5.3 Factors That Affect Test Results & Tests That May Be Unnecessary at Periodic Inspection
Key Takeaways
- Assessment criterion 4.7 names five factors that affect a result: circuit length, temperature, parallel conductors, loading and preparation (nulling)
- Copper resistance rises about 0.4% per degree C, so a circuit measured cold at 10 C reads roughly 20% lower than the same circuit at its 70 C operating temperature - this is where the 0.8 rule comes from
- Parallel earth paths through conduit, trunking, armour or bonded pipework make R2 and Zs read LOWER than the true protective conductor value and can mask a broken cpc
- Connected load distorts insulation resistance downwards and standing earth leakage shortens the measured RCD operating time; disconnect loads where practicable and record what could not be disconnected
- At periodic inspection some tests may be unnecessary or impracticable - the justification is recorded on the report as a limitation, never left blank
Why This Sits in the 56% Block
Assessment criterion 4.7 asks you to identify factors that affect the result of each test, and criterion 4.8 to interpret the result. You cannot do the second without the first. A measurement that looks like a fail may be a sound circuit measured badly, and — far more dangerous — a measurement that looks like a pass may be a defective circuit measured through a parallel path.
The specification names five factors. Learn them as a set, and learn which direction each one pushes the reading.
| Factor | Affects | Direction of error |
|---|---|---|
| Circuit length | R1+R2, Zs, voltage drop | Longer circuit → higher reading (a genuine effect, not an error) |
| Temperature | R1+R2, Zs | Cold measurement → reads low against a 70 °C tabulated limit |
| Parallel conductors | R2, Zs, continuity | Parallel path → reads low, can mask a broken cpc |
| Loading | Insulation resistance, RCD time, loop impedance | Connected load → IR reads low, RCD time reads short |
| Preparation (nulling) | Continuity, R1+R2 | Un-nulled leads → reads high |
1. Circuit Length
Conductor resistance is proportional to length: R = ρL / A. Doubling the run doubles R1 and R2, and therefore raises Zs and voltage drop. This is not a measurement error — it is why the R1+R2 and Zs values are taken at the furthest point of the circuit, where they are worst-case.
Two practical consequences:
- A long circuit may satisfy overload protection comfortably and still fail on Zs or voltage drop. The remedy is a larger conductor, a shorter route, or a different protective device — not a different test.
- An unexpectedly high reading at an intermediate accessory, on a circuit you believe to be short, suggests a longer cable route than the drawings show, or a joint you have not found.
2. Temperature
Copper has a temperature coefficient of resistance of about 0.004 per °C, so resistance changes by roughly 0.4% for every degree.
The tabulated maximum Zs values in Tables 41.2, 41.3 and 41.4 are calculated at the conductor’s maximum operating temperature — 70 °C for thermoplastic (PVC) insulation. But you test on site at ambient, typically around 10 °C, on a circuit that has carried no load.
Over a 60 °C span the resistance difference is about 1 + (0.004 × 60) ≈ 1.24, which is why a cold measurement can look comfortable and the same circuit fail once it is warm. Turned around, the on-site limit is roughly 1 ÷ 1.24 ≈ 0.8 of the tabulated figure — the origin of the 0.8 rule covered in section 8.2.
The corollary matters too: if you measure a circuit that has been carrying load all day, it is already warm and the 0.8 rule is over-conservative. In that situation the full Appendix 3 adjustment gives a fairer answer.
3. Parallel Conductors and Parallel Paths
This is the factor that produces genuinely dangerous false passes, and it is heavily examined.
A protective conductor is rarely the only metallic route back to the main earthing terminal. Any of these can sit in parallel with it:
- Metallic conduit, trunking or ducting enclosing the circuit
- The steel wire armour of an SWA cable
- Bonded extraneous-conductive-parts — water pipes, gas pipes, structural steel — that also run past the accessory
- The cpc of another circuit connected to the same metalwork
Because parallel resistances combine to less than the smallest of them, the measured R2 or Zs comes out lower than the true value of the protective conductor alone. A circuit protective conductor that is broken part-way along the run can still give a healthy-looking Zs, because the current is finding its way home through the conduit.
How to catch it:
- Compare the dead R1+R2 with the live Zs. If Zs is materially lower than Ze + (R1+R2), something is in parallel.
- Compare R2 against the expected value for the conductor size and route length. A 1.5 mm² cpc on a 20 m run should read around 0.24 Ω; a reading of 0.05 Ω is not a very good cpc, it is a parallel path.
- Where practicable, disconnect the parallel path at the main earthing terminal for the duration of the continuity test — with the installation isolated and the disconnection secured and restored immediately afterwards.
4. Loading and Connected Equipment
Equipment left connected changes what your instrument sees.
- Insulation resistance — every connected load provides a parallel leakage path, so the measured value falls. Ten circuits each at 20 MΩ measure as 2 MΩ together. This is why a whole-installation figure below 1 MΩ is investigated by testing in sections, and why electronic equipment, dimmers, LED drivers and surge protective devices are disconnected or linked out. Applying 500 V DC to them can also destroy them.
- RCD operation — standing earth leakage from connected equipment adds to your injected test current, so the device operates sooner than it would on the leakage you applied. Disconnect loads where practicable; where you cannot, note it.
- Loop impedance — a heavily loaded circuit gives a slightly unstable reading, and testing under load risks nuisance tripping.
Whatever cannot be disconnected is a limitation and is recorded as such.
5. Preparation — Nulling the Leads
A pair of standard test leads has roughly 0.05 to 0.20 Ω of resistance, and a 10 m wander lead can add 0.3 to 0.5 Ω. Against an R1+R2 of 0.30 Ω, that is not a rounding error — it is the whole measurement.
Null the leads (or measure and subtract them) before every set of continuity measurements, and re-null when you change leads. Failing to null pushes every reading high, producing false fails, unnecessary remedial work and an inaccurate schedule of test results. It is one of the most common practical-assessment failures.
A lighting circuit wired in steel conduit gives a measured Zs of 0.28 Ω. Ze is 0.35 Ω. What does this tell you?
A whole-installation insulation resistance test on a domestic property gives 0.7 MΩ. Every circuit measured individually gives more than 20 MΩ. What is the explanation?
Tests That May Not Be Necessary at Periodic Inspection (AC4.4)
Assessment criterion 4.4 asks you to explain why certain tests may not be necessary at periodic inspection and test. The reasoning is that periodic inspection assesses condition, on a sample, on an installation that is in use — so a test that would tell you nothing new, or that would cause more risk or disruption than the information is worth, may legitimately be omitted.
| Test | Why it may be reduced or omitted |
|---|---|
| Ring final circuit continuity | The full three-step test proves the ring was built correctly. Where the previous certification shows a satisfactory ring and there is no evidence of alteration, damage or additional spurs, a full re-test of every ring may not add information — though the ring is still verified where records are absent or alterations are suspected |
| Insulation resistance | Full-value testing requires isolation and the disconnection of equipment. In occupied premises with essential loads, or where the wiring serves electronic equipment that cannot be disconnected, the test may be limited to a reduced number of circuits or carried out at a reduced voltage on a specific circuit, with the reason recorded |
| Polarity | Often confirmed as part of the loop impedance measurement and by visual inspection, rather than by a separate dead test on every accessory |
| Earth electrode resistance | Not applicable at all where the installation is TN-S or TN-C-S |
| Phase sequence | Not applicable to a single-phase installation |
| Voltage drop | Rarely verified at periodic inspection unless there is a specific complaint or a change of load, because it is a design check rather than a condition check |
The rule that makes this safe
Omitting a test is a decision the inspector must justify and record, not a gap left in the schedule. Anything not tested is either N/A (genuinely does not apply) or LIM (a limitation), and the limitation must appear on the front of the report where the reader will see it.
The failure mode the examiner is testing for is the candidate who treats "not necessary" as "optional". It is not. A test is omitted because it is not applicable, because the information already exists and nothing has changed, or because carrying it out would create more risk or disruption than the finding is worth — and in that last case the client has agreed to the limitation in advance.
During a periodic inspection of an occupied office, a server room cannot be isolated, so its final circuits cannot be insulation-resistance tested. What is the correct treatment on the EICR?