Estimated reading time: 5 minutes

Flexible Thinking: Beyond DFM—PCB Design and Manufacturing as a Cooperative Effort

The concept of co-designing PCBs by incorporating input from all areas of manufacturing has grown significantly over the past several years, bringing a sharper focus on the early contributors to a collaborative electronic system. From those pioneers, we have learned that the PCB is no longer simply the platform on which a design is manifested; it is now an active part of the electrical, thermal, mechanical, and economic architecture of the product. As operating frequencies rise, component pitches shrink, power densities increase, and product schedules tighten, decisions made by one discipline can no longer be isolated from the consequences they create for others.

In my past PCB workshops, I have referred to PCB designers as the “drum majors” of the industry. They lead the electronic manufacturing “parade” and the decisions by designers can make the difference between stirring marching music and chaotic cacophony. Yes, the designer must still translate logical component connectivity into a manufacturable physical structure, but that task requires earlier access to system requirements, greater awareness of manufacturing capabilities, and continuous interaction with specialists who once entered the process much later.

A central question, in my opinion, is whether this evolution will simply lead us back to another discussion of DFM or whether the industry is ready to recognize that successful PCB development depends on something broader that integrates the process and its executors to create perfect or near-perfect printed circuit assemblies every time. Don’t misunderstand me. DFM remains essential and won’t go away, but as I’ve written about in past columns, DFM alone is not designing with manufacturing (DWM); it is not cooperative design.

PCB Project Steps
Every PCB project has a natural process hierarchy, whether the organization formally recognizes it or not. There are myriad elements to the process, starting with program management, which defines the product objectives, including the usual suspects: size, weight, cost, schedule, performance, reliability, regulatory obligations, expected production volume, acceptable risk, and others. These requirements inevitably flow down to the team of individuals responsible for turning the concept into a working product.

In the PCB design environment, the work branches into several distinct but interdependent families of activity, including design-rule development, logical design and schematic capture, component selection, box design, material selection and stackup, component placement and routing, signal and power integrity analysis, HDI planning, documentation and configuration control, thermal consideration, and many others, often product-specific considerations. You probably have your own concerns to add to the list here.

Each of these activities often has its own terminology, analytical tools, constraints, and measures of success. Yet, importantly, none of these activities can or should operate independently. A decision that optimizes one activity may well create serious difficulties elsewhere.

Here are some examples that might help. A smaller component may reduce total PCB area but may require finer fabrication features, which can limit the number of potential vendors. A lower-loss laminate may improve RF performance but create procurement or processing challenges. A buried-via structure may simplify routing while increasing fabrication complexity, inspection difficulty, and cost. A mechanically convenient connector location may force longer high-speed routes or disrupt the power-distribution network.

It must be conceded that, at times, overall requirements make such decisions unavoidable. Ultimately, the project’s success or failure often hinges on whether these decisions are considered together.

The PCB Designer as Drum Major
The PCB designer, like the drum major, is uniquely positioned within the hierarchy, as nearly every project requirement eventually becomes a physical constraint on the board.

Electrical intent becomes placement, routing, impedance, spacing, and return-path requirements. Mechanical intent becomes board outlines, keep-out regions, mounting features, connector positions, and component-height restrictions. Thermal requirements become copper distribution, thermal vias, heat spreaders, airflow considerations, and component separation. Manufacturing requirements become feature sizes, annular rings, aspect ratios, material selections, panel-utilization choices, and process tolerances.

The PCB database is like a musical score, where these otherwise separate engineering intentions meet.

This does not mean that the PCB designer should be expected to independently solve every electrical, mechanical, manufacturing, and reliability problem. Like the drum major, it is not his role to play every instrument; the PCB designer ensures that all band members know the music and follow instructions to create a satisfying musical experience. The designer must have timely access to the people who own those problems.

I was discussing this with a well-known and respected friend in the industry, who observed that one should not ask a piccolo player to sound like a tuba. He was spot on!

Too often, real collaboration begins only after the layout is substantially complete. At that point, the manufacturing engineer reviews the design for fabrication concerns, the assembler checks component clearances and solderability, the SI specialist identifies problematic interconnects, and the mechanical team discovers enclosure conflicts. Unfortunately, you get a predictable result: redesign, schedule pressure, strained relationships among team members, and too often, costly compromises that don’t always deliver on their promises. The argument here is that earlier collaboration changes the process from error correction to much more cost-effective error prevention.

Some Thoughts on RF and HDI
RF and HDI designs are often treated as specialized classes within the larger PCB discipline. That classification is useful, but it can also create artificial boundaries.

We have learned through hard lessons that RF performance depends not simply on trace geometry, but also on laminate properties, base copper roughness (top and bottom), plating thickness and uniformity, connector selection, fabrication tolerances, and assembly consistency, along with other more arcane concerns. The RF designer, PCB designer, material supplier, fabricator, and test engineer must therefore participate in a shared and interactive design process.

HDI (and now more frequently UHDI) presents a similar challenge. Microvias, sequential laminations, few-micron lines, few-micron-thick dielectrics, microvia-in-pad structures, and stacked and/or staggered interconnects collectively expand routing capability, but they also affect cost, yield, reliability, and perhaps most importantly, supplier availability. There are simply not that many suppliers for such a product, especially in North America.

An HDI architecture cannot be responsibly selected solely because it makes routing easier. Nor should it be rejected solely because it appears more expensive on a per-panel basis. The architecture must be evaluated, again in terms mentioned earlier, in the context of total product size, layer count, assembly methods, yield, electrical performance, testing, reliability expectations, and production volume.

This same principle applies to almost every advanced PCB technology. Its value cannot be understood from the perspective of a single discipline; thus, manufacturing must become a true design participant, not an observer.

Conclusion

Manufacturing is frequently positioned at the end of the design chain, where its role is to determine whether the completed design can be built. This is the traditional home of DFM. Sadly, the fabricator is only asked to approve or reject a finished design and is being used as an inspector rather than as an engineering resource. They can offer so much more.

Experienced, top-tier fabricators are invaluable resources, possessing detailed knowledge of material behavior(s), registration capabilities, drilling and plating limits, imaging processes and their limitations, material movement during lamination, copper distribution concerns and recommendations, impedance control, and myriad other potential defect mechanisms. In addition, and not yet covered but very important, assembly specialists understand stencil design, solderability, package behavior, component warpage, thermal profiles, cleaning constraints, rework risks, and inspection limitations. That combined knowledge (manufacturing and assembly) provides its greatest value before the design architecture is fixed, which nicely sums up the purpose of this (hopefully) brief commentary.

At the end of the day, unless we collectively understand the limitations of DFM and the potential of DWM, we will continue to miss the target. DFM asks whether a design can be manufactured. DWM teaches how a product can be jointly optimized for performance, manufacturing, assembly, reliability, inspection, cost, supply resilience, and lifecycle support. In short. DFM rules are applied to the design, and DWM, or co-design, shapes the design to get it right the first time. The distinction is fundamental.

Joe Fjelstad is founder and CEO of Verdant Electronics and an international authority and innovator in the field of electronic interconnection and packaging technologies with more than 185 patents issued or pending. To read past columns or contact Fjelstad, click here. Download your free copy of Fjelstad’s book Flexible Circuit Technology, 4th Edition, and watch his in-depth workshop series “Flexible Circuit Technology.”

This column originally appeared in the August 2026 issue of I-Connect007 Magazine.