Training teaches current-carrying capacity first, and it sticks. A table, an installation method, a cross-section. In plant practice, however, that is rarely the point where a design fails.
Two conditions, one of them binds
A conductor cross-section has to satisfy two conditions at once. It has to carry the operating current thermally without the insulation exceeding its limit temperature. And it has to keep the voltage drop over the cable length small enough that sufficient voltage still arrives at the load.
For the first condition there are tables: IEC 60364-5-52, published in Germany as DIN VDE 0100-520, assigns a current-carrying capacity to every installation method. For the second there is no table but a recommendation — Annex G of the same standard gives 3 % for lighting and 5 % for other loads, measured from the origin of the installation. In industrial plants the internal limits are often tighter.
Over short distances inside a cabinet, the first condition binds. As soon as a cable leaves the building, the balance shifts. An example from conveyor technology: 32 A operating current, 45 m to the drive, 400 V, cos φ 0.9. Current-carrying capacity would be satisfied with 6 mm² depending on installation method. For 3 % voltage drop, 3.3 mm² is required by calculation — 4 mm² in the standard series to IEC 60228. Both hold.
Extend the same run to 120 m, not unusual in a larger hall and voltage drop already demands almost 9 mm², so 10 mm². The current has not changed. The distance has.
Where the calculation flatters
The usual formula assumes resistive behaviour at 20 °C. Both are optimistic. A conductor in operation is warmer, and resistance rises with temperature. Copper has a temperature coefficient of about 3.9 · 10⁻³ 1/K; from 20 °C to 70 °C that is close to 20 % more resistance. The actual voltage drop is then around a fifth above the calculated value. At large cross-sections the inductive component adds to this, and it does not appear in the formula at all.
So anyone calculating against a 3 % limit and landing at exactly 2.9 % has no margin, only a rounding result.
What has to survive in the documentation
The cross-section ends up in the cable schedule; the reasoning does not. That takes its revenge at the first extension: if nobody knows whether 10 mm² came from ampacity or from voltage drop, the next station gets added by guesswork.
So the governing condition belongs in the plant documentation, as a note on the circuit, not in a separate calculation file nobody can find three years later.
Check it yourself
Both directions take seconds: the required cross-section from a permissible voltage drop, or the actual voltage drop for a cross-section already fixed.