PCB copper weight is quoted in ounces, which is a mass per unit area rather than a dimension, and the familiar thickness figures are simply that mass spread evenly over a square foot. What ends up on your finished board differs from the nominal because etching removes copper from the sides of every conductor and plating adds copper to outer layers and via barrels. Designing as though the number is a guaranteed thickness overstates what you have.
Where the ounce figure comes from
Historically, copper foil was specified by the weight of copper covering one square foot, and one ounce per square foot corresponds to a uniform thickness of roughly thirty-five micrometers, or about one and a half thousandths of an inch. Two ounces is double that, half an ounce is half. The convention persists because foil is still manufactured and purchased that way, and because the conversion is straightforward. What the figure describes is the starting foil, before any of the processes that make it into a circuit. Fabricators also increasingly quote thickness directly in micrometers, which removes the ambiguity, and where a datasheet or a fabrication note offers both it is worth confirming that they agree.
Etching narrows the conductor as it defines it
Making a trace means removing the copper around it, and the etchant attacks sideways as well as downward. The result is a conductor with a trapezoidal cross-section rather than a rectangular one: wider at the base where it meets the laminate and narrower at the top. Thicker copper requires longer etching, so the sideways loss is greater and the trapezoid is more pronounced, which is why heavy copper boards need wider spacing between conductors and cannot achieve the fine features that thin copper allows. The practical consequence for design is that a trace drawn at a given width has a slightly smaller cross-sectional area than the drawing implies, and therefore slightly more resistance and slightly less current capacity than a naive calculation suggests.
Outer layers get plated, inner layers do not
This distinction surprises people. On a multilayer board the inner layers are the original foil, etched and laminated, with nothing added. The outer layers go through the plating process that forms the via barrels, and that process deposits copper across the whole outer surface as well, so outer conductors finish thicker than the foil they started as. A board specified with a particular starting weight on the outer layers therefore ends up with more copper there than the number suggests, and how much more depends on the fabricator’s process. That is why fabrication notes often distinguish between base copper and finished copper, and why a current capacity calculation should use inner-layer figures for inner traces and finished figures for outer ones.
Why the distinction matters for real designs
Anything depending on conductor cross-section is affected. Current capacity and temperature rise depend on area, and inner layers have both less copper and worse heat removal, which is why published guidance treats them separately. Resistance matters for voltage drop and for any current-sensing arrangement that relies on a copper element. Controlled impedance depends on conductor geometry, so the etched trapezoid and the plated thickness both feed into the impedance the fabricator achieves, which is one reason impedance should be specified as a requirement for the fabricator to meet rather than calculated once and assumed. Thermal spreading through copper pours likewise depends on real thickness, and a design relying on a pour to cool a component should use realistic figures.
Specifying so the fabricator delivers what you need
Say what you need rather than what you assume. If current capacity depends on it, state the required finished copper thickness on inner and outer layers explicitly in the fabrication notes. If impedance matters, specify the target impedance and tolerance and let the fabricator adjust geometry to achieve it with their stack-up. Ask for the stack-up drawing and check it. For heavy copper work, confirm the minimum trace and space the fabricator can achieve at that weight before completing the layout, because the design rules tighten considerably. And when a design is transferred between fabricators, re-verify rather than assuming the new supplier’s process produces the same finished result, since plating practice differs between shops.
What to check before you act
- State finished copper thickness, not just an ounce figure, where it matters.
- Use inner-layer values for inner traces and finished values for outer traces.
- Confirm minimum trace and space at your copper weight before routing.
- Request the stack-up drawing and verify it against your assumptions.
Common questions
Is two-ounce copper always twice as good for current?
It roughly doubles the cross-section, which helps considerably, but heat removal, spacing rules and etching tolerance all change too. Expect improvement rather than exact proportionality.
Does copper weight affect impedance?
Yes. Conductor thickness is part of the geometry that sets characteristic impedance, so changing copper weight changes impedance unless the trace width and stack-up are adjusted to compensate.
The practical takeaway
Treat the ounce figure as a starting foil specification rather than a finished conductor dimension. Where current, resistance or impedance depend on the copper, state the requirement in the fabrication notes, distinguish inner from outer layers, and confirm the achieved stack-up with the fabricator before committing to a build.
References and further reading
- NFPA: electrical safety research and guidance on overheating equipment and fire risk
- OSHA: wiring design and protection requirements for electrical installations
- Department of Energy: how electronics losses turn into wasted energy and heat
Numerical scenarios are illustrative unless identified otherwise. Follow the exact product instructions; component ratings and local installation requirements can differ.
