Double-Sided Rigid-Flex PCB: Design and Build

What the Double-Sided Construction Adds

A rigid flex board combines rigid sections that carry the components with flexible sections that fold, replacing connectors and cables with copper on polyimide. In a single-sided rigid flex construction the flexible part carries one conductor layer, which limits it to a few signals and makes it impossible to route a controlled impedance line or to place components on both sides.

A double-sided rigid flex board puts copper on both faces of the flexible section as well. The flexible area can then carry more signals, can support a ground plane, and can be used where the product needs impedance controlled lines between the two rigid sections. It is the common construction in handheld and wearable products, in instruments built into curved housings and in camera and sensor modules.

The Stack in Detail

A typical double-sided rigid flex board has six layers, arranged symmetrically. Reading from the top: a coverlay over the exposed outer surface of the flex section, then a copper layer, an adhesive or a no-flow prepreg, a polyimide core, another prepreg, a second copper layer, and a second coverlay. The rigid sections use the same flexible core as their base, with additional prepreg and copper layers bonded on each side to bring them to the required thickness and layer count.

Three details decide whether the board works.

The neutral axis. In the bend area, the copper should sit as close as possible to the middle of the stack, where the bending strain is lowest. A layer that is far from the neutral axis is stretched or compressed on every bend, and copper cracks after a few thousand cycles.

The coverlay. A flexible circuit is covered by a laminated polyimide film with an adhesive rather than by solder mask, because solder mask cracks when it bends. The coverlay has openings for the pads, and those openings are part of the fabrication rather than the assembly.

The stiffeners. Component areas need support, because a flexible board cannot hold a connector or a large package flat. Stiffeners of FR-4, polyimide or aluminium are bonded to the flexible core, and the connector stiffener also carries the insertion force. Our notes on flex PCB assembly describe how the stiffened areas are handled in production.

Designing the Bend

The bend area is designed before the routing, not after it.

Bend across the traces. Traces run perpendicular to the bend line. A trace that runs along the bend direction is stretched along its length and cracks first.

Use rolled annealed copper. Electrodeposited copper has a grain structure that resists bending poorly; rolled annealed foil is ductile in the plane of the bend and survives far more cycles.

Keep the bend area free of vias and plated holes. A plated barrel is a stress concentrator and the first thing to fail. Where a via is unavoidable, it goes outside the bend.

Give the bend a generous radius. The practical minimum is a multiple of the stack thickness, and the multiple is larger for a dynamic bend, where the part flexes repeatedly, than for a static one that is folded once and fixed.

Anchor the copper. Copper features at the edge of a bend are tapered or rounded rather than square, because a square corner concentrates the stress and starts the crack.

Plan the cut-out. Where the flex section leaves the rigid section it is narrower, and the transition is a cut-out in the rigid material rather than a tapered strip. The tooling that produces that shape is part of the design.

Rigid Flex Against Connectors

The reason for a rigid flex construction is usually to remove a connector and a cable. The comparison is worth making explicitly, because a rigid flex board is expensive and a connector is cheap.

• Reliability. A rigid flex board removes a mechanical contact that can vibrate loose and a cable that can be mis-mated or damaged. In a product that is dropped or that vibrates, that is the main argument.
• Space. The fold occupies less volume than two boards and a connector, and it can be routed around a battery or a mechanical part.
• Assembly. Removing a connector removes an assembly step and a potential error, which is worth a great deal at high volume.
• Cost. The rigid flex board costs several times the two rigid boards and the connector, and it is justified by the volume, the reliability requirement or the space.
• Serviceability. A rigid flex assembly is harder to replace, which matters in products that are repaired rather than replaced.

Our notes on PCB design and layout cover the layout consequences of the construction.

Manufacturing

The flexible core is processed like a flexible circuit: the copper is imaged and etched, the coverlay is laminated with openings for the pads, and the stiffeners are bonded. The rigid sections are then built up with additional layers and laminated, which is where the process discipline matters, because the flexible material and the rigid prepreg have different flow characteristics and the two must bond without voids at the transition.

Two process points decide the yield. The no-flow or low-flow prepreg that bonds the rigid layers must not flow into the bend area, because cured resin there makes the section brittle and it will crack. And the cut-out in the rigid material must be clean, since a rough edge or a delaminated corner is a crack starter. Our notes on PCB manufacturing describe the lamination and routing controls involved.

Assembly

Assembly is where a rigid flex board is most often damaged. The handling jig supports the flexible section during paste printing and placement, the boards are panelised with the flex area held flat, and the reflow fixture holds the assembly so that the flex section is not stressed while the solder is molten.

Components are placed only on the rigid sections. Where a small part must sit on the flexible area, it is placed on a stiffened island, and the stiffener is added specifically for it. Cleaning, coating and testing all follow the same rule: support the flex section, and never let it be pulled or twisted. Our notes on PCB assembly describe the fixtures and the handling.

Reliability Testing

Three tests prove the design. A bend test flexes the flexible section to the specified radius for the specified number of cycles, and it is the only test that reveals a poor neutral axis or a square corner. A thermal cycling test verifies the plated through holes in the rigid sections and the bond between the flexible and rigid materials. And a pull test on the stiffener and the connector confirms that the mechanical attachment survives the handling the product will see.

The bend test is run on a coupon from the panel, built with the same stack and the same coverlay as the product, because the result depends on the construction rather than on the geometry of the product outline. Our notes on quality management describe how the qualification is documented.

Cost and Lead Time

A rigid flex board is expensive relative to rigid boards of the same area, because the flexible material, the coverlay, the stiffeners and the extra lamination and routing steps all add cost, and the yield is lower. In small quantities a double-sided rigid flex board commonly costs tens of dollars per piece, falling with volume but remaining several times the price of an equivalent rigid board.

Lead time is longer as well. A prototype typically takes two to four weeks rather than one, because the flexible core, the coverlay and the rigid lamination are separate process stages and each has to be scheduled. Volume production takes longer again. The schedule should be built around that rather than discovered at the first order.

FAQ

How many layers can a rigid flex board have? Six to eight layers are common, with more in complex designs. The flexible section is usually limited to two or four conductors so that it remains bendable.

What is the minimum bend radius? It depends on the stack and on whether the bend is static or dynamic. The fabricator quotes the value for the specific construction, and a dynamic bend needs a much larger radius than a fold that is made once.

Why use a coverlay instead of solder mask? Because solder mask is brittle and cracks when the board bends, while a laminated polyimide coverlay flexes with the copper.

Is a rigid flex board more reliable than a connector? In products that vibrate or are dropped, usually yes, because it removes a mechanical contact and a cable that can be damaged or mis-mated.

Conclusion

A double-sided rigid flex board carries copper on both faces of the flexible section, which adds signals, a ground plane and impedance control to a construction that folds to fit the product. Design the bend first, keep the copper near the neutral axis, use rolled annealed foil and a coverlay, stiffen every component area, and support the flexible section through assembly. The board costs several times a rigid one, and it removes a connector, a cable and the failure modes that come with them.