Achieving design collaboration: driving design intent through the PCB layout process is a highly collaborative activity between experts in performance and functionality and the layout staff--and their tools. A look at the tools and processes that make for happy relationships. (Cover Story).

Time was, component placement and signal routing was performed by a single person and consumed most of the design time. Thanks to advancements in design automation technologies such as constraint-driven design and shape-based autorouting, these same steps typically account for no more than 30% of the overall design time. The remaining 70% is consumed by such activities as component selection, design tradeoffs, transmission line analysis, design verification and supply chain management.

Design collaboration is the essence of these activities--a process based on the dynamic exchange of design data and expertise across organizational and geographical boundaries. There are two basic forms of collaboration: synchronous, which entails all individuals participating at the same time, and asynchronous, when individuals look at the data on their own tools and then pool the results. In either case, the design engineer feels most comfortable directing this process from within the schematic environment while having unrestricted access to the physical layout details--without having to become an expert in the layout tool itself.

Driving design intent through the PCB layout process is a highly collaborative activity between individuals and their tools--those who have expertise in design performance and functionality vs. those who have the expertise in implementing layout details. This activity is based on interaction between two traditionally separate design abstractions contained within separate ECAD tools and expressed through incompatible data models: schematic and layout. The intent of collaboration is to merge the two so that they don't step on each other's toes.

Because a PCB is part of a larger product, system-level considerations play a key role. Experts in areas such as RF design, thermal management, mechanical design, stress analysis, etc., have a vested interest in monitoring and influencing PCB layout details. Intermediate PCB layout results are analyzed from a particular perspective and feedback is provided regarding appropriate layout modifications. The result is a steady stream of often-conflicting requests to the PCB layout designer for data access and data modifications.

Traditional ECAD tools were not designed to handle such dynamic and interdisciplinary collaboration, and market opportunities for new solutions were created. These opportunities are actively pursued with innovative approaches to enterprise-wide ECAD data accessibility and browsing--all without the use of native ECAD systems. These same opportunities are also of intense interest to the providers of product data management/product lifecycle management (PDM/PLM) solutions.

Conforming to manufacturing requirements is an integral part of the PCB design process. Such considerations as manufacturing cost, production schedule, target yields and parts availability are increasingly finding their way into the early stages of the PCB design process--creating yet another layer of design collaboration.

Supply chain collaboration is based primarily on allowing individuals to understand what design decisions were made and what is coming down the pipe in terms of BOMs, board structures, etc. This collaboration is focused on tracking and interpreting the data from the perspective of the supplier and the manufacturer and is often the domain of a PDM/PLM system.

Development and implementation of test and repair strategies, along with ECOs, is the last area of design collaboration. While the basic design is already done, the board-level and system-level manufacturing and assembly segments of the organization must be able to work with the design data and the design organization to troubleshoot the systems and help in developing ECO instructions. This requires unrestricted access to the original schematic and layout databases at the time when they are vaulted and no longer readily accessible.

A successful approach requires an understanding of collaboration styles and the functional areas to which collaboration is applied. A significant percentage of collaboration occurs while the design data are fluid and not ready for an official handoff to the next design stage. This process is informal and unstructured, and the result does not need to be tracked.

A known and tracked state of the design data is made available to the enterprise for investigation or verification of a specific issue. Is there a problem with placement of the critical parts? Is the resulting interconnect topology of a clock signal acceptable? Are the critical signals shielded appropriately? Collaboration occurs against a known state of the design and the results of the collaboration must be tracked.

All appropriate issues are resolved and the team is invited to certify the design for a handoff to the next step. Each approval and signature must be tracked as part of the overall design activity. The flexibility and dynamics of collaboration differ depending on whether it is internal or external to the organization, and include such issues as data security, commitment of resources and overall formality of the collaboration itself.

Recent Approaches

As the significance of collaboration increases so do innovative approaches to it. There are at least three emerging technologies that target the limitations of traditional approaches.

Intelligent ECAD data browsers. This class of tools is designed specifically for collaboration using live and fully intelligent ECAD data. They are equally adept at handling and correlating schematic, PCB layout and manufacturing Gerber level abstractions within the same tool. They protect the data against modifications while permitting a high degree of annotations layered on the top of the design. They eliminate the complexity of a full-blown ECAD system. They are ECAD database-neutral; they automatically handle all native ECAD databases. They permit significant freedom through interactive or API-based data mining. In some cases, the tools permit the import of the resulting markups directly back into the native ECAD system for implementation. Finally, they offer communication and integration benefits that make it easy to integrate them with established ECAD and PDM/PLM tool flows as well as existing intranet and Internet infrastructures--making them fully compatible with the synchronous and asynchronous collaboration styles.

Web browser desktop sharing. This technology permits users to link over a network in a synchronous session that shares a client desktop with all participants. Although readily available to all, the technology seems to be used primarily by ECAD vendors for tool support and pre-sales activities. A likely reason for the technology not being widely used for actual design collaboration is the cost of the service as well as the inherent conflict with the ECAD licensing model of the individual ECAD tools. ECAD licenses are typically tied to a specific geographical location and sharing one license on multiple desktops violates the ECAD license agreement.

PDM/PLM systems. Originally developed to manage mechanical design data, vendors of PDM/PLM systems are extending the model into the ECAD environment. While PDM/PLM vendors have a leg up for effectively managing enterprisewide information and collaboration infrastructures, much is being learned about the adaptability of the mechanical data management and collaborative solutions to the EDA space. Many questions remain unanswered: Must a company use the same PDM/PLM solution/vendor in the ECAD space that they use in the mechanical space? What about interpretation of the ECAD data structure, which is so different from the 2D and 3D mechanical rendering? Is it sufficient to map all ECAD structures into a 2D visualization? How will the market resolve conflicts between PDM/PLM vendors who want to control ECAD data flow and ECAD vendors who want to incorporate PDM/PLM for ECAD within their own portfolio? Who owns the ECAD library? How useful is the native ECAD database once it is vaulted with a PDM/PLM system?

Design data make up the foundation on which design collaboration is based and therefore design data formats are a key consideration. The fundamental question to ask is, which is more critical to the end-user: the technology (technical particularities of a specific format) or the solution (overall relevance of the format/tool packaging to the task at hand).

The truth is that data formats are useful only if they can be generated and used for the task at hand. Gerber is a great example. Despite Gerber's limitations, every ECAD system generates it, every DfM tool reads it, and every manufacturing imaging process can be driven by it. The issue is not the format; the issue is the solution.

The users in the context of the above solution model readily embrace even proprietary formats. This is exactly what Adobe did with the PDF format and Acrobat Reader. Since the overall packaging solves the problem, users don't mind the fact that the PDF format is closed, binary and proprietary.

Collaboration around ECAD data faces a similar dilemma: Which format should I base my solution on?

* If the collaborative solution is based on the native ECAD format, it automatically limits the solution to that tool. What about license availability and cost across the entire enterprise? What about the future migrations in the tool flow? What about heterogeneous ECAD tool environments on the enterprise level?

* If the collaborative solution is based on open ASCII formats such as EDIF or GenCAM, is the format available from various ECAD tools? What is the collaborative toll performance vs. file size and parsing of the underlying structures? What about a neutral API into the file's data model?

* If the collaborative Solution is based on a readily available but content-limited format such as Gerber, then what about the intelligent access to the overall design intelligence?

The answer is in letting the solution providers come up with the best format/tool combination for the design collaboration. Details of the format itself are of no concern as long as the solution includes:

* Unrestricted availability of data from the native ECAD systems.

* Sufficient performance and functionality within the tools that use the data.

* Data modeling capabilities that account for all related design abstractions (die packaging, schematics layout, manufacturing, etc.).

* A well documented and external API to the file's data model and to the collaborative technology. An API is important in permitting further integration of the data file and the collaborative technology with the ever-changing needs of the enterprise.

Helpful Hints

Implementation of a collaborative solution is not a straightforward task but a certain process can help you get there faster and with less pain. The following suggestions are based on personal and professional experience and reflect a healthy dose of reality:

1. As with any project, be sure to clearly identify the problem, the solution and the expected improvements. Do this in terms of short-term goals that fit an overall long-term strategy. Long-term visions are a hard sell, but short-term goals carry more weight if they are presented in a context of a broader strategy. For example, tying your success criteria to the overall success of PDM/PLM deployment is a big mistake. EDA collaboration benefits from PDM/PLM but does not require it.

2. Establish the source of the initial funding, because it's a long haul. There are costs associated with purchase, deployment, training, etc., and you'll lose credibility initiating something that cannot be funded. The good news is that the initial funding can come from other departmental budgets as long as the short-term savings are meaningful to that department.

3. Document the current process vs. the target process. This will serve as a sanity check of where you are vs. where you want to be. Besides, if you can't document the process in a way that makes it clear to the others, you don't understand it yourself. This has a lot to do with the areas and styles of collaboration you are focusing on.

4. Evaluate what is already available within your enterprise and the other aforementioned alternatives. Don't get sidetracked by particular format issues that have nothing to do with the overall effectiveness of the solution. Evaluate all underlying data abstractions of interest (schematic, PCB layout, die packaging, manufacturing, etc.), and test average vs. largest possible data sets. Select the best solution.

5. Evaluate how collaboration results will be delivered back to the native ECAD tool responsible for the final implementation. This key feedback loop is often overlooked.

6. Make sure that the ECAD department is willing to support the underlying data generation process and implement it as soon as possible. One of the biggest stumbling blocks in deployment of collaborative tools is a lack of data. People get quickly discouraged if they can't access the data in the first place.

7. Make sure that everyone involved understands what is being evaluated and why it will benefit them. Extra tools and extra steps within an existing workflow are never a source of excitement to the end-user.

8. The project must have a champion, someone who is willing to create yet another training video, stay after hours to solve a data compatibility issue or trace down a license failure. Is that individual enabled by the manager to spend the time attending to these tasks? Initial tool/process rollouts are always support-intensive and not all of the support can be expected to come from the vendor, or handled by the users.

9. Evaluate the overall results of the project on the basis of the short-term goals. This was your only realistic target.

PAWEL CHADZYNSKI is vice president of product development and operations at Ohio Design Automation Inc. (ohio-da.com) He holds a BSEE and a master's in technology management from the Polytechnic Institute of Brooklyn. Chadzynski has held technical, marketing and managerial positions at Photocircuits, Kollmorgen, Algorex, Cadence, and an Al startup. He can be reached at [e-mail].