Skip to main content

What Size Wire Do I Need? Voltage Drop Over a Long Run

The current goes out and comes back, so a 100 ft run is 200 ft of copper — and moving to 240 V helps twice over.

By Mohamed Zakrya

Updated · 9 min read

Share
Voltage drop, and the two multipliers in it What size wire do I need Two multipliers decide it, and people get them wrong in opposite directions. 1 · THE CURRENT What the load actually draws, not the breaker. 20 A Drop scales with it. 2 · THE RUN, DOUBLED Out on one conductor, back on the other. 100 ft → 200 ft Enter one way; it doubles. 3 · THE COPPER NEC Chapter 9, Table 8, uncoated copper at 75 C. 12 AWG = 1.93 ohms per 1,000 ft. 4 · THE LOSS 2 × 100 × 20 × 0.00193 7.72 V 6.43% 112.28 V at the load. AND THE MULTIPLIER NOBODY COUNTS TWICE 240 V is worth four, not two The same 2,400 W load draws half the current, which halves the volts lost. That smaller loss is then measured against a supply that has doubled. Two independent halvings. 6.4333% → 1.6083%, on the same cable Turn it around for the reach L = percent × V ÷ (200 × current × ohms per foot). At 20 A on 120 V inside 3 percent, 12 AWG reaches 46 ft and 8 AWG reaches 117. The same load at 240 V takes 12 AWG to 186. four times the percentage, four times the distance This settles one constraint. It does not size a circuit. The 3 and 5 percent figures are NEC Informational Notes, and NEC 90.5 states those are not enforceable as requirements. Ampacity, temperature correction, conduit fill and overcurrent protection are separate rules. A licensed electrician sizes and installs.
The current goes out and comes back, so a 100 ft run is 200 ft of copper — and moving to 240 V helps twice over.

Voltage drop happens along the cable, not at the equipment using the electricity. Every foot of conductor adds resistance, and current flowing through that resistance consumes part of the available voltage before it reaches the load.

A meter at the panel can therefore show the full supply voltage while the far end of a long circuit reads low. The panel measurement describes the starting point, not what survives the journey through the conductors.

The calculation is Ohm's law applied to the cable instead of the load, and two multipliers control the result. The current travels both ways, which makes the electrical path twice the run length you measured. And for the same load, moving from 120 V to 240 V improves the percentage drop by a factor of four rather than two.

The current goes out and comes back

Cable runs are normally described by their one-way physical distance. A load 100 ft from the panel has 100 ft of outgoing conductor and another 100 ft carrying the current back, so the calculation has to account for 200 ft of copper. That return path is the factor of two at the front of the equation, and leaving it out models half the circuit and reports half the real loss.

voltage drop = 2 × one-way distance × current × ohms per foot
ohms per foot = ohms per 1,000 ft ÷ 1,000
percent drop = voltage drop ÷ supply voltage × 100
voltage at load = supply voltage − voltage drop

The default case uses a 20 A load, a 100 ft one-way run, 12 AWG copper and a 120 V supply. The resistance is 1.93 ohms per 1,000 ft, taken from NEC Chapter 9, Table 8 for uncoated copper at 75 C.

ohms per foot = 1.93 ÷ 1,000
ohms per foot = 0.00193

voltage drop = 2 × 100 × 20 × 0.00193
voltage drop = 7.72 V

percent drop = 7.72 ÷ 120 × 100
percent drop = 6.43%

voltage at load = 120 − 7.72
voltage at load = 112.28 V

The load receives 112.28 V after 7.72 V has been lost in the conductors, and at 6.43 percent that is graded HIGH because it is past both the 3 and the 5 percent figures. Those verdict labels belong to the voltage drop calculator, which grades 3 percent or less as GOOD, over 3 through 5 as ACCEPTABLE, and over 5 as HIGH. The bands are not themselves code requirements.

The 7.72 V is spread along the whole current path rather than lost at any one point. A measurement beside the panel is taken before that resistance has accumulated, so a normal reading there establishes nothing about the voltage available at the far end.

The same load at 240 volts

The comparison that matters keeps the load, the cable and the distance unchanged. A 2,400 W load draws 20 A at 120 V but only 10 A at 240 V, which is the same relationship the watts to amps calculator works in.

SupplyLoadCurrentVoltage lostDropAt the loadVerdict
120 V2,400 W20 A7.72 V6.4333%112.28 VHIGH
240 V2,400 W10 A3.86 V1.6083%236.14 VGOOD
120 V current = 2,400 W ÷ 120 V = 20 A
240 V current = 2,400 W ÷ 240 V = 10 A

Current creates the first improvement. Voltage drop across a fixed resistance is proportional to current, so halving the current halves the absolute loss from 7.72 V to 3.86 V. That change on its own is worth a factor of two.

Supply voltage creates the second, and it is a different mechanism rather than more of the same one. The percentage compares the lost voltage against the starting voltage, and that reference has doubled, so the already smaller loss is halved again as a proportion.

absolute-loss improvement = 7.72 ÷ 3.86 = 2
percentage improvement = 6.4333 ÷ 1.6083 = 4

The two effects are independent, so they compound. Half the current produces half the voltage loss, and twice the supply voltage then makes that smaller loss half as large again in percentage terms, which leaves the combined figure at one quarter of the original.

A second comparison exposes the half-measure and is worth running deliberately. If the current stays at 20 A while only the reference voltage changes, the cable still loses the full 7.72 V, and the percentage falls only because the same loss is divided by twice the supply.

SupplyCurrentImplied loadVoltage lostDropVerdict
120 V20 A2,400 W7.72 V6.43%HIGH
240 V20 A4,800 W7.72 V3.22%ACCEPTABLE

That is not the same-load comparison, because holding 20 A at 240 V means the load has doubled to 4,800 W. Changing the reference voltage alone is worth two; halving the current as well is what makes the improvement four.

For the unchanged 2,400 W load the verdict moves from HIGH to GOOD without touching the 12 AWG cable or the 100 ft route. That compounding is the electrical reason long feeds to outbuildings are so often run at 240 V.

Two halvings, one on top of the other Same load, same cable, same 100 ft A 2,400 W load on 12 AWG copper. Only the supply voltage changes — and the current follows it. AT 120 V, DRAWING 20 A 7.72 V lost — 6.4333% HIGH AT 240 V, DRAWING 10 A 3.86 V lost — 1.6083% GOOD The bar is one quarter as long. The cable never changed. HALVING ONE · THE CURRENT 2,400 W needs 20 A at 120 V, but only 10 A at 240 V. 7.72 → 3.86 V (÷2) HALVING TWO · THE REFERENCE That smaller loss is measured against 240 V, not 120 V. 6.4333 → 1.6083% (÷4)
The same 2,400 W load on the same 12 AWG cable: 6.43 percent at 120 V and 1.61 percent at 240 V, a factor of exactly four.

Turn the formula around

The usual equation starts with a distance and returns a percentage. Rearranged, it answers the question people actually have when planning: the longest one-way run that stays inside a chosen percentage at a given voltage, current and conductor resistance.

percent = 2 × L × current × ohms per foot ÷ voltage × 100
percent × voltage = 200 × L × current × ohms per foot
L = percent × voltage ÷ (200 × current × ohms per foot)

At 20 A on 120 V, these are the longest one-way runs that stay inside 3 percent. Each figure is cut down to whole feet rather than rounded, because rounding up would cross the limit the number exists to respect.

Copper gaugeResistance per 1,000 ftMaximum one-way length
14 AWG3.07 ohms29 ft
12 AWG1.93 ohms46 ft
10 AWG1.21 ohms74 ft
8 AWG0.764 ohms117 ft
6 AWG0.491 ohms183 ft
4 AWG0.308 ohms292 ft

This is a voltage-drop comparison and not an installation recommendation. It shows how resistance changes the distance available under one electrical target, with current and supply voltage held fixed.

For 12 AWG the default 100 ft run is well past the 46 ft that limit allows at 20 A on 120 V. Moving the same 2,400 W load to 240 V changes the current and the percentage reference together.

120 V reach = 3 × 120 ÷ (200 × 20 × 0.00193) = 46.63 ft
240 V reach = 3 × 240 ÷ (200 × 10 × 0.00193) = 186.53 ft

The whole-foot figures are 46 ft and 186 ft. Before that cut the formula gives exactly four times the reach, for the same reason the percentage fell to a quarter: the current halved while the supply voltage doubled. A circuit losing a quarter of the percentage per foot can run four times as far before it meets the same boundary.

How far each gauge gets before 3 percent The reach of each copper gauge Longest one-way run inside 3 percent, at 20 A on 120 V. L = 3 × V ÷ (200 × I × ohms per foot). GAUGE 14 AWG 29 ft 12 AWG 46 ft 10 AWG 74 ft 8 AWG 117 ft 6 AWG 183 ft 4 AWG 292 ft the 100 ft default run Or change nothing but the supply: 12 AWG goes from 46 ft to 186 ft. The same 2,400 W load at 240 V draws half the current and is judged against twice the voltage — four times the reach.
How far each copper gauge reaches at 20 A on 120 V while staying inside 3 percent — 46 ft on 12 AWG, 117 on 8.

The current is the load, not the breaker

Voltage drop is calculated on the current the circuit actually carries. A breaker rating sets an overcurrent-protection boundary; it does not mean the circuit sits at that current, and entering it when the load never approaches it makes the run look worse than it is.

Actual currentVoltage lostDropVerdict
20 A7.72 V6.43%HIGH
16 A6.18 V5.15%HIGH
12 A4.63 V3.86%ACCEPTABLE
8 A3.09 V2.57%GOOD

At 20 A the circuit is firmly in the HIGH band, at 16 A it is still just above 5 percent, 12 A moves it into ACCEPTABLE, and 8 A brings it comfortably inside GOOD. The same cable crosses two verdict boundaries across that range with nothing about the conductor or the distance changed.

Sizing to a breaker rating the load never reaches therefore points at heavier cable than the situation calls for. The current to use comes from the load you expect, subject to the separate rules that govern circuit sizing.

What a step up in gauge buys

Larger copper has lower resistance, so every step up the gauge scale cuts the loss for the same current over the same distance. On the default 20 A, 100 ft, 120 V run the progression looks like this across six common sizes.

Copper gaugeVoltage lostDropVoltage at loadVerdict
14 AWG12.28 V10.23%107.72 VHIGH
12 AWG7.72 V6.43%112.28 VHIGH
10 AWG4.84 V4.03%115.16 VACCEPTABLE
8 AWG3.06 V2.55%116.94 VGOOD
6 AWG1.96 V1.64%118.04 VGOOD
4 AWG1.23 V1.03%118.77 VGOOD

Each step cuts resistance by close to the same proportion: 37.1 percent from 14 to 12 AWG, 37.3 from 12 to 10, 36.9 from 10 to 8, and 35.7 from 8 to 6. Because the reduction is proportional rather than fixed, the same fraction of a large loss saves more volts than the same fraction of an already small one.

The first step, from 14 to 12 AWG, saves 4.56 V on this run. The last, from 6 to 4 AWG, saves 0.73 V. The ladder flattens as it descends, so the early steps carry nearly all the benefit, while moving a suitable fixed-power load to 240 V attacks the current and the percentage reference at once instead.

A 100 ft run is 200 ft of copper You measured one way. Current goes both. The resistance in the formula belongs to every foot of conductor the current passes through. PANEL 120 V LOAD 112.28 V out — 100 ft back — 100 ft 200 ft of copper in the circuit 2 × 100 ft × 20 A × 0.00193 = 7.72 V drop the 2 and you get 3.86 V — exactly half the real loss Enter the one-way distance. The engine doubles it for you. The loss is spread along the whole path, which is why a meter at the panel reads normal on a run that is failing.
A 100 ft run is 200 ft of copper, because the current returns on the second conductor.

What this does not decide

NEC 210.19(A) carries an Informational Note recommending that branch-circuit voltage drop at the farthest outlet not exceed 3 percent, with feeder and branch-circuit drop combined kept to 5 percent, and NEC 215.2(A) carries a parallel note for feeders.

NEC 90.5 states that Informational Notes are informational only and are not enforceable as requirements. Those percentages are therefore recommendations rather than code, and the GOOD, ACCEPTABLE and HIGH labels are this calculator's own grading of them.

Energy codes and local amendments often make the 3 percent figure binding anyway, and staying inside it is defensible engineering regardless, because equipment at the far end then receives voltage closer to what it was designed for.

The figures here are copper only, and aluminium needs its own conductor data. The engine uses the NEC Chapter 9, Table 8 DC resistance for uncoated copper at 75 C rather than modelling AC impedance or the operating conditions of a particular installation.

Voltage drop settles one constraint out of several. Ampacity, temperature correction, conduit fill, overcurrent protection and applicable local rules are separate questions that a favourable percentage cannot answer or override, so treat this as a way to compare options rather than as a specification. The circuit has to be sized and installed by a licensed electrician.

Questions people ask

Why does the formula multiply the run length by two?

Because the current has to come back. It leaves the panel on one conductor and returns on the other, so a load 100 ft away sits at the end of 200 ft of copper, and the resistance the circuit fights is the resistance of all of it. That is the factor of two at the front of the equation. Leave it out and the answer is exactly half the real loss: 3.86 V instead of 7.72 V on the default run. Enter the one-way distance and let the engine double it.

Why is 240 V four times better than 120 V, and not twice?

Because two separate halvings compound. A 2,400 W load draws 20 A at 120 V but only 10 A at 240 V, and since drop is proportional to current, the volts lost fall from 7.72 to 3.86 — that is the first factor of two. The percentage then compares that smaller loss against a supply that has doubled, which halves it again. The result is 6.4333 percent against 1.6083, exactly four times better, and the verdict moves from HIGH to GOOD without changing the cable.

How far can I run 12 AWG copper at 20 amps?

On 120 V, about 46 ft if you want to stay inside 3 percent. Rearranging the formula gives the reach directly: length equals percent times voltage divided by 200 times current times ohms per foot, which for 12 AWG at 1.93 ohms per 1,000 ft works out to 46.63 ft, cut to 46 to stay under the limit. Ten AWG reaches 74 ft and 8 AWG 117 ft on the same terms. The same 2,400 W load at 240 V draws 10 A and takes that 12 AWG out to 186 ft.

Should I enter the breaker rating or the actual load?

The actual load. A breaker rating is an overcurrent-protection boundary, not a statement that the circuit runs at that current, and voltage drop is caused by the amps that genuinely flow. On the same 12 AWG over 100 ft at 120 V, 20 A gives 6.43 percent and a HIGH verdict, while 8 A gives 2.57 percent and a GOOD one. The same cable crosses two verdict bands across that range, so sizing to a rating the load never approaches points you at heavier cable than the situation needs.

Is the 3 percent rule actually in the code?

Not as a requirement. The 3 percent branch-circuit figure and the 5 percent combined figure appear in Informational Notes attached to NEC 210.19(A) and NEC 215.2(A), and NEC 90.5 states that Informational Notes are informational only and are not enforceable as requirements. So they are recommendations. In practice energy codes and local amendments often make the 3 percent figure binding anyway, and staying inside it is defensible engineering regardless, because the equipment then receives voltage closer to what it was designed for.

What does one step up in wire gauge save?

Close to a constant proportion of the resistance, which means a shrinking number of volts. Each step in this range cuts resistance by roughly 36 to 37 percent: 37.1 percent from 14 to 12 AWG, 37.3 from 12 to 10, 36.9 from 10 to 8 and 35.7 from 8 to 6. Because the cut is proportional, the absolute saving falls away as you climb. On a 20 A, 100 ft, 120 V run the first step saves 4.56 V while the step from 6 to 4 AWG saves 0.73 V, so the early steps carry nearly all the benefit.

Can I measure voltage drop at the panel?

No, and that is what makes the problem hard to spot. The loss accumulates along the conductors rather than appearing at the source, so a meter at the panel is reading the supply before any of the resistance has been passed through and shows a perfectly healthy figure. On the default run the panel reads 120 V while the load sees 112.28. The measurement that tells you anything is taken at the far end, under load, since an unloaded circuit carries no current and therefore drops no voltage.

Does a good percentage mean the wire is the right size?

No. Voltage drop settles one constraint out of several, and a favourable percentage cannot override the others. Ampacity, temperature correction, conduit fill, overcurrent protection and whatever your jurisdiction has adopted are separate rules, and a conductor can be comfortable on drop while failing one of them. The figures here are also copper only, and they use the tabulated DC resistance at 75 C rather than AC impedance. Treat the result as a way to compare options, and have a licensed electrician size and install the circuit.