The Pen-Fault Boundary Constraint: Open-PEN Detection Engineering and Earth Electrode Realities

A Low-Probability Fault with a High-Consequence Geometry

Most of the electrical safety measures in a UK dwelling are directed at faults that originate inside the installation: an insulation failure, a damaged cable, a wet appliance. Residual current devices and overcurrent protective devices disconnect these faults rapidly because the fault current has a defined path back to the source and is large enough to be detected. The open-circuit failure of the combined protective earth and neutral (PEN) conductor in the public network belongs to a different category. It occurs outside the installation, nothing inside the dwelling causes it, no protective device is guaranteed to see it, and its principal hazard is not an overcurrent but a dangerous voltage on exposed metalwork that appears to be perfectly safe. The introduction of the domestic electric vehicle, which places a large conductive body outdoors, in contact with the ground through its tyres and often handled by occupants in wet conditions, has made this long-recognised weakness of Protective Multiple Earthing (PME) a practical regulatory issue.

This audit characterises the physics of the open-PEN fault, examines the requirements of BS 7671 Regulation 722.411.4.1 and the voltage envelope they establish, quantifies why a conventional earth electrode is so difficult to achieve in urban soils, and describes the architecture of the electronic isolation devices that have become the dominant compliance route. It is intended for electrical contractors, retrofit coordinators and distribution planners, and it identifies a number of points on which the commonly repeated summaries of the topic need qualification.

The Physics of an Open-PEN Shock Vector

In a TN-C-S supply arrangement, the form most widely known in the UK as PME, the network operator provides a single conductor which serves as both the neutral and the protective earth over the distribution main, and which is connected to earth at multiple points along its length. At the consumer's cutout, the PEN conductor divides: one part continues as the neutral of the installation and the other as the main protective earthing conductor, connected to the main earthing terminal. From that terminal, the circuit protective conductors (CPCs) extend to every exposed conductive part in the installation and, through the charging equipment, to the body of any connected vehicle. In normal operation the protective earth and the neutral are at the same potential, which is close to that of the general mass of earth, and the vehicle body is therefore safe to touch.

Suppose that the PEN conductor in the street main fractures upstream of the property, through corrosion, mechanical damage, a failed joint or excavation. The network downstream of the break has lost its neutral return path to the source transformer. The line conductor remains live. Current can now return to the source only through whatever path to earth remains: the local earth electrodes of the PME system itself, the earthing of neighbouring properties and the general mass of earth. The potential of the whole downstream PEN conductor, and therefore of the neutral and every CPC and exposed conductive part connected to it, is no longer anchored to earth by the source. It is determined by the voltage divider formed by the loads connected between line and PEN on the one hand and the resistance to true earth of the disconnected section on the other.

The Voltage Divider

The simplified relationship is V_PEN = V_L × R_E / (R_E + R_L), where V_L is the line voltage relative to true earth (approximately 230 V), R_E is the effective resistance to earth of the isolated section of PEN and the paths associated with it, and R_L is the equivalent resistance of the loads connected between the line and the PEN. As R_L becomes large compared with R_E, which is the situation when little is connected, the PEN potential approaches the full line voltage. A charging point delivering 7.36 kW has an equivalent resistance of 230² / 7,360, or approximately 7.2 ohms. With an effective R_E of 20 ohms, the PEN is raised to 230 × 20 / 27.2, or approximately 169 V. With R_E of 50 ohms the value is approximately 201 V, and with 10 ohms approximately 134 V. For a household load of 1 kW, with R_L of approximately 53 ohms, the same 20 ohm resistance gives approximately 63 V. With no load at all, the PEN potential is the full line voltage.

This produces the counter-intuitive and dangerous result that the lightest load conditions are the worst. When the vehicle is parked and not charging, and the household is quiet, the PEN, and with it the CPC, the charger's earth terminal and the vehicle body, can sit at or near mains potential relative to the ground on which the occupant is standing, with no overcurrent, no tripping and no visible sign. The path of the return current is through the load, the neutral-earth bond at the cutout and out through the PEN into the CPC network. The charger does not need to be energised for the vehicle body to be raised: the body is connected to the CPC through the charging cable, whenever the cable is plugged in.

The person who touches the vehicle while standing on damp ground or in contact with a grounded structure forms the final element of the circuit: from the raised body, through the hand, the torso and the legs, to the earth. Body impedance depends on contact area, moisture and voltage, and for a wet hand-to-foot path it is commonly taken as the order of 1,000 ohms or lower. At 230 V this implies currents of the order of 100 to 200 mA. Under the effects described in IEC 60479, currents above approximately 30 mA flowing for a second or more can cause ventricular fibrillation, and currents above about 10 mA can prevent the victim from releasing their grip. The hazard is therefore lethal, not merely unpleasant, and the exposure is not a rare event of a broken neutral at the instant of touching: because a PEN fault persists until detected and repaired, the vehicle can remain dangerous for hours. Open-PEN faults are infrequent, but their consequences are such that the regulations treat them as a design basis rather than as an extraordinary event.

BS 7671 Regulation 722.411.4.1 and the Voltage Monitoring Envelope

Section 722 of BS 7671, which covers electric vehicle charging installations, addresses the issue in Regulation 722.411.4.1. The principle is that the PME earthing terminal of the supply must not be used as the sole means of earthing for an EV charging point unless additional protective measures are taken. The regulation, as amended, sets out alternatives. As I understand the current text, these include a device that monitors the supply and disconnects the charging point when the voltage deviates from the permitted range, an earth electrode of sufficiently low resistance that the vehicle body cannot rise above a specified voltage in an open-PEN condition, and the avoidance of PME altogether, by providing a TT earthing arrangement for the charging point. The precise wording and any exceptions, which include some installations in which the vehicle is not accessible to a person in contact with the ground, should be checked against the current edition and amendments before specifying.

The Voltage Window

For the electronic option, the device is required to monitor the voltage between line and neutral, and to disconnect the line, the neutral and the protective conductor of the charging point if the root-mean-square (RMS) value rises above 253 V or falls below 207 V. These thresholds correspond to the nominal 230 V plus 10% and minus 10%, respectively: 230 × 1.10 = 253 V and 230 × 0.90 = 207 V. The disconnection must take place within a specified time, which I understand to be five seconds. The use of an RMS measure is significant. The device must evaluate the effective heating value of the waveform rather than its instantaneous peak, and so must sample at a rate adequate to resolve the waveform, typically several kilohertz, computing the RMS over a whole number of cycles, and applying a filtering interval so that a brief transient does not cause disconnection.

It is worth noting how these thresholds relate to the statutory supply envelope. The Electricity Safety, Quality and Continuity Regulations 2002 require a declared voltage of 230 V with a tolerance of plus 10% and minus 6%, that is, from approximately 216 V to 253 V. The 207 V lower threshold used in the device is wider than the statutory lower limit by about 9 V. The difference is deliberate: legitimate low-voltage events on loaded networks, such as the starting of a large motor or a heavily loaded feeder, can approach the lower statutory limit, and a threshold set at exactly 216 V would cause nuisance tripping, with the consequence that users disable or bypass protection. The 207 V figure is a compromise between sensitivity and availability.

What the Voltage Window Can and Cannot See

An engineering assessment of this approach must acknowledge its limitations. The device measures the voltage between line and neutral, whereas the hazard is a voltage between the vehicle body and true earth. The two are related only through the fault. When the PEN conductor opens upstream, the neutral potential shifts, and the line-to-neutral voltage seen by the device changes according to the load balance: a heavily loaded charging point would pull the PEN towards the line potential and the voltage across the load would fall, perhaps below 207 V, and a lightly loaded one might not. Conversely, an unbalanced neighbouring load, connected to the same section of the main, can push the neutral away and produce an overvoltage. In both situations the device detects a deviation and operates. There exists, however, a set of conditions in which the PEN rises with respect to earth while the line-to-neutral voltage remains within the window, notably when the load at the property is small and the network downstream is balanced in such a way that the line and neutral move together. In that case, a device relying exclusively on line-to-neutral measurement is blind to the elevated potential.

This is not a defect peculiar to any manufacturer but an inherent property of detecting a ground-referenced hazard with a measurement that is not ground-referenced. The regulatory and product-standard frameworks accept it on the basis of a risk assessment, the probability of the coincidence of conditions, and the fact that a vehicle is often charging, and thus loading the circuit, during exposure. Some designs add a measurement of the line-to-earth voltage, with an internal reference electrode or the vehicle itself as a reference, or evaluate the asymmetry between the line-to-earth and neutral-to-earth voltages, which widens the range of conditions detected. Practitioners should consult the product standard to which the device is certified, which is understood to be the BS 8300 series or an equivalent, and the manufacturer's declared detection performance, instead of assuming that all devices are equivalent.

Earth Electrode Impedance and Soil Physics

The alternative to electronic detection is to provide the installation with its own earth electrode of low enough resistance that the vehicle body cannot rise to a dangerous potential. The mechanics follow directly from Ohm's law and the divider relationship above. For the body potential to be held at or below a touch voltage limit V_t with the PEN raised to the line voltage, the electrode resistance must satisfy R_E ≤ V_t × R_L / (V_L − V_t). For a conventional limit of 50 V a.c. and the 7.2 ohm equivalent resistance of a 7.36 kW load, R_E must not exceed 50 × 7.2 / 180, or approximately 2.0 ohms. For a limit of 70 V, which I understand to be the value used for this purpose in the regulation, it is approximately 3.2 ohms. These calculations are deliberately simple, and treat the electrode as the only return path. In practice the remaining PME electrodes and the neighbouring properties provide parallel paths, which is part of the reason that guidance documents quote less demanding figures, with a maximum electrode resistance of around 20 ohms commonly cited, derived from assumptions about those parallel paths. The designer should work from the current guidance and not from the single-load calculation, but the single-load figure illustrates the order of the problem.

Rod Resistance

The resistance of a single vertical driven rod of length L and diameter d in soil of resistivity ρ is given by the Dwight formula, R = ρ / (2πL) × [ln(8L / d) − 1]. For a 2.4 m rod of 16 mm diameter, ln(8 × 2.4 / 0.016) is approximately 7.09, so the bracket is 6.09. In soil of resistivity 30 ohm-metres, representative of wet clay, the resistance is 30 / 15.08 × 6.09, or approximately 12 ohms. In soil of 100 ohm-metres, typical of mixed or drier ground, it is approximately 40 ohms. In soil of 200 ohm-metres, which is representative of gravel, sand and rubble fills that are common beneath urban driveways, it is approximately 81 ohms. Soil resistivity varies seasonally with moisture and frost, so a value measured in wet winter conditions will rise in a dry summer.

Reducing the resistance requires multiple rods, which must be spaced at least their driven length apart, and preferably twice that, if they are not to share the same volume of soil and so lose effectiveness. With a spacing of twice the length, the utilisation factor is of the order of 0.8 to 0.9, and the resistance of n rods is approximately R / (n × 0.85). To reach 20 ohms at 100 ohm-metres requires about 2.4 rods, so three. To reach 2 ohms at 100 ohm-metres requires 40 / (2 × 0.85), or about 24 rods, at a spacing of 4.8 m, which cover an area of several hundred square metres. At 200 ohm-metres the number exceeds 47. These figures explain why the electrode route is usually described as unviable to the standard required for a strict touch-voltage limit in an urban plot: it is not that the physics is obscure, but that the geometry cannot be accommodated in a driveway or a small garden, and that the cost of drilling and trenching through concrete, block paving and services is out of proportion to the cost of a protective device.

It should also be recognised that a rod of low resistance is only one component of a safe electrode system. The conductors connecting the rod to the main earthing terminal must be of adequate size and protected against damage, the joint to the rod must be accessible for inspection, and the resistance must be tested periodically, since corrosion and seasonal drying cause it to drift. An electrode that is installed, tested once and then forgotten may no longer provide the protection it was designed for within a few years.

The TT Alternative

A further route avoids the problem by separating the charging point from the PME terminal. In a TT arrangement the exposed conductive parts of the installation are connected to a local earth electrode, not to the supply's earthing terminal, and fault protection is provided by a residual current device. Here the electrode resistance requirement is far less onerous, because it is governed by the RCD rating: for a 30 mA device the relationship R_A × IΔn ≤ 50 V gives a maximum value of approximately 1,667 ohms, though a practical value of well below 200 ohms is normally recommended to allow for seasonal variation. The difficulty is that the installation must be arranged so that the PME terminal is not bonded to the charger's exposed metalwork, which requires a separate sub-installation, a supply and cable arrangement that complies with the separation requirements, and careful consideration of the extraneous conductive parts and of the bonding to the vehicle. TT conversion also does nothing for the rest of the dwelling. It is a viable approach for a detached charging point, such as a post on a driveway or a wall unit in a separate building, and less so for a charger integrated into a house with an existing bonded earthing system.

Solid-State Separation and Hardware Architecture

The open-PEN detection device is, in effect, a monitoring and switching module that sits in the supply path to the charging point, either as a separate unit or integrated into the charger itself. The architecture comprises three functional blocks: a measurement chain, a decision unit and a power switching stage.

Measurement and Decision

The measurement chain derives scaled and electrically isolated copies of the line-to-neutral voltage, and in some designs the line-to-earth and neutral-to-earth voltages, through resistive dividers or transformers, and digitises them with an analogue-to-digital converter at a sample rate of several kilohertz. A microcontroller calculates the true RMS value over a window of one or more mains cycles (20 ms per cycle at 50 Hz) and compares it with the upper and lower thresholds. To prevent spurious operation due to transients, the decision logic applies a debounce or confirmation interval, and to prevent chattering around a threshold it applies hysteresis when re-enabling. The overall response time, from the onset of a sustained deviation to the opening of the contacts, is determined by the filter interval plus the mechanical operating time of the relay, and for products on the market it is generally specified in the range from several hundred milliseconds up to the permitted maximum of a few seconds. The statement that such devices act within milliseconds is true of the relay operating time itself, which is of the order of ten to thirty milliseconds, but not of the complete detection chain, and the regulatory allowance is of the order of seconds. A faster response reduces exposure but increases the risk of nuisance tripping.

Reliability of the decision is critical because the device is a safety function. Well-designed units use two independent measurement and processing channels that cross-check each other, a watchdog timer that forces a safe state if the processor ceases to execute correctly, and supervision of their own supply voltage. A single-channel design with no diagnostic coverage can fail silently, which is a worse condition than a device that fails open.

Switching Stage

The switching stage uses a multi-pole relay or contactor to isolate the charging point. When it operates, it opens the line and the neutral conductors, and also the protective conductor that connects the charging point and the vehicle to the installation's earthing system. Opening the protective conductor is what separates the vehicle body from the elevated PEN potential: with the live conductors disconnected, the body is no longer driven by any source, and with the CPC disconnected, it is no longer bonded to the raised installation earth. Depending on the design, a two-pole device switches line and neutral and a further stage switches the protective conductor, while four-pole contactors are used in three-phase equipment.

The relay is designed to fail safely. The contacts are normally open, closed by the coil against a spring, so that loss of control power or a processor failure causes the contacts to open. Forced-guided, or mirror, contacts allow the controller to verify that all poles have opened and, equally important, that a contact has not welded closed, which is detected by a self-test at each plug-in and by monitoring auxiliary contacts. The contactor is rated for the full charging current, 32 A continuous, with a margin for the inrush and the breaking of load, and an endurance rating consistent with the number of operations over the life of the equipment. After a trip, the device re-enables the supply only when the voltage has been inside the permitted window for a defined period, so that the vehicle is not repeatedly connected and disconnected during an unstable supply condition.

A further design feature concerns the coordination with the charging controller. When the device opens under load, the vehicle's on-board charger experiences an abrupt loss of supply. Where possible the device signals the charging controller to withdraw the pilot signal, which asks the vehicle to stop drawing current, shortly before the contacts open, which reduces arcing at the contacts, but the protective action must not depend on that signalling.

The matrix is a summary for comparative purposes. The thresholds and times refer to the regulation as understood by the author at the time of writing, and the installer must confirm the requirements applicable to the specific installation, edition and amendment of BS 7671, and the declared performance and certification of the product selected.

Specification and Commissioning Guidance

For the majority of domestic installations on PME supplies, the practical conclusion is that electronic open-PEN detection is the proportionate solution, because the alternatives require groundwork that is either impossible or uneconomic in the plots on which most chargers are installed. That conclusion does not mean the choice of device is a formality. The specifier should confirm the product standard and certification, understand the detection principle used and its stated coverage, and, where possible, prefer designs that add a ground-referenced measurement or an equivalent diagnostic to the line-to-neutral window. The installation should be tested at commissioning in accordance with the manufacturer's instructions and BS 7671, including verification of the trip thresholds if the product supports it, and the device should be left in a condition in which its self-test status is visible to the user.

Where the supply arrangement is known, the contractor should record whether it is TN-C-S, TN-S or TT. A TN-S supply, in which a separate earth conductor runs back to the source, does not present an open-PEN risk of the same kind. A TT supply has its own electrode. Only the TN-C-S case invokes Regulation 722.411.4.1, and the contractor should establish the arrangement with the DNO or by measurement and not by assumption. Records of the chosen protective measure, the devices and the test results form part of the Electrical Installation Certificate and are the evidence that the hazard has been considered.

Finally, it is worth communicating the nature of the residual risk to the customer in proportionate terms. An open-PEN fault is rare, and a compliant installation reduces its consequences substantially, but no protective measure is perfect, and an owner who notices a tingle or shock from the vehicle or the charger, or any unusual behaviour of the equipment, should stop charging, avoid touching the vehicle and call a qualified electrician and the DNO. The aim of the regulatory regime is to make that scenario sufficiently improbable that the technology can be deployed at scale, and the engineering reality described in this article is the reason that the requirement exists in the form that it does.