How to tackle rigid

Unless properly managed, can lead to design failures, re-spins, and long-term reliability issues. Below are some of these challenges and how to tackle them with Siemens's Digital Industries Software Xpedition

Rigid flex designs include several regions, each with a different layer stack-up. These stack-ups must be defined during design so that all downstream operations, including electrical/manufacturability analysis, and fabrication output can use it. With Xpedition, all you need to do to create a rigid-flex design is to draw the multiple board outlines and assign layer stack-up to each. Each board outline is given a name so that it can be easily identified when outlines overlap partially or completely.

A cover layer is a form of protective coating on top of the metal foil in a flex design. It offers better protection against wear and scratch compared to solder mask, and also helps the metal foil stick to the base material, providing improved adhesion. You can use a sheet of cover layer that covers the entire design and becomes embedded in the rigid stack-ups of the rigid-flex design. This is called "Embedded cover layer," as it is embedded into the rigid sections. In Xpedition, the cover layer is part of the layer stackup, creating an embedded cover layer.

Stiffeners are pieces of rigid material bonded to flex regions to "rigidize" a section of a flex design, to allow components to be assembled on the flex area, or to provide rigid mounting holes. The stiffener material can be conductive, such as metal, or non-conductive, such as plastic or FR4. In Xpedition, you define the stiffener as a layer in the layer stack-up, but you also draw the actual stiffener as a stiffener-shape on that layer.

You will need to define the adhesive layer in the layer stack-up but, as with stiffeners, you can draw one or more "adhesive shapes" on the adhesive layer.

The key reason for making a flex design is so that it can be bent or folded. As a designer, you need to define where your design bends, how it bends, and what design "quirks" you can accept in the zone impacted by the bend. Xpedition includes a draw object, called a "bend area," the location of which defines where the bend happens. Properties on the bend area define how much it bends: angle, direction (+ or – angle), and how sharply it bends (the bend radius). While the center of a bend can be drawn as a line, the zone on the flex board that is impacted by the bend is a wider region. As expected, the bend zone is wider with large bend angles and large bend radius.

When you have a curving flex cable and as many signals that can fit with a tight squeeze, you need a special routing algorithm that can follow the contour of the board outline and automatically insert all the signals required.

Xpedition has several ways to solve this challenge. It's vital that all curved traces be routed with true arcs rather than with approximated segmented arcs, as was common in the past. Even cases where arcs were segmented into as many as 64 segments can still get stress fractures due to the arc not being completely round. In Xpedition, all curved traces use true arc primitives.

It's generally not acceptable to have solid plane fill in bend regions in a flex design. Instead, the typical pattern is cross-hatched. However, for best reliability and to prevent metal fatigue, the cross-hatched pattern must be rotated 45 degrees off the bend line.

Note that the board strip that bends may be at any angle and the bend line itself can be angled, so the 45 degrees off the bend line can actually be any odd angle value. In Xpedition, you can set a unique hatch angle for every fill that requires it using the "Plane Classes and Parameters" dialogue.

Xpedition is unique in its ability to generate and maintain these dynamically and have a DRC that reports if a tear drop fails.

Most designs today require some level of signal integrity analysis as part of the design process. This is advanced enough in single stackup designs, and now in flex-rigid designs, that we have boards with multiple rigid stack-ups, multiple flexible stack-ups, partial adhesive in the stack-up, and stiffeners that must be properly modelled in order to get correct analysis results.

Co-developed with Xpedition for analysis of flex-rigid designs, HyperLynx understands how interconnects pass different stack-up scenarios and applies proper modelling in each section. You can use this analysis to ensure a functioning design with complex stack-up.

Flex circuitry is more difficult to manufacture than rigid PCBs, and the best way to ensure that designs are manufacturable is to run DFM checks throughout the layout process as well as running DFM on the manufacturing release process. Running DFM checks on PCBs with Valor NPI can reduce respins on average by 57%. For rigid/flex, it is even more critical to ensure manufacturability before sending the product model data to the fabricator. Applying DFM to PCB designs results in more time spent designing and less time fixing.

With a complex design that is a mix of rigid and flexible boards, each of which might have a unique layer stack-up, it's clear that board fabricators have a challenge in figuring out what the designer wants to build. Even educated guesses are bound to lead to misunderstandings that can lead to very costly situations. Avoid the guesswork by using ODB++ and its built-in constructs as a fabrication data format to safely communicate all flex-specific information.

Find all the references and more information at Siemens Digital Industries Software