In my last column, I discussed how manufacturers of rigid-flex boards use techniques similar to those used by manufacturers of hard boards and flexible circuits, and how techniques vary. That column’s discussion was for a standard, straightforward rigid-flex design. This column will talk more about non-standard designs, which can present process difficulties and require extra care for effective yields.
Asymmetrical Rigid-flex Material Layups
Asymmetrical rigid-flex designs are fairly common but not recommended (IPC-2223A.8.1), as they can be very difficult, and depending on the design, impossible to build. Using materials with differing properties without balance and symmetry can cause manufacturing issues with your board, and there are two common types.
Figure 1: Six-layer rigid-flex with flex layers off-center of the neutral axis.
One is where the flexible layers are off center of the neutral axis in the material layup, putting more of the glass-reinforced layers on one side versus the other (Figure 1). Manufacturing panels with this construction are prone to warp during manufacturing. Since most PWB processes are planar, a warped panel can be very difficult to drill, image, and plate correctly, resulting in reduced yields. Once removed from the production panel, those parts can present great difficulty at assembly because the processes are planar as well. Thus, it is wise to avoid these constructions if at all possible to improve both manufacturing and assembly yields.
Figure 2: Pressure points at the rigid-to-flex interface.
The second one is where the flexible layers are placed on the external layers of the board. These designs are rarer, but they do come in from time to time. Rigid-flex boards with flexible layers as the external layers share the same risk of warp as any asymmetrical board construction due to the CTE mismatch between flexible materials and glass-reinforced materials.
They also struggle in photoimaging and plating. Depending on the thickness of the base flexible laminate(s), the distance between the rigid boards, and the width of the circuits being imaged, these designs will have significant yield loss—typically, over 50%. Also, depending on your fabricator, these designs might be unmanufacturable.
The reason is that the flexible areas of the board deform during dry-film lamination, which is crucial to faithfully reproduce the circuits on your design. A material layup with the flexible layers on the outer layers is shown in Figure 2; layers four and five are flexible laminate. During dry-film lamination, the laminator applies pressure to the dry film, so it adheres to the base laminate. The rigid layers remain stable, but the flexible layers conform to the pressure, creating pinch points at the rigid-to-flex interface. The strain deforms the dry-film photoresist and can even deform the flexible layers at the same time. The copper on the flex layers is very soft—rolled annealed copper foil—and is prone to stretching easily when stressed.
Rigid-flex Designs With Flex Arms
A very common technique in rigid-flex design is to have one or more of the flex arms terminate in a flexible board rather than a rigid board. These are often used to accommodate ZIF connectors, hot bar solder joints, conventional connectors, through-hole assembly, etc.
Figure 3: Rigid-flex with flex arm (L) and conventional rigid-flex (R).
An example of a conventional rigid-flex board and a rigid-flex where one arm is a flexible board are shown in Figure 3. The board on the right is a conventional rigid-flex board where each arm terminates in a rigid board. The board on the left is very similar, but the center arm terminates in a flexible board designed to mate with a ZIF connector. This flex arm was designed to accommodate in-circuit test of the final assembly and was then removed and discarded.
Rigid-flex designs similar to the board on the right are generally very high yielding and run through the manufacturing sequencing with few or no issues, which results in lower costs for your design. Rigid-flex designs with one or more flex arms that terminate in flex rather than hardboard take longer to build and generally have lower yields.
Figure 4: Rigid-flex board with pouch construction.
To build rigid-flex boards with flex arms, we have to build the flexible layers to completion somewhat akin to building a whole board. Then, we take that flex circuit and bury it within the rigid board, using a common technique called “pouching.” The pouch protects the inner flexible circuit during outer layer processing. If we were to leave the flexible circuit exposed during outer layer processing, the etching and plating chemistries would attack the circuits and pads on the flexible circuit. Figure 4 shows how a typical rigid-flex with a flex arm is processed.
The core material on the external layers forms a protective pouch for the circuitry on the internal flex layers. This pouch remains throughout the entire rigid-flex manufacturing process and isn’t removed until the board is completed through all imaging, drilling, etching, plating, solder mask, etc. It is usually removed just before electrical test so that the connections on the flexible circuits can be tested at the same time as the rest of the board.
Figure 5: Strain-relief beads applied to the edge of rigid areas to protect flex.
Although these designs are very common, their manufacturing sequencing is slower than conventional rigid-flex boards. First, we have to build the flex layers to completion—drilling, imaging, plating, coverlayer lamination, laser route, etc. Often, buyers think we build the whole board all at the same time and don’t understand why these designs take longer. The only way to get them to yield effectively is to build them in sequence.
Generally, pouched constructions don’t cost much more than conventional rigid-flex, but they are much slower builds. The pouches are usually removed by hand, though under certain parameters can be removed by laser or controlled depth routing. If you have a design with four flex arms that are pouched, that is eight pieces of material that are being removed on each part. If your order is for 1,000 pieces, that is 8,000 pieces of core material that must be removed. Again, pouch removal is done at the end of manufacturing just before electrical test where, often, a buyer won’t understand why their parts are taking so long to complete.
Pouches also leave an edge along the rigid-to-flex transition area that is exposed glass fabric. It is customary to bead these edges—especially with high-reliability applications—to protect the flexible arm from coming into contact with the glass fabric edge and to protect it from abrasion (Figure 5). The beading material is usually epoxy-based but is soft, similar to silicone. The beading material is also placed by hand, resulting in longer lead times. In the previous example, each flex arm will have two beads placed by hand for a total of 16,000 beads, which can take time to process as well.
One other issue with pouched constructions that needs to be accommodated is dimensioning of the rigid-to-flex transition area—the line of which is defined by no-flow prepreg. The no-flow prepreg actually does flow, and that flow can vary from manufacturer to manufacturer and lot to lot. There will be some variability, which means that the line of your transition can be up to ±0.030”. If that dimension is critical to your design, it would be best to consider another method for that flex arm other than pouching it.
In my next column I will discuss other rigid-flex design and manufacturing techniques that may not be the most straightforward, but can definitely be new tools in your toolkit for answering challenging customer design requirements.
Bob Burns is national sales and marketing manager for Printed Circuits LLC.
This column was originally published in the April 2019 issue of Flex007 Magazine.