Rigid Flex Stackup Design: Layers, Bend Areas and Stiffeners

Why the Stackup Is the Hard Part

A rigid flex board is a single assembly that contains rigid sections carrying the components and flexible sections that fold to connect them. It replaces cables and connectors with a continuous piece of circuitry, which removes the most common failure points in a conventionally wired system and saves the space the connectors would have occupied.

The difficulty is that the same printed circuit has to satisfy two incompatible sets of requirements. The rigid sections behave like an ordinary multilayer board, with planes, impedance control and via structures. The flexible sections have to bend, which rules out most of the materials and structures that the rigid sections depend on.

Designing the transition between the two is what rigid flex stackup design really means. Getting it wrong produces boards that pass electrical test and fail at the first fold, or that develop a crack in a flex section after thermal cycling in the field.

How the Stack Is Assembled

The construction starts from a flexible core: a thin polyimide film, typically with rolled annealed copper bonded to it. Rolled annealed copper is used because it is more ductile than electrodeposited copper and tolerates bending without cracking, which is the whole point of the flex portion.

From there, the stack is built with the flex material in the bend regions and rigid material added in the stiff sections.

• Adhesiveless flex core: copper bonded directly to polyimide without an adhesive layer. Thinner, more dimensionally stable and better for fine line work and for impedance control. Also more expensive.
• Adhesive based flex core: a layer of adhesive between the copper and the polyimide. Thicker, with less predictable dielectric properties, and the adhesive is often the weak point at high temperature and in flexing.
• Coverlay or covercoat: in the flex regions, the equivalent of solder mask is a coverlay, a polyimide film with an adhesive layer that is laminated over the outer circuits. It protects the conductors and defines the bend properties. A photoimageable covercoat is sometimes used instead, with better registration but less mechanical protection.
• Bond ply: the adhesive material used to laminate layers together, in a no flow or controlled flow grade depending on whether it must not flow into the bend area.
• Rigid material and stiffeners: FR-4 or another rigid laminate added in the stiff sections, and local stiffeners where a component, a connector or a press fit part needs mechanical support.

The Bend Area Rules

The bend region has its own design rules, and they exist because the flex section is being asked to deform permanently.

• Bend radius: typically at least ten times the total thickness of the flex section for a static bend, and considerably more for a dynamic application where the board flexes repeatedly. The ratio applies to the finished thickness, including coverlay and adhesive.
• Copper orientation: copper in a bend should ideally be in the neutral axis of the stack. If a single conductor layer is present, centring it in the flex section minimises the strain it experiences.
• Cross hatch or solid ground in the bend: a solid copper plane through a bend, or cross hatched copper, changes the mechanical behaviour. Hash or cross hatch patterns are often used to keep the stack flexible while maintaining an impedance reference.
• No vias in the bend: a plated via is a rigid feature and will crack if it is placed in a region that bends.
• Trace direction: traces in a bend should run perpendicular to the bend line, so that a bending deformation stretches them along their length rather than across their width. A trace that runs parallel to the bend line is a crack waiting to happen.
• No plated through holes or component pads in the bend, and no coverlay openings that create a stiffness discontinuity at the bend line.

Layer count in the flex section is usually kept as low as possible, often a single conductor layer or two, because each additional layer increases the thickness and therefore the minimum bend radius. A common practical configuration is an unbalanced stack in the flex area, with copper on one side only, which bends more easily than a symmetric one.

Electrical Considerations

• Impedance control in flex: possible, but the dielectric thickness is thin and the material properties of polyimide differ from FR-4, so the geometry has to be calculated for the flex material rather than reused from a rigid stackup.
• Loss: polyimide has higher dielectric loss than the low loss rigid laminates, which matters on high speed signals that pass through the flex section. Where the requirement is severe, the routing strategy has to keep the critical high speed nets in the rigid sections and use the flex only for lower speed connections.
• Reference continuity: a signal crossing from a rigid section to a flex section changes dielectric environment, which is an impedance discontinuity. It has to be designed rather than ignored, particularly on fast signals.
• Ground structure in the bend: keeping a continuous reference through a bend usually means a cross hatched or narrow trace structure rather than a solid plane, which changes the impedance and has to be accounted for.

This is where the layout and the stackup interact most directly. The decision about which signals cross the flex region is an electrical decision with mechanical consequences, and it is made at the point where the stackup is defined rather than after the routing exists.

DFM Constraints That Decide Yield

• Bond ply flow: the adhesive used to laminate the rigid sections must not flow into the bend area during pressing, since it would change the bend thickness and the mechanical properties. No flow or controlled flow materials are used, and the stackup has to give the adhesive somewhere to go.
• Registration between rigid and flex sections: the flex core is thin and can move during lamination, so the alignment of the flex layer to the rigid sections requires control that a standard multilayer process does not.
• Coverlay registration: the coverlay opening has to align to the pads, and coverlay is more difficult to register accurately than photoimageable solder mask.
• Stiffener attachment: where a stiffener is added, its adhesion and its effect on the local thickness have to be considered in the assembly plan.
• Handling: a board with a thin flex section is fragile until it is assembled into its enclosure, which affects packing and transport.

None of these is unusual for a supplier who builds rigid flex routinely, and all of them are difficult for one who does not. The fabrication capability question on a rigid flex program is specific: how many flex layers, what minimum bend radius the process supports, and what the coverlay registration tolerance is.

Verification and Reliability

• Bend testing: cycling the flex section through the specified bend radius or angle and checking for conductor cracking. This is the direct mechanical verification and it is essential on any dynamic flex application.
• Cross sectioning: confirming the flex stack thickness, the copper position relative to the neutral axis, and the coverlay adhesion.
• Thermal cycling: relevant at the rigid to flex transition, where the materials change and thermal expansion is mismatched.
• Impedance verification on coupons where the flex section is impedance controlled.
• Adhesion testing of the coverlay and the stiffener bonds, which are mechanical joints that have to survive assembly and service.

The reliability evidence on a rigid flex board has a strong mechanical component that a conventional multilayer program does not require, and that evidence is part of what a supplier’s quality system should be able to produce for a comparable construction.

Cost and Where the Money Goes

Rigid flex costs more than an equivalent rigid board for several reasons: the flexible core materials are more expensive, the lamination is more complex because two different material systems are being joined, the coverlay adds process steps, stiffeners add assembly work, and the yield is lower because the flex sections are fragile and the registration requirements are tighter.

The comparison that matters is against the system the rigid flex replaces. Where it replaces a cable, a connector set and a separate board, the saving in connectors, assembly labour and the connector failure modes often exceeds the board cost difference. Where it replaces nothing and is chosen for its own sake, it is simply an expensive board. The economics work when the integration is real.

Assembly Notes

The rigid flex assembly is a conventional surface mount process applied to the rigid sections, with the flex sections adding handling constraints. The board is often held in a carrier during assembly to keep it flat, and the sequence of assembly and folding has to be planned so that the board is not folded until the components are placed.

Where the assembly is done at a partner facility, the flex construction affects how the board can be transported, panelised and handled, which is a conversation for the flex assembly process rather than a standard SMT flow.

Frequently Asked Questions

What is the minimum bend radius? Commonly around ten times the total flex thickness for a static bend, with a larger ratio for dynamic applications where the board flexes repeatedly. The ratio applies to the finished thickness including coverlay.

Can vias be placed in the bend area? No. A plated via is rigid and will crack when the flex region bends. Vias belong in the rigid sections or in flat areas of the flex.

How should traces be oriented in a bend? Perpendicular to the bend line, so the bending deformation acts along the trace length. A trace running parallel to the bend line is at high risk of cracking.

Can impedance be controlled on the flex section? Yes, but the geometry has to be calculated for the polyimide and the thin dielectric, and the reference structure is usually a cross hatched or patterned plane rather than a solid one.

Why is rigid flex so expensive? The combination of flexible core materials, complex lamination of two material systems, coverlay processing, stiffeners and lower yield. It pays back where it replaces cables and connectors rather than purely adding cost.

Summary

A rigid flex stackup is two designs in one board: rigid sections that behave like an ordinary multilayer circuit, and flexible sections that have to bend without cracking. The stackup is built from a polyimide flex core with rolled annealed copper, joined to rigid laminate in the stiff areas, with coverlay protecting the flex conductors and bond ply holding the layers together.

The bend area has its own rules, and they are mechanical rather than electrical: a bend radius of at least ten times the flex thickness, copper placed near the neutral axis, no vias, traces perpendicular to the bend line, and no stiffness discontinuities at the bend boundary. Most rigid flex failures originate from breaking one of those rules rather than from an electrical problem.

Electrically, the flex section introduces a material change, so impedance has to be calculated for the polyimide, reference structures in bends are patterned rather than solid, and high speed nets are usually kept in the rigid sections where the laminate can support them. The stackup decision therefore determines both whether the board survives being folded and whether the signals that cross it behave, which is why it belongs with the design review rather than after it.