Continuity testing proves that every circuit protective conductor and bonding conductor runs unbroken from the earth bar to the part it protects, and Regulation 643.2 of BS 7671:2018+A4:2026 puts it first in the test sequence. There are two accepted methods. The R1+R2 method links line and cpc at the board and measures the pair in series at each point, which gives a figure you reuse for Zs. The wander lead (R2) method measures the protective conductor alone from the main earthing terminal. As a yardstick, 2.5 mm² twin and earth with a 1.5 mm² cpc is 19.51 mΩ per metre at 20 °C, so a 25 m radial should read about 0.49 Ω.
- Continuity comes first in the Regulation 643 sequence because a circuit with a broken cpc passes insulation resistance perfectly and is still dangerous.
- Use an instrument that delivers at least 200 mA at 4 V to 24 V, and null the leads before you start. Unnulled leads add 0.1 to 0.3 Ω to every reading.
- The R1+R2 method proves the cpc, confirms polarity and feeds Zs = Ze + (R1+R2) × 1.2. The wander lead method proves the protective conductor only.
- Work out the expected reading from cable size and length before you measure. Too low is as suspicious as too high.
- Record the highest R1+R2 on each circuit, taken at the electrically furthest point.
- Never disconnect the earthing conductor or main bonding while the installation is energised.
What does a continuity test prove, and why is it done first?
A continuity test measures the resistance of a conductor to confirm it is unbroken and soundly terminated along its whole length. Regulation 643.2.1 requires it on protective conductors, including main and supplementary bonding conductors, and on the live conductors of ring final circuits. The cpc is the route earth fault current takes to operate the protective device. If it is broken or high resistance, a fault to a metal case never clears, and the case sits at mains voltage until somebody touches it.
It is first in the sequence for a practical reason: every test that follows assumes the protective conductors are there. A lighting circuit with its cpc snapped in a junction box under the loft insulation will read over 200 MΩ on insulation resistance and look perfect. Only continuity finds it, which is why it turns up so often in loft and cellar wiring.
Amendment 4 does not alter any of this. The published summaries of BS 7671:2018+A4:2026, including SELECT's overview of the changes, list nothing new for the continuity requirements of Regulation 643.2, so the method you already use still stands.
What instrument and test current do you need?
Use the low-resistance ohmmeter range of an MFT or a dedicated instrument made to BS EN 61557-4. IET Guidance Note 3 gives the two numbers that matter: a no-load voltage between 4 V and 24 V, a.c. or d.c., and a short-circuit current of not less than 200 mA. It also expects a resolution of 0.01 Ω, because the readings you are judging are often only a few tenths of an ohm.
A multimeter on its buzzer or ohms range is not a substitute. It passes a few milliamps at most, which will bleep through a single remaining strand or a film of corrosion that a 200 mA test shows up as a high or drifting reading. Check the instrument against a checkbox or known resistors at regular intervals and keep the record. A continuity range that reads 0.10 Ω high condemns good circuits, and one that reads low passes bad ones.
How do you null the test leads?
Nulling, also called zeroing or lead compensation, removes the resistance of the leads from the reading. Clip or press the probes firmly together, press null or zero, and confirm the display settles at 0.00 Ω. A pair of standard leads is typically 0.1 to 0.3 Ω, and a 50 m wander lead can be 0.5 Ω or more, which is larger than most of the circuits you are about to measure.
Three habits keep the null honest:
- Null again whenever you change anything: a different lead set, a wander lead added, crocodile clips swapped for probes, or a fused probe fitted.
- Null the link too. If your shorting lead at the board is a metre of flex with two clips, include it in the null or use a link short and heavy enough to ignore.
- Re-check during the job. Some instruments hold the null until they are switched off, others drop it when the dial moves. Touch the probes together now and then and make sure you still see 0.00 Ω.
If every circuit on a new board reads about 0.25 Ω whatever its length, the leads were not nulled.
How does the R1+R2 method work?
The R1+R2 method measures the line conductor (R1) and the circuit protective conductor (R2) in series. You join the two at one end and measure across them at the other, so the instrument sees the same loop that fault current would take within the installation.
Step by step
- Safely isolate the board or circuit, lock off and prove dead.
- At the board, link the circuit's line conductor to its cpc, either at the outgoing terminal of the protective device or with the line conductor lifted. Use a short link with two good clips.
- Null the leads.
- At each socket-outlet, switch, luminaire and accessory, measure between the line terminal and the earth terminal.
- Note the highest reading. That is the R1+R2 for the circuit, and it is normally found at the point electrically furthest from the board.
- Remove the link and reinstate the line conductor before you move to the next circuit.
On lighting circuits the switch must be closed when you measure at the luminaire. With two-way and intermediate switching, test in each switch position so the reading includes the longest strapper route. On a three-phase board the method is the same for each circuit, one line conductor at a time; the three-phase testing guide covers the extra steps.
The test gives you three results for one set of readings. It proves the cpc. It confirms polarity, because a reading at the line terminal shows that the switched and fused conductor really is the line. And it gives the figure for calculating Zs at the far end without a live loop test at every point, which matters on RCD-protected circuits and wherever live testing is awkward.
How does the wander lead (R2) method work?
The wander lead method measures the protective conductor on its own. One lead goes to the main earthing terminal or the earth bar, and a long trailing lead, usually 20 to 50 m on a reel, is taken to each exposed-conductive-part in turn. Null the instrument with the whole wander lead in circuit, fully unwound.
It suits anything that has no line conductor to link to: main bonding to gas, water and oil pipework, supplementary bonding in a bathroom, steel trunking and conduit, structural steel and metal equipment cases. Guidance Note 3 accepts both methods as proof of continuity, but R2 alone cannot be used to calculate Zs.
| R1+R2 method | Wander lead (R2) method | |
|---|---|---|
| What is measured | Line and cpc in series | cpc or bonding conductor alone |
| Connection at the board | Line linked to cpc, circuit isolated | One lead on the MET or earth bar |
| Confirms polarity | Yes | No |
| Gives Zs by calculation | Yes, Zs = Ze + (R1+R2) × 1.2 | No |
| Best for | Final circuits on an EIC, Minor Works or EICR | Bonding, containment, parts with no line conductor |
| Parallel path risk | Lower: the R1 half of the loop has no parallel path | High: pipework and steelwork can carry the reading |
What should the readings be?
Expected values come from the resistance of copper per metre. The figures below are the standard values for copper conductors at 20 °C given in the IET On-Site Guide and Guidance Note 3.
| Conductor csa (mm²) | Resistance at 20 °C (mΩ/m) |
|---|---|
| 1.0 | 18.10 |
| 1.5 | 12.10 |
| 2.5 | 7.41 |
| 4.0 | 4.61 |
| 6.0 | 3.08 |
| 10.0 | 1.83 |
| 16.0 | 1.15 |
Add the line and cpc figures together to get R1+R2 per metre for a cable.
| Twin and earth (line / cpc, mm²) | R1+R2 at 20 °C (mΩ/m) | 20 m run | 30 m run |
|---|---|---|---|
| 1.0 / 1.0 | 36.20 | 0.72 Ω | 1.09 Ω |
| 1.5 / 1.0 | 30.20 | 0.60 Ω | 0.91 Ω |
| 2.5 / 1.5 | 19.51 | 0.39 Ω | 0.59 Ω |
| 4.0 / 1.5 | 16.71 | 0.33 Ω | 0.50 Ω |
| 6.0 / 2.5 | 10.49 | 0.21 Ω | 0.31 Ω |
| 10.0 / 4.0 | 6.44 | 0.13 Ω | 0.19 Ω |
| 16.0 / 6.0 | 4.23 | 0.08 Ω | 0.13 Ω |
Use the table as a sanity check, not a pass mark. BS 7671 sets no maximum R1+R2; it limits the Zs that results. Two checks are worth making every time.
First, is the reading too high? Two or three times the expected figure on a short run means a loose terminal, a damaged conductor or a poor joint. Bear in mind that older 2.5 mm² twin and earth has a 1.0 mm² cpc, which gives 25.51 mΩ/m, so a 1970s socket radial reads about 30 per cent higher than modern cable of the same length.
Second, is it too low? The line conductor has no parallel path, so R1+R2 can never honestly be less than R1 alone. A 30 m run of 1.0 mm² lighting cable has 0.54 Ω of line conductor in it. If the instrument shows 0.20 Ω, you are not measuring the circuit you think you are.
The table also works backwards. Divide a good reading by the mΩ/m figure and you have the cable length: 0.59 Ω on 2.5/1.5 is 30 m, which is handy for a volt drop check on a circuit you did not install.
Ring final circuits use a three-stage test with r1, rn and r2, covered in the guide to ring final circuit testing. The R1+R2 for a sound ring is (r1 + r2) ÷ 4.
How does temperature change the reading?
Copper resistance rises by about 0.4 per cent for every degree Celsius, and that matters twice.
The first effect is the temperature on the day. The tabulated values are for 20 °C. The On-Site Guide gives multipliers for other ambient temperatures, so you can adjust the expected figure before comparing it with the instrument.
| Ambient temperature when testing | Multiply the 20 °C value by |
|---|---|
| 5 °C | 0.94 |
| 10 °C | 0.96 |
| 15 °C | 0.98 |
| 20 °C | 1.00 |
| 25 °C | 1.02 |
On a 0.50 Ω circuit that is only 0.02 to 0.03 Ω, so it rarely changes a verdict, but it explains why an unheated garage in January reads slightly under the book figure.
The second effect is the one that counts for disconnection times. Under fault conditions the conductors are assumed to be at their maximum operating temperature, not at 20 °C, so the cold reading is scaled up before it is used for Zs.
| Cable insulation | cpc in the same cable or bunched | cpc run separately |
|---|---|---|
| 70 °C thermoplastic (PVC) | 1.20 | 1.04 |
| 90 °C thermosetting (XLPE) | 1.28 | 1.04 |
The familiar 1.2 is therefore the figure for ordinary PVC twin and earth. Thermosetting cable run at its full rated temperature needs 1.28.
How do you get Zs from R1+R2?
Add the corrected R1+R2 to the external impedance:
Zs = Ze + (R1+R2 measured) × 1.2
Take a 20 A Type B radial in 2.5/1.5 twin and earth, 28 m long, in a garage at about 10 °C. The expected R1+R2 is 28 × 19.51 mΩ, which is 0.55 Ω at 20 °C or 0.52 Ω on the day. You measure 0.54 Ω at the last socket, which is believable. With a measured Ze of 0.22 Ω:
Zs = 0.22 + (0.54 × 1.2) = 0.87 Ω
The maximum in Table 41.3 for a 20 A Type B device is 2.19 Ω, so the circuit disconnects in time with plenty in hand. For a sub-board, use the Zs at that board (Zdb) in place of Ze.
The alternative route is to measure Zs live and compare it with 0.8 times the tabulated maximum, which is how the On-Site Guide tables are built. Both routes make the same temperature allowance, so do not apply 1.2 and 0.8 to the same figure. The full list is in the guide to maximum Zs values, the live test is in the guide to earth fault loop impedance testing, and if the answer comes out high, work through high Zs troubleshooting.
What should bonding conductors read?
Work it out from length and cross-section, the same as a cpc. A 10 mm² main bonding conductor is 1.83 mΩ/m, so 25 m comes to about 0.05 Ω, and so does 15 m of 6 mm². That is where the 0.05 Ω figure long quoted from Guidance Note 3 comes from. It is an indication for a typical domestic run, not a limit in BS 7671, and the instrument's own accuracy is of the same order. What you are looking for is the reading that does not fit: 0.8 Ω on a 6 m run means a corroded clamp, a clamp fitted over paint, or a conductor nicked during other work.
Put the wander lead probe on the pipe itself, just beyond the clamp, so the clamp connection is inside the measurement. If you want the conductor on its own, without the pipework acting as a parallel path, one end has to come off, and that is only ever done with the whole installation isolated.
Supplementary bonding has a limit you can calculate. Regulation 415.2.2 requires the resistance between simultaneously accessible parts to be no more than 50 V ÷ Ia, where Ia is the operating current of the protective device. For a 32 A Type B device Ia is 160 A, giving 0.31 Ω; for a 30 mA RCD it is 1,667 Ω. The bathroom zones guide shows where that test still applies.
Steel conduit and trunking used as the protective conductor are tested with the same instrument, but the reading alone is not enough. Inspect every joint you can reach. Rusty couplers and loose locknuts are where steel containment fails, and a reading that jumps about when you flex a joint is a failed joint, whatever the number says.
What does a wrong reading tell you?
| What you see | Likely cause | What to do |
|---|---|---|
| Open circuit | Switch open, cpc broken or never terminated, link on the wrong circuit | Check the link and switches, then halve the circuit to find the break |
| Two to three times the expected value | Loose terminal, damaged conductor, longer route than assumed, older cable with a reduced cpc | Re-terminate at the high point and the one before it; confirm the cable type |
| Reading drifts or will not settle | Poor connection, oxidised terminal, dirty probe tips | Clean and re-make the joint; re-null and repeat |
| Far lower than the length allows | Parallel path through pipework, containment, another cpc or a shared back box | Find and separate the parallel path with the installation isolated |
| The same small value at every point | Leads not nulled | Null and start again |
| Reading rises steadily along the circuit, then drops | You have passed the furthest point; the circuit is wired in a different order from the one assumed | Record the highest value, wherever it is |
How is continuity tested on an EICR?
On a periodic inspection you rarely get to unpick everything. The extent of testing and any sampling are agreed with the client and recorded before you start, as the guide to sampling and limitations explains. Within that, the aim is the same: confirm an earth connection at the accessible exposed-conductive-parts and earth terminals of the circuits you test.
Where a circuit can be isolated, use R1+R2 or the wander lead as you would on new work. Where it cannot, a measured Zs at the point shows that an earth path exists, but not whether that path is the cpc or a gas pipe. Say which test you did, and record the rest as a limitation.
One rule has no exceptions. Do not disconnect the earthing conductor or main bonding conductors while the installation is energised. Under fault conditions, or with a damaged PEN conductor on a TN-C-S supply, those conductors carry current, and the end you have just removed can become live in your hand.
When continuity is missing, Electrical Safety First's Best Practice Guide 4 is the usual reference for the code.
| Finding | Typical code |
|---|---|
| No earth continuity at a socket-outlet | C2 |
| Lighting circuit with no cpc supplying Class I fittings or metal switch plates | C2 |
| Lighting circuit with no cpc where every fitting and switch is Class II or all-insulated | C3 |
| Main protective bonding absent where it is required | C2 |
| Main bonding undersized, but at least 6 mm² with no sign of thermal damage | C3 |
| Continuity not confirmed because the circuit could not be isolated | Agreed limitation; FI only where there is a specific reason to suspect a defect |
The reasoning behind each code is in the guide to EICR observation codes.
How are continuity results recorded?
On the schedule of test results, enter the highest R1+R2 in ohms for each radial circuit. For each ring, enter r1, rn and r2 as measured end to end, plus the R1+R2 found at the sockets. Where the wander lead method was used, fill in the R2 column and leave R1+R2 blank. The model schedule in Appendix 6 has both columns and asks for at least one to be completed, so either method can be shown as it was done.
Continuity of the earthing conductor and of each main bonding connection is confirmed on the certificate itself, in the particulars of the installation. Do not record a test you did not carry out. An accessory you could not reach is a limitation, and a figure copied from the last report is a guess.
What are the common mistakes?
| Mistake | What happens | How to avoid it |
|---|---|---|
| Not nulling the leads | Every reading is 0.1 to 0.3 Ω high, and short circuits look poor | Null at the start and after any change of leads |
| No estimate made first | A parallel path or a missed link goes unnoticed | Work out length × mΩ/m before you measure |
| Measuring at the nearest point only | The recorded R1+R2 is not the maximum, so the calculated Zs is optimistic | Test every point and record the highest |
| Switch open on a lighting circuit | Open circuit at the luminaire | Close the switches; test each position on two-way circuits |
| Applying 1.2 and 0.8 together | The circuit is penalised twice, or not at all | 1.2 for calculated Zs, 0.8 for measured Zs, never both |
| Leaving the link in place | Line and cpc are joined when the circuit is re-energised | Make removing the link a fixed step before the next circuit |
Parallel paths deserve the most attention. With metal back boxes, steel conduit and bonded pipework there are many routes back to the earth bar, and a cpc broken at a junction box can still read low through a back box that touches a bonded pipe. If you expect 0.6 Ω and see 0.08 Ω, investigate before you write anything down. The same discipline catches borrowed neutrals and crossed circuits on the following tests.
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 £4.99 a month.