It is also the requirement people skip. Ampacity gets checked because it is a safety rule with a table behind it; drop gets ignored because it is a performance recommendation. Then the shop lights dim when the compressor starts and nobody can explain why the wire that passed every check is not doing its job. This guide covers what the figure is, how it is estimated, and where it starts to govern.
What is actually lost
Push current down a conductor and the conductor's own resistance consumes some of the voltage. Whatever is consumed on the way out and on the way back is not available at the load. A 240 V circuit losing 12 V delivers 228 V, and every piece of equipment on it operates at 228 V, not at its nameplate.
The consequences are unglamorous but real. Motors are the most sensitive: a motor at reduced voltage draws more current to produce the same torque, which heats it and shortens its life, and starting current makes the drop transiently worse at exactly the moment the motor needs voltage most. Resistive heating loads underperform roughly with the square of the voltage. Electronics mostly cope until they suddenly do not. Lighting flicker on motor start is the classic complaint and is very often a voltage-drop problem in a circuit that is otherwise entirely legal.
Two properties of the loss are worth internalising. It scales with current, so a lightly loaded long run may be fine while a heavily loaded one of the same length is not. And it scales with the round-trip length of the conductor, which is why a one-way distance of 300 feet is a 600-foot conductor path as far as the arithmetic is concerned.
The 3% and 5% figures are recommendations
The familiar targets - 3% on a branch circuit, 5% total from the service to the far end including the feeder - are recommendations in the code rather than requirements. They appear in informational notes about conductors sized for reasonable efficiency of operation, not in an enforceable rule about conductor size.
That does not make them optional in practice. Project specifications routinely require them, some jurisdictions enforce them through amendments, and equipment manufacturers state their own minimum voltages. More to the point, they are good engineering targets: they were chosen as the point where the efficiency loss and the performance penalty stop being worth the copper you would save.
This site sizes to a 3% branch-circuit target by default and reports the drop it achieves, so you can see how much margin a given conductor has rather than just whether it passed.
How the estimate is made
The method used here is the circular-mil estimate, the same one used on paper in the field. Drop is proportional to the current, to the one-way length, and to a resistivity constant for the conductor material, and inversely proportional to the conductor's circular-mil area. Single-phase circuits carry a factor of two for the out-and-back path; three-phase circuits carry the square root of three instead.
Two things follow directly from the shape of that formula. Doubling the length doubles the drop. Going up a conductor size cuts the drop by whatever ratio the circular-mil areas have - which is why a single size increase often solves a marginal run outright, and why paralleling has the effect it does.
Material matters too. Aluminum's resistivity constant is substantially higher than copper's, so an aluminum conductor of the same size drops more voltage carrying the same current. That is a large part of why aluminum runs are typically specified a size or two above the copper equivalent, quite apart from the ampacity difference.
The estimate deliberately ignores conductor reactance and power factor. On small branch circuits that is a rounding error. On large or long AC feeders it is not, and the figure this site reports should be treated as a first pass rather than an engineering study.
Where drop takes over from ampacity
Ampacity and voltage drop are independent requirements and the larger conductor wins. On short runs ampacity governs almost always. Somewhere out along the distance axis the two cross, and past that point the conductor is being chosen for performance rather than for heat.
A 50 A circuit at 240 V shows the crossover cleanly. Out to roughly 75 feet one way, the conductor that satisfies the 75°C ampacity requirement also stays inside 3%. At 100 feet it does not, and the sizing steps up. By 150 feet it steps up again, and past 250 feet the conductor is two or three sizes above what heat alone would demand. The published scenario pages work every combination out and state which requirement governed.
Low voltage makes this much worse, because the drop is a percentage of a smaller number. The same wire, the same current and the same distance produce twice the percentage drop at 120 V that they produce at 240 V. A 20 A load on 12 AWG copper at 120 V is already past 3% at 50 feet one way and well past 6% at 100 feet - which is why long 120 V runs are so often the ones that misbehave.
- Drop scales with current, with one-way distance, and inversely with conductor area.
- At the same size, current and distance, a 120 V circuit shows double the percentage drop of a 240 V circuit.
- Aluminum drops more voltage than copper at the same size, on top of its lower ampacity.
- Long runs are where drop, not heat, picks the conductor - and the code's ampacity rules will not catch it for you.
Fixes that are not simply bigger wire
Upsizing works and is usually the answer, but it is not the only lever, and on a long feeder it is the expensive one. Raising the system voltage is the most effective change available: the same load at twice the voltage draws half the current and the percentage drop falls by a factor of four. This is the reason detached buildings are so often fed at 240 V and stepped down locally.
Shortening the run helps more than people expect, since drop is linear in length - relocating a subpanel closer to the load it serves can do more than two conductor sizes would. Splitting a heavy load across separate circuits reduces the current in each. And on a feeder, moving the panel rather than the branch circuits attacks the longest part of the path.
One caution about upsizing. Where circuit conductors are increased in size for voltage drop, the equipment grounding conductor generally has to be increased in proportion, which is a calculation this reference does not perform. It is easy to upsize the phase conductors, forget the grounding conductor, and end up with an installation that fails inspection for a reason unrelated to the problem you set out to solve.
Frequently asked questions
Is voltage drop a code requirement?
The 3% branch-circuit and 5% total figures appear as recommendations in informational notes about conductors sized for reasonable efficiency of operation, not as enforceable ampacity rules. Project specifications and some local amendments do make them mandatory, and equipment manufacturers set minimum voltages of their own.
How much voltage drop is too much?
Beyond about 3% on a branch circuit, motors run hotter, heating loads underperform and lighting is visibly affected by starting currents on the same circuit. The 5% figure covers the whole path from the service through the feeder to the far end of the branch circuit.
Does voltage drop matter on a short run?
Rarely. Drop is linear in one-way distance, so a 25-foot branch circuit carrying a normal load loses a fraction of a percent. It becomes the governing requirement on long runs, on heavily loaded circuits, and on 120 V circuits, where the same loss is twice the percentage it would be at 240 V.
Why does aluminum drop more voltage than copper?
Aluminum has a higher resistivity, so its constant in the circular-mil estimate is substantially larger than copper's. At the same conductor size, current and distance, an aluminum run loses more voltage - which is why aluminum is usually specified a size or two above the copper equivalent.
Does going up one wire size fix a voltage drop problem?
Often, yes. Drop is inversely proportional to circular-mil area, so a single size increase cuts it by the ratio of the two areas, which is enough for most marginal runs. Raising the system voltage or shortening the run attacks the problem harder on long feeders.