Multi-disciplined collaboration for efficient product development

Part 1 - Multi-disciplined collaboration for efficient product development

Developing and delivering a product requires teams of designers and engineers across multiple disciplines, local or globally dispersed, that must efficiently work together.


A recent report published by Aberdeen Group, “Why PCB Design Matters to the Executive” (Reference 1), highlighted strategies and best practices used by best-in-class electronics companies to meet their aggressive business goals. These top 20% of all companies surveyed pointed to “improve communication and collaboration across R&D” as the most important strategy to help get their most competitive products to market faster and at lower costs. They understand that developing and delivering a product requires teams of designers and engineers across multiple disciplines, local or globally dispersed, that must efficiently work together.


If my primary area of responsibility is PCB (printed-circuit-board) design, I understand that many dependencies affect what my organization is developing. These might include the design of an IC/Package or an FPGA. I understand that my PCB(s) will be integrated in an enclosure designed by a mechanical engineer. I must collaborate and communicate with my supply chain and the PCB manufacturer. I may even have a specialist, such as RF designers, outside of my organization that will add RF or analog circuitry to my digital design.


And, I have two very basic challenges. The first is: How do I communicate, bi-directionally, with these other disciplines efficiently? Sending paper back and forth is no longer acceptable in terms of efficient time use and end product quality. I must be able to communicate electronically. But often these other disciplines have their own systems and languages so I must be able to communicate in a manner where the teams receive information in their own context and are not required to learn different set of software tools.


The second challenge is design concurrency: How do we insure that the different teams and team members act in parallel versus serially? If members have to wait for edits to be made prior to them continuing with their work, design cycle times will increase and cost will be added to development as members wait for progress. Conversely, a more parallel process compresses design times so we can meet time-to-market goals.


Let’s look at some of the areas that require collaboration, and discuss the challenges and the solutions to those challenges that can be implemented in today’s product development processes.


IC/package/ FPGA with PCB collaborative co-design

Today, a high percentage of products contain either a custom-designed integrated circuit or an FPGA—or both. In the past, it was common practice to design the custom IC or the FPGA first and then deliver it to the PCB designer with the parameters and pinouts cast in concrete. However, this process can negatively affect the performance and cost of the product. If the pinout of the IC or FPGA package are not defined in the context of the PCB, it can result in excess trace lengths as we connect the IC/FPGA to the rest of the circuitry in the PCB. Extra trace length compromises performance (high-speed route delays) and can add extra PCB layers due to routing congestion.


However, if the IC/FPGA designers collaborate with the PCB designer and perform their design process in parallel (versus serially), we can arrive at a more optimal design. Software exists that enables teams to communicate bi-directionally, negotiating the pin-outs of the packages, to arrive at a solution that meets the requirements of both design teams and arrives at a product design that is optimized for performance and cost.





ECAD/MCAD collaboration

PCBs are integrated into an enclosure that is designed and analyzed by a mechanical engineer. Typically these processes are performed in parallel after some initial design is done of the enclosure. This initial enclosure design determines the board outline, mounting hole locations, etc., which are passed to the PCB designer. Both the PCB and the enclosure designs proceed in parallel with changes being made on both sides. Some changes affect the other discipline, some do not. So the ability to communicate those changes, and only the changes, that affect the other discipline is important.


But in the past, interfaces such as IDF was used to communicate between the ECAD (electronic CAD system) and the MCAD (mechanical CAD system) and transfer complete design databases. This method is very effective for initial and final data transfer, but not terribly effective for incremental changes. In 2007, a new standard was approved by ProSTEP that enabled transfer of only those incremental changes. This standard became the vehicle for implementing bi-directional collaboration. Either discipline can propose a change and have that communicated to the other in the context of their native system (no new software needs to be learned). This initiates an electronic negotiation process where counter proposals can be made, etc. Once a change is agreed to, both databases (ECAD and MCAD) are updated and design can continue. This replaces paper methods of change proposals that can be error prone and time consuming.


Another common collaboration opportunity between mechanical and electronic design is thermal management. While thermal analysis can be performed on the IC package and PCB by the electronics designer, comprehensive analysis and accurate representation of the product requires that the PCB(s) be integrated in the enclosure and analyzed as a full product. This form of virtual prototyping (prototyping in software instead of actually building and testing a physical model) requires action by the mechanical designers since they are the people defining the enclosure, fans, or whatever cooling methods are used at the full product level.



Multi-technology (RF, digital, analog) design collaboration

RF circuit design remains what some might consider an art performed by highly-technical engineers. But with today’s explosion of wireless applications, more and more PCBs need to be designed with a mix of RF, digital, and analog circuitry. This circuitry needs to be jig-sawed into small form factors, which means that there must be interaction between the disciplines as the PCB is being designed. The various disciplines can no longer afford to design their circuitry in a vacuum and then expect to have them magically fit together on the PCB.


In the past, the design of these PCBs was substantially a serial process performed in separate design and analysis software tools, with interfaces such as IDF to communicate between those tools. Today this has changed. Although RF circuit analysis continues to be different than digital circuit analysis and still requires specialized software, companies have partnered to integrate their respective systems and provide capabilities to design mixed technology in the context of the complete PCB. The RF circuitry can be design in either system and then simulated in the RF system’s specialized software.


Using this level of integration eliminates the need for many design iterations and leads to a higher quality, and a more compact product. The engineers are able to see and design their portion in the context of the rest of the PCB as well as run analyses to determine how the circuits interact.



Collaborating with the electronics supply chain and manufacturing

Attention is generally focused on designing the best PCB without much consideration of efficient manufacturing, at high yields, at the lowest cost, and on time to meet the market window. Some of this responsibility lies with the manufacturer and how efficient they run their operation. But much of the burden falls on the designer. If the PCB is designed collaborating with the supply chain and manufacturing the design data can be delivered with the assurance that it will meet production-ready requirements.


It starts with the electronics supply chain. As parts are chosen, communicating with procurement to assure that the parts will be available at target volumes and cost can avoid expensive re-design iterations late in the process. Early and frequent communication of BOM (bills of material) enables the procurement department to find parts supplies, suggest alternate parts, check costs, make sure that the volumes are available, and pre-order the parts in preparation for immediate production.


The next consideration is DFM (design for manufacturability). The manufacturer can supply an extensive set of rules. Some of these rules highlight manufacturing violations that produce a hard failure. Others will communicate best practices that increase yields and improve production efficiency. These rules can be used to analyze the design throughout the design process, from beginning to data release to the PCB fabricator. By using this checking software from beginning to end, you insure that your design is making forward progress and avoiding costly redesigns. And, when you hand the data to manufacturing, it will be correct and you won’t be required to implement changes to improve the manufacturability.


The third step in a design-through-manufacturing process is the design data handoff to your board manufacturer. That manufacturer will perform the same checking you have performed to insure that there are no manufacturing violations as well as ensure high yields. This data transfer can take several forms. In the past, your intelligent design data was broken down into several files with no associations—Gerber, Excellon drill, BOMs, neutral placement file, etc. A manufacturing engineer then had to “reverse engineer” these files to add intelligence back into the data. Now, using a format like ODB++, this intelligence can be maintained in the data eliminating the costly step of reverse engineering, and in a single file.


The importance of intellectual property management to collaboration

Key to the success of design team collaboration is the ability to create, control, and leverage their static (e.g. parts libraries and reuse data) and work-in-progress (design data, high speed constraints), intellectual property. Members of the team may be local or dispersed around the globe. But management of this data is a very complex issue due to its nature and the design process. Members work in parallel and the design data (schematic, constraints, design) can get out of synchronization. The various data sets have interdependencies very particular to PCB constructs. The intellectual property needs to be a “the click of a mouse” away from the designers and not require access through complex corporate infrastructures.


The IP typically takes two forms, static and dynamic. Static IP is usually corporate approved and qualified component parts libraries. These are created by librarian specialists, organized, and stored in a central infrastructure, and accessed by the design teams around the world. Also contained as static IP is design data qualified for re-use. These blocks of previous designs can be accessed and dropped into current designs being developed. The advantage of these re-useable schematics and design data are that they have been proven in previous products and thus represent a productivity and quality advantage.


Dynamic IP is the “work-in-progress” design data that is created and shared by the design teams as the design proceeds. This is typically the schematics, the design (high-speed and manufacturing) constraints, the actual PCB design data (layout, etc.), and the results of virtual prototyping simulations and analyses. This data is extremely complex and requires a PCB focused IP management systems to provide the creation, storage, control, and access required by the design teams.


As the IP is developed for a product, snapshots are archived in corporate PLM and ERP systems. These corporate systems are not appropriate for managing the work-in-progress IP but necessary to archive the product and feed into the corporate supply chain systems as the product proceeds through its production and life cycle.



Collaboration is necessary

With today’s extremely complex PCB designs, it is impossible to meet an electronics company’s aggressive business goals without efficient, multi-disciplined collaboration and design systems that support that collaboration. In this article we have touched on just a few of the areas where this collaboration can make a difference in getting the most competitive product to market faster and at lower costs. In the future, these designs and processes will only get more complex and the need for collaboration supportive design solutions a make or break situation. 


John Isaac, Director of Market Development, Systems Design Divisions, Mentor Graphics Corp -- EDN, August 24, 2011


Reference 1.

Boucher, Michelle, "Why Printed Circuit Board Design Matters to the Executive: How PCBs are a Strategic Asset for Cost Reduction and Faster Time-to-Market," Aberdeen Group, Boston, MA, February 2010