The core difficulty in C&G 2396 design work is that no calculation stands alone. Design current feeds device rating; device rating and correction factors feed cable size; cable length feeds voltage drop; and earth fault loop impedance feeds both disconnection time and protective conductor sizing. This guide teaches that chain as one connected procedure. Work through the two scenarios, the decision table and the final exercise, and practise producing every value from a single design brief, checking each figure against the current BS 7671 tables rather than from memory.
Starting With Cable Size Instead of Design Current Breaks the Chain
Define four currents in order: design current Ib, protective device rating In, tabulated current-carrying capacity It, and installed capacity Iz. Each step constrains the next, so a cable size is the output of the chain, never the starting point.
Begin every circuit with the load. Design current Ib comes from the power the circuit is intended to carry, at the nominal voltage, adjusted where the design deliberately applies diversity. The protective device rating In is then chosen so that In is at least Ib, using a standard device rating. Only after fixing Ib and In do you enter the current-carrying capacity tables to find It, the tabulated value for the cable in its installation method.
Keep the definitions distinct in your notes: It is the value printed in the table for the reference method, while Iz is what the cable can actually carry once correction factors are applied. A cable is acceptable when Iz still meets or exceeds In. Writing these four symbols and their inequalities at the top of every practice brief, before any calculation, forces the correct sequence and prevents the habit of reaching for a familiar cable size first and justifying it afterwards.
Voltage Drop Along the Submain and Final Circuit Must Be Added Together
The voltage drop allowance runs from the origin of the installation to the point of use, so every series segment spends part of the budget. Check the final circuit only after subtracting what upstream cables consume.
Treat the total drop allowance as a budget shared by the intake, any submain, and the final circuit. A final circuit that passes in isolation can still leave the load undervolted because the submain consumed a share. In a worked example, a 32 A radial runs 25 m in 4 mm² two-core copper with an illustrative millivolt-per-ampere-metre value of 11: drop is 11 × 30 × 25 ÷ 1000 = 8.25 V, comfortably inside a 11.5 V allowance for a 230 V power circuit. On that basis alone the circuit looks compliant.
Now include the submain: the distribution board is fed by a cable already dropping an illustrative 4.0 V at the load it carries. The true total is 8.25 + 4.0 = 12.25 V, which exceeds 11.5 V, so the design fails even though each figure looked fine separately. The better decision is to target the segment contributing most per metre: upsizing the final circuit to 6 mm² (illustrative value 7.3) gives 7.3 × 30 × 25 ÷ 1000 = 5.5 V, total 9.5 V, compliant without touching the submain. Always confirm every figure against the current BS 7671 appendix values rather than the numbers used here.
Reduced CPC Sizes Depend on a Condition Many Designs Skip Checking
Standard protective conductor sizes can be selected from the reduced-size table only where the protective device operates in the required time under fault conditions. If that condition is not met, the adiabatic equation must be applied instead.
Two distinct routes exist for protective conductor sizing, and they are not interchangeable. The reduced-size route lets a conductor smaller than the line conductor be used, for circuits up to the stated line size, provided the design ensures the protective device disconnects within the specified time for the prospective fault current. The adiabatic route is a calculation: the conductor cross-section must satisfy S ≥ √(I²t) ÷ k, using the fault current, the device's operating time at that current, and the k value for the conductor's insulation and type.
Worked scenario: a designer selects a 1.5 mm² CPC alongside a 4 mm² line, citing the reduced-size table, because the circuit is within the line size limit. The circuit is long and the external loop impedance is high, so the illustrative fault current is only about 200 A and the device takes around 6 s to operate — longer than the condition allows. The table route is therefore unavailable. The adiabatic check with illustrative k = 115 gives √(200² × 6) ÷ 115 ≈ 4.3 mm², so the 1.5 mm² CPC is undersized. The better decision is to verify the operating condition before citing the table, and where it fails, increase the CPC or select a device and arrangement that clears faster. All values here are illustrative; use the current tables and k values in real practice.
Correction Factors Multiply Against It — and Each One Comes From a Different Table
Ambient temperature, grouping and thermal insulation each have their own factor and their own table, and the factors multiply together in the denominator: It must be at least In divided by the product of all applicable factors.
Write the requirement as It ≥ In ÷ (Ca × Cg × Ci) so no factor is lost. The ambient temperature factor Ca depends on the conductor's operating temperature as well as the ambient temperature, so the same 35 °C ambient gives different factors for 70 °C thermoplastic and 90 °C thermosetting cables. The grouping factor Cg depends on how many loaded circuits are bunched together. Omitting or misreading any one of these silently inflates the apparent capacity of the chosen cable.
Scenario: a designer selects 70 °C thermoplastic cable, looks up the ambient factor in the 90 °C column, and concludes the selected size carries the required It with margin to spare. Recomputing from the correct column reduces the factor, the denominator shrinks, the required It rises, and the next size up is needed. The better decision is to fix the cable insulation type first, then read every factor from the column matching that type. In practice briefs, annotate each factor with the circuit count, ambient temperature and insulation type it came from; if you cannot name all three for a factor, the value is not yet defensible.
Verification Results Only Make Sense When Traced Back to the Design Record
Measured values are compared with values the design predicted: maximum permitted loop impedance, expected conductor resistance, and the disconnection time the design assumed. Documentation is the thread that connects them.
A design record should show, for each circuit, the design current, device rating, cable size and route, applied correction factors, calculated voltage drop, and the maximum earth fault loop impedance assumed for compliance. Verification then measures loop impedance and conductor resistance and records them on schedules. The comparison is the learning point: a measured loop impedance is acceptable when it does not exceed the design maximum, and an expected conductor resistance can be predicted from the cable's per-metre values multiplied by length.
Practise reading discrepancies rather than reacting to them. A measured loop impedance lower than the design maximum is expected, because tabulated maxima assume conductors at operating temperature while measurements are usually taken on cooler conductors. A measured conductor resistance far above the calculated value points to a loose connection, a wrong route length, or a parallel path misread. In paper scenarios, ask of every figure: which design value does this correspond to, and is the difference explained by temperature, parallel paths or an error? A schedule entry you cannot trace back to a design value signals a gap in either the design or the record.
Deciding Between Upsizing, Regrouping, or Redesigning the Circuit
When a check fails, three broad responses exist: increase a conductor size, change the installation arrangement, or change the protective device. Choose by targeting the constraint that actually failed, at the least cost to the design intent.
The most efficient fix addresses the segment or factor that dominates the failure. Upsizing a final circuit where it spends most of the drop allowance, splitting a group to ease a severe grouping factor, or selecting a device characteristic that clears faults faster are each targeted responses. Recalculating the whole chain after any change is essential, because a larger conductor changes voltage drop per metre, fault current, and sometimes the reduced-size options for the CPC.
Use the table below as a drill: for each failure in a practice brief, write your chosen response and the reason before recalculating. The discipline of naming the failed constraint first prevents the shotgun response of upsizing everything, which can create its own problems such as larger CPC requirements and harder-to-achieve terminations. Every value you apply afterwards must come from the current BS 7671 edition.
| Failed check | First targeted response | Why this response |
|---|---|---|
| Total voltage drop exceeds allowance | Upsize or shorten the segment contributing most per metre | Fixing the dominant segment costs less than upsizing the whole route |
| Grouping factor drives required It very high | Regroup or split circuits across separate containment | Dividing a group can restore capacity without a large cable jump |
| Measured loop impedance above design maximum | Recheck connections, lengths and temperature assumptions before redesigning | Several benign causes explain differences between design and measurement |
| Adiabatic check fails for the CPC | Increase CPC size, or select a device that clears faster at the fault current | The product k × S must exceed the fault energy I²t |
| Ambient factor applied from the wrong insulation column | Fix insulation type first, then reread every factor | Ca differs by conductor operating temperature, not ambient alone |
A Full-Chain Practice Exercise With a Self-Check Rubric
Build one complete design from a single brief: derive the current chain, apply factors, check voltage drop, disconnection and the CPC, then grade yourself against a rubric of observable checkpoints rather than a feeling of completion.
Exercise brief (all values illustrative): a 230 V single-phase distribution board is fed by a submain dropping 4.0 V at full load. One final circuit carries a 7 kW load, is 30 m long, is grouped with four other circuits (illustrative Cg 0.7), in a 35 °C ambient (illustrative Ca 0.94 for the chosen 70 °C cable), with no thermal insulation contact. Produce, in order: Ib, In, the required It, the cable size, total voltage drop, maximum acceptable loop impedance, and the CPC size with your chosen sizing route justified. Then mark yourself: two points per step where the value is correct and the source table is named.
Self-check rubric: pass at each checkpoint only if (1) Ib is derived from the load, not assumed; (2) In is a real standard rating at or above Ib; (3) the required It divides In by every applicable factor, each annotated with its source; (4) voltage drop includes the submain's 4.0 V; (5) the CPC route cites its precondition before the size. A clean full run, repeated on fresh briefs, is a meaningful readiness milestone — it confirms process control, not a prediction of the assessment result. An adaptable sequence: week one and two, chain drills on single circuits until every inequality is automatic; week three, vary factors, lengths and device types so each check can fail; week four, full briefs under a self-set time limit, then documentation and reconciliation practice. Registration and administrative details for the qualification are set by City & Guilds and approved centres; confirm current specifics on the City & Guilds website.
- Ready when you can reproduce the full chain on an unseen brief without notes
- Ready when every value in your working names the table it came from
- Ready when you can explain a measured-versus-design difference in paper scenarios
- Ready when two fresh briefs in a row pass every rubric checkpoint within your self-set time
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.