Developing trends within the Industrial, Consumer Electronics and Aerospace & Defense arenas have increased the demand for high-speed, high-density flexible printed circuit solutions. A few examples include: center dot High-end servers and supercomputing platforms that consistently implement faster processor-to-processor interface speeds and require increased 10 bandwidths. center dot Cost-sensitive consumer electronics and gaming consoles that continue to improve chip-to-chip bandwidth and speed for graphics-intensive applications. center dot Defense applications such as missile guidance systems, unmanned drones, avionics and radar systems requiring low-loss interconnects to maximize density, flexibility, and functionality while minimizing attenuation and weight. As a result, there is renewed interest to improve conventional flexible interconnect solutions and provide new alternatives for integrated high-speed applications. The ability of flexible printed circuits to provide integrated 10, power distribution, decoupling and electro-magnetic compatibility solutions make them prime candidates for chip-to-chip 10, high-speed serial links and for direct-attach first-level packaging. However, most conventional multi-layer flexible printed circuit cross-sections cannot provide loss characteristics required by high-speed applications with interconnect distances of more than a few inches. Therefore, more costly bundled solutions, such as '' ribbonized '' coax or fiber optics are commonly employed to meet electrical requirements of high-speed applications requiring mechanical flexibility. The goal of this paper is to describe the loss contributors within conventional flexible printed circuit constructs and how their content can be reduced to improve insertion loss and reduce attenuation. Reduction (or eventual removal) of conventional bonding film adhesive and potential future changes to existing dielectric and/or adhesive materials are considered to reach desired performance targets. Based on simulation results and laboratory measurements, this paper will describe significant loss improvements to conventional flex cross sections and outline plans for future improvements. Moreover, we will describe why this research may be of value to the Signal Integrity (SI) engineer working within a cost-sensitive and performance oriented application. Finally, we will discuss what new applications are enabled through the reduction of lossy conventional flexible circuit cross-section dielectric materials.
