A 32 A type B MCB on a 230 V supply needs a Zs of no more than 1.37 Ω at conductor operating temperature to disconnect in 0.4 s, and GN3's on-site rule is to compare the cold measured value against 80% of that, so 1.10 Ω. When a socket circuit reads 1.4 Ω the first question is not "what code is this" but "is the reading real". Most high Zs readings on domestic work come from one of four places: a high Ze at the origin, a long or undersized circuit, a poor connection somewhere in the earth path, or the tester itself. This guide works through them in the order that finds the cause fastest.
- Compare measured Zs against 80% of the Table 41 maximum (or use the tester's temperature-corrected figure), not the raw table value.
- Zs = Ze + (R1+R2). Measure both parts separately and the arithmetic tells you where the problem is.
- A Ze above 0.35 Ω on TN-C-S or 0.8 Ω on TN-S is a DNO matter, not a wiring fault.
- Loose CPC terminations, unsleeved earths under the wrong screw, and long runs in 1.0 mm² are the usual circuit-side causes.
- A 30 mA RCD can provide fault protection where the MCB cannot, but the reason must be recorded.
Is the reading actually high?
Before fault finding, check the arithmetic. BS 7671 Tables 41.2 to 41.4 give maximum Zs values for each device type and rating at 230 V, and they assume the conductors are at their operating temperature of 70 °C. A cold installation measures lower than that, so GN3 applies a correction: the measured value should not exceed 0.8 × the tabulated maximum. Most modern MFTs can display the corrected limit for the selected device.
| Device (230 V, 0.4 s) | Table 41 max Zs | 80% on-site limit |
|---|---|---|
| 6 A type B | 7.28 Ω | 5.82 Ω |
| 16 A type B | 2.73 Ω | 2.18 Ω |
| 20 A type B | 2.19 Ω | 1.75 Ω |
| 32 A type B | 1.37 Ω | 1.10 Ω |
| 40 A type B | 1.09 Ω | 0.87 Ω |
| 32 A type C | 0.68 Ω | 0.55 Ω |
| 32 A RCBO type B | 1.37 Ω | 1.10 Ω |
Full tables, including 5 s values for distribution circuits, are at Maximum Zs Values BS 7671: Tables, 80% Rule and Ze. A reading of 1.20 Ω on a 32 A type B is over the 80% limit and needs investigating, even though it is under the raw 1.37 Ω.
Then check the tester. A no-trip loop test on a circuit without an RCD is less accurate than the standard high-current test, and can read high at low impedances. Nulled leads that have not been re-nulled after a lead change add a few tenths of an ohm. A test taken at the socket with a plug-in lead includes the lead's own resistance if it has not been zeroed. And a socket with a worn earth contact will read high while the ring behind it is fine, so test at a second point on the same circuit before believing a single reading.
How do you separate Ze from R1+R2?
Zs at any point is the external loop impedance Ze plus the circuit's line and CPC resistance R1+R2. Measure both:
- Ze at the origin with the main earthing conductor disconnected from the MET (installation isolated, no parallel paths through bonding).
- R1+R2 during the dead tests, with a link between line and CPC at the board and a low-ohms reading at the far end.
Add them. If Ze + R1+R2 is close to the measured Zs, the reading is real and the next question is which of the two parts is too big. If the measured Zs is much higher than the sum, something changed between the dead test and the live test: usually a CPC that was making contact under the low-ohms test's small current but has a poor joint that the loop tester's higher current exposes. If the measured Zs is much lower than the sum, parallel paths through bonding, metal conduit or SWA armour are carrying part of the fault current, which is fine but worth noting.
Is the problem the supply?
Typical maximum values quoted by distribution network operators for Ze are 0.35 Ω for TN-C-S (PME), 0.8 Ω for TN-S, and 21 Ω for TT (where the electrode resistance dominates). If Ze at the origin is above the TN figure for the earthing type, every circuit in the house will struggle, and no amount of rewiring will fix it. On a TN-S supply a Ze creeping towards 1 Ω often means the lead-sheath earth on an old cable is corroding at the cut-out, which is a DNO fault to report. On TN-C-S a high Ze may mean a poor PEN connection, which is also a DNO matter and potentially dangerous.
Record Ze on the certificate with the method (main earth disconnected) and, if it exceeds the DNO's declared maximum, note it as an observation and report it to the DNO. What Ze is and how to measure it is covered at What is Ze? External earth fault loop impedance.
On TT systems the arithmetic is different: Zs is dominated by the electrode resistance and the MCB will never meet its disconnection time, so a 30 mA RCD provides fault protection and the requirement is RA × IΔn ≤ 50 V. See Earth Electrode Testing for TT Systems: Methods and Limits.
Is the problem the circuit length or cable size?
If Ze is normal, the excess is in R1+R2. Approximate resistances per metre at 20 °C, from the IET On-Site Guide:
| Cable (line / CPC) | R1+R2 per metre | Length for 1.0 Ω R1+R2 |
|---|---|---|
| 1.0 / 1.0 mm² | 36.2 mΩ | 28 m |
| 1.5 / 1.0 mm² | 30.2 mΩ | 33 m |
| 2.5 / 1.5 mm² | 19.5 mΩ | 51 m |
| 4.0 / 1.5 mm² | 16.7 mΩ | 60 m |
| 6.0 / 2.5 mm² | 10.5 mΩ | 95 m |
A 2.5/1.5 mm² radial for an outbuilding socket run 50 m from a TN-C-S board with Ze at 0.30 Ω will land at about 1.28 Ω, over the 80% limit for a 32 A type B. Nothing is faulty; the circuit was simply never going to meet the requirement with that breaker. The fixes are a smaller breaker (a 20 A type B allows 1.75 Ω on site), a larger CPC, or RCD fault protection with the reason recorded. Cable sizing and volt drop go together here, see Cable Sizing and Volt Drop Basics: Ib, In, Iz and Table 4D5.
Lighting circuits in 1.0 mm² are the classic case. A long loop-in run around a large house with switch drops adds up quickly, and a 6 A type B has plenty of headroom (5.82 Ω on site), but a 10 A or a type C on the same circuit may not.
Is the problem a bad joint?
When Ze is fine and the circuit length does not explain the reading, there is a high-resistance connection in the earth path. Where to look:
- CPC terminations at the board. Several earths under one screw in the earth bar, or an earth under the clamp plate rather than in the tunnel.
- Ring final earth continuity. A break or poor joint in the CPC of a ring final will not show up on Zs at most sockets because the other leg carries it, but it shows up in the r2 figure during ring testing. If r2 is much higher than r1 and rn, or open, fix that first. Ring Final Circuit Testing: r1, rn, r2 and R1+R2 Explained covers the method.
- Junction boxes and spurs. Earths that were sleeved and pushed into the box but never terminated. Common behind kitchen units.
- Metal back boxes. A CPC terminated to the box lug only, relying on the accessory's fixing screws for continuity.
- Old MICC or conduit installations relying on the sheath or conduit as CPC, with corroded couplers or a broken run.
- Heat-damaged terminals. A neutral or line connection that has been hot also tends to have a heat-affected earth next to it.
The method is the same as for insulation resistance faults: split the circuit and test R1+R2 to each section until the high section is isolated. A wander lead to the MET makes this quick on a lighting circuit.
What do you do when the circuit cannot meet the maximum Zs?
BS 7671 offers three routes, and all of them need recording on the certificate rather than quietly applied:
- Change the protective device. A lower-rated MCB or a type B in place of a type C raises the permitted Zs. Check the cable's current-carrying capacity still suits the load.
- Improve the earth path. Increase the CPC size, add a supplementary CPC, or fix the joint. On an outbuilding supply a larger SWA or a separate earth may be the answer.
- Use a 30 mA RCD or RCBO for fault protection. Regulation 411.4.204 and 411.5.2 permit an RCD to provide fault protection where the disconnection time cannot be met by the overcurrent device. The MCB still provides overload and short-circuit protection for the cable. The RCD test results and the reason must go on the certificate, and the maximum Zs for the RCD (1667 Ω for 30 mA at 50 V) is what is entered on the schedule.
Route 3 is not a way of ignoring a bad joint. If the reading is high because of a poor connection, that connection is a fire risk and an RCD does not address it.
How do you code a high Zs on an EICR?
| Situation | Suggested code |
|---|---|
| Zs above maximum, no RCD, circuit in use | C2 |
| Zs above maximum, 30 mA RCD provides fault protection, RCD tests pass | C3 with observation, or no code where the design is documented |
| Zs very high or open circuit (no effective earth) | C1 if exposed metalwork could become live; otherwise C2 |
| Zs above 80% rule but below table max, no obvious cause | Note in observations; consider C3 |
| Ze above DNO declared maximum | Observation, report to DNO; code by risk |
Write the observation with the numbers: "Circuit 5 (upstairs sockets, 32 A type B): measured Zs 1.52 Ω, maximum 1.10 Ω on site; Ze 0.28 Ω; R1+R2 measured 1.21 Ω against expected 0.55 Ω for length. High resistance in CPC suspected; 30 mA RCD present and tested satisfactorily. C3, recommend investigation." Observation coding is covered in EICR observation codes explained: C1, C2, C3 and FI.
How Certio helps
Certio produces the EICR, EIC, Minor Works certificate and PAT records on the BS 7671 model forms and turns them into branded PDFs. Photograph the consumer unit and the circuit schedule drafts itself; point the phone camera at a Megger MFT-X1 and the reading goes straight into the schedule of test results, or dictate it while your hands stay on the probes. Suggested observation wording with the regulation reference is there for you to check and confirm. The judgement stays yours; the typing goes. Starter is free for 7 days, then £5 a month.