PCB Copper Weight vs Trace Width: A Current Sizing Guide

Copper weight and trace width are two ways of solving the same problem: getting current from one place to another without overheating the conductor or dropping so much voltage that the load misbehaves. Copper weight sets how much metal is available per unit of width, trace width sets how much of that metal the current actually uses, and the pair together determine resistance, temperature rise and the space the routing consumes.

How Copper Weight Is Defined

Copper weight is quoted in ounces, meaning the weight of copper spread over one square foot. One ounce works out at roughly 35 microns or 1.4 mil, half an ounce is about 17 microns, and the values double in thickness as the weight doubles: two ounce is around 70 microns, three ounce about 105 microns, and the heavy end of the range runs to six and ten ounces.

One ounce is the industry standard and the value most boards are built with. Two ounce appears wherever a supply rail carries real current, and three ounce and above is the territory of power converters, battery management and motor drives. The manufacturing consequences grow faster than the number, because etching a thick layer takes longer and removes more material sideways, so minimum width and spacing both widen.

The Relationship Between Weight and Width

For a given current and a given allowed temperature rise, doubling the copper thickness roughly halves the width required. A design that needs 120 mil of one ounce copper to carry ten amps can use about 60 mil at two ounce and roughly 40 mil at three ounce. That is the whole reason heavy copper exists: it makes high current routing possible in a space that a thin layer could not serve.

The relationship is not exactly linear in practice, because a wider trace has more surface area to lose heat while a thicker one loses heat mainly through the laminate, and because the temperature rise depends on the surrounding copper and the airflow. Modern current capacity tables account for those effects rather than assuming a simple geometric ratio, and the calculation method is set out in trace width and current.

Using Capacity Tables

A current capacity table is a starting point, not an answer. The published figures assume a stated temperature rise, usually ten degrees Celsius, an external layer and still air. Move the same trace to an inner layer and it loses the convection path, so the required width increases noticeably. Add airflow or a large copper area and the requirement falls. The table tells you the order of magnitude; the environment decides the final number.

Voltage drop is the check that is most often forgotten. A trace that is thermally comfortable can still lose enough voltage over a long run to push a load outside its specification, and that failure has nothing to do with temperature. For a low voltage rail it is usually the binding constraint, so the calculation should be run for resistance first and temperature second whenever the supply is below a few volts.

IPC-2152 and the Older Method

The earlier standard modelled current capacity from a small set of experiments and treated the board as a fixed thermal environment. The newer method derives the values from a much larger body of measured data and, crucially, allows for the effect of copper area, board thickness, airflow and the presence of adjacent conductors. For modern dense boards the difference is significant, often in the direction of allowing a narrower trace than the old charts implied.

That does not mean the older figures are wrong; they are conservative. Where a design is cost sensitive and space constrained, using the newer method can save board area or allow a thinner copper weight. Where the product is safety related, the margin that the older method provides may be exactly what the design wants. The important thing is to state which method the calculation used.

Choosing Copper Weight Against Width

Widening a trace is usually cheaper than increasing copper weight. Extra width costs board area, which may or may not be available, while extra copper weight changes the whole panel: etching, lamination, solder mask and cost all move together. Where the layout has room, use it. Where it does not, and the current is genuinely large, heavy copper is the tool that solves the problem.

The practical sequence is to start with one ounce, size the traces for the current and the allowed temperature rise, and check the voltage drop. Only where the required width does not fit, or where the trace has to be short and very wide for thermal reasons, is it worth moving to two or three ounce. Spreading current between two layers is a third option that is often overlooked and costs nothing but via area, and the trade between a solid and a meshed plane is described in copper flooding practice.

Design Practices That Save Cost

Size for the real current plus a margin, typically 120 to 150 percent of the calculated value, and then stop. Over-sizing a trace costs area that could carry another signal, and on a dense board the second signal is often more valuable than the extra thermal margin. Where the current is intermittent, the thermal time constant of the trace means short pulses can be carried by a narrower conductor than a continuous load, provided the duty cycle is documented.

Work on the constraints together rather than in sequence. The copper weight, the trace width, the allowed temperature rise, the voltage drop budget and the available area are one calculation, and fixing any of them before the others are known usually forces a redesign later. Where the design has a manufacturability review, the review is the right place to confirm the numbers, because a fabricator can point out that a required width is below the minimum for the chosen copper weight.

Verifying the Result

Simulation is cheap and worth the time on any board with a supply rail. A thermal solver can show the temperature of a trace under load with the actual copper area and airflow of the design, which is far more informative than a table entry. Where a solver is not available, an infrared measurement on a prototype under load gives the same answer for the cost of one board.

It is also worth checking the result against the fabrication limits before the layout is released. A width that is electrically correct can still be below the etching minimum for the chosen copper weight, and the fabricator will either reject it or widen it without asking. Reviewing the design against the process capability, as described in manufacturable design guidelines, catches that conflict while it is still cheap to fix.

FAQ

How much current can a 10 mil trace carry? On one ounce external copper with a ten degree rise, roughly one amp. The figure falls on an inner layer and rises with airflow or adjacent copper area, so the environment has to be stated before the number means anything.

Does doubling copper weight halve the required width? Approximately, for the same current and temperature rise. The relationship is not exactly linear because heat loss depends on the surroundings, but it is close enough to use as a first estimate.

Should I use the newer current capacity standard? Yes for new designs, because it accounts for copper area, airflow and board thickness that the older method ignored. Keep the older figures as a conservative cross-check on safety related designs.

Why does heavy copper cost so much more? Because etching, plating and lamination all slow down and lose yield as the layer thickens, and the minimum line width and spacing that can be produced both increase. The material premium is usually the smaller part of the increase.