Voltage drop and cable size calculator
| AWG | Diameter (mm) | Area (mm²) | Nearest metric | Typical use |
|---|---|---|---|---|
| 24 | 0.511 | 0.205 | 0.25 mm² | Signal wiring |
| 22 | 0.644 | 0.326 | 0.35 mm² | Signal and data |
| 20 | 0.812 | 0.518 | 0.5 mm² | Low-current DC |
| 18 | 1.024 | 0.823 | 1.0 mm² | Light fittings |
| 16 | 1.291 | 1.31 | 1.5 mm² | Lighting circuits |
| 14 | 1.628 | 2.08 | 2.5 mm² | Socket circuits |
| 12 | 2.053 | 3.31 | 4 mm² | Cookers, showers |
| 10 | 2.588 | 5.26 | 6 mm² | Heavy circuits, EV chargers |
| 8 | 3.264 | 8.37 | 10 mm² | Sub-mains |
| 6 | 4.115 | 13.3 | 16 mm² | Meter tails, feeders |
Current has to get to the load and back, so a 30 metre run is 60 metres of conductor. That is why the single-phase formula carries a factor of two. Three phase is different. The return currents largely cancel in the star point, so the factor is the square root of three instead, which is why three-phase distribution wins over long distances.
Bands follow the common 3% lighting / 5% total limits. Your local regulations set the figure that actually applies.
Estimates for planning and study. Sizing here is for voltage drop; a real cable selection also has to carry the current after derating for grouping, insulation, ambient temperature and installation method. The ampacity rows are free-air comparisons, not ratings, and IPC-2221 is a conservative guideline that IPC-2152 supersedes for serious thermal work.
Voltage drop on single phase is 2ρLI ÷ A, with ρ = 0.0172 Ω·mm²/m for copper. Sixteen amps over 30 m of 2.5 mm² copper drops 6.6 V: 2.9% of a 230 V supply. The same panel sizes a cable to a drop limit, converts AWG against mm², and sets a PCB trace width to IPC-2221.
How to calculate voltage drop
Voltage drop rarely trips anything; it quietly degrades everything downstream instead. Lamps run dim, motors run hot and draw more current to make the same torque, and heating elements deliver less than their rating. The classic 3% and 5% limits exist because equipment is designed to tolerate roughly that much and no more. On low-voltage systems the problem is far worse in proportion: a 2 V drop is under 1% at 230 V but nearly 17% at 12 V, which is why campervan and solar wiring is so much thicker than it looks like it needs to be.
Sizing a cable is two separate tests and the bigger answer wins. The first is current-carrying capacity: can the conductor carry the load without its insulation exceeding its rated temperature, after derating for how it is installed, how many cables share the route and how hot the surroundings get. The second is voltage drop, which is what this page computes. On short runs the current test almost always dominates; past twenty or thirty metres drop takes over and starts driving the size up quickly, because drop is linear in length.
The two size systems never line up, which is the whole reason the gauge conversions are here. AWG is a geometric series, d = 0.127 × 92^((36−n)/39) millimetres, so the number counts drawing passes and runs backwards, and every three gauges roughly doubles or halves the area. Metric sizes are a preferred-number series instead, so they interleave rather than match: AWG 14 is 2.08 mm² against a metric 2.5, and AWG 12 is 3.31 against a metric 4. The safe substitution is always toward the thicker conductor, which means the metric size above or the lower AWG number. That is the whole content of a wire size chart, computed here rather than looked up.
The trace-width mode is the same problem on a board. IPC-2221 sizes copper from cross-sectional area, using a constant of 0.048 for an external layer and 0.024 for an internal one, because a buried trace cannot shed heat to air. The temperature rise is the design choice and it is more forgiving than instinct suggests: 20 °C above ambient is normal and damages nothing on its own. What the formula does not capture is context. A trace beside a hot regulator starts from a higher ambient, and the vias carrying the same current are a far more constrained path than the trace either side of them.
What people use it for
- Sizing a cable for a long garden or outbuilding run
- Checking a 12 V or 24 V circuit in a van or boat
- Diagnosing lights that dim when a load switches on
- Comparing copper against aluminium for a long feed
- Finding the smallest cable that holds a 3% drop limit
- Converting a wire gauge into a cross-section and resistance
- Reading an American spool label in millimetres
- Specifying an American gauge for a metric cable
- Sizing a PCB power trace for a temperature rise
- Checking a conductor against the gland, crimp or conduit it has to fit
- Deciding whether cable already in the wall can take a new load
- Choosing between 1 oz and 2 oz copper for a board
- Estimating the resistance of a cable run you already have
- Choosing speaker or low-voltage DC wire from a gauge number
Questions
Commonly 3% for lighting and 5% total from the origin to the furthest point, though the exact figure is set by your local wiring regulations.
Current flows out and back, so a 30 m run is 60 m of conductor. Three phase uses a factor of √3 instead, because the return currents largely cancel.
Take the larger of two answers: the size that carries the current after derating, and the size that keeps voltage drop inside the limit. The cable size mode answers the second.
Metric runs 1, 1.5, 2.5, 4, 6, 10, 16, 25, 35, 50 mm² and up, in fixed steps, so a calculated minimum almost never lands on a real cable. Suppliers write that size as mm2 where a superscript is awkward, and it means the same thing. The theoretical figure is shown beside the standard size so you can see how much headroom the next step up buys. AWG runs the other way: a smaller number is a thicker wire.
For voltage drop, yes, and linearly. Doubling the length doubles the drop, so it doubles the area needed to hold the same percentage.
3.31 mm². The nearest metric size up is 4 mm², which is the safe substitution.
About AWG 13.2. A fractional answer is normal, because the two scales interleave and a metric size usually falls between two gauges rather than on one. Which whole gauge to buy is a separate question from which is closest. AWG 13 is 2.62 mm² and is the safe substitution, but odd gauges are rarely stocked, so the real choice is between AWG 12 at 3.31 mm² and AWG 14 at 2.08 mm². Take AWG 12: AWG 14 is thinner than the cable it is replacing, and substituting it on a circuit carrying the same current gives up the margin the original size had.
The number counts the drawing passes used to make it: more passes, thinner wire. What the gauge actually fixes is the diameter, with the cross-section following from it, and every three gauges roughly doubles or halves that area. Below 1 the scale keeps going as 0, 00, 000 and 0000, written 1/0 to 4/0 and read as "one aught" upward.
Toward the thicker conductor: the metric size above, or the lower AWG number, if the replacement has to carry the same current.
The conductor cross-section matches, but the bundle is thicker overall, more flexible, and slightly more resistive per metre. That overall diameter, insulation included, is the figure that decides whether a cable fits a gland, a crimp or a conduit; the gauge on its own does not tell you.
It depends on insulation, grouping, ambient temperature and how it is installed. The ampacity rows here are free-air estimates for comparison only.
Yes. The drop times the current is power turned into heat in the cable, and the tool shows it. Over a long run it can be tens of watts, continuously.
It has about 1.6 times the resistivity, so an aluminium conductor needs roughly 1.6 times the cross-section for the same drop. It is lighter and cheaper, which is why long feeds still use it.
Enormously. The same 2 V loss is under 1% at 230 V and nearly 17% at 12 V, which is why low-voltage cabling is so much heavier.
For 2 A on external 1 oz copper at a 10 °C rise, about 0.78 mm. Doubling the copper weight roughly halves the width. The rise is your choice: 10 °C is conservative, 20 °C is ordinary practice, and the real ceilings are the board Tg and whatever heat-sensitive part sits next to the trace.
One ounce of copper spread over a square foot, which comes out at 1.378 mil, about 35 microns thick. It is the default on most fabricators’ stackups; 2 oz is the usual step up for power boards.
They are buried in FR4 and cannot shed heat to air, so IPC-2221 halves the constant. Halving it does not halve the width, though: area carries an exponent of 0.725, so the width goes up by 2^(1/0.725), which is about two and a half times. On this panel’s defaults that is 30.8 mil external against 80.0 internal — size an internal trace at “double” and it is 23% narrow.
IPC-2152 supersedes it and is more accurate, particularly for thick copper and dense boards. IPC-2221 remains the common quick estimate.