Designing a Heavy Copper PCB With 4 oz Copper

Heavy copper boards exist because current has to go somewhere. When a design has to carry tens of amperes, the trace widths required on one-ounce copper become impractical: a 20 A trace at a 10 degree rise would be wider than the connector pitch that feeds it. Thicker copper solves the problem by increasing the cross-sectional area of the conductor without increasing its footprint, and it brings a set of manufacturing constraints along with it.

This article covers what 4 oz copper means in practice, how it changes current capacity and thermal behaviour, and which design rules have to be adjusted for the process.

What 4 oz Copper Means

Copper weight is specified as the mass of copper per unit area before etching. One ounce per square foot corresponds to a thickness of about 35 microns, so 4 oz copper is roughly 140 microns, or 0.14 mm, of copper on the finished layer. That is four times the conductor cross-section of a standard board.

The distinction between the starting weight and the finished thickness matters. Heavy copper is usually plated up as well as laminated thick, so the finished conductor can be thicker than the starting foil, and the trace width after etching depends on the amount of sidewall etching, which is proportionally larger for thick copper.

Current Capacity and Temperature Rise

The current a trace can carry is set by the allowable temperature rise, not by a fixed number. A wider trace dissipates more heat and runs cooler, and a thicker trace has more copper to heat. Doubling the copper thickness roughly doubles the current for the same rise, and so does doubling the width.

In practice this means a 4 oz trace about 2.5 mm wide carries on the order of 20 A at a 10 degree rise on an outer layer, while the same current on 1 oz copper requires roughly 10 mm of width. On inner layers the figure falls, because the laminate conducts heat away less effectively than air, and the derating is commonly 30 to 50 percent.

Thermal Spreading Under Power Devices

Heavy copper is as much a thermal decision as an electrical one. A wide 4 oz plane under a power device spreads heat laterally far better than a thin one, which lowers the junction temperature without a heatsink and reduces the thermal gradient across the board.

The thermal benefit is greatest in surface-mount power stages, where the copper area is the only heatsink available. It also helps in boards with buried resistors or shunts, where the heat generated inside the laminate has to be conducted out to the surfaces before it can escape.

Minimum Width and Spacing Limits

The cost of heavy copper is resolution. Etching 140 microns of copper leaves more undercut and produces a less precise edge than etching 35 microns, so the minimum trace width and the minimum spacing both increase. A process that holds 0.1 mm on 1 oz copper may only hold 0.2 mm on 4 oz, and the tolerance on that width is wider.

The same applies to drilling. Thick copper is harder to drill cleanly, and the plating current has to be spread more carefully to fill a via barrel surrounded by a large copper mass. Vias in heavy copper areas frequently need a thermal relief, or the copper sinks heat away during soldering and produces a cold joint.

Etching and Registration Challenges

Uniform etching across a panel is difficult when the copper weight is high and the pattern is uneven. Areas with dense fine features etch differently from areas with large solid pours, because the etchant becomes locally depleted. The usual remedy is to balance the pattern by adding thieving bars and by keeping the copper distribution as even as possible.

Registration follows from the same problem. If the copper thickness varies across the panel, the etch factor varies with it, and the finished trace width varies as well. Fabricators who build heavy copper regularly will specify a wider tolerance on trace width and will expect the designer to accept it, which in turn means the current calculation should be made with the worst-case width rather than the nominal one.

Solder Mask, Finish and Assembly

Solder mask does not cover heavy copper as evenly as it covers thin copper. On a thick trace the mask is thinner at the edges of the trace and thicker in the middle, and on very thick copper it may not cover the sidewall at all. Surface finish choice is affected too: some finishes, such as thin immersion coatings, do not build up enough to protect the exposed copper edges.

Assembly is where heavy copper boards most often surprise designers. Large copper areas draw heat away from a solder joint, so the reflow profile has to be adjusted and the soak time extended. Hand soldering onto a heavy plane is impractical without a high-thermal-mass iron, and wave soldering may need preheating on both sides. Thermal reliefs on pads connected to heavy planes solve most of these problems and should be planned from the start.

Mixed Copper Weights in One Stack

Not every layer has to be heavy. A common construction uses 4 oz copper on the outer layers for current and thermal spreading, with 1 oz or 2 oz on the inner layers for signal and control. This keeps the layer count and the cost down while giving the power path the copper it needs.

The stack has to be balanced for lamination, however. A board with 4 oz on one outer layer and 1 oz on the other will bow during reflow, so heavy copper is usually applied symmetrically. Where the electrical requirement is asymmetric, a heavy copper plane can be added on the opposite side and left unconnected as a balancing layer.

Design Rules to Adjust

The three rules that change most are minimum width, minimum spacing and via geometry. Current calculations should use the trace width and current relationship with the worst-case etched width, not the nominal value, and the derating for inner layers should be applied explicitly rather than assumed.

Where large areas of heavy copper are used for current rather than for shielding, the choice between a solid pour and a meshed one is worth considering: a mesh is easier to process and adheres better through thermal cycling, while a solid pour carries more current and spreads heat further. The comparison of mesh and solid flooding sets out the trade-offs. Beyond current and thermal, the manufacturable design rules still apply, with the dimensions relaxed to match the heavy copper process.

Cost and When Heavy Copper Pays

Heavy copper costs more for the material, for the slower etching, for the extra care in drilling and plating, and for the reduced yield. A 4 oz board can cost two to four times as much as an equivalent 1 oz board, and the difference grows quickly beyond 6 oz.

The justification is usually system level. Replacing a busbar, a copper strap or a heatsink with copper on the board reduces assembly steps, removes a mechanical joint and improves thermal performance at the same time. Where the alternative is a hand-assembled cable harness or an aluminium spreader, the heavier board is frequently the cheaper solution, and it is almost always the more reliable one.

FAQ

How much current can a 4 oz trace carry? As a rough guide, an outer-layer trace 2.5 mm wide carries about 20 A at a 10 degree rise, and an inner-layer trace of the same size carries roughly half that. The figures depend on the allowable rise and the surrounding copper, so the calculation should be repeated for each design.

Can heavy copper be combined with fine-pitch components? With difficulty. The minimum width and spacing on 4 oz copper are larger than the pads of a fine-pitch device, so the fine-pitch area is usually placed on a layer that carries 1 oz or 2 oz copper and connected to the heavy layer with thermal-relieved vias.

Is 4 oz copper suitable for high-frequency signals? Generally no. Thick copper has a rougher surface after etching, which increases conductor loss at high frequency, and the wider minimum geometry makes controlled impedance difficult to hold. Heavy copper is best reserved for power, return and thermal layers, with the signal layers kept thin.