Grounding, bonding, lightning and surge protection

Grounding and bonding are among the most misunderstood topics in electrical engineering, partly because one word, "ground", is used for several different things. Lesson 4 covered system grounding, the connection of a power system's neutral to earth. This lesson covers the rest: the bonding that makes faults clear quickly and keeps metal surfaces safe to touch, grounding electrodes and their resistance to earth, the step and touch voltages that make substations dangerous during faults, the bonding networks that protect telecommunications and IT equipment, and the lightning and surge protection that guards a facility against the largest currents it will ever see.

What bonding is for

When a phase conductor faults to a metal enclosure, two things must happen: the fault must be cleared quickly by the overcurrent device, and nobody touching the enclosure in the meantime should receive a dangerous shock. Both depend on a low-impedance path from the enclosure back to the source, provided by the equipment bonding (grounding) conductor, metal raceways and the main bonding jumper that connects that system to the neutral at the source. The earth itself is not part of this path: its resistance is far too high to carry enough current to trip a breaker.

One neutral-to-ground bond

The neutral and the bonding system are connected at one point only: the service, or the source of each separately derived system such as a transformer or generator. Additional neutral-to-ground connections downstream create parallel paths, so normal neutral current flows on bonding conductors, raceways, cable trays and building steel. This "objectionable current" causes nuisance tripping of ground fault protection, magnetic fields that disturb equipment, and shock and fire hazards. Hunting down improper bonds is a common commissioning and troubleshooting task, using clamp meters on bonding conductors (which should normally carry almost nothing).

Generators and transfer switches require particular care. If a transfer switch does not switch the neutral (a three-pole switch), the generator neutral must not be bonded at the generator; if it does (a four-pole switch), the generator is a separately derived system with its own bond. Getting this wrong defeats ground fault protection on the service or the generator.

Grounding electrodes

The grounding electrode system connects the electrical system to the earth to limit voltages from lightning, surges and accidental contact with higher-voltage lines, and to stabilize voltage to earth. Electrodes include ground rods, concrete-encased electrodes (the reinforcing steel in footings), ground rings around the building, metal water pipes and building steel, all bonded together.

The resistance of an electrode to remote earth depends mainly on the soil resistivity \(\rho\) and the electrode's size. For a single rod of length \(L\) and radius \(a\):

Soil resistivity is measured on site before design with the Wenner four-pin method: four probes are driven in a line at equal spacing \(a\), a current is passed between the outer pair and the voltage measured between the inner pair, and the apparent resistivity is \(\rho = 2\pi a R\), where \(R\) is the measured voltage-to-current ratio. Repeating at several spacings reveals how resistivity changes with depth. For example, \(R = 5.3\) Ω at \(a = 3\) m gives \(\rho = 2\pi \times 3 \times 5.3 = 100\) Ω·m.

The Canadian Electrical Code (Section 10) specifies acceptable grounding electrodes and their installation; utility substations and the grounding grids of large facilities are designed to IEEE 80.

Step and touch voltages

During a ground fault at a substation, thousands of amperes flow into the ground grid, raising its potential relative to remote earth (the ground potential rise, GPR). The voltage differences that matter for safety are local: between a person's hand on a grounded structure and their feet (touch voltage), and between their two feet a metre apart (step voltage). IEEE 80 sets tolerable limits based on the current a body can survive for the fault duration. For a 50 kg person, the tolerable touch voltage is

where \(\rho_{s}\) is the resistivity of the surface layer, \(C_{s}\) a derating factor for its thickness and \(t_{s}\) the fault duration in seconds.

Bonding for IT and telecommunications

Electronic equipment is sensitive to voltage differences between interconnected devices. A telecommunications bonding network, designed to ANSI/TIA-607 and related standards, ties every rack, cable tray, ladder rack, raised-floor pedestal and equipment ground terminal into a low-impedance, equipotential system:

• A primary bonding busbar (PBB) in the entrance room, bonded to the building's electrical service grounding.
• Secondary bonding busbars (SBB) in each telecommunications space and data hall, connected to the PBB by a telecommunications bonding backbone.
• Within data halls, a mesh common bonding network: a grid of copper conductors, often under the raised floor or overhead, to which each rack is bonded with a short conductor.

The aim is not a "clean" isolated ground, an idea that has caused many problems when it led to equipment being disconnected from the safety bonding system, but a single, interconnected, low-impedance system in which everything rises and falls together during a surge or fault.

Lightning protection

A lightning stroke carries a peak current of typically 30 kA, and sometimes over 200 kA, rising in a few microseconds. A lightning protection system intercepts strokes with air terminals on the roof, carries the current down the building through down conductors (or the structural steel) and disperses it into the earth through the grounding system, without letting it flash over to internal wiring or damage the structure. Design follows NFPA 780 in North America and IEC 62305 internationally, starting with a risk assessment that weighs lightning frequency, building size, contents and the consequences of damage; for data centres the risk assessment almost always justifies protection.

Surge protection

Surges reach equipment from lightning, utility switching and switching within the facility. Surge protective devices (SPDs) divert surge current to the bonding system and limit the voltage reaching equipment. They are applied in coordinated layers:

• At the service entrance (Type 1 or Type 2 SPDs), handling the largest currents from outside.
• At distribution boards and switchboards (Type 2), limiting surges generated within the facility and residual surges from outside.
• At or in the equipment (Type 3 and built-in protection).

SPDs are rated by their maximum continuous operating voltage, their nominal discharge current and their voltage protection rating (let-through voltage). Their effectiveness depends heavily on installation: long or looped connecting leads add inductance, and since \(V = L\,di/dt\), a surge rising at 10 kA/µs through 1 µH of lead inductance (roughly a metre of wire) adds 10 kV. SPD leads must therefore be as short and straight as possible.

In Canada, SPDs must be certified to CSA C22.2 No. 269 series standards, and the Canadian Electrical Code sets requirements for their installation.