Material and Layer Planning During PCB Design for Rigid-Flex Boards
A rigid-flex board is two products in one: a stiff, densely populated rigid section and a thin, bending flex tail that has to survive every fold, every connector insertion and every thermal cycle the product sees. Long before the first trace is routed, the material set and the layer stack have already decided how the board will behave. Getting those two decisions right is where rigid-flex projects succeed; getting them wrong is why boards that pass electrical test still crack at the bend line or fail after a few hundred flex cycles.
The practical question is not which material is best in the abstract, but which combination of film, copper, adhesive and stiffener delivers the mechanical behaviour the product needs at a cost the program can carry.
Selecting Base Materials and Copper Weights for the Flex Portion
The flexible core of a rigid-flex board is almost always a polyimide film. Kapton and comparable polyimide materials dominate because they hold their mechanical and dielectric properties across a wide temperature range and tolerate repeated bending. Thickness is the first lever available. Standard films run from 1/2 mil (12.5 μm) through 1 mil (25 μm), 2 mil (50 μm), 3 mil (75 μm) and 5 mil (125 μm). Thin film bends more easily but carries less mechanical strength; thicker film is tougher but pushes the bend radius out. Choosing film thickness is therefore not really an electrical decision at all — it is a decision about how tight the finished product needs the fold to be.
Coverlay, typically 1 mil, 2 mil or 3 mil, protects the flex traces and is usually specified selectively. Selective or partial coverlay — covering only the areas that need protection and leaving others open for stiffeners or contacts — is one way to control both stiffness and cost. Adhesive layers, available from 1/2 mil to 4 mil, add thickness and compliance but also introduce a material that must survive thermal cycling; where possible, modern practice leans toward adhesiveless substrates, and IPC-2223C specifically calls for eliminating or minimizing adhesive within rigid areas. Pressure-sensitive adhesives at 1 mil to 5 mil serve a different purpose, bonding stiffeners and shielding films rather than laminating the core.
Copper weight matters just as much. The available range spans 1/4 oz (9 μm) through 1/3 oz, 1/2 oz, 1 oz, 2 oz, 3 oz, 5 oz, 7 oz and 10 oz (356 μm). In a flex region, heavy copper is a liability: it is stiff and it fatigues. A common approach is to hold flex copper to 1/2 oz or 1 oz and reserve heavier copper for the rigid areas, since the two zones do not have to share a copper weight. Stiffeners complete the material picture: polyimide from 1/2 mil to 10 mil, FR4 from 0.005 in to 0.125 in, stainless steel up to 20 mils, plus copper and aluminum options. A polyimide stiffener behind a ZIF finger area, an FR4 stiffener under a component area, EMI or RF shielding films, and PSA attachment are all standard additions to an otherwise simple flex core.
Layer Stack-Up Planning: Keeping Rigid and Flex Sections in Balance
Stack-up planning for rigid-flex starts with a cost rule that is easy to remember: an even number of layers is the most cost-effective construction, and every rigid portion of the board should carry the same number of layers and the same stack-up. Break either rule and the transition zones become asymmetric, which shows up later as bow, twist, or a lamination that will not sit flat in the assembly fixture.
Standard constructions scale from a simple two-layer build and a three-layer build up to four, five and six layers, then into HDI territory: a six-layer any-layer HDI stack-up, an eight-layer HDI build with blind and buried vias, an eight-layer any-layer HDI stack-up, and a twelve-layer any-layer HDI construction that includes ZIF contacts. Each step up adds routing capacity but also adds thickness in the flex, which converts directly into a larger minimum bend radius. That trade-off is far cheaper to settle on paper than in a re-spin.
Symmetry is the second principle. Copper distribution, dielectric thickness and coverlay placement should mirror around the neutral axis wherever the geometry allows, so the flex does not curl on its own when it is released from the press. Where the flex carries fewer layers than the rigid sections — a common arrangement, since many flex tails are single-layer or two-layer — the transition zone needs explicit attention. Copper should be balanced within the flex itself, and any plane acting as an impedance reference must remain continuous through the bend region, or the return path breaks exactly where the signal is most vulnerable.
For products that need a sharp fold, a bookbinder construction with unbonded flex layers can compress below the usual radius-to-thickness ratio, but it requires progressive layer lengths calculated per IPC-2223 and it is expensive to tool, difficult to process and lower yielding. It is a deliberate answer to a deliberate constraint, not a default. On the HDI side, microvia stacking should be defined so the drilling and plating sequence stays within the fabricator’s process, and blind or buried vias should terminate in the rigid sections rather than crossing into the flex window.
Design Rules and Mechanical Constraints That Protect the Flex Region
Once the stack is settled, geometry takes over. A small set of spacing rules keeps mechanical stress away from the places where the flex meets the rigid board. Holes should sit at least 0.9 mm from the flex area, and the flex area itself should be at least 2.0 mm long so the bend has somewhere to happen. Pads on the rigid side need at least 0.8 mm of clearance from the flex boundary, while pads inside the flex need at least 1.5 mm from the rigid material. These are not stylistic preferences; they describe where laminate, adhesive and copper will delaminate first if the transition is crowded.
Bend radius is the other governing number. For a single-layer rigid-flex the minimum radius is 6 × T, for two layers it is 10 × T, for multilayer constructions 10–15 × T, and for a dynamically loaded design that flexes repeatedly, 25 × T. Here T is the total flex thickness — roughly 150 μm for a single-layer board and 200 μm for a two-layer board. A two-layer flex at 200 μm therefore needs about 2000 μm, or 2 mm, of bend radius. Treating those figures as hard constraints during pcb design is what separates a board that survives qualification from one that passes first article inspection and then fails in the field.
Copper planes inside the flex deserve their own decision. A solid copper plane is the least flexible option but gives the best shielding and the cleanest impedance reference; a cross-hatched plane is more flexible but can leak EMI depending on hatch pitch. Hatched planes are common in flex and rigid-flex work precisely because they raise flexibility without abandoning the reference plane entirely — the designer simply has to accept the shielding trade-off knowingly rather than by accident.
Finally, define constraint areas explicitly instead of letting the layout imply them. A rigid-flex board contains several distinct mechanical zones — rigid core, stiffened contact area, bend region — and each carries its own requirements that the mechanical designer should document before routing begins. Because material availability drives so much of this, it pays to confirm the stack against the fabricator’s listed films, coverlays, adhesives, stiffeners and copper weights early. Where a required thickness falls outside a standard range, the question is settled in a short conversation rather than a re-spin, and the flex region keeps the geometry it needs.
Bucharest cybersecurity consultant turned full-time rover in New Zealand. Andrei deconstructs zero-trust networks, Māori mythology, and growth-hacking for indie apps. A competitive rock climber, he bakes sourdough in a campervan oven and catalogs constellations with a pocket telescope.