A cable can carry its rated current perfectly safely and still be the wrong cable. If the run is long enough, the voltage arriving at the far end sags below what the equipment needs — motors run hot, drives trip, LED drivers flicker, and the electrician gets called back to a circuit that "passed" every ampacity check.
That is why cable sizing is always two separate checks: current-carrying capacity first, voltage drop second — and on long runs, voltage drop is usually the one that decides the size. This guide covers the limits, both common calculation methods, and a worked example.
The Limits: 3% and 5%
Most codes converge on similar recommendations. The NEC's informational notes suggest limiting voltage drop to 3% on a branch circuit and 5% total from the service to the furthest outlet. BS 7671 takes a similar position for installations supplied from the public network — around 3% for lighting and 5% for other uses, measured from the origin of the installation. IEC-based practice elsewhere follows the same shape.
Note the nuance: in the NEC these figures are recommendations rather than hard requirements (specific equipment rules aside), but specifying engineers treat them as design limits because the cost of ignoring them arrives as nuisance faults after handover.
Check One: Ampacity
Before any voltage drop math, the cable must carry its design current continuously without exceeding its insulation temperature rating. The catalogue ampacity is only a starting point — it must be derated for the real installation:
- Ambient temperature above the reference value
- Grouping — multiple circuits sharing a tray, conduit or trunking
- Installation method — clipped direct, in conduit, buried, on ladder
- Insulation type — PVC (70°C) vs XLPE (90°C) ratings differ substantially
After derating, the cable's corrected ampacity must still exceed the circuit's design current and coordinate with its protective device. Only then does voltage drop enter the picture.
Check Two: Voltage Drop
The Imperial Method (NEC Practice)
North American practice uses the circular-mil formula:
VD = 2 × K × I × L ÷ CM — single phase; use 1.732 instead of 2 for three phase. K ≈ 12.9 Ω·cmil/ft for copper, ≈ 21.2 for aluminium
where I is the load current in amperes, L the one-way run length in feet, and CM the conductor's circular-mil area. The result is volts dropped; divide by system voltage for the percentage.
The Metric Method (BS/IEC Practice)
Metric practice is table-driven. Manufacturer and BS 7671 tables list each cable size's voltage drop in millivolts per ampere per metre (mV/A/m):
VD (volts) = mV/A/m × I × L ÷ 1,000 — I in amperes, L the one-way route length in metres
For larger cables the tables split the figure into resistive and reactive components, but for typical building sub-circuits the single impedance figure is used directly.
A Worked Example
A 32 A single-phase circuit at 230 V feeds equipment 60 m from the panel, run in copper thermoplastic cable:
- Try 6 mm² — tabulated drop ≈ 7.3 mV/A/m: VD = 7.3 × 32 × 60 ÷ 1,000 = 14.0 V, which is 6.1% of 230 V — fails the 5% limit
- Try 10 mm² — tabulated drop ≈ 4.4 mV/A/m: VD = 4.4 × 32 × 60 ÷ 1,000 = 8.4 V, which is 3.7% — passes
Both sizes carry 32 A comfortably. The run length, not the current, set the cable size — which is exactly the trap of sizing by ampacity alone.
The Same Logic in Three Phase
Three-phase circuits follow the identical procedure with the three-phase column of the tables (or the 1.732 factor in the imperial formula), and they enjoy a structural advantage: for the same power, current per conductor is lower and the tabulated drop per ampere-metre is smaller, so three-phase distribution tolerates distance far better. That is one reason long feeder runs — risers, car parks, external plant — are kept three-phase as far as possible, with the single-phase split made close to the load.
A Note on Temperature
Conductor resistance rises with temperature — roughly 0.4% per °C for copper — and the tabulated mV/A/m figures assume the conductor running at its full rated temperature. A lightly loaded cable runs cooler and drops less than the table suggests, which means the standard method errs on the safe side. Precision designs can take credit for this; everyday building circuits simply use the tables and keep the margin.
Cable sizing is two independent checks — derated ampacity and voltage drop — and the larger answer wins. On any run beyond roughly 30–40 metres, calculate voltage drop before committing the cable schedule; it is the check most likely to govern.
Where It Bites in Real Buildings
Voltage drop problems cluster in predictable places: car park and external lighting circuits, rooftop plant fed from basement switchrooms, pump and fan motors at the far end of long risers, and anything fed from a generator, where the source impedance is already higher. In motor circuits, low terminal voltage raises current and heat — torque falls with the square of voltage, so a 5% sag costs nearly 10% of torque.
These long-run circuits are also where the electrical design earns its fee: a cable schedule produced with proper voltage drop checks per circuit, as part of a coordinated electrical design, costs a fraction of one post-handover callback. It is standard scope in the drawing packages we deliver.
When a Circuit Fails the Check
A failed voltage drop check has more than one fix, and the biggest cable is not always the right one:
- Upsize the conductor — the default fix, but copper is expensive and larger cables need larger containment and glands
- Move the distribution board closer — a sub-board near a cluster of distant loads often beats a dozen long branch circuits
- Run the distance at a higher voltage — keep the feeder three-phase and transform or split down near the load
- Rebalance the design current — on marginal fails, confirming the realistic continuous load rather than the breaker rating sometimes passes the circuit honestly
On a full project, these trade-offs ripple through the panel schedule, containment design and budget — which is why voltage drop is checked during design, circuit by circuit, rather than discovered during commissioning.
Steady-State Drop vs Starting Dip
One final distinction keeps diagnoses honest. The 3%/5% limits govern steady-state voltage drop at the running current. Motor starting is a different phenomenon: for a few seconds, a direct-on-line motor draws six to eight times its rated current, and the momentary dip it causes is assessed separately — typically against what the motor can accelerate through and what nearby equipment, especially lighting and electronics, can ride out. A circuit can pass its steady-state check and still dim the floor every time the pump starts; the fixes there are starting methods (star-delta, soft starters, VFDs) and source stiffness, not just cable size. On generator-fed buildings this check matters doubly, because the generator's higher source impedance amplifies every dip.
Conclusion
Size for ampacity with the real derating factors, then check voltage drop against the 3%/5% limits with the method your code uses — and let the bigger cable win. It is simple discipline, but it is the difference between an installation that merely passes inspection and one that performs for its whole life.