Rigid-Flex Bend Radius: Designing the Transition Zone

A bend radius is the radius of the curve a flexible circuit is formed into, and it is the single number that most strongly determines whether the circuit survives. Bending stretches the copper on the outside of the curve and compresses the copper on the inside, and the strain that results is what limits the life. The number is quoted by the designer, and it is only achievable if the stack and the process allow it.

The radius is a design decision, and it interacts with the thickness of the base material, the thickness of the coverlay and the position of the conductors. Changing any of the three changes the strain, which is why a bend radius quoted without the rest of the stack means very little. The same circuit formed over a smaller mandrel fails sooner, and the failure is not always immediate.

What Happens When a Flex Circuit Bends

The layers of a bent circuit follow arcs of different lengths. The layer on the outside travels further than the layer on the inside, and the difference has to be absorbed as strain in the materials. That difference in arc length is proportional to the distance from the neutral axis.

Copper tolerates far less strain than polyimide before it cracks, so copper is the layer that fails first. The failure is a fatigue crack that grows over cycles, and it appears at the point of maximum strain rather than at a weak spot in the pattern. A crack in a flexible circuit is difficult to see, because the copper is covered on both sides.

Strain and the Neutral Axis

The neutral axis is the surface within the stack that is neither stretched nor compressed. A conductor placed on that axis experiences no bending strain at all, which is the ideal case for a dynamic bend. Placing the conductors there is a stack-up decision rather than a routing decision.

The position of the axis is set by the thickness of the layers on each side of it. A coverlay thicker than the base moves the axis toward the coverlay, and a thicker base moves it toward the base, and the two cases give different strain in the copper. A coverlay that is too thin leaves the copper near the outside surface of the bend.

Why the Transition Zone in a Rigid-Flex Is Critical

The transition zone is where a rigid section meets a flexible one. The rigid material has a much higher stiffness, so bending is concentrated in the flexible part immediately beside it. The concentration is a function of the stiffness ratio between the two sections.

That concentration means the effective radius at the transition is smaller than the radius of the free flexible section. A stack behaves in the same way in the rigid region, and the difference is the stiffness change at the boundary. The rigid side of the stack is described in the notes on layup symmetry control. The remedy is to move the bend away from the boundary rather than to reduce the radius.

Coverlay and Bend Radius

The coverlay takes part in the calculation twice, through its thickness and through its stiffness. A stiff coverlay close to the neutral axis has little effect, while the same coverlay further from the axis increases the strain in the copper beneath it. Thickness and stiffness are sometimes traded against each other by changing the material.

Where the bend is dynamic, the coverlay is chosen to be as thin as the electrical requirements allow. A coverlay bonded with voids behaves differently again, because the voids remove the constraint the film would otherwise apply. A void under a coverlay at a bend is also a starting point for a crack in the film itself.

Copper Orientation and Grain

Rolled annealed copper is used in flexible circuits because its grain structure tolerates bending far better than electrodeposited copper. The difference matters more in a dynamic bend than in a static one. The rolling direction is set by the foil supplier and it is not changed by the circuit design.

The direction of the grain relative to the bend matters as well. Bending across the grain and bending along it give different fatigue lives, and the preference is stated in the material specification rather than in the design rules. A bend along the grain is often preferred for a dynamic application, and the drawing should say so.

Static Versus Dynamic Bends

A static bend is formed once and held, so it only has to survive one excursion. A dynamic bend is cycled continuously, and the number of cycles is the specification, which may be in the thousands or in the millions. The cycle count is often the most demanding requirement in the whole product specification.

The two cases have different rules. A static bend tolerates a radius that would fail immediately in a dynamic one, and a design that treats the two as the same will be either over-constrained or short lived. A design that meets the static requirement and is then cycled in service fails in the field.

Design Rules and Their Limits

A common rule states the minimum bend radius as a multiple of the total thickness of the flexible stack. The multiple is between five and ten for a static bend and much larger for a dynamic one. It also depends on whether the stack includes adhesive layers that are softer than the film.

The rule is a starting point, because it excludes copper thickness, coverlay stiffness and the position of the copper. A stack that satisfies the rule and places the copper far from the neutral axis can still fail.

Test Methods

Flex life is measured by cycling a test coupon around a mandrel of the specified radius and counting cycles to failure. The coupon has the same stack as the product, because a coupon with a different stack measures a different circuit. The mandrel diameter is measured rather than assumed, because the surface of a mandrel wears.

The failure criterion has to be stated as well. A resistance rise of ten percent is common, an open circuit is the other option, and the two give very different life figures.

Records and Qualification

The qualification record holds the stack, the radius tested, the failure criterion and the cycles achieved. With those fields a design change can be assessed against a real result rather than against a rule of thumb. The result is specific to the coupon, so the coupon drawing belongs in the record too.

The assembly side of the same product family is described in the notes on multilayer flex assembly, and the material storage that affects bond quality in the notes on prepreg and laminate storage. A qualification that is not repeated after a stack change is a qualification for a different product.

FAQ

Is a smaller bend radius always worse? Yes for a given stack. The radius and the thickness together set the strain, so a thin stack tolerates a tighter radius.

Can the radius be reduced by thinning the coverlay? It can, when thickness is what moves the neutral axis away from the copper. The position of the copper matters as much.

How is flex life specified? As a number of cycles at a stated radius with a stated failure criterion. A cycle count without those two is not a specification.