Copper thickness on flexible circuits is rarely the first parameter designers debate, yet it is one of the most frequent sources of reliability problems once the boards reach dynamic bend testing or high-current operation. Selecting the wrong FPC copper thickness forces compromises later that no amount of coverlay adjustment or stiffener placement can fully recover.
Most production flexible PCBs use one of three common weights: 1/3 oz (approximately 12 µm), 1/2 oz (18 µm), or 1 oz (35 µm). Heavier foils appear occasionally, but they are the exception and usually limited to static or low-cycle applications.
What Each Common Copper Weight Actually Delivers
1/3 oz copper is the default for high-cycle dynamic flex. The reduced cross-section keeps neutral-axis stress low and allows tighter bend radii without early fatigue cracking. It is also the foil of choice when the design already pushes the limits of polyimide thickness or when the stackup must stay under 0.1 mm total.
1/2 oz sits in the middle. It still flexes well enough for most consumer and industrial dynamic applications while offering measurable improvement in current-carrying capacity and etch uniformity on fine-pitch traces. Many mid-volume designs land here because it balances the two dominant constraints.
1 oz copper is selected when continuous current or low DC resistance is non-negotiable. The penalty is stiffness. Once copper reaches 35 µm, the minimum recommended bend radius increases and the number of reliable flex cycles drops sharply unless the copper is rolled-annealed and the coverlay is carefully matched.
How Copper Thickness Changes Bend Performance and Fatigue Life
Copper is the highest-modulus layer in a typical single- or double-sided FPC. When the circuit is bent, the outer fibers of the copper experience the highest tensile strain. Strain scales directly with thickness for a given bend radius. Move from 12 µm to 35 µm and the peak strain roughly triples under the same geometry.
That is why dynamic applications almost always start with the thinnest practical foil. Electrodeposited copper can be used, but rolled-annealed copper is preferred once the cycle count exceeds a few thousand. The grain structure of RA copper accommodates repeated plastic deformation far better than ED copper of the same thickness.
A common failure mode appears when a designer keeps 1 oz copper "for current capacity" and then specifies a 3–4× thickness bend radius. The copper cracks after a few hundred cycles, usually initiating at the edge of a plated through-hole or at a sharp change in trace width. Coverlay cannot stop the crack once it has started in the copper itself.
Current-Carrying Capacity Versus Flexibility Trade-off
Thinner copper forces wider traces or parallel paths if the same current must be carried. A 12 µm foil carrying 1 A continuous may need 1.5–2 mm of trace width depending on temperature rise limits and ambient conditions. The same current on 35 µm copper can often be handled with half that width.
Designers sometimes default to 1 oz simply because the IPC-2152 charts look more comfortable. In flexible circuits the charts still apply, but the thermal path is different: heat must conduct laterally through the thin polyimide or into the coverlay and then into air or a stiffener. Local hot spots develop faster than on rigid boards of equivalent copper weight.
When current and flex life both matter, the practical solution is often to keep the dynamic region in 1/3 oz or 1/2 oz and transition to heavier copper only in the static zones. That transition must be gradual; an abrupt step in copper thickness creates a stress riser that fails early.
Impedance Control on Thin Copper Foil
Controlled-impedance traces on flexible circuits already operate with limited dielectric thickness. Switching from 1 oz to 1/3 oz copper changes the etch factor and the final conductor cross-section. The thinner starting foil etches with less undercut, so the finished trace is closer to the drawn width. That can move characteristic impedance several ohms if the stackup was originally calculated for 35 µm copper.
Most fabricators will compensate by adjusting the artwork, but only if the copper weight is declared early. Changing copper thickness after the impedance model is locked almost always requires a new calculation and often a new dielectric thickness or trace width.
Differential pairs are especially sensitive. The narrower gap that becomes possible with thinner copper improves coupling, yet the reduced copper height also lowers the effective dielectric constant contribution from the coverlay. Both effects must be modeled together.
Dynamic Flex Versus Static Flex: Matching Copper Thickness to the Real Use Case
Dynamic flex (repeated motion, hinge, or continuous rolling) almost always demands the thinnest copper that still meets electrical requirements. 1/3 oz RA copper with a well-designed coverlay and a bend radius of at least 10× total thickness is the common starting point. Cycle life targets of 50 000–100 000 or higher become realistic only when copper is kept thin and the neutral axis is centered.
Static flex (one-time or occasional fold, installation bend, or fixed curve) tolerates thicker copper. 1 oz is routinely used when the bend occurs only during assembly and the finished product never moves again. Even here the bend radius should still follow the copper thickness guideline; a 1 oz static fold into a 1 mm radius will still crack if the copper is ED rather than RA.
A frequent DFM observation is a design that claims "static only" yet places the fold line across a region that experiences vibration or thermal expansion mismatch in the field. The copper then sees low-cycle fatigue that was never accounted for in the original thickness selection.
Practical Sequence for Selecting Copper Thickness in the Design Phase
Start with the mechanical requirement. Determine the minimum bend radius the product must survive and the expected number of cycles. That immediately narrows the copper choices. If the radius is tight or the cycle count is high, 1/3 oz or 1/2 oz is the only realistic starting point.
Next calculate the continuous and peak current for every critical net. Use the actual thermal environment of the flex region, not the rigid-board charts. If the required trace width exceeds the available routing space, either accept a higher temperature rise, add parallel paths, or move the high-current section out of the dynamic zone and use thicker copper there.
Then lock the impedance model using the chosen copper weight. Do not assume the fabricator's default etch compensation will match your original calculation. Ask for the process capability data for the specific copper thickness and dielectric combination.
Finally, confirm the copper type (RA versus ED) and the grain orientation relative to the primary bend axis. Orientation is often overlooked and can cost thousands of cycles of life.
Flexible PCB copper thickness selection is a compromise between mechanical endurance, electrical performance, and manufacturability. The thinnest foil that still meets current and impedance targets almost always produces the most reliable dynamic circuit. Heavier copper is justified only when the mechanical duty cycle is low enough to tolerate the added stiffness. Getting that judgment right early avoids the majority of copper-related failures that appear in later DFM and reliability reviews.
