not automatically identify a network neutral defect. Second is neutral current sharing through parallel bonded paths with a continuous PEN. Third is increased diversion caused by a high-resistance or open PEN. Fourth is externally driven current or voltage arriving through shared conductive services. Several can coexist.
Correct investigation identifies which mechanism is present. A clamp reading is evidence about current in the conductor being measured at that time. It does not by itself identify every source, determine the potential of every accessible surface or certify that opening a connection would be safe. The absence of a reading during a quiet period cannot establish what will happen when loads elsewhere change.
The difference between touch voltage and conductor current is equally important. Prospective touch voltage is the potential difference between accessible parts before a person touches them. The voltage actually applied across a person during contact also depends on the resulting circuit, including contact and body impedances. A bonding conductor can carry substantial current with a relatively small voltage drop while its continuity remains sound. Interrupt that path and the voltage across the interruption can become very different. Conversely, a hazardous potential can exist with little current until a person or another conductive route completes a circuit.
The local neutral and MET may rise together during a PEN problem. A low neutral-to-earth reading between those two locally connected points does not prove that either is close to remote Earth potential. Both instruments and protective devices must be assessed against the references they actually use. A measurement can be accurate yet answer the wrong safety question.
What the Evidence Shows and What It Does Not
The IET's 2025 research update reported approximately 76% TN-C-S, 22% TN-S and 2% TT among the reported earthing arrangements. It recorded an average measured current of 12 A and a highest reported current of 67 A. About 49% of respondents reported contacting the DNO. These are findings from submitted incidents, not a random survey of every UK installation. [2]
Those figures must not be converted into claims that 76% of homes have DNC, that 12 A is normal or that 67 A is a maximum possible current. The denominator and selection method matter. People reporting an unusual finding are not a representative sample of all electricity users. Equally, limited prevalence data do not make the physical mechanism hypothetical. They limit what can responsibly be said about its national frequency.
There is also a classification problem. An installation recorded as TN-S may be connected through a network containing combined neutral and earth sections. National Grid Electricity Distribution explains that separate neutral and earth service conductors can be supplied from a combined network. It distinguishes that arrangement from a genuinely separate system maintained along the whole route. [4]
For an article about solutions, the most useful response to incomplete statistics is to specify what evidence a proposed intervention needs. Does it reduce hazardous touch voltage? Does it reduce heating at vulnerable connections? Does it detect events missed by current processes? Does it remain effective when neighbours change their loads? Does it protect workers who disturb a shared service? Those are measurable questions.
The evidence should also distinguish an event from a location. One unresolved joint can generate many readings over time. One incident can affect numerous premises. Counting every reading as a separate network fault inflates one measure; counting only one address can understate exposure. A useful reporting system records the event, affected locations, time history and confirmed cause separately.
The Legal and Technical Framework
In Great Britain, the Electricity Safety, Quality and Continuity Regulations 2002 apply to relevant network and supply activities. Regulation 3 addresses equipment adequacy and prevention of danger so far as reasonably practicable. Regulation 7 requires all reasonable precautions to maintain supply-neutral continuity. Regulations 8 and 9 address earthing and PME. Regulation 8(4) prohibits a consumer combining neutral and protective functions in one conductor within the consumer installation. Regulation 28 requires specified written supply information for a person who can show reasonable cause for requiring it. [5]
These duties do not mean that every open PEN proves an identical legal breach regardless of circumstances. Nor do they support leaving a known dangerous network condition unresolved because a customer has installed mitigation. The condition, responsibility, response and applicable duty require evidence. An engineering article should explain those duties without pretending to decide individual liability from a photograph or a single reading.
The Electricity at Work Regulations 1989 are particularly relevant to employers and people undertaking electrical work. They address safe systems, earthing or other suitable precautions, suitable connections, isolation, prevention of re-energisation and competence. Their isolation provisions concern separation from sources of electrical energy, not simply moving a familiar switch to OFF. [6]
BS 7671 is a technical standard rather than an Act of Parliament. HSE describes it as a widely recognised code whose application is likely to support compliance with relevant aspects of the Electricity at Work Regulations. Contractual requirements, connection conditions and other legislation can also make its application significant. It should neither be presented as the whole law nor dismissed as optional advice without consequences. [7]
At the research date, 27 October 2026, the IET identifies BS 7671:2018+A4:2026 as the current edition following the October transition. BSI identifies BS 7430:2026 as the current protective-earthing code. Older articles remain useful explanations of principles but their clause references, terminology and edition-specific recommendations require checking against the current documents. Existing installations do not become unsafe solely because a newer edition has appeared. [8][9][31]
Northern Ireland has its own Electricity Safety, Quality and Continuity Regulations (Northern Ireland) 2012 and Electricity at Work Regulations (Northern Ireland) 1991. The physics is unchanged, but the correct legislation and network procedures must be used. A UK-wide campaign should not describe the Great Britain regulations or reporting arrangements as automatically covering every jurisdiction. [10][11]
Full current standards and current manufacturer instructions are necessary for project design. This article uses accessible primary guidance and published product documents; it is not a clause-by-clause conformity assessment of a proposed installation. That distinction matters most where protective-conductor switching, source transfer and special earthing arrangements are involved.
Proposal One: An Earth Rod at Every Home
The proposal is often expressed as installing millions of rods so that every home's neutral is connected to an electrode below 200 ohms. Its attraction is obvious: provide a local connection to Earth so a damaged neutral cannot leave metalwork floating. Additional electrodes can be useful. The difficulty is that the stated resistance alone does not establish the voltage that will arise under the relevant fault conditions.
There are actually two different proposals hidden in that sentence. One is an additional electrode connected to the MET of an installation that remains TN-C-S. The other is conversion to TT, with the consumer's protective arrangement appropriately independent of the distributor's PME earthing facility. They have different current paths and protective requirements. Adding a rod while retaining the PME connection does not, by itself, create TT.
Nor should a consumer interpret the proposal as permission to add an arbitrary neutral-to-earth link downstream. That can introduce additional operating-current paths and interfere with protective arrangements. The connection point, ownership boundary and intended system must be designed rather than inferred from the phrase earth rod.
What 200 Ohms Does and Does Not Mean
The familiar 200-ohm discussion in TT guidance concerns electrode performance and stability. The IET explains the relationship between electrode resistance, RCD operating current and disconnection, including the condition RA multiplied by IΔn being no greater than 50 V in the cited general arrangement. Here RA includes the electrode and protective-conductor resistance to the exposed parts; IΔn is the RCD rated residual operating current. Required disconnection times must also be met. The inequality does not mean the electrode voltage can never exceed 50 V before disconnection. It also highlights the concern that electrode resistance above 200 ohms may not be stable. These are not a guarantee that a PME electrode below 200 ohms controls open-PEN voltage. [12]
Consider an original simplified calculation. If 10 A actually flows through an electrode resistance of 200 ohms, Ohm's law gives 2,000 V across that resistance. A normal 230 V domestic source will not necessarily sustain that assumed current through such a path. The calculation is therefore not a prediction of 2,000 V at a house. It demonstrates why selecting a resistance without solving the complete circuit is invalid.
For a physically bounded illustration, take an ideal 230 V RMS AC source supplying a fixed 23-ohm resistive load. All calculated AC voltages and currents in this example are RMS values. The normal metallic neutral return is open. The only return is through a 200-ohm local electrode plus a deliberately idealised 1-ohm source electrode. Ignore cable impedances and other paths. The series resistance is 224 ohms, so current is approximately 1.027 A. The local electrode voltage rise relative to remote Earth is approximately 205.4 V. The load receives only about 23.6 V. The 205.4 V result is electrode potential rise relative to remote Earth, not a calculated voltage across a person. Ground potential near the electrode and the position of contact would affect an actual touch exposure.
That is not a prediction for every installation. It is a counterexample to the blanket claim that an electrode below or near 200 ohms guarantees safe touch conditions. Change the load and electrode values and the result changes. Introduce neighbours, pipework or three phases and the model must expand. The important point survives: the electrode is a resistance in a circuit, not a perfect zero-volt anchor.
If the design aim in another simplified case were to keep electrode rise below 50 V while 20 A actually flowed into that electrode, its resistance would need to be no more than 2.5 ohms. At 100 A, the corresponding arithmetic gives 0.5 ohm. These are conditional calculations, not design targets or universal safe-touch limits. They show why a domestic electrode test value cannot be assessed without the current and protective strategy.
Why the Rod Does Not Necessarily Replace the Pipe
Suppose a diverted current has two parallel return branches between the same electrical nodes: a 200-ohm earth path and a 0.5-ohm metallic-services path. For an illustrative total current of 20 A, the earth branch carries approximately 0.050 A while the metallic branch carries approximately 19.95 A. The potential difference across the parallel branches is approximately 9.98 V. Almost all the current still uses the metal route.
The assumed branch resistances represent complete paths between common nodes, not just the resistance of one rod or a short length of pipe. Real earth paths include spreading resistance and other electrodes; real metallic paths include connections and network conductors. Nevertheless, this simple current-divider example disproves the idea that providing a rod instructs fault current to avoid pipework.
Reducing the earth-path resistance changes the sharing. It may lower a particular potential difference while increasing the current carried by a local earthing conductor. That can be beneficial in one respect and demanding in another. The conductor, joints and electrode connection must remain suitable for the currents and durations the new arrangement makes possible. Lower resistance is not a substitute for analysing the whole system.
Multiple rods also do not necessarily behave like identical isolated resistors in parallel. Their surrounding soil regions interact. Closely spaced electrodes may offer less improvement than a simple resistance-divided-by-number calculation suggests. Soil conditions vary. A programme that counts installed rods but does not verify performance, location and continuing condition could create an impressive installation total without a demonstrated reduction in risk.
The Practical Scale of the Idea
The headline of 60 million rods should be treated as a proposal, not an established count of UK homes or required electrodes. Official statistics distinguish dwellings, households and countries. England alone had approximately 25.8 million dwellings in March 2025; a UK programme should use the relevant UK dwelling-stock dataset and a defined scope. Neither count tells us how many premises use PME or how many independent electrode locations are feasible. [14][15]
A flat in a tower block is not equivalent to a detached house with accessible ground. A shared intake may serve many dwellings. Some sites have existing foundation electrodes. Others have restricted ground, buried services or neighbouring earthing systems. The policy question is therefore what verified protective outcome is needed at each installation, rather than how many identical rods can be purchased.
Ground penetration introduces its own hazards. HSE's underground-services guidance emphasises planning, locating services and safe working. Installing an electrode beside an intake is not automatically safe simply because the visible work area is small. Electricity cables, gas pipes and other services may be close together. [13]
The defensible conclusion is that additional electrodes can form part of an engineered solution. They may improve the local reference or support another protective arrangement. A universal sub-200-ohm prescription does not demonstrate control of diverted current, imported voltage or broken-PEN exposure. The performance target must come first, with electrode design following it.
Proposal Two: PNB for Every Property
Protective neutral bonding is a legitimate network arrangement. The problem is the claim that implementing it at every house on an existing feed would make faults go down rods instead of pipes. That description combines a real arrangement with an incorrect account of current flow. PNB must be understood from the source, neutral, electrode and customer connections together.
National Grid Electricity Distribution describes PNB where customer connections occur at one position on a network, commonly with a dedicated transformer. Its current installer guidance says the arrangement is normally considered TN-S but is classified as TN-C-S where a combined neutral and earth conductor lies between the electrode connection and the customer's connection. That makes the actual conductor arrangement decisive. [4]
The current BS 7671 terminology also matters. SELECT's Amendment 4 guide describes PNB as a TN-C-S arrangement and identifies the new Figure 3.9B, with PME conditions potentially applying on public networks. That description and NGED's broader use of PNB must not be collapsed into a claim that all PNB supplies are TN-S or exempt from PME precautions. Establish the actual conductor arrangement, applicable standard and DNO declaration. [31] Electrical Safety First also provides an Amendment 4 guide covering supply characteristics and earthing; it explicitly does not replace the standard or engineering judgement. [30]
What PNB Can Change
Consider a dedicated supply whose neutral reference to Earth is established at the customer connection position rather than through a remote earth connection followed by a long combined protective and neutral section. A break upstream of that local reference can have a different effect from a break in a conventional PME PEN supplying many installations. The position of the reference and the conductors serving protective functions determine what rises in potential.
This can be valuable, especially where a suitable dedicated transformer and connection layout already exist. The engineering attraction is a controlled topology: the designer can identify the source, the earthing point and the customer conductors. It is not an ability to direct current selectively into soil. The electrode still has resistance. Other conductive connections still exist unless the design establishes otherwise.
The statutory background is specific. ESQCR regulation 8(3)(b) allows the neutral connection to Earth at the consumer-connection point, or nearer the source, where consumer installations connect to a single source at only one point on the network. That is a defined exception to the usual source-earthing location. It does not authorise householders to redesign an unchanged shared PME network independently. DNO-specific conditions still need confirmation. [16]
The distinction also explains why the phrase for everyone needs scrutiny. A street with repeated neutral-to-earth connections along a shared distribution route is not converted into a collection of independent single-point PNB systems just by changing labels. The shared neutral and its multiple Earth connections remain part of the same electrical network. The physical topology has to change if the intended protective behaviour is to change.
What Would Universal Conversion Involve?
A serious PNB proposal would need to specify the source arrangement for each group of customers. Would there be dedicated transformers? One common connection position? Separate protective conductors beyond that position? How would shared building structures and utility services be treated? Which existing connections to Earth would remain? Who would own and maintain the altered system?
Those questions expose the practical difference between a local earthing upgrade and network reconstruction. If the proposal requires new transformer locations, revised feeders, redesigned protection and new land rights, it should be costed and assessed as such. Calling the work PNB does not remove the construction, access and outage implications.
There is a potentially useful targeted application. Where a particular site already has a dedicated source and suitable layout, the DNO and designer can examine whether a well-defined PNB arrangement offers advantages over the existing arrangement. The answer may differ for a rural isolated property, a private industrial site and a terrace supplied along a common street main. A single national slogan cannot replace those distinctions.
Why Pipework Remains Part of the Assessment
Imagine a redesigned property still connected by a metallic service to a neighbour whose MET can rise during a network event. The shared service can import that potential. A local electrode may reduce the resulting voltage in some conditions, but it also creates a current path. The consequences depend on all relevant impedances and the reference points involved.
Removing required protective bonding is not an acceptable way to make that connection disappear on a drawing. Bonding addresses differences of potential within the installation. Where a shared conductive service undermines an intended separation, the engineering problem must be resolved with the relevant parties and utility requirements. It cannot safely be solved by leaving accessible conductive parts at unrelated potentials.
PNB therefore deserves assessment where the network arrangement supports it. It does not support the claim that current will choose an electrode instead of a pipe and it cannot be created independently at every house while retaining an unchanged shared PME topology. The worthwhile question is where a different source-reference arrangement demonstrably reduces exposure.
Proposal Three: Replace TN-C-S with TN-S
Of the seven ideas, genuine separation of neutral and protective functions addresses the combined-conductor dependency most directly. In a true TN-S system, neutral carries operating return current while the protective conductor provides the protective connection. A neutral-only break does not, by itself, sever the protective conductor or force the installation MET to follow the disconnected load neutral.
That is a real architectural advantage. It is also narrower than saying TN-S eliminates every electrical hazard. The separate protective path must remain continuous and suitable. Connections can deteriorate. Equipment can develop insulation faults. Shared services can import potentials from other systems. A broken neutral can still produce dangerous supply-voltage conditions for equipment even where the protective earth remains intact.
Take a simplified two-load example. Two resistive loads of 100 ohms and 400 ohms are connected from two different phases to a common neutral point. If that neutral point becomes disconnected from the source neutral and no other paths are present, the loads become series-connected across an ideal 400 V line-to-line supply. The current is 0.8 A. The loads receive 80 V and 320 V respectively. Separate protective conductors do not correct that voltage division.
The example is not intended to describe every three-phase fault. It illustrates a distinction: neutral integrity and protective-conductor integrity are different design concerns. TN-S separates the functions so one conductor is not relied upon for both, but continued neutral monitoring, suitable protection and sound maintenance remain necessary.
The System Has to Be TN-S All the Way That Matters
A separate earth at the service head is insufficient evidence. If that protective connection derives upstream from a combined neutral and earth section, a defect in that section can still affect it. Replacing only the final service cable may improve the cable without changing the fundamental upstream dependency.
The IET's 2025 research discussion highlights hybrid networks containing older separate-conductor sections and later combined-conductor sections. That makes supply records and DNO confirmation important, particularly when an installation's design depends on avoiding PME conditions. A visual assessment of the intake cannot reveal every joint and cable type beneath the street. [2]
For new privately controlled networks, or suitably bounded redevelopment projects, an end-to-end separate arrangement can be specified from the outset. The boundary is clearer and the owner can maintain records and control later alterations. Retrofitting a mixed public network requires the same clarity, but on a much larger scale with more interfaces and more parties involved.
Must Every Road Be Dug Up?
Not necessarily. Some areas may already contain usable separate infrastructure. Some routes may offer ducts or other construction options. Others may require substantial excavation and replacement. A credible programme would start with network condition and topology rather than assuming either that every road must be excavated or that conversion can be achieved through a few intake changes.
The cost is more than additional conductor material. It includes investigation, design, traffic management, cable installation, joints, service transfers, outage planning, reinstatement, records and continuing maintenance. Existing customer installations may need assessment at the interface. A national figure quoted without a defined scope would provide little basis for comparison with alternative risk-reduction programmes.
Conversion also creates a transition period. A street partly converted to TN-S can remain connected to PME premises through metallic services. Temporary arrangements and the order of works therefore matter. The completed architecture cannot be assumed to exist throughout the construction programme. Each stage needs a safe operating arrangement and an accurate record of what has actually changed.
The strongest case for TN-S is its removal of the normal combined-conductor dependency within the defined system. The strongest reservation is the practical scale of achieving and maintaining that separation, especially across existing mixed networks. Targeted conversion during renewal or redevelopment may be more defensible than either dismissing the option as impossible or presenting universal conversion as an immediate remedy.
Proposal Four: Install matt:e Guardian Devices
Monitoring and automatic protective action can provide a valuable layer between a network problem developing and a person encountering its consequences. But product evaluation must distinguish sensing, decision logic, switching hardware and the installation being protected. A named device cannot be assessed as though every configuration has identical capabilities.
matt:e describes Guardian as a monitoring platform using its O-PEN technology. The published product page lists the Guardian TP, GES, GPC and GPC-A packages. It describes monitoring, alerts and control of connected loads, with different packaging and power arrangements. These are manufacturer descriptions of intended functions, not independent evidence that every possible current path at every connected building is interrupted. [17]
The Trumeter Guardian data sheet is particularly important. It identifies three-phase applications and states that single- and split-phase use does not provide the VN-N and VN-E measurements, meaning measurements relative to the virtual neutral. The sheet is marked September 2024. Its limitation must be checked against the exact proposed product and current documentation rather than ignored when discussing a typical single-phase house. [18]
There is a documentation discrepancy to resolve. The November 2024 instructions, version 1.2 for firmware 006_001_23, list virtual-neutral measurements for the 1P3W configuration, whereas the September data sheet excludes them for split-phase use. The instructions also distinguish 1P2W from 1P2W + E. Manufacturer confirmation of the exact hardware, firmware and reference arrangement is therefore needed before making a single- or split-phase performance claim. Neither document alone establishes every current variant's capability. The instructions identify configuration-dependent events and relays controlling external contactors or undervoltage devices. A monitor and a complete protective installation remain different things. [18][19]
What a Suitable System Could Achieve
A system that recognises the relevant abnormal condition and operates suitable switching can reduce the duration for which specified loads remain energised. Alerts and event records can help responsible people respond. Repeated events may provide evidence of an intermittent problem that is difficult to observe during a short visit. Those are credible functions to investigate and verify.
The actual safety outcome depends on what the detector recognises. Does it detect elevated protective-conductor potential relative to a suitable reference? Abnormal line-to-neutral voltage? Loss of neutral? Some combination? How does it behave with balanced loads, phase loss, high-resistance faults or an externally imported potential? A successful demonstration of one condition does not answer all those questions.
It also depends on what opens. Disconnecting a property's line and neutral removes its own connected loads as contributors through that supply path. It does not necessarily remove current or voltage imported through the DNO earthing connection, shared pipes, structural metalwork or another supply. The question is not whether a contactor operates, but which hazardous paths remain after it operates.
For example, suppose house A has its live supply conductors disconnected while house B continues to drive current through shared pipework. House A's appliances may be de-energised from their normal supply, yet the pipe entering house A can remain involved in the external circuit. The protective outcome for house A must be evaluated against that imported source as well as its own load.
This is a circuit-level observation, not a claim that a particular Guardian package necessarily uses an inadequate arrangement. The public documents examined do not justify assigning one universal switching topology to all Guardian installations. A fair article should neither claim comprehensive protection without evidence nor allege a product defect from an unspecified schematic.
Is It Only a Temporary Fix?
Not inherently. A properly specified monitoring or protective system can remain a permanent part of an installation's safety strategy. However, it remains mitigation if it does not correct the defective network connection. The distinction is between the purpose of the measure and the duration for which it is installed. Calling everything temporary obscures the legitimate role of continuing protection.
Equally, the presence of local protection cannot be treated as permission to leave a known dangerous supply fault unresolved. A trip should lead to a defined response, investigation and appropriate repair. A device that repeatedly operates without anyone addressing the cause is evidence of an unresolved operational problem, not a successful long-term outcome merely because it keeps switching.
What about neighbours? The device does not automatically extend its protective boundary to another house. Removing a local load can reduce one contribution to a shared problem. Depending on network conditions, changing the connected loads can also alter voltages and current sharing elsewhere. The correct conclusion is that neighbouring properties are not thereby proven protected. An absolute claim of no effect on neighbours would be too strong.
What Should Be Established Before Specifying It?
The evidence should identify the model, supply configuration, sensing references, supported fault conditions, switching arrangement and application limits. It should establish behaviour following loss of control power, sensor failure, a failed switching contact, communication loss and restoration of the network. Any automatic reconnection needs a defined basis that does not confuse a temporarily normal reading with proof of repair.
Local protective action should be distinguished from cloud notification. Receiving an email is not the same event as interrupting a hazardous circuit. The designer must establish which safety functions operate locally and which services depend on external communications. This is a general requirement for evaluating connected protection equipment, not an assertion about an undocumented weakness in Guardian.
EV open-PEN protection also has a defined application context. The IET's OPDD specification was developed for devices associated with EV charging. Compliance claims within that scope should not automatically be extended into certification of an entire home's earthing, bonding and alternative-supply system. The complete application needs its own engineering justification. [20]
Guardian is therefore a serious option to evaluate as part of a specified protective system. The defensible position is conditional: establish what the selected arrangement detects, what it disconnects and what remains connected. The evidence reviewed supports discussing its monitoring and control capabilities. It does not support a blanket promise that fitting a device makes every homeowner, pipe and neighbour safe from every form of DNC.
Proposal Five: Generate Electricity and Operate as an Island
Local generation and storage can provide resilience when a public supply is interrupted or deliberately disconnected. They do not automatically protect against every hazard associated with a PME network. The critical distinction is between being independent for energy and being independent of externally imposed electrical potential. Those are different properties of an installation.
A property can obtain all its operating power from an inverter while remaining connected to shared metallic services and a distributor's earthing facility. Its energy meter may show no import, but those conductive connections still exist. Electricity generation behind the meter does not erase the connections represented by the earthing and bonding system.
Likewise, ordinary grid-connected solar generation is not necessarily a backup supply. The equipment must be designed to establish and sustain the required island, with appropriate transfer, earthing, protection and controls. A solar installation that disconnects when the grid is lost may be doing exactly what its connection requirements intend. Backup capability has to be specified, not assumed from the presence of panels or a battery.
Does Bonding Make Island Operation Impossible?
No. The statement that island operation is possible only where nothing requires bonding is too broad. Protective bonding remains an important part of many correctly designed islanded installations. What can be difficult is achieving the intended separation from an external PME system where shared conductive parts cross the boundary.
That distinction matters in a normal home. A metallic service might require protective bonding because it can introduce a potential from outside the installation. Removing the bonding to make the island appear independent can create a dangerous potential difference between the service and locally earthed equipment. The engineering objective is a safe relationship between accessible conductive parts, not a drawing with fewer green-and-yellow lines.
The earlier IET island-mode guidance makes a particularly useful point: when the distributor's earthing connection remains connected, PME conditions can continue to apply in island operation even though the local supply does not rely on that connection for its own fault protection. A backup supply and an imported-PEN-voltage solution are therefore not automatically the same project. [21]
The Local Source Needs Its Own Protective Arrangement
The IET's July 2026 guidance, scoped to a single residential dwelling with one local power-conversion source, explains the common local TN-S arrangement, a system referencing conductor linking the island neutral and protective system, a local electrode and coordinated switching for islanding. It also explains why limited inverter fault current changes the protection problem. Appropriate RCD provision and its position relative to the source reference require attention. [22]
The practical implication is straightforward. A circuit-breaker that operates satisfactorily on a strong public supply cannot simply be assumed to achieve the same disconnection performance from a current-limited inverter. The protective arrangement must be checked for each operating mode. A successful mains-mode test does not establish the island-mode result.
The MCS battery-installation standard similarly addresses island earthing, protection of maintained circuits and the suitability of switching arrangements. Its requirements belong within their stated scope and edition. They are further evidence that backup operation is a complete electrical design task rather than a convenience setting in an app. [23]
Resilience Has an Energy Limit
Even a correctly protected island cannot supply unlimited energy. An illustrative usable battery output of 8 kWh would support a constant 1 kW load for about eight hours if that output figure already accounts for usable capacity and conversion losses. A 4 kW average load would reduce that arithmetic duration to about two hours. Real operation also depends on reserve settings, temperature, ageing and load changes.
Power capability matters separately. A system may contain enough stored energy for an evening but be unable to start a motor or support several high-power appliances simultaneously. Essential-load selection and automatic load management can therefore be part of a sensible resilience design. They do not alter the need to address imported current or potential through conductive services.
Generation and storage also interact with network connection requirements. ENA guidance identifies G98 and G99 as relevant connection frameworks. National Grid Electricity Distribution describes G99 procedures for generation above the relevant threshold or outside G98 requirements. The appropriate route depends on the proposed equipment and operating arrangement; a grid-parallel system with backup is not exempt merely because it is on private premises. A purely switched alternative requires a different assessment from equipment that can run in parallel. [25][26] In Great Britain, ESQCR regulation 21 addresses switched alternatives and requires prevention of parallel operation; regulation 22 addresses sources capable of parallel operation. Northern Ireland has corresponding provisions in regulations 22 and 23 of its 2012 regulations. [5][10]
Island capability can be useful where loss of supply has serious consequences or where local energy use provides other benefits. Its limitation as a DNC answer is that it solves continuity of energy only when properly designed and electrical independence only to the extent actually established by the whole installation. Shared metalwork remains part of the assessment.
Proposal Six: An Open-PEN Detector, an N/E Relay and a Rod
This proposal is attractive because it combines detection, switching and local supply. However, saying that an open-PEN detector switches a neutral-to-earth relay so the system becomes TT mixes different functions. The detector identifies a condition. Transfer equipment separates sources. A referencing link establishes the local source relationship to the protective system. An electrode establishes a connection to the general mass of Earth. None of those functions can be inferred from the presence of the others.
Start with the earthing terminology. In a common islanded inverter arrangement, one source conductor is referenced to the local protective system and the installation's protective conductors return faults to that same source reference. The neutral and protective conductors are separate downstream. That is the essential local TN-S relationship. The presence of an earth rod does not turn it into TT.
In TT, the installation's exposed-conductive-parts are connected to an electrode arrangement independent of the supply's protective earthing connection, with the earth path forming part of the fault circuit. If a local source reference and equipment protective conductors are directly connected together, the fault-return topology is different. Classifying the system from the label on a relay instead of tracing that topology is a serious design error.
There are specialised configurations and broader standards discussions beyond a typical domestic inverter installation. That is not a reason to label an unspecified arrangement TT. The actual diagram and applicable requirements must establish the classification. A comparison table in Schneider Electric's installation guide illustrates that permitted connected-mode and island-mode combinations depend on the starting system and source arrangement. It does not support arbitrary conversion from TN-C-S to TT by adding a relay. [24]
Why Coordination Matters More Than Component Count
Consider what could go wrong if a local neutral-to-earth reference were established while an unintended connection to the grid neutral remained. The installation could acquire an additional parallel neutral path. Current could appear in conductors intended for protection. Conversely, if a local source energised circuits without the intended reference and protective arrangement, a fault might not produce the response assumed by the design.
The engineering task is therefore to define valid operating states and prevent invalid combinations. Normal grid operation, transition, stable island operation, shutdown and restoration must each have an understood source reference and protective path. A drawing showing two relays is insufficient unless the assembly's interlocking, ratings and failure behaviour support those states.
Current IET guidance distinguishes the system referencing conductor switching device from the switching device for islanding. The functions are coordinated but are not interchangeable. The article also identifies product standards for these roles. This is a useful framework for asking suppliers the right questions; it is not a generic wiring recipe for combining unrelated devices. [22]
The open-PEN detector must also be suitable for its detection task in the proposed configuration. If it depends on a particular reference that changes during transfer, the behaviour in every state must be understood. A protection function should not become blind or unstable simply because its own output changed the circuit it measures. That needs explicit system-level verification.
What Happens to the DNO Earth and Shared Services?
This is the question most easily omitted. Separating the grid's live conductors does not necessarily separate its earthing connection. Separating one earthing connection does not necessarily remove a shared pipe or another metallic link. If an external fault can still impose a potential on the local protective system, the island may continue operating while that imported hazard remains.
The design must therefore state which connections remain in each mode. It must explain how required bonding is preserved and how any boundary between different earthing arrangements avoids hazardous simultaneous contact. The presence of local generation makes this more important: local circuits may remain energised when the grid is intentionally unavailable or unsafe.
This proposal can develop into a legitimate engineered protection-and-resilience system. But the useful concept is coordinated detection, source transfer and an appropriate local protective arrangement. The statement that a detector, relay and rod automatically produce TT is incorrect. Treating that combination as a ready-made domestic remedy would omit the very interactions that determine its safety.
Proposal Seven: A Three-Pole Switch for Complete Isolation
Before evaluating this idea, ask which three conductors the switch is intended to interrupt. A three-pole device commonly switches three line conductors in a three-phase circuit. That would not automatically disconnect the neutral. In a single-phase proposal, the intended poles might instead be line, neutral and a protective connection. Those are fundamentally different arrangements.
Pole count is therefore not a description of protective performance. The conductors, switching duties, timing, device suitability and external connections must all be identified. A switch that successfully disconnects the intended live supply can still leave conductive parts connected to another source of potential through bonding or shared infrastructure.
Line and Neutral Disconnection
Disconnecting both live conductors of a single-phase supply can separate the property's normal load from that supply when the device and arrangement provide suitable isolation. It does not automatically interrupt a DNC path involving the DNO earthing facility and shared services. The path can be outside the part of the circuit that the switch opens.
This is why a local main switch being OFF does not establish that all associated metalwork is free from externally driven current or voltage. The Electrical Safety First safe-isolation guide explicitly recognises that diverted neutral currents can originate elsewhere and remain present after local isolation. Its guidance supports additional precautions and prompt DNO reporting rather than treating switching as complete proof of safety. [27]
Switching a Protective Connection
It would be inaccurate to state that UK practice never permits protective-conductor switching under any circumstances. Defined exceptions and arrangements exist. It would be equally inaccurate to turn that observation into general permission to put an ordinary spare switch pole into the earthing conductor. The conditions governing linked operation, protective continuity, device suitability and the whole installation must be satisfied.
A technical review published by WorkSafe New Zealand expressly notes that the United Kingdom permits protective-conductor switching under specified conditions. Its Appendix D examines the pre-Amendment-4 edition of BS 7671. It establishes that exceptions have existed, not approval of a proposed circuit under the current edition. Current protective-conductor continuity and isolation requirements must be checked for the actual design. The protective path must not be lost while related live conductors remain connected unless the applicable protective arrangement expressly permits and controls that state. [29]
There is a further distinction between a separate protective conductor after the supply functions have been separated and the distributor's combined PEN conductor. A PEN performs a neutral function as well as a protective function. Interrupting or modifying it is not ordinary consumer-side isolation work. No proposal here authorises alteration of sealed DNO equipment or the network PEN.
Why Opening One Protective Connection May Not Isolate the Property
Suppose a specially engineered arrangement disconnects the property's grid line, neutral and designated supply-earthing connection. A continuous metallic water service still connects the property to neighbouring bonded systems. Another route might exist through structural steel, cable armour or a second electrical supply. The property is not electrically independent merely because three named conductors have opened.
An external potential on a pipe can remain accessible inside. Required bonding may connect it to the local MET. Alternatively, removing that bond can leave the pipe at a different potential from other accessible parts. Neither outcome can be declared safe from the switch position alone. The shared-service boundary must be investigated and resolved within the overall protective design.
This does not mean that switching is useless. It means its benefit must be described precisely. An arrangement can remove specified supply connections, stop the local load contributing to a fault path or support a controlled change of source. Complete separation from every external electrical influence is a much stronger claim requiring much more evidence.
Failure States and Restoration
A protective switching assembly must also be assessed when something fails. A contact can fail to open. A control supply can disappear. A sensor can give an invalid signal. A neutral can reconnect in an unsuitable relationship to other conductors. These are foreseeable design questions, not exotic objections to innovation.
Restoration deserves equal attention. If the network appears healthy for a short interval, does the system reconnect automatically? Could reconnection recreate the hazardous path? Is a manual reset appropriate after an event? What information reaches the person responsible for investigation? The required answer depends on the application, but it must be defined rather than left to an undocumented default.
The three-pole proposal is consequently neither a universal cure nor something that can be evaluated solely by a blanket prohibition. It needs a complete circuit and a demonstrable protective strategy. An ordinary switch is not enough to establish that strategy and a property cannot be declared wholly isolated while relevant shared conductive paths remain unexamined.
Where a Properly Designed TT Conversion Fits
TT deserves separate treatment because it is repeatedly invoked without actually being created. A genuine conversion can avoid relying on the distributor's PME earthing facility for the consumer's protective arrangement. It can therefore address a particular route by which an open-PEN condition would otherwise raise the consumer's bonded metalwork. That is a meaningful benefit when the installation can be designed appropriately.
The conversion is not simply an electrode plus removal of a wire. Fault protection, RCD arrangements, electrode performance, conductors, enclosures, surge protection and every connection crossing the intended boundary need assessment. The IET's TT guidance emphasises electrode stability and separation from nearby earthing systems and buried conductive parts. Those considerations explain why a drawing that looks independent may not behave independently on a crowded site. [12]
A common metallic service can reconnect the intended TT installation to PME metalwork through the very bonding required to control local potential differences. An electrode can also lie within the potential gradient of another earthing system. Independence is therefore a property to establish through design and site information, not a label to attach after installing a rod.
The distinction between exposed-conductive-parts and extraneous-conductive-parts matters here. An exposed-conductive-part is a touchable conductive equipment part that is not normally live but can become live if basic insulation fails. An extraneous-conductive-part is not part of the electrical installation but is liable to introduce a potential, generally Earth potential. These definitions concern electrical behaviour, not simply whether an object looks metallic. Not every visible piece of metal automatically belongs to either category. Equally, a plastic service section does not establish the status of every other metallic connection in a building.
For a detached property with suitable ground and well-understood services, TT may be practical. For one flat in a building with extensive shared metalwork, a local conversion may be far more difficult. That is a reason for a building-level assessment, not permission to omit bonding or to rely on a small island of nominally separate earthing without examining simultaneous contact.
The continuing condition matters as well. Later building alterations can add a metallic connection across a separation boundary. An outbuilding can be supplied with a cable whose armour reintroduces an unwanted link. A new energy system can change source referencing. Documentation and periodic review are therefore part of keeping the design effective, not merely paperwork after installation.
Why Ordinary RCDs and Circuit-Breakers Are Not a Complete DNC Answer
An RCD responds to an imbalance between the currents passing through its monitored live conductors. It does not directly measure every accessible metal surface against remote Earth. If a local load's current passes out through the RCD line and back through its neutral before diverting elsewhere upstream, those monitored currents can still balance. The RCD has not failed merely because it does not identify that external diversion.
Similarly, current arriving through a shared pipe and leaving through another protective connection may bypass the RCD's monitored conductors entirely. A hazardous imported potential can therefore exist without the residual-current condition that would operate the device. That does not make RCDs unimportant: they remain essential components of many protective arrangements for the faults they are designed to detect.
A fuse or circuit-breaker also responds according to its current-time characteristics and the current in its own path. A harmful temperature rise at a poor connection or in a diverted path need not correspond to an overcurrent sufficient to operate the local installation's protective device. The location of heating and the location of measurement matter.
For an illustrative fixed-current calculation, 20 A RMS through a 0.1-ohm resistive connection dissipates an average 40 W at that connection. At 40 A, the same resistance would dissipate 160 W. These are calculations of electrical power, not predictions of a particular pipe temperature or ignition event. Heat dissipation, connection construction, surrounding materials and duration determine the thermal consequence. They explain why sustained local heating can matter even when the overall supply current does not look exceptional.
Comparing the Seven Proposals Fairly
The comparison should ask what changes in the circuit rather than which option has the most reassuring name. The following table summarises the engineering assessment. Each potential benefit assumes a suitable design and installation; none is a universal product endorsement.
| Proposal | Potential contribution | Main limitation |
|---|---|---|
| Additional electrodes on retained PME | Can improve the local Earth reference and support a defined protective strategy | Does not guarantee safe open-PEN voltage or remove lower-impedance metallic paths |
| PNB in a suitable supply arrangement | Can change the source-reference topology and reduce particular combined-conductor exposures | Requires an appropriate network arrangement; adding rods along a PME street is not universal PNB |
| Genuine end-to-end TN-S | Separates neutral and protective functions within the defined system | Requires verified continuity and separation; shared services and neutral-voltage faults remain relevant |
| Guardian in a specified protective system | Can monitor supported conditions and initiate configured alerts or disconnection | Capability depends on model, supply, settings and switched paths; it does not repair the network |
| Local generation with island capability | Can maintain selected loads during a supply interruption | Energy independence does not establish independence from imported potentials |
| Detection, source transfer and N/E referencing | Can form part of a coordinated local-source protective arrangement | Does not automatically create TT; incorrect coordination can introduce new hazards |
| Three-pole switching proposal | May interrupt specified conductors within an engineered arrangement | Pole count does not establish complete isolation or address every shared conductive path |
It is also important to compare like with like financially. A network renewal programme is not directly comparable with one domestic protective device. The first may change the underlying architecture for many customers; the second may reduce a defined exposure at one installation. Costs and benefits need the same population, hazard definition and time horizon before a meaningful comparison is possible.
Whole-life cost includes investigation, construction, commissioning, inspection, maintenance, component replacement and response to faults. It also includes outages and the consequences of an unsuitable intervention. A low purchase price can be misleading if the installation requires substantial additional work. Conversely, a high initial infrastructure cost can have benefits extending beyond DNC. Neither direction should be assumed without evidence.
Three Situations That Expose Weak Assumptions
A Terrace with a Shared Metallic Service
An electrician identifies unexpected current in bonding at one house. The property is locally isolated but an external contribution remains. A proposed extra rod may add another return branch without removing the shared metal route. A proposed live-conductor disconnector may remove the property's load but leave the externally driven path. The immediate priority is a safe response and investigation of the actual sources and connections.
The lesson is that the unit of analysis is not always the individual customer account. Electrically connected premises can form one relevant problem area. Neighbours need not share the same supplier for the physical connections to exist. Commercial billing boundaries do not define electrical fault boundaries.
A Detached Home with Battery Backup
The owner wants essential circuits available during an outage and protection against an external PEN problem. Those are two requirements to specify separately. The designer needs to identify the island source, maintained loads, protective arrangement and external conductive connections. Demonstrating that the lights remain on after transfer proves continuity of energy, not protection from every external potential.
A successful project would document both functions and their limits. If external conductive connections prevent the intended separation, that must be resolved or clearly reflected in the protection strategy. It should not be hidden behind the word backup.
A Block of Flats Undergoing Electrical Renewal
One flat cannot necessarily establish an independent protective environment while remaining connected to shared risers, structural steel and services. The building's intake, distribution and bonding arrangements may need coordinated assessment. Individual devices might still contribute useful protection, but their boundaries must be clear and their interaction with the common installation understood.
The useful opportunity is to address topology, records and monitoring while renewal work is already being planned. A building-level solution may deliver benefits that isolated alterations cannot. Conversely, installing many devices without understanding the common connections may reproduce the same unresolved assumption at every front door.
What a Credible Improvement Programme Should Do
The first priority is to identify and rectify actual network defects. Where a high-resistance or open PEN is confirmed, restoring a sound intended return path addresses that defect directly. Protective measures at customer installations can support safety during and after the response, but they do not physically repair a damaged conductor or joint.
Repair is not the whole answer to every form of DNC. Parallel sharing can remain on a continuous network. A broader programme should therefore distinguish defective-neutral events from persistent parallel-current conditions and from imported hazards associated with shared services. Each category needs an appropriate investigation and response rather than an identical instruction to fit a rod.
Network information should be made usable. Designers need reliable declarations of supply earthing and relevant characteristics. Operators need records that reflect repairs and conversions. Where a future design depends on continuing TN-S conditions, those assumptions need to be recorded and managed. A change buried beneath the street should not silently invalidate the protective assumptions of connected equipment.
Local protection can then be selected against a defined hazard. A design brief should state who is being protected, which conductive parts are included, which sources are present and what acceptable operation means. It should identify any residual exposure and the action required when a device operates. This creates a basis for selecting equipment rather than allowing the chosen equipment to define the problem after purchase.
Trials should measure outcomes that matter. A pilot should distinguish detected events, confirmed defects, false alarms, interruptions, residual touch voltages and currents in relevant paths. It should record whether an intervention changes exposure elsewhere. A before-and-after clamp reading from one conductor is not enough if the current has simply moved into another path that was not monitored.
The trial population should also be described. A group of detached houses with plastic incoming services cannot establish performance for blocks with extensive shared metalwork. A three-phase demonstration cannot automatically establish the behaviour of a single-phase variant. Clear limits make research more useful because readers can see where further evidence is needed.
What Commissioning Should Demonstrate
A useful commissioning plan separates evidence of construction from evidence of protective performance. The equipment must be the specified version, connections must match the approved design and the installation must have the required inspection and test results. Beyond that, the responsible designer needs evidence that the intended detection and switching functions produce the defined outcome under the relevant conditions. An illuminated display or successful remote command proves only a small part of that chain.
Testing should use authorised methods appropriate to the equipment and the hazards. It must not involve deliberately breaking a live network PEN or creating dangerous current in occupied pipework. Suitable laboratory facilities, validated models, manufacturer test evidence and controlled commissioning procedures serve different purposes. The evidence should say which method established each claim and which conditions were outside its scope.
Acceptance criteria should be written before the trial results are known. Otherwise a system that behaves unexpectedly can be declared successful by changing the objective afterwards. If the objective is to limit exposure at specified accessible parts, record that outcome. If it is to detect particular network anomalies, record detection coverage and missed cases. If it is to support resilience, record transfer performance and maintained-load capability separately.
The handover information should explain what the user sees when the system operates, what action to take and which functions depend on maintenance or communications. It should identify who can change settings and who investigates an event. Uncontrolled changes made to stop inconvenient trips can undermine the original design even though the equipment remains physically installed.
Finally, the protective assumptions need a review trigger. Changes to the supply, a new outbuilding, replacement services, added storage or a revised distribution arrangement can affect them. A well-documented system makes those assumptions visible so later work can be assessed against them. That is more durable protection than relying on the original installer remembering why a particular link or separation was provided.
Training belongs within this programme. Electricians, assessors, designers and people working on gas or water services need to understand why local isolation may not remove an externally driven hazard. Simulation can help people explore these current paths without creating hazardous mains conditions for learners. Its scenarios should be checked against circuit models and real equipment behaviour, with clear boundaries between educational practice and permission to undertake work.
Safe Investigation and Useful Reporting
This article does not provide a live fault-finding procedure for householders. Unexpected voltages, shocks, arcing or heating associated with an intake or metallic service require a competent response. People should not disconnect bonding, loosen an earthing terminal or cut a pipe to see whether the current stops. Removing a current-carrying connection can change the hazard immediately.
Electrical Safety First's safe-isolation guidance includes DNC checks within the broader precautions for safe work. These checks complement suitable voltage testing and secure isolation; a clamp meter is not a replacement for proving dead at the point of work. Instruments have different functions and limitations. No single reading proves every relevant conductor safe. [27]
There is no universal bonding-current number that can be presented here as a safe-to-touch threshold. Body current is not the same quantity as current measured in a bonding conductor. Contact conditions, potential difference, duration and the path through a person affect shock risk. Likewise, a low reading does not justify opening a connection when the consequences of interruption are unknown.
Suspected DNC associated with the network should be reported promptly to the DNO. In Great Britain, 105 provides the power-network reporting route. Northern Ireland Electricity Networks publishes 03457 643 643 for faults and hazardous situations. An immediate threat to life or a fire also requires the appropriate emergency response. [1][28]
Useful records describe what was observed without overstating what was proved. They identify the location, supply arrangement as known, conductors measured, relevant load conditions, instrument used and whether readings changed during authorised investigation. Photographs and time-stamped logs can support the record where safely obtained. Unverified assumptions should be labelled as such.
Completion should include a clear account of the cause established, work undertaken and checks performed after repair or alteration. If the cause remains uncertain, that uncertainty should not disappear from the record merely because a reading returned to normal. Intermittent conditions require a response suited to their nature.
The Solution Is a Verified Outcome
The seven proposals are not interchangeable alternatives on a shopping list. They operate at different levels. Earth electrodes alter connections to the ground. PNB and TN-S concern supply topology. Protective devices detect and respond to defined conditions. Island systems provide local energy and require their own protective design. Switching interrupts specified conductors but does not erase other conductive connections.
Some can be combined effectively. Some are suitable only in particular settings. Some versions of the proposals need correcting before they can become designs. The practical question is always what the complete arrangement does during normal operation, the relevant fault, transition and restoration, including conditions outside the property boundary.
For PENfault.com, the strongest case is therefore for better evidence and clearly defined protection. Investigate and repair actual defects. Make network arrangements understandable. Preserve required protective measures. Assess shared conductive paths. Verify local protection against its stated scope. Treat backup power as a separate requirement unless the design proves that it also addresses the external hazard.
Millions of rods, a new label at the intake or a reassuring switch position cannot substitute for that work. Equally, difficulty and uncertainty are not reasons to dismiss practical improvements. They are reasons to define the problem precisely enough that an improvement can be demonstrated. A credible solution shows which hazardous condition it controls, how it does so and what still needs attention.
Practise Safe Isolation and Testing with Tradefox
Every proposal above depends on the same thing: competent people who measure before they assume. Tradefox lets apprentices and electricians rehearse safe isolation, earth fault loop and RCD testing in realistic interactive simulations, so the routine is second nature before it matters on site. Explore the full range of Tradefox courses to keep your skills sharp.
Sources and Scope Notes
Primary guidance and public product documentation checked for this article. Publication dates and limits matter: older guidance is used for principles, not as a substitute for current design requirements.
- IET: Neutral current diversion industry research, July 2024
Background mechanism and reporting. Not a national prevalence estimate. - IET: Neutral current diversion industry research update, May 2025
Reported incidents, measured currents and hybrid networks. Survey percentages are not percentages of UK homes. - IET: Protective earthing webinar questions and answers
Confirms that broken PEN conductors are not the only possible cause of DNC. Older edition references require care. - National Grid Electricity Distribution: Information for electrical installers
DNO explanation of CNE/SNE networks and topology-dependent PNB classification. - Electricity Safety, Quality and Continuity Regulations 2002, revised text
Great Britain legislation. See particularly regulations 3, 7, 8, 9 and 28. - Electricity at Work Regulations 1989, Part II
Great Britain workplace duties including isolation, referenced conductors and competence. - HSE: HSR25, guidance on the Electricity at Work Regulations
Explains the relationship between BS 7671 and statutory duties. - IET: Ensure you are up to date with BS 7671
Current-edition and transition information checked 27 October 2026. - BSI: BS 7430 current release
Identifies BS 7430:2026. The full paid standard was not reproduced or independently audited for this article. - Electricity Safety, Quality and Continuity Regulations (Northern Ireland) 2012
Separate Northern Ireland network legislation. - Electricity at Work Regulations (Northern Ireland) 1991
Separate Northern Ireland workplace legislation. - IET: TT earthing considerations, September 2021
Background to electrode stability, RCD coordination and separation. Uses an earlier BS 7671 edition; not a substitute for current design requirements. - HSE: Avoiding danger from underground services, HSG47
Planning and safe working where ground penetration could affect buried services. - ONS: Dwelling stock by tenure, UK
Appropriate statistical series for defining a UK housing scope. Dwelling counts do not establish PME connection counts. - MHCLG: Dwelling stock estimates, England, 31 March 2025
Approximately 25.8 million dwellings in England. Not a UK total. - ESQCR 2002, regulation 8: connection with Earth
Paragraph 3(b) provides the single-connection-point exception relevant to PNB. Read with the applicable DNO requirements. - matt:e: Guardian product information
Manufacturer descriptions and package names. Not independent product certification. - Trumeter: Guardian data sheet, September 2024
Includes the footnote limiting VN-N and VN-E measurements in single- and split-phase use. Confirm exact current product applicability. - matt:e: Guardian user instructions V1.2, November 2024
V1.2 for firmware 006_001_23. Its 1P3W virtual-neutral information appears inconsistent with the September 2024 data-sheet footnote. Confirm applicable capability with the manufacturer. - IET: Open combined protective and neutral conductor detection devices, IET 01
Defines the EV-charging application context. Does not establish whole-premises approval. - IET: Island mode earthing arrangements, March 2021
Explains retained PME connections and local source referencing. Historical guidance read alongside the 2026 update. - IET: Automatic disconnection of supply for island-mode prosumer installations, July 2026
Current technical discussion for single residential dwellings with one local power conversion source. - MCS: MIS 3012:2025, Battery installation standard
Island-mode earthing and protection provisions within the document's stated scope and edition. - Schneider Electric: Specific protection of prosumer electrical installations
Technical explanation of source referencing and connected/island earthing combinations. Not a replacement for UK requirements. - Energy Networks Association: Energy storage
Connection framework for storage, including G98 and G99. - National Grid Electricity Distribution: G99 connection procedures
Operator guidance for generation and storage connection applications. - Electrical Safety First: Best Practice Guide 2, issue 4.3, August 2026
The PDF cover identifies issue 4.3 despite its older issue-4-2 filename. Guidance recognises externally sourced DNC and safe-isolation precautions. - Northern Ireland Electricity Networks: Powercheck and emergency contact
Published 24-hour fault and hazard reporting number: 03457 643 643. - WorkSafe New Zealand: G Kenyon Technology technical review, March 2026
Appendix D discusses BS 7671 before Amendment 4; Appendix E notes UK exceptions to protective-conductor switching. Historical/comparative evidence, not approval of a current design. - Electrical Safety First: BS 7671 Amendment 4 pocket guide, 2026
Publisher landing page replaces the broken PDF link. Overview of supply characteristics and earthing; not a replacement for BS 7671. - SELECT: Amendment 4, what members need to know
Transition guidance and the distinction between an older installation and an unsafe one.
