An earthing arrangement settles where the neutral is bonded and which conductor carries fault current back to the source. It does not settle whether that path is good enough. That is the electrode’s job, and on a generator site the electrode is often asked to work in dry sand, saline sabkha or imported fill at a resistance the protection settings silently assume. An electrode that is never measured leaves the assumption untested; one measured in the cool month leaves it tested under the wrong conditions.
The electrode closes the fault loop that protection sees
A fault loop is a circuit and has to be read as one: the source winding, the phase conductor to the fault, the protective conductor or the mass of earth back to the source, and the source’s neutral point. On a TN-S or TN-C-S installation the return leg is metallic, so the electrode does not carry fault current and the disconnection time depends on the impedance of copper, not on soil. The electrode still matters, because it holds exposed metalwork and the main earthing terminal at the same potential and gives surge current somewhere to go.
On a TT installation the picture changes completely. There is no metallic return, and the fault current has to travel through the installation electrode, the mass of earth and the source electrode before it reaches the neutral. Electrode resistance is then a term in the loop equation, not a background detail. A value that looked acceptable on a drawing can leave a residual-current device unable to operate inside the required time, and no amount of cable oversizing will correct it.
Two jobs are therefore easy to confuse. Bonding makes metalwork equipotential; the electrode connects the system to earth. A site can have excellent bonding and a poor electrode, or a good electrode and a missing bond on a fuel tank. Protection responds to the loop, and the loop contains both.
| Function | What it depends on | Evidence at handover |
|---|---|---|
| Fault-current return on a TT system | Electrode resistance in series with the fault loop | Measured resistance at the electrode, with method and date |
| Touch-voltage control | Bonding continuity and electrode resistance together | Continuity results for every bonded item |
| Equipotential bonding of extraneous metal | Main earthing terminal and the bonding conductors | As-built drawing with each connection marked |
| Surge and lightning current diversion | A short, direct path to earth and coordinated surge devices | Electrode inspection and device status |
| Reference for insulation monitoring on an IT system | Deliberate separation from earth, continuously monitored | Insulation-monitoring record and alarm test |
Soil resistivity is a measurement, not a table lookup
Electrode resistance is governed by the soil around it, and soil is the least stable material on the site. The usual field method is the four-electrode array: four probes driven in a straight line at equal spacing, current injected through the outer pair and voltage measured across the inner pair. Apparent resistivity is 2πaR, where a is the spacing and R the measured resistance. Changing the spacing changes the depth the measurement sees, so one reading at one spacing says very little. Take readings at several spacings, in two directions at right angles, and record the date, the weather and the moisture state.
In this climate the dry season is the design case, not the commissioning week. A 40 °C ambient dries the surface layer, cracks clay and lowers the water table; a reading taken after rain can be a fraction of the value the site will show in August. How large that change is depends on the soil profile and is established by repeat measurement, not by a rule of thumb. The same applies to salinity: sabkha and coastal fill can be aggressive to the electrode and conductive to the soil at the same time, which is a corrosion problem dressed up as a good reading.
Use the survey to choose a depth and a geometry, then check the choice against a layered soil model rather than a single number. If the shallow layer is dry and the deeper layer is moist, a deeper electrode does more than a longer horizontal run. If the whole profile is dry, the answer may be a larger array, a different location, or an earthing arrangement that does not depend on a low-resistance electrode for disconnection.
The survey also has a second use: it tells the installer what the ground will accept. Rocky fill, buried rubble and old foundations change the driving route, and a rod that stops short of its design depth has a resistance that was never calculated. Recording refusal depth is as important as recording resistivity.
| Ground condition | Behaviour | Design response | Verification |
|---|---|---|---|
| Dry sand and imported fill | High resistivity, poor moisture retention, unstable contact | Deep-driven electrodes and a larger array; moisture-retaining backfill only where the supplier supports it | Dry-season measurement |
| Sabkha and saline coastal soil | Low resistivity but aggressive chemistry | Material selected from a soil analysis, with the connection protected | Electrode condition, not only resistance |
| Clay | Moderate resistivity, shrinks and cracks when dry | Depth below the seasonal shrinkage zone | Repeat measurement in the dry season |
| Gravel and rubble | Poor contact, difficult driving | Plate or ring electrode, or a different location | Contact resistance and continuity |
| Rock | Very high resistivity; drilling may be needed | Borehole electrode, or a design that does not rely on local soil | Survey before construction |
Electrode geometry: what each type buys you
A vertical rod reaches deeper, wetter layers with a small footprint, and a driven rod is usually the first choice where there is room to drive it. Its resistance falls roughly with the logarithm of length, so doubling the length does not halve the resistance; the gain has to be calculated. A rod in dry sand may need to be deep, and the driving depth is limited by what the ground will accept without damaging the rod or its coupling.
Horizontal ring and radial electrodes suit sites where driving is impossible and a large area is available. They are shallow, so they are the most exposed to drying and to later excavation. Foundation and structural electrodes use reinforcement already in the ground and can be very effective, but only if the connection is brought out during construction and protected afterwards; a foundation electrode discovered after the slab is poured is not an electrode.
Chemical and backfill electrodes are sold as a remedy for high resistivity. They can help, but they add a maintenance item and a service life that depends on the backfill staying in place and moist. They are not a substitute for a soil survey, and the manufacturer’s data, not a catalogue claim, should set the expected performance. The same caution applies to plate electrodes: surface area helps, but the installation detail decides whether the plate stays in contact with moist soil.
| Type | Best suited to | Limits | Service-life risks |
|---|---|---|---|
| Driven rod | Dry surface layer over moist subsoil; limited space | Depth limited by ground conditions; coupling damage | Mechanical damage and corrosion at the soil line |
| Sectional deep rod | Deep moist layer below a high-resistivity surface | Needs driving equipment and a clear route | Joint integrity and corrosion |
| Plate | Shallow soil with useful contact area; rubble | Vulnerable to excavation; limited depth | Contact loss and corrosion |
| Ring or radial counterpoise | Large site, no driving, lightning dissipation | Shallow, seasonal variation, land-use conflicts | Damage by construction and drying |
| Foundation or structural | New build with reinforcement available | Must be planned before concrete is poured | Continuity to the earthing terminal |
| Backfill or chemical | Difficult soil on a constrained site | Maintenance dependency and environmental limits | Backfill depletion and moisture loss |
Bonding, the single bond rule and where the electrode connects
Every electrode needs a reliable route to the main earthing terminal, and every extraneous conductive part needs a bond to that terminal. On a generator site the list is longer than it looks: the set frame, the alternator frame, the fuel tank and its fill point, the enclosure and doors, cable trays and armouring, water and gas pipes, and the structural steel. Each bond needs a test joint or a disconnection point that can be opened for measurement without leaving equipment unbonded while the site is live.
The neutral bond is a separate decision and follows the earthing arrangement rather than the electrode. One interconnected system has exactly one neutral-to-earth bond. A standby set connected to a TN-C-S supply normally keeps its neutral unbonded so the bond stays at the origin; a set operated as a separately derived system carries its own bond at the alternator and needs a changeover that switches the neutral. The argument is set out in the earthing and neutral arrangement guide; what matters here is that the electrode is connected on the correct side of that decision.
Bond conductor size, test-joint accessibility and the acceptable resistance of each connection are set by the project specification and the installation standard that governs the site. Two practical points survive every specification: clean the metal to bare, sound material before making the connection, and protect the joint afterwards. A bond painted over during finishing is a bond that will be missed at the next inspection.
| Item | Why it matters | What the inspection looks for |
|---|---|---|
| Generator and alternator frames | Fault current and static discharge path | Bond to the main earthing terminal with a visible test joint |
| Neutral-to-earth bond | Defines the system type and the fault path | Located exactly as the single-line diagram shows |
| Fuel tank, fill point and pipework | Static and fault current; ignition risk | Bonded across flexible sections and isolated joints |
| Enclosure, doors and panels | Touch voltage on accessible metal | Braided jumpers across hinges; paint removed at contact faces |
| Cable trays, armouring and glands | Earth continuity of the cable system | Gland and tray bonds checked after every cable change |
| Structural steel and extraneous pipes | Equipotential bonding of the installation | Bond at the point of entry, with a test joint |
Earth-fault loop verification on a generator source
A generator is a weaker source than the mains. Its fault current is limited by the subtransient reactance of the machine, by exciter response and by cable impedance, so the loop impedance at the far end of a circuit is higher than the same circuit would show on the utility supply. A loop test taken at the mains intake and copied into the generator section of the report proves nothing about the generator. The test has to be made with the set running, at the set terminals and at the remote end of the circuit, with the set at the load condition the design assumes.
For a TT site the measurement that matters is the electrode resistance, and the protection calculation has to be checked against the measured value rather than a remembered one. Residual-current devices need their trip time and trip current verified at the point of use, not only at the board, and an earth-fault relay on the generator breaker needs its settings confirmed against the coordination study. Where a neutral earthing resistor is fitted, the resistor limits the earth-fault current and the relay must be sensitive enough to see that limited current; the resistor also needs monitoring, because an open resistor turns a restricted fault current into no fault current at all.
Record the instrument, its calibration status, the ambient and soil condition, the test point and the result. A number without those details cannot be compared with the next survey, and comparison is the whole point of the exercise. The same discipline applies to the mains-side test: it is useful as a reference, but it does not transfer to the generator case.
| Measurement | Where | Condition | Acceptance basis |
|---|---|---|---|
| Earth electrode resistance | At the electrode test joint, by fall-of-potential | Dry season, electrode isolated from the installation | Measured value against the design calculation |
| Protective conductor continuity | Main earthing terminal to each bonded item | Installation isolated as required | Continuity and joint condition recorded |
| Earth-fault loop impedance | Set terminals and the far end of each circuit | Set running at the design load condition | Compared with the protection settings |
| RCD trip time and current | At the device and at the point of use | Load connected, instrument calibrated | Device data and project specification |
| Earth-fault relay operation | Generator breaker | Secondary injection or primary test | Coordination study settings |
| Neutral earthing resistor | Resistor and its monitoring circuit | Set isolated | Continuity and alarm function |
Hot, dry and saline: what service does to the electrode
An electrode is installed once and then left in a changing environment. The water table moves, landscaping and irrigation change the moisture, new construction covers the ground or cuts through a radial conductor, and salt attacks the connection rather than the rod. The reading that mattered at handover is a baseline, not a permanent property of the site.
Two failures are common and both are invisible from the control room. The first is a connection that corrodes under protective tape and opens slowly; the generator still runs, the protection still holds, and the fault path has quietly gone high-resistance. The second is a bond removed during maintenance and never replaced, often on a door, a panel or a pipe that was painted or renewed. Neither shows up in a fuel or load trend, which is why the earthing inspection belongs in the maintenance schedule rather than in the fault-finding procedure.
Lightning adds a separate duty. A strike on the building or on the incoming supply sends current through the earthing system, and the electrode and bonding have to share it without flashing across a joint. Surge protective devices need a short, direct connection to the earthing terminal; a device wired with a long loop is a device that lets the voltage rise it was meant to clamp.
| Check | When | Trigger for action |
|---|---|---|
| Electrode resistance | Dry season and after any ground works | Rise beyond the design margin, or a step change |
| Bond continuity at test joints | Each scheduled maintenance visit | Loose, corroded or missing connection |
| Neutral bond location | After any change to the changeover or supply | Bond in the wrong place, or a second bond |
| RCD and earth-fault relay test | At the interval set by the project | Failure to trip, or a trip time outside the recorded baseline |
| Surge device status | Each inspection | Indicator showing end of life |
| Earthing drawing and test records | At handover and after every modification | Drawing that no longer matches the site |
Records that survive a change of contractor
The earthing system is the part of an installation most likely to be altered without a drawing update. A useful handover pack contains the as-built single-line diagram with the bond points marked, a site plan with every electrode located and described, the soil survey with its spacings and dates, the test sheets with instrument details, and photographs of each test joint before it was closed up. That pack is what lets the next contractor extend the system without cutting a conductor nobody knew was there.
The annual load test is a convenient moment to re-verify the parts that can be checked with the set running. The test is run at 50% to 100% of rating, and the fuel bill is real: a 500 kVA set uses about 101 litres an hour at full load, 80 at 75% and 60 at 50%, against a planning basis of 0.24 to 0.29 litres per delivered kWh. Adding a loop-impedance check and a bonding inspection to that day costs little extra time and leaves the site with a dated record.
None of this replaces the design calculation. It confirms that the installation still matches the calculation, which is the only claim an earthing system can honestly make after a few years in service.
Does a standby generator need its own earth electrode?
It depends on the earthing arrangement, not on the generator. A set that operates as part of a TN-S installation with a metallic fault path may need no separate electrode for disconnection, but it still needs bonding to the installation earth. A separately derived system or a TT site needs its own electrode and its own protection calculation. The drawing should state which case applies.
What earth resistance should an electrode achieve?
There is no universal figure. On a TT system the resistance must be low enough, in combination with the protective device’s rated residual current, to clear the fault inside the required time and to keep touch voltage within the limit the project adopts. That calculation, the soil survey and the project specification set the target; a number copied from another site does not.
Why does the electrode reading rise in summer?
Surface soil dries, clay cracks and the water table falls, so the contact between electrode and soil worsens. The change can be large and is site-specific. Design for the dry-season value and repeat the measurement at the same test points, in the same season, so the trend is comparable. A single commissioning reading taken after rain proves very little.
Can a generator use the building’s existing earthing system?
Often yes, provided the set is part of the same earthed system and the single neutral-to-earth bond rule is respected. What cannot be assumed is the loop impedance: the generator is a weaker source than the utility, so the loop has to be verified with the set running. The building’s original test certificate is a reference, not evidence for the generator case.
How often should earth resistance and bonding be tested?
At commissioning, after any change to the earthing, the changeover or the ground around the electrode, and then at the interval set by the project specification. Linking the check to the annual load test at 50% to 100% of rating keeps it from being forgotten. Records matter as much as the test: the same points, the same method and the same season make the trend readable.
Do chemical or backfill electrodes solve a high-resistivity site?
They can improve contact in difficult soil, but they are not a substitute for a soil survey or for a correctly sized array. They add a maintenance dependency, and their performance depends on the backfill remaining in place and moist. Where the whole soil profile is dry, a deeper electrode or a different earthing arrangement is usually the more durable answer.
Send the earthing arrangement, the soil survey, the single-line diagram and the protection settings. The reply sets out an electrode scheme, the bonding points to verify and the records needed at handover.
Written and maintained by the KVA Grid sales engineering desk. The planning figures are published elsewhere on this site: the 0.24-0.29 litres per delivered kWh basis and the 500 kVA figures of 101, 80 and 60 litres an hour at full, 75% and 50% load from fuel consumption; the annual load test at 50% to 100% and the 10% to 20% light-load range from maintenance scheduling; the 40 °C reference ambient and derating above it from hot and humid climates; the single neutral-to-earth bond and the solid or switched neutral decision from generator earthing and the neutral; the handover record from standby generator commissioning; and transfer timing from generator and UPS interaction. Electrode resistance targets, touch-voltage limits, bonding conductor sizes, test intervals and acceptance tolerances are determined by the protective device data, the equipment manufacturer’s data and the project specification. Nothing on this page is a quotation for a particular project. Reviewed 5 November 2026.
