Why isn’t a 3-layer rigid flex less expensive than a 4-layer rigid flex?
Designing flex and rigid flex can be tricky. Not only do you need to design for a product that will be bending, folding and flexing, the material sets are different, fabrication processes are different, and it is difficult to communicate clearly how the part is intended to be flexed in final application. Typically, the goal is to use the most cost-effective stack up that will meet both the electrical and mechanical performance needs. Logically, wouldn’t a 3-layer design be less expensive than a 4-layer design? Rigid flex can be a funny thing. Let me share an example of an application that proved that logic to be incorrect.
This specific medical application was originally designed as a 3- layer rigid flex. The construction had flexible layers on the outside layers with a rigid layer sandwiched between. The original intention of this stack up was to keep the layer count as low as possible to keep costs down. This is something that we all learn early; the higher the layer count, the higher the cost. So, a 3-layer rigid flex seemed logical as it was being designed. The design had a dense component area, that was being supported by the rigid inner-layer, and a flex area that needed to be dynamically flexing with a tight bend radius. Once the design was complete it was sent to the fabricator for quotation. Pricing came back higher than expected, but this was one of the first rigid flex designs this company had done, so it was assumed this was the typical price of rigid flex.
As this project was going through tooling and set up at the fabricator, the was asked if flexible photoimageable solder mask could be used to help define the surface mount pads in the dense component area. Drilled polyimide coverlay was not able to support the density. This was approved, product was built, shipped and assembled. As the flex was mounted in the unit and powered up, it was quickly failing. Failure analysis showed the traces in the flex tail were cracking in use, causing the failure.
Communicate how the flex is going to be used in the end application
Working with the fabricator, it became obvious that the flexible solder mask selected to define the surface mount pads was not flexible enough to stand up to the end-use flexing requirements. One thing that is rarely communicated well between design and fabrication is how the flex is going to be used in the end application. Gerber files are two dimensional and the end use is three dimensional. Had the fabricator realized that this was intended to be dynamically flexing, flexible solder mask would not have been recommended.
After several discussions, it was decided that the flexible solder mask would be changed to a polyimide coverlay. This brings us back around to the initial issue with the surface mount pad density and definition. As a solution to that challenge, laser drilling the polyimide coverlay was recommended. Also discussed at that time was changing the fabrication to button plate rather than panel plate. Button plating adds copper only to the pad area and plated through holes and eliminates electrodeposited copper from the construction, keeping the layers as flexible as possible.
The project was again released for fabrication, assembled, installed and tested successfully. Perfect! Right?
How did the 3-layer rigid flex become 20% more expensive than a 4-layer rigid flex?
Here is what was going on behind the scenes to fabricate this part number. Flex on the outer layers of a rigid flex construction is not and in this case was driving up costs significantly. The dense component areas that pushed this design to require laser cutting of the coverlay, not only required a more expensive process to create those openings, but because of the tight registration needed, the fabricator needed to reduce the panel size and subsequently the number of pieces running per panel, which drives cost up. In addition, the button plating process itself adds extra cost.
Ultimately, this product was redesigned to a more conventional, 4-layer rigid flex construction with rigid layers on the outer layers and the flex layers sandwiched between. Those dense component areas were then able to be defined with a standard solder mask process and the button plating process was eliminated. Once standard processes were used, the panel size was increased and the number per panel increased as well. Even though the raw material costs were higher with a 4-layer design, the overall cost of the part decreased 20% with this redesign.
Cost of Rigid Flex
First, the initial 3-layer rigid flex design was ultimately successful, and the end unit was functioning as needed. There was nothing “wrong” with that design. But, when cost is a significant factor in the design, which it almost always is, there was a less expensive way to approach the rigid flex design. Lesson learned.
Could this have been avoided? Maybe. Part of the issue was the learning curve that goes along with a new technology, and part of the issue was communication. I sometimes feel that I have become like a broken record, stuck on the same few words, but my best advice for designers new to rigid flex is always to work with your fabricator, over-explain what you are trying to accomplish and ask for their advice. They work with the technology every day and are happy to recommend best practices and help avoid introducing unnecessary costs.