Four Layer Flex and Rigid Flex Boards

Flexible circuits are used because they bend, and bending is exactly what limits how many layers they can carry. A four layer flex is routine in modern designs, but it is a different construction from a two layer flex with extra copper, and the differences show up in the bend radius, in the materials and in what the finished assembly can survive.

What Changes When a Flex Circuit Gains Layers

A single sided or two sided flexible printed circuit bends around a thin core, which is the property that makes it flexible. Adding layers means adding dielectric and copper, and copper does not stretch. The neutral axis of the stack moves, the material on the outside of a bend is placed in tension, and the achievable bend radius increases.

The practical consequence is that a four layer flex is usually used as a rigid flex, in which the bending is confined to designated regions where the stack is reduced to fewer layers. The layers that carry the extra routing are kept out of the bend, and the flexible area behaves like a two layer circuit.

Rigid Flex Construction

In a rigid flex design, the stack includes rigid sections with all layers and flexible sections where only some layers continue. The transition between them is where most of the mechanical risk lives, because the copper and the dielectric change thickness across a short distance and the bend concentrates there.

The rigid sections are laminated like a conventional board, and the flexible sections are covered with an adhesive-backed coverlay instead of solder mask, because a mask does not tolerate repeated bending. Our notes on rigid flex layer stackup design describe how the transition is constructed and where the copper has to be kept away from the bend.

Bend Radius and Layer Count

Bend radius is specified as a multiple of the total thickness, and the multiple depends on whether the bend is static or dynamic. A circuit bent once during assembly can tolerate a tighter radius than one that flexes in service, and a dynamic application requires a much larger radius as well as thinner materials.

Adding layers increases the thickness, so a four layer flex with the same materials has a larger minimum radius than a two layer circuit of the same overall area. Where the mechanical envelope is fixed, the answer is usually to reduce the number of layers in the bend rather than to accept a radius the material cannot hold.

Materials and Adhesives

The flexible core is typically polyimide, chosen for its temperature range and its mechanical durability. The adhesive used to bond the layers, however, is often the limiting element: an acrylic adhesive is more compliant than an epoxy but has a lower temperature rating, and the choice affects both the bend behaviour and the assembly process.

Adhesiveless constructions are used where the thickness has to be minimised or the thermal performance maximised, and they behave differently in bending because the copper is bonded directly to the polyimide. The choice is made with the fabricator, since the availability of the material constrains the design.

Copper Thickness and Bend Performance

Copper is the least compliant material in a flexible stack, and its thickness is the single largest influence on bend performance. A one ounce copper layer will crack at a radius where a half ounce layer survives indefinitely, which is why flexible circuits sometimes use thinner copper than the current calculation alone would suggest.

Where the current requires thick copper, the usual solution is to keep the thick copper out of the bending area and to use a thinner layer, or a wider trace, in the region that flexes. Rolled annealed copper is also used in dynamic applications because its grain structure tolerates repeated bending better than electrodeposited copper.

Coverlay and Access Openings

The coverlay protects the conductors and defines the access openings for pads, and it is thicker and less precise than a solder mask. The openings have to be larger than the mask openings on a rigid board, and the registration tolerance is wider, which affects how tightly the pads can be spaced in the area where the coverlay ends.

Coverlay also has to be kept away from the flexing region. An opening in the coverlay near a bend creates a step that concentrates stress, so pads and their openings are placed in the rigid section or in a flat area of the flexible section.

Designing the Transition Zone

The transition from rigid to flexible is the most heavily engineered part of the design. Copper is removed from the flexible layers near the transition, the material is stepped gradually rather than abruptly, and stiffeners are sometimes added to control where the bend begins. Traces crossing the transition are routed perpendicular to the boundary so that they are not loaded in shear.

Plated through holes are a particular concern. A via in the flexible section near a bend is a stress concentrator, and the barrel can crack. The conventional rule is to keep all plated holes in the rigid sections and to make any connection in the flexible area a simple trace transition.

Assembly Considerations

A flexible circuit is usually assembled flat and bent afterwards, or held flat in a fixture for assembly and formed at the end of the process. The sequence matters, because a circuit bent before reflow will not sit flat on the stencil and the paste deposit will be uneven. Our notes on flexible PCB copper pour techniques describe how the copper pattern in the flexible area is arranged to keep the circuit flat and to avoid stiffening it unnecessarily.

Where the circuit will be formed repeatedly during test, the handling method should be designed rather than improvised. Bending a flex circuit by hand at a radius tighter than the specification is one of the most common ways to destroy an otherwise sound assembly.

Dimensional Stability and Processing

Polyimide moves more than FR-4 during processing, and a flexible circuit with more layers accumulates more of that movement. Registration between layers therefore has tighter limits, and the pad sizes have to absorb the variation. Our notes on PCB dimensional stability and expansion describe the same effect in rigid boards, and the flexible case is a more extreme version of it.

Handling is part of the process as well. A thin flexible circuit is easy to crease, and a crease cannot be removed. Specifying how the circuits will be packed, supported and moved between operations is a small administrative step that prevents a large yield loss.

When Four Layers Are the Right Answer

Four layers are justified when the routing cannot be completed on two, or when a controlled impedance or a reference plane is required in a circuit that must still flex. They are not justified by a desire for a ground plane in a circuit that can be routed on two layers with a wider design.

The decision should be made with the mechanical requirement in hand. If the bend radius is generous and the circuit is static, four layers may be entirely practical. If the circuit must flex in service, the layer count and the materials should be reviewed together with the fabricator before the design is frozen.

FAQ

Can a four layer flex bend like a two layer flex? Not at the same radius with the same materials. The extra copper and dielectric increase the thickness, and the minimum bend radius grows with it.

Are vias allowed in the flexible area? They are best avoided. Plated holes concentrate stress where the material bends, and the usual practice is to keep them in the rigid sections and connect through the flexible area with traces.

Why use coverlay instead of solder mask? Because solder mask cracks when it bends, while a coverlay is a thicker film that tolerates flexing. It costs resolution and registration tolerance, which is why pad spacing in the flexible area is more generous.