How to Arrange Layers in a Rigid-Flex PCB Stackup

A rigid-flex board is built by laminating flexible polyimide layers into rigid sections, and the stackup is what makes that possible or impossible. Unlike a conventional board, the layer arrangement has to satisfy two different sets of mechanical conditions at once: a rigid section that behaves like an ordinary multilayer board, and a flexible section that has to bend without cracking the copper.

What the Construction Combines

The rigid sections are built from cores and prepreg in the usual way, so they can carry fine traces, dense component footprints and blind or buried vias. The flexible sections use polyimide film, typically one to three copper layers, and they connect the rigid islands while allowing the assembly to fold into the enclosure.

The result removes connectors and cables from the product, which is the main reason the technology is adopted. A design that was previously three boards and two cable assemblies becomes one part with no mating interfaces, which improves reliability and removes assembly labour. That benefit is paid for in material cost and process complexity, and in a design cycle that has to be longer because the mechanical behaviour has to be settled before the layout can be finalised. A rigid-flex board is rarely the fastest option to first prototype, and the schedule should be planned around the additional lamination and mechanical review steps from the beginning.

How the Layers Are Arranged

The flexible layers must run continuously through both the flexible and the rigid regions, because they are what ties the assembly together. The rigid layers are added only where components will be mounted, and they are bonded to the flexible core by prepreg during the lamination cycles.

A typical six layer rigid-flex construction might use four layers in the rigid areas and a two layer flexible core that passes from one rigid island to the other. The flexible core carries the signals that cross between the rigid regions and provides the mechanical hinge, while the rigid layers carry the local density around the components.

Material Choices

The rigid sections use an ordinary epoxy glass laminate for cost and mechanical strength, while the flexible sections use polyimide for its temperature tolerance and toughness. The adhesive that bonds them is the least reliable interface in the construction, so adhesiveless polyimide laminate is preferred for high reliability products.

Prepreg selection is more critical than on a rigid board, because it has to flow enough to fill the rigid region without flowing into the flexible region and stiffening the bend. Low flow prepreg and controlled layer thicknesses are used to keep that boundary sharp, and the flexible region is usually protected by a coverlay of polyimide film with a flexible adhesive.

Where the Bend Should Be

The bend region should be a dedicated area with no components, no plated through holes and no stiffeners, and its length should be defined so that the bend radius never goes below the minimum for the construction. As a rule, the radius should be at least ten times the total thickness of the flexible portion for a single fold, and more for a dynamic application.

Signal traces should cross the bend perpendicular to the bend line rather than running along it, because a trace running parallel to the bend is flexed over its whole length and fails earlier. Ground planes in the flexible region are usually cross hatched rather than solid, because a solid plane both reduces flexibility and is prone to cracking at the fold, and it also stiffens the hinge that the design depends on.

The Neutral Axis Matters

When the assembly bends, the material on the outside of the curve is in tension and the material on the inside is in compression. Somewhere between them is a neutral axis where the strain is zero, and conductors placed on or near that axis last far longer than conductors placed elsewhere.

Designing for the neutral axis means planning the layer order of the flexible portion so the signal layers sit near the middle of the stack, with coverlay on both sides. An unbalanced flexible section moves the neutral axis away from the conductors and concentrates strain in the copper, which is a common cause of cracking in flex circuits that are exercised.

Dielectric and Impedance Considerations

Controlled impedance is harder in a rigid-flex construction because the flexible dielectric is thinner than the rigid one and the two regions transition. The width that gives fifty ohms in the rigid section will not give fifty ohms in the flexible section, so the trace has to be adjusted at the transition or an impedance discontinuity will reflect energy.

In practice the transition is handled by tapering the trace width across a short distance, with the geometry computed from the two stackups. Simulation of the whole channel, including the transition, is the only reliable way to confirm that the impedance stays inside tolerance. Related via considerations are described in blind and buried via stack selection.

Common Stackup Mistakes

The most frequent mistake is an unbalanced construction, where the number or thickness of layers above the centre differs from the number below. That produces a board that curls during lamination and changes shape when it is heated, and it forces the fabricator to compensate with process adjustments that reduce yield.

The second is placing a plated through hole in the bend region. A barrel is a rigid cylinder inside a material that has to flex, and it cracks. The third is using adhesive bonded laminate in a dynamic application because it is cheaper, which fails at the adhesive interface rather than in the copper. A fourth is leaving the flexible region unprotected where the assembly will be handled, since a crease introduced during assembly is indistinguishable from a design fault once the board is in service.

Protection in Service

Once the assembly is folded, the flexible region is usually held in position and may be coated or potted for environmental protection. Coating a flex region adds stiffness and can change the bend behaviour, so the choice has to be made with the mechanical requirement in mind rather than applied uniformly across the board, as discussed in conformal coating and board protection.

Where a potting compound is used to protect the joints, its modulus matters: a stiff compound transmits mechanical stress directly into the solder joints it was meant to protect. The selection principles are set out in potting and dispensing adhesives.

FAQ

How many layers can a rigid-flex board have? Constructions of twenty layers and above are producible, but the flexible portion normally stays between one and three copper layers regardless of the rigid layer count, because every added flexible layer reduces the achievable bend radius.

Can rigid-flex be used for a dynamic application? Yes, with a dedicated design: rolled annealed copper, conductors near the neutral axis, a generous bend radius and a defined cycle life. A construction intended for a single fold should not be used where the assembly will move continuously.

Why is the stackup agreed with the fabricator so early? Because the layer arrangement determines the lamination cycles, the material consumption and whether the construction is producible at all. Changing it after routing has begun invalidates the geometry and the impedance calculations.