Coding is where two competent inspectors most often disagree, because the code follows the circumstances rather than the defect. Each of these gives you the circumstances. Decide on a code before you open the answer.
Several of these deliberately hang on one detail: a document that does or does not exist, a pipe made of the wrong material, where a board happens to be sitting. That is not a trick. It is the actual reason coding is hard, and it is why a code copied from a previous report is worth so little.
9. No main bonding to the water pipework, and a plastic incoming main
An EICR on a 1980s house on a TN-C-S supply. There is no main protective bonding conductor on the copper water pipework. In the intake cupboard the incoming service is 25 mm blue plastic pipe, changing to copper at the stopcock. Measured between the internal copper pipework and the main earthing terminal you get 0.45 megohms.
What code goes on the report for the missing bonding?
Show the answer and the reasoning
Answer: No code, it complies
Main protective bonding is required for extraneous-conductive-parts: metalwork that is not part of the electrical installation and that is liable to bring an outside potential, usually earth potential, into the building.
The incoming service here is plastic, so the copper inside the house is not connected to anything outside it. The 0.45 megohm reading confirms that, sitting far above the 22 kilohm figure used as the threshold for deciding whether a part is extraneous.
So there is nothing to bond. Fitting a bonding conductor to pipework that is not extraneous is not a neutral act either: on a PME supply it gives diverted neutral current a route onto the plumbing if the network PEN conductor is ever lost.
The judgement worth taking away is that bonding is decided by measurement and by what the incoming service is made of, not by whether the pipe in front of you is metal.
Why the others are wrong
- C2 assumes the pipework can bring in an outside potential, and it cannot.
- C3 implies something falls short of the standard, and nothing does.
- FI is for when you cannot determine the position, and here you measured it.
Where this is written down
BS 7671 Regulation 411.3.1.2 requires main protective bonding for incoming metallic services that are extraneous-conductive-parts. The 22 kilohm figure used to decide whether a part is extraneous comes from IET Guidance Note 8, not from BS 7671 itself.
Settled rule
Go deeper: The EICR coding framework, for how C1, C2, C3 and FI are decided, and where the extraneous-part threshold comes from.
10. No RCD on workshop sockets, and no risk assessment either
A periodic inspection of a light industrial workshop. Twelve 13 A double sockets sit on two 32 A radial circuits installed in 2012, none of them with 30 mA RCD protection. Asked for the documented risk assessment that would allow the omission, the dutyholder confirms none has ever been done.
What code applies to the sockets?
Show the answer and the reasoning
Answer: C2, potentially dangerous
In premises other than dwellings, sockets rated up to 32 A need additional protection by a 30 mA RCD unless a documented risk assessment determines it is not necessary. That exception is the whole of the argument here.
There is no risk assessment, so the exception does not apply, and the installation does not meet the requirement. In a workshop with portable tools and extension leads in daily use, that is a foreseeable shock risk rather than a paperwork failing, which puts it at C2.
Note what would change the answer. A dated risk assessment by a competent person, setting out why RCD protection is not needed and what is in place instead, moves this to no code or C3. The presence or absence of that document is doing the work, not the presence or absence of the RCD.
Why the others are wrong
- C3 would be reasonable only if a documented assessment existed and stood up.
- C1 needs danger present now, and no live part is accessible.
- FI is for an unknown, and the dutyholder has told you the answer.
Where this is written down
BS 7671 Regulation 411.3.3 requires 30 mA RCD additional protection for AC socket-outlets rated up to 32 A, with an exception in premises other than dwellings where a documented risk assessment determines it is not necessary. The coding convention follows Electrical Safety First Best Practice Guide 4.
Engineering judgement
Go deeper: The EICR defect finder, for how socket rating, premises type and the risk assessment exception combine into a code.
11. An open knockout in the top of a metal board
A commercial sub-distribution board in a store cupboard, mounted at 1.4 metres. On the top surface of the steel enclosure there is an open 25 mm knockout where a cable was removed. Looking down into it you can see insulated tails and shrouded main switch terminals about 75 mm below. Nothing bare is reachable with a test finger.
What code applies to the opening?
Show the answer and the reasoning
Answer: C2, potentially dangerous
Accessible top surfaces of an enclosure have to reach IP4X, which keeps out solid objects of 1 mm and above. That is a higher bar than the IP2X applied to the sides, and it exists for one obvious reason: things fall downwards.
A 25 mm hole in the top will take a washer, a screw, a strip of swarf or a small tool, and drop it straight onto the busbars. That is a short circuit and an arc inside a live board, not a shock risk to a finger.
So the danger is real but not present at the moment you look at it, which is the definition of C2. The distinction between C1 and C2 here is not about severity, it is about whether the danger is live now.
Why the others are wrong
- C1 needs something dangerous now, and nothing bare is touchable.
- C3 treats a fire and arcing route as a nicety.
- No code ignores the higher IP requirement that applies to top surfaces specifically.
Where this is written down
BS 7671 Regulation 416.2.2 requires readily accessible top surfaces of enclosures and barriers to provide at least IP4X or IPXXD. Regulation 416.2.1 covers the general IP2X or IPXXB requirement.
Settled rule
Go deeper: Arc flash risk and PPE, for what happens inside a board when something conductive lands across live parts.
12. Zs is above the table value, but the circuit is on an RCBO
A commercial radial socket circuit on a TN-S system, protected by a 20 A Type B RCBO with a 30 mA residual rating. Measured Zs at the furthest outlet is 2.52 ohms against a tabulated maximum of 2.19 ohms for that device. The RCBO trips in 28 ms at its rated residual current. Prospective line to neutral fault current at the same point is 210 A, comfortably above the 100 A the magnetic element needs.
How should the Zs reading be treated on the report?
Show the answer and the reasoning
Answer: C3, improvement recommended
The table value is the figure that makes the overcurrent device clear an earth fault in time on its own. Exceed it and that particular mechanism no longer works.
It is not the only mechanism. Where an RCD provides fault protection on a TN system, the loop impedance limit becomes 50 divided by the residual operating current, which for a 30 mA device is over 1600 ohms. At 2.52 ohms you are nowhere near it, and the device is proven to operate in 28 ms.
Short circuit protection is separately satisfied, because 210 A operates the magnetic element of a 20 A Type B device.
So nobody is in danger, but the circuit is not as designed and it now depends on a single electronic device rather than on the cable. That is a C3, and it is worth writing the reason on the report rather than the code alone.
Why the others are wrong
- C2 would say disconnection is not assured, and it demonstrably is.
- C1 requires danger present.
- FI is for an unknown, and everything here has been measured.
Where this is written down
BS 7671 Regulations 411.4.204 and 411.4.5 cover the use of an RCD for fault protection in a TN system where the loop impedance exceeds the Table 41.3 value. The coding convention follows Electrical Safety First Best Practice Guide 4.
Engineering judgement
Go deeper: Maximum Zs and temperature correction, for where the tabulated values come from, and what changes when an RCD is doing the disconnecting.
13. A plastic consumer unit under the stairs
An EICR on a two-storey house wired in 2011. The consumer unit is a plastic one, in the under-stairs cupboard, directly beneath the timber staircase that is the only way out of the upper floor. It is intact: no discolouration, no heat damage, no loose terminations, blanks all in place, grommets all present.
What code applies to the enclosure?
Show the answer and the reasoning
Answer: C3, improvement recommended
The requirement for a non-combustible enclosure on a domestic consumer unit came in after this board was installed, and BS 7671 is not applied retrospectively. An installation that met the standard of its day is not automatically unsatisfactory.
Nor is an intact plastic board dangerous in itself. What made the requirement necessary was loose connections overheating inside a plastic box, and there is no sign of that here.
What does matter is where it is. A fire in this cupboard is a fire under the only staircase, which is why the improvement is worth recommending rather than merely noting. That is a C3.
The general point: a C3 is the code for something that would be done differently today and would be better if it were, not for something that is failing.
Why the others are wrong
- C2 treats an undamaged enclosure as potentially dangerous, which over-codes it and devalues the report.
- C1 is for a board that is actually overheating or has exposed live parts.
- No code ignores the escape route, which is the one thing that makes this worth raising.
Where this is written down
BS 7671 Regulation 421.1.201 requires the enclosure of a consumer unit in domestic premises to be non-combustible, or to be enclosed in a non-combustible cabinet. It applies to installations from its introduction and is not retrospective. The coding convention follows Electrical Safety First Best Practice Guide 4.
Engineering judgement
Go deeper: The EICR coding framework, for when a change in the standard becomes a C3 and when it becomes nothing at all.
14. A borrowed neutral between the upstairs and downstairs lighting
Periodic inspection of a domestic split-load board. You switch off the upstairs 6 A lighting MCB to do an insulation resistance test, and when you lift the upstairs neutral off the bar you get a spark and the downstairs lights flicker. Tracing it out, a two-way landing switch takes its line from the upstairs circuit and its neutral from the downstairs one.
What code applies?
Show the answer and the reasoning
Answer: C2, potentially dangerous
Isolating a circuit is supposed to make it safe to work on. A borrowed neutral breaks that: with the upstairs circuit off at the board, its neutral is still carrying return current from the downstairs lights.
Anyone who lifts that neutral, having done everything right, is in series with a live load. The spark you got is the warning, and the next person may be holding the conductor rather than a screwdriver.
It also causes trouble with RCDs, because the outgoing and returning currents no longer belong to the same device, but that is a nuisance beside the shock risk.
It is C2 rather than C1 because nothing is dangerous while the installation is left alone. It becomes dangerous the moment somebody does the right thing.
Why the others are wrong
- C3 would treat a shock hazard created by correct isolation practice as an improvement.
- C1 needs danger present in normal use, and in normal use the conductors are enclosed.
- FI is for an unknown, and you have traced it.
Where this is written down
BS 7671 Regulation 132.14 requires every circuit to have effective means of isolation. Regulation 14 of the Electricity at Work Regulations 1989 covers working on or near live conductors. The coding convention follows Electrical Safety First Best Practice Guide 4.
Settled rule
Go deeper: Safe isolation in the real world, for why proving dead at one board is not proof, and how borrowed neutrals defeat isolation.
15. Metal light fittings on a circuit with no earth
An EICR on a house from 1962. The ground floor lighting is original twin cable with no circuit protective conductor. Somebody has fitted decorative metal wall lights in the living room and hall. Each one has an earth terminal on its chassis with nothing connected to it.
What code applies to the fittings?
Show the answer and the reasoning
Answer: C2, potentially dangerous
Exposed-conductive-parts of Class I equipment have to be connected to a protective conductor, because that is what makes the protective device operate when a live conductor touches the metalwork.
With no protective conductor, an internal fault puts 230 V on the outside of the fitting and leaves it there. Nothing disconnects, and the first indication is somebody touching it.
It is C2 rather than C1 because the fault has not happened yet. The metalwork is not live today.
What would change the code: if every fitting and accessory on the circuit were Class II and a warning notice were fitted at the board, the absence of a protective conductor on its own is a C3. It is the Class I metalwork that makes the difference here, not the missing earth by itself.
Why the others are wrong
- C3 applies only where everything on the circuit is genuinely double insulated and labelled as such.
- C1 would need the metalwork to be live now.
- FI is for an unknown, and both the missing conductor and the Class I fittings are confirmed.
Where this is written down
BS 7671 Regulation 411.3.1.1 requires exposed-conductive-parts to be connected to a protective conductor. The coding convention follows Electrical Safety First Best Practice Guide 4.
Settled rule
Go deeper: The defect finder reference, for older circuits without a protective conductor, and where the Class I and Class II line falls.
16. A locked board you agreed to inspect
An EICR on a logistics warehouse. A 63 A three phase sub-main feeds a detached modular office on the same site. The board inside is padlocked by a tenant who is away, so you cannot see it, open it or test anything beyond it. The agreed extent and limitations for the report did not exclude that building.
What goes on the report?
Show the answer and the reasoning
Answer: FI, further investigation required
You have not found a defect. You have failed to establish whether there is one, within a scope you agreed to cover, and those are different findings that need different treatment.
FI is exactly that: the safety of part of the installation could not be determined. It is not a soft option, because in normal practice a report carrying an FI is an unsatisfactory report until the item is resolved.
The tempting answer is the fourth one, recording it as a limitation. That would be legitimate if the building had been excluded when the work was agreed. It was not, so writing it in afterwards converts your problem into the client believing their installation was checked.
The client needs telling in plain words: this board has not been inspected, and its condition is unknown until access is arranged.
Why the others are wrong
- C2 asserts a danger you have not seen.
- C3 recommends an improvement to something you have not looked at.
- A limitation written in after the event misrepresents what was inspected.
Where this is written down
BS 7671 Chapter 65 covers periodic inspection and testing, including recording the agreed extent and limitations and reporting where safety cannot be determined. The convention that an FI produces an unsatisfactory report comes from IET Guidance Note 3 and from scheme provider rules rather than from BS 7671 itself.
Engineering judgement
Go deeper: The standards directory, for what the periodic inspection and reporting requirements actually ask of you.