Issue link: https://iconnect007.uberflip.com/i/344372
18 The PCB Design Magazine • July 2014 sign options) and actual alignment wasn't de- termined until physical assembly. Today, PCB design systems include 3D visualization to en- able optimization for the enclosure—and when collaboration is required to optimize the elec- tro-mechanical product, dynamic communica- tion with the mechanical engineer is provided via tools leveraging the ProSTEP IDX standard. Given basic outline constraints, design of the PCB can now proceed in parallel with refine- ment of the mechanical enclosure. Integration of FPGAs on PCBs provides an opportunity for pin swapping to optimize the PCB for performance and cost—so efficient col- laboration between FPGA and PCB engineers is critical. The cross-domain interchange has been vendor-specific, requiring flow-dependent processes and limiting optimization passes. Im- proved integration has enabled bi-directional collaboration and even multi-FPGA optimiza- tion based on PCB placement. Most PCBs don't live by themselves; they're coupled with daughter cards or backplanes for cost-efficiency and flexibility. Definition of that multi-board PCB system has typically been iso- lated from the PCB design environment, requir- ing re-capture at the PCB stage (much like the original collaboration between PCB engineers and layout), increasing design time and poten- tial for re-design errors, and also minimizing opportunities system optimization. Emerging environments support dynamic collaboration between system-level and PCB-level design, fa- cilitating quick ECO updates and re-partitions for optimized system performance, size and cost. Another system context is the design of IC through package to PCB. These three disciplines are often supported with disparate, disconnect- ed toolsets. Tying the disciplines together (or even leveraging common tools) improves col- laboration, resulting in optimal system perfor- mance, size and cost. Recent advances have au- tomated definition of the connectivity between platforms, and also leveraged common design technologies to minimize the system design learning curve. Mid-to-large enterprises achieve economies of scale by leveraging common design data (e.g., component libraries, simulation models, and reuse blocks) across multiple projects. They typically manage this data, as well as control ac- cess to work-in-process designs, via convention- al PLM systems. While functional, this process views PCB data as black boxes, with no visibil- ity into the actual data, limiting design reviews and collaboration opportunities. Emerging standards such as the EDX format coupled with PCB-level design data management enable effi- cient collaboration with enterprise-level teams, and provide optimal access control and visibil- ity for PCB design data. Intra-Discipline Collaboration We've covered the parallelization of inter- discipline tasks, but what about concurrency within a design task like schematic or layout design? Given that legacy schematic entry and layout tools were designed for a single user, it's no surprise that a multi-user process is cumber- some, though some enterprising teams haven't created workarounds. Engineers have long teamed up on a schematic design, enabled by either a hierarchical or flat multi-sheet struc- ture that allows them to partition and distrib- ute schematic blocks, and later stitch them back together. This "split and join" process requires manual communication regarding system up- dates (e.g., a new global signal name), dimin- ishes team visibility into each other's work, and provides fun alignment challenges when the blocks are reintegrated. Now imagine do- ing that on a layout (it's often done by region or layer). How do you ensure that cross-region traces or vias will line up when the design is reintegrated? How do you partition the design when you've got cross-board (side-to-side or front-to-back) connectivity? How do you decide partition areas if not all parts are placed? To enable efficient concurrent intra-disci- pline collaboration, the tools must be re-ar- chitected to support the team model. A client/ server architecture enables multiple team mem- bers to concurrently access and edit a design, with dynamic updates so everyone is in sync. The communication between clients and server must also be optimized to deal with network la- tency issues if the team is distributed. Applied to the schematic stage, multiple en- gineers could define logic in different sheets of OPTIMIzINg COLLABORATION FOR PCB SySTEMS DESIgN continues feature