When designing a rigid-flex, start with flex in the middle of a stack-up and move outward.

You’re designing a new rigid-flex. Devices are getting fanned out, via structures defined, and layer count is becoming clearer. You have determined how many flex layers you need from rigid section to rigid section. There are competing considerations on how those flex layers are configured: foil and dielectric thickness, bonded or unbonded, and where they will be in the stack-up of layers. All this impacts flexibility and how the part will bend in the installed application.

For today, let’s concentrate on where the flex layers land in the stack-up and the effect that can have on manufacturing and end-application use. Several strategies have rational logic and can be successful in select situations.

The most common and lowest impact is to place the flex at the middle of the stack-up. There are several advantages. First, it permits symmetry of the stack-up. Symmetry provides the opportunity for the flattest stack-up with the least tendency for bow and twist. This is more and more critical as component pitches get denser, and the size of BGAs and FPGAs gets larger. This is also the easiest to fabricate with the lowest cycle time, resulting in the lowest cost. If you can put the flex in the middle of the stack-up, this is the best option.

Increasing distances between rigid areas helps prevent potential damage.

When designing a rigid-flex that needs to bend 90° in the flex area, what is the minimum flex length (distance between rigid sections) one should allow?

That is a loaded question without knowing the overall thickness and width of the flex layers. For this column, I will assume only a couple of flex layers, and the flex width is 2" or less. Several issues come into play when determining the minimum distance between rigid areas on a rigid-flex. Some will affect the supplier’s cost because of yield reductions, and others may affect the mechanical function.

Manufacturing issues. Fabricating a rigid-flex circuit means juggling a number of technical issues to get everything to work. First, the rigid material must be removable in the flexing areas. (The rigid material is applied in full sheets.) This can be done by pre-scoring the rigid-flex interface lines part way through and removing the adhesive in the flex areas used to bond the rigid material to the stack.