C&G 2391-52 common mistakes
C&G 2391-52 · Understand the traps and how to avoid them.
Performs the dead test but does not re-prove the voltage indicator on a known source after testing the conductors, or locks off the isolator without retaining the key/warning notice.
Candidates rehearse the 'prove–test–prove' sequence as 'prove once, then test', and treat lock-off as a courtesy rather than part of safe isolation. Under exam pressure the final prove step is dropped because the circuit has already 'read dead'.
What to do instead
Practise the fixed GS38 sequence until it is automatic: identify the correct isolation point and all sources (including secondary supplies, PV, generators), isolate, lock off and retain the key, post a warning notice, prove the indicator on a known live source or proving unit, test for dead between all conductor combinations (L-N, L-E, N-E and phase-phase), then re-prove the indicator. Never accept a single dead reading as proof.
Measures R1+R2 and r1/r2 without nulling the test leads, or with the main protective bonding and parallel earth paths still connected.
Lead resistance of 0.2–0.5 Ω looks small next to a loop impedance but is significant against a 1.0 Ω circuit limit, and parallel paths through bonding, pipework and metal containment pull the reading below the true value, hiding a high-resistance circuit protective conductor.
What to do instead
Null/zero the leads as a pair, including the wander lead and any extension, before every set of readings and again after a battery change or lead swap. If nulling is not possible, record the lead resistance and subtract it. Disconnect main protective bonding where the measurement demands it, or state the parallel-path limitation in the report.
Carries out insulation resistance testing with equipment still connected (dimmers, neons, SPDs, IT and control gear, capacitors) or only performs a single L-N test and records the result as the insulation resistance.
Fitting the whole test into limited time encourages testing with the board still 'as found', and a single good L-N reading feels like the box has been ticked. Candidates also forget that a low reading is often the connected load rather than a fault.
What to do instead
Isolate and disconnect sensitive equipment (or use the permitted 250 V DC test where 500 V would cause damage), link L and N for the L-PE and N-PE tests, and record which tests were made. Test L-N, L-E and N-E, treat the lowest reading as the result, and reinstate and functionally check all disconnected equipment afterwards.
Compares a measured Zs directly with the tabulated maximum in BS 7671 and passes it without applying the 0.8 factor, or compares against the table for the wrong protective device rating or curve.
The tabulated values assume conductors at their maximum operating temperature, so a cold measured value must be corrected; candidates either do not know the rule of thumb from the guidance notes or read the wrong row of the table under time pressure.
What to do instead
Identify the device type, rating and curve from the actual device, look up the correct maximum Zs for the required disconnection time, and confirm that the measured value does not exceed 0.8 × that tabulated maximum. If it does, investigate (loose connection, undersized cpc, long run) rather than recording a pass.
Tests the RCD only at 1 × IΔn and records a single trip time, skipping the 0.5 × no-trip test, the 5 × IΔn test and testing at both half-cycles (0° and 180°).
Many candidates think 'it tripped, so it's fine', and the plug-in tester defaults to a single 1 × test with one polarity. The requirement for the no-trip test and both half-cycles is easily forgotten.
What to do instead
For each RCD record: no trip at 0.5 × IΔn; trip at 1 × IΔn at both 0° and 180° (recording the longer time); and trip at 5 × IΔn (for general, non-delay types). Apply the correct time limits for the device type (including Type S), test at the device and, for 30 mA socket protection, at the furthest point, and record actual measured times and the multiplier/phase used.
Completes the schedule of inspections from memory or from the previous certificate rather than physically checking each item - for example, not verifying that main protective bonding is connected within 600 mm of the meter on the consumer side of the incoming service, or not checking enclosure IP ratings and conductor identification.
The visual inspection is perceived as the easy part and gets compressed into the last ten minutes of the visit; on familiar installations the candidate assumes the condition matches what the paperwork says.
What to do instead
Walk the installation with the schedule of inspections open and physically confirm each item, including bonding connection points and clamps, labels and warning notices, IP ratings, cable routing and support, signs of thermal damage, and correct conductor identification. Note anything not inspected as a limitation.
Applies a fixed 10% sample of accessories regardless of the condition found, and does not increase the sample size when a defect is discovered.
Candidates look for a single numeric rule for sampling, but the guidance is risk-based; when a defect appears mid-inspection the candidate finishes the original sample rather than expanding it.
What to do instead
Plan the sample on age, use, environment and previous results, explain the sampling approach to the client in advance, and escalate - widening the sample, or testing the whole circuit - whenever a defect is found until the extent of the problem is established. Record what was and was not sampled.
Records a measured value that fails the required limit as a pass - for example a 0.6 MΩ insulation resistance or an over-limit Zs - and fails to assign an observation code, or codes a departure from a previous edition of BS 7671 (such as absence of 30 mA RCD protection not required when installed) as C2 rather than C3.
Candidates fear that a failing result means the whole report is 'unsatisfactory' so they rationalise it, and they confuse 'does not comply with the current edition' with 'potentially dangerous'.
What to do instead
Compare every result against its limit before it reaches the certificate and investigate anything at or beyond it. Remember that BS 7671 is not retrospective, so a departure from a previous edition that is not unsafe attracts C3 at most, whereas genuine danger is C1 and potential danger is C2. If the condition cannot be established without further work, use FI.
Issues a certificate or EICR that omits the schedule of circuit details and test results, the schedule of inspections, the extent and limitations, the next inspection date, the departures or the signature, or uses the wrong model form (an EIC where an EICR is required).
Candidates treat the certificate as the paperwork that follows testing rather than as part of the deliverable, and rush the document at the end of a long day; the difference between the model forms is not always well understood.
What to do instead
Before leaving site, check that the chosen model form matches the work (EIC for new work, MEIWC for minor work, EICR for periodic inspection) and that every mandatory section is complete: extent and limitations, schedule of inspections, schedule of circuit details, all test results, departures, next inspection date and a signed declaration. Use the current IET model forms rather than an old template.
States that the existing installation must be brought up to the current edition of BS 7671, or quotes a fixed inspection interval from the Regulations as if it were mandatory, instead of explaining that the interval is set by the client's risk assessment, insurance or licensing requirements.
Candidates conflate 'must comply with the current edition' for new work with retrospection, and look for a simple table of intervals that the Regulations do not provide.
What to do instead
Be clear that BS 7671 applies to the design and construction of new work and to alterations, and that earlier compliant installations are not required to be upgraded; reclassification of an existing condition is about safety, not age. Advise the client in writing that the interval between inspections is a matter for them on the basis of use, environment and insurance, and record the recommended date on the report.
Arrives on site without the previous certificates, schedules, extent agreement or a test plan, and then has to negotiate access and limits mid-inspection.
Candidates focus on the test instruments and forget that a periodic inspection is a planned exercise; the client is often unavailable to discuss limitations once testing has started.
What to do instead
Before the visit, obtain previous certification and schedules, agree in writing the extent of the inspection, the limitations (for example areas that cannot be isolated or accessed) and whether the supply can be interrupted, then build the test plan and sampling strategy around that agreement.