Revisiting the EDA pyramid: are you a power user, a mainstream user or a late adopter? Your company's infrastructure and your engineering organization may help make that differentiation for you.

For some reason, the pyramid has become ubiquitous in market research. Maybe it's just an electronics industry thing, but it keeps popping up every time someone wants to classify the market into different areas.

One that's become common on the PCB side of the fence is Gartner Dataquest's FPGA/PCB tool user pyramid. It classifies designers as power users, upper and lower mainstream, and later adopters. Problem is, most designers consider themselves power users because of the challenges they face in designing their PCBs, particularly when the other end of the spectrum is known as the late adopter (also known as "laggard"), which is not a term many want to be associated with.

There is, however, value in characterizing designers by characteristics of their company infrastructure, the breadth of their design responsibilities, the complexity of their technology and their organizational structure. This discussion will review the trade-offs to be made and the value of leveraging the environment that you are in.

Segments and Buckets

The Dataquest percentages for EDA systems design seat counts and revenue are shown in model A of FIGURE 1. Another model, Geoffrey Moore's technology adoption life cycle, is shown in model B and looks more like a seasoned PCB designer than a pyramid, but the segmentation roughly maps to the Dataquest pyramid. A common segmentation relating to company and design team size is shown in model C.

[FIGURE 1 OMITTED]

All of the models can be generalized to say that the top is represented by low seat volume with high investment per seat and the bottom reflects the opposite. All of these models point to one primary characteristic (design environment) and three factors that can be used to differentiate all design environments, including engineering organization, infrastructure investment and design complexity.

I don't expect anyone to look at those categories and immediately recognize where they fit. There's no simple place for anyone because multiple factors are involved. So it's really just a spectrum - call the data points along the way whatever you like. As differentiators for the spectrum are explored, you may identify with some characteristics on one side and some on the other.

Engineering Organization

The organization in which you work is the biggest environmental differentiator. You're obviously the same person regardless of where you work, but the size of the engineering organization in which you work can offer significant benefits, as well as a few hurdles.

For example, is your engineering team concentrated at a single location, or is it spread across multiple sites and multiple continents? A single location facilitates direct communication but can't leverage multiple time zones to get a critical design done in an "around the globe" operation. Meanwhile, a multi-site organization (FIGURE 2) has to work a lot harder to provide library and process consistency to maximize design quality and designer productivity.

[FIGURE 2 OMITTED]

Critical to the equation is the size of the design team. At the far end of the spectrum are the maverick individuals who do everything themselves. They create the logic, layout the board, verify it for signal integrity and manufacturability, and manage the libraries and data. They may even be building their own FPGAs to go along with their PCBs and if they're real superhumans (or gluttons for punishment), they'll even design embedded software and the mechanical enclosures. Some call these folks generalists because they touch so many tools. I'd say they redefine the concept of the power user. One benefit here is that the individual designer is usually also the business owner, meaning he gets to make the rules, break them when he wants and make his own tool-buying decisions.

On the other side of the spectrum are product area specialists, such as layout designers, design engineers, mechanical designers, librarians, RF/analog/digital designers and high-speed engineers. These folks know everything there is to know about their design space and the tools they use to get the job done. They live, breathe and abuse their tools day-in, day-out. They know every short-cut and workaround to maximize their productivity and they're willing to put up with pain to leverage design techniques and design tools to get the job done. This is why these experts are a different brand of power user. A typical extension of the specialist is the specialist team - a group of individuals concurrently working on the same design. This includes the traditional design engineer/layout designer combo, multiple engineers creating the same schematic, and multiple designers working on the same layout.

So on one side you have the "generalists" who cover a broad set of design disciplines and tools to get the job done, and on the other the "specialists" who focus on a narrow discipline and tool set. Because the generalists have to span such a broad range of tools, they put a higher premium on tool usage model consistency, integration and overall ease of use than does the specialist.

In today's business climate, larger engineering organizations are typically the result of company or division mergers/acquisitions. They also exist at large design shops (e.g., service bureaus, EMS or ODM) serving diverse customers. They more than likely have multiple tool flows to manage, requiring extremely structured library, design data and process management to facilitate leveraged reuse of the engineering team's IP (e.g., libraries, tool expertise, processes, manufacturing output automation, enterprise integration). Just to make life more interesting, large teams also get to deal with multiple hardware and OS platforms designed to maximize productivity.

Therefore, on one side of the spectrum is a homogenous hardware/software environment (single tool flow and hardware/OS platform), and on the other side a heterogeneous environment (multiple tool flows and platforms). Designers in a homogenous environment have the benefit of a clean, simple-to-manage design flow. Designers in a heterogeneous environment have ultimate flexibility to address complex design challenges with a variety of tools and platforms, but have to tightly manage the flows to ensure that chaos doesn't break loose. Note that by nature you can have homogeneous environments within heterogeneous environments (e.g., engineers designing prototypes or reference boards), but they often make no attempt to leverage common IP, such as share a common library or reuse prototype design data in the production design.

Infrastructure Investment

It should be no surprise that multisite, global engineering organizations make significant investments in their infrastructures. Individuals or small teams tend to leverage the standard libraries that come with the tools, while large organizations invest in central libraries custom-built to their own standards. An industry infrastructure has emerged in support of both organization types that can supply standard and custom libraries for most design disciplines, such as schematic symbols, footprints, simulation models, parametric data and common board outlines/stackups. However, while large organizations will certainly leverage these industry resources, they also continue to maintain their own custom libraries to ensure process consistency and protect IP.

Another aspect of infrastructure is the existence of a support team. Large organizations tend to have a dedicated CAD/IT team that provides primary support and training to the design teams. This team acts as a "firewall" between the design team and the outside world, creating a consistent environment across the organization. The support team is very deliberate in its adoption of new software releases because of the impact of the change on the consistency and quality of the engineering team's environment. The CAD team is also responsible for creation of custom processes, tools and interfaces required to optimize the performance of the design team for the company's particular product. Large companies also leverage their purchasing power by centralizing standard functions like part selection under the finance department. This requires integration with that dimension as well.

A small organization doesn't have the resources to create this kind of support infrastructure. As such, they typically use the product as it was delivered, or "out of the box." This is where the term ready-to-use (RTU) has emerged. The installation, licensing and start-up time for RTU tools is expected to be easy because the smaller firms don't have the internal resources/expertise to do this. That doesn't mean large organizations don't also want this too--they are making a trade-off against the benefits discussed earlier, such as customization. Small organizations typically don't customize their software and don't even want to spend the time and money to get trained on the product. In this case they are making a trade-off that may limit their time-to-productivity but will ultimately be more cost-effective. EDA vendors are addressing the needs of these organizations with online tutorials delivered with the product, Web-based live or recorded training, and Web-based support mechanisms.

As an example of these trade-offs: I purchased and installed Adobe Photoshop because I wanted to edit some graphics supplied by an outside vendor. Getting the product loaded and running was simple and painless, as was opening the graphics design I was interested in. That's when the pain started. I didn't know where the commands I thought I wanted were, and I had no clue how to accomplish what I was trying to do. Since I'm stubborn, I ultimately persevered, but I'm sure my productivity was not even in the ballpark relative to that of the experienced, full-time graphics designer who created the original image. If I used the tool full-time, or if I had invested enough time up front to adequately train myself, my productivity and the resulting product would have been much higher.

This is exactly the situation faced by those who do everything in the design process but don't spend much time on any particular tool. Even if they were trained on every tool in their process, these users would have to possess exceptionally large brains to remember all the nuances of every tool if they didn't use them every day. Not only must the tools be easy to learn the first time, they must be easy to relearn every time they're used.

Design Complexity

I've talked a lot about organization and infrastructure investment as the big differentiators between design teams. I started with those because most people think design complexity is the best differentiator. The reality is that the ability to design complex boards does not scale directly with organization size. I've seen some monster designs built by one-man shops. But, like everything else, trade-offs are involved.

With increased end-product complexity come increased tool and process complexity to get the job done. The designer who prizes the ease of implementation available with simpler tools must now adopt advanced tool capabilities. Additional constraints must be defined and adhered to for things like high-speed signals, microvia/HDI structures and manufacturability, or tight mechanical enclosures. Simulation also becomes a necessity to verify that the constraints were followed and tighter processes must be adopted to ensure that consistent design steps are followed. In addition, the supply chain must be leveraged to get access to advanced components and their simulation models. All of this requires a resource investment, which can limit the size of companies capable doing these designs. Early adopters by nature invest as much as possible in new technology to maintain a competitive edge. Again, it boils down to a simple tradeoff of resources vs. design complexity.

Another dimension of that investment is horsepower--whether it's in multi-GHz parallel CPUs or high-end software automation. You can find independent power user designers who know all the right buttons to push to get the most out of a tool during interactive design, but with the complexity of today's designs, horsepower is winning out.

It's not simple to categorize designers and the environment in which they work. And if there's one thing I've learned about PCB designers, it's that they're not fond of being labeled and put into buckets. With that said, designers are differentiated by the type of organization they work in, their infrastructure needs and the performance requirements for their designs.

In fact, depending on what bucket designers fall into or how they are labeled, any competitive edge they achieve is a result of trade-offs. Ease of use vs. design complexity, or budget vs. advanced tool technology, or customizability vs. ease of learning are just a few of these trade-offs. In the end, leveraging their environment and making the necessary trade-offs are extremely important to designers.

DAVE WIENS (david_wiens@mentorg .com) is director of business development, Systems Design Division, Mentor Graphics Corp.