DFX--facilitating EMS/OEM communication: effective DFX can translate to big savings in cost and process inefficiencies.

In the electronics manufacturing services (EMS) world, price is a combination of visible cost, waste and profit. The challenge for many EMS providers is that, while most of their original equipment manufacturer (OEM) customers expect regular cost reductions, many are reluctant to participate in teaming efforts that focus on reducing either the cost or waste components of the equation.

Visible costs include: material costs, assembly time, test time, logistics, handling and administration. Waste can include function redundancy between EMS provider and OEM, inefficiencies and quality issues caused by poor design, process inefficiencies at the EMS provider or OEM, and quality or delivery issues caused by poor supply chain performance. In all cases, achieving significant cost reduction requires focus from both parties. The term design for X, or DFX, represents any design enhancement that contributes to reduced cost. For this article, representatives from a broad range of EMS providers were polled for their perspectives on common design flaws they encountered, impediments to DFX initiatives and strategies for achieving DFX-related cost reduction with their customers. Their conversations follow.

Typical Design Issues

"90% of cost is determined by design," said Scott Morey, president of the Morey Corp. (Woodridge, IL). "We take overall responsibility in the development process not only for the design itself, but for component selection, overall board layout and product packaging."

Morey was also quick to point out that an additional benefit of the EMS company taking the leadership role in product development is that EMS companies are expert in product manufacturing and, therefore, see product development as a cost.

"Many OEMs don't understand the cost of the development or don't see the relationship between development time and development cost or the opportunity cost of lost sales when product windows are missed" Morey added. "They are focused on activity vs. results. Our focus is on identifying the critical path and staying focused. Our people understand that the scope of our responsibilities includes managing the customer."

Genesis Electronics Manufacturing (Oldsmar, FL) also believes in a strong EMS role in both design and redesign.

"When we are building the entire assembly, we attempt to provide cost reduction opportunities in the mechanical as well as electronic components," said Robert Stoller, president, Genesis Electronics Manfacturing. "There are frequently opportunities to combine daughter boards and smaller boards onto one main board through the use of sophisticated components. Some believe that conversion from PTH [plated through hole] to surface mount reduces cost, but that is generally not the case. But, when an entire assembly can be reduced by integrating its function on the main assembly, there is significant savings." Stoller added that significant cost reduction opportunities can be found in enclosures, including unnecessary sheet metal tolerances that impact both cost of the sheet metal product and final assembly efficiency. He also has seen many OEMs specifying the same finish both inside and out.

"These items are 'low hanging fruit' and can be resolved without too much technical manpower," Stoller said. "Overall reductions can amount to 50% of the original cost."

"A design's ability to allow automated and optimized component placement can cut a tremendous amount of manufacturing labor," said Steve Trnka, project coordinator at Winland Electronics, Inc. (Mankato, MN). "We just recently had an example in which a minor modification to the PCB layout allowed a different connector type to be used that could be placed with automated rather than manual equipment. Not only did this result in a 4% cost reduction, it also reduced the quality concerns that can be associated with manual operations,"

Danny Baudoin, program manager, and Jim Reeder, director of engineering, at Skyline Electronics, LLC (Colorado Springs, CO), had an exhaustive list of typical design flaws that impact quality and manufacturability. "Typical design flaws include: silk screen reference designators inside the component footprint that are hidden once the component is placed; no solder mask between pads, which can be particularly bad on QFP [quad flat pack] and BGA [ball grid array] pad/via traces; solder mask over test nodes; solder mask over tin/lead that blisters at wave solder; solder mask registration problems; unbalanced copper density that causes thermal expansion and warp; no thermal relief around pads; uneven BGA pads caused by hot air solder level [HASL]; poor choice in plating such as electroless tin on low-volume boards; poor lead/pad/hole diameter ratios; poor component orientation and placement relative to reflow or wave solder processing; edge connectors with rails for support during wave solder; breakaway tabs under connectors and negative break perforations that break into traces or pads," said Baudoin.

"There is a tendency for OEM designers to specify components and PCB form factors that are well-suited for the performance desired but not so well suited for yield, availability or adequate process window limits," said Russ Steiner, manager, marketing and business development, TotalEMS (Logansport, IN). "That includes PCB size and shape. For example, the size and shape (or form) often play a big role in how effective panel yield is. We have been able to get an extra row or column (sometimes both) in a process panel or solder pallet by changing the length or width of an end product by as little as a few thousandths of an inch. Think of a PCB form that might fit one column wide by one row deep on a process panel due to its size or aspect ratio. If that shape can be modified enough to yield two columns by two rows, yield dramatically increases. It is not the cost of the square inches of bare PCB where the money is saved. It's the reduced amount of handling each end item due to increased panel yields. Many incremental improvements add up to big process improvements. That is why DFX is an embedded value adding NPI procedure that TotalEMS employs without premium."

"It is hard to pinpoint one particular design flaw," said Dennis Gleason, VP sates/marketing, Wolf Electronix, Inc. (Orem, UT). "We tend to see them all frequently. For example, we see things such as lack of panelization consideration, missing fiducials, the use of extremely small components to be leading edge even when there are no footprint issues, designing double-sided assemblies when single-sided would suffice, still designing labor intensive versus automated, and a 'biggy' is still designing through hole instead of surface mount."

According to Jim Lanigan, director, engineering and operations at Preco Electronics (Morton, IL), manufacturability and cost of assembly are not the only sources of cost reduction to be considered in design.

"The most common design flaw that drives cost is sole sourced components," said Lanigan. "If the design group has enough foresight to look at the high dollar components (typically the 80/20 rule) and select at least two sources, the product can be more competitively bid. Following closely behind are tight process control windows. Special assembly tooling, inspection and test processes can be minimized if the OEM can design and comply with standard IPC guidelines."

Greg Hessler, program manager at Kimball Electronics Group, also feels materials selection should be a key area of focus.

"Common design flaws include inflexibility regarding component values," said Hessler. "Also, there may be a lack of component commonality with existing product. I also see designers reinventing the wheel versus choosing off-the-shelf technology."

Why DFX Attempts Fail

For a variety of reasons, OEMs do not embrace all EMS provider DFX suggestions. The top three are: lack of internal resources to validate or implement the suggested change, cost of agency or other third-party approvals and the time factor involved in implementing a change.

"Occasionally, there is an issue with an engineering department's perception that an outsider's expertise infringes on its domain," said Stoller. "One would like to think that this mentality has gone by the wayside, but there remain pockets out there, especially in small to medium-sized companies and sometimes in older companies where knowledge and expertise is power."

"In most cases it comes down to resources and time needed to review these recommendations," added Trnka. "There tends to be reluctance because many of the people that would approve such changes are not the ones tasked with finding the cost savings."

Selling DFX Recommendations

"In the development phase of a project, these types of recommendations are much easier to implement," said Lanigan. "By having a partnership mentality in the EMS/OEM relationship and working together, many of these design issues will and can be addressed. In the production phase of a project, lack of resources and insufficient ROI [return on investment] for the redesign/requalification are the key elements that hold back suggested changes. The EMS provider can help alleviate these roadblocks by sharing in some of the redesign and qualification effort and sharing the savings with the OEM."

"We often create a virtual prototype of the entire proposed circuit card design," said William Johnson, engineering staff lead, Sypris Electronics (Tampa, FL). "Straightforward design characteristic outputs supplied by the customer allow Sypris to create the virtual prototype. We then simulate the complete design, fabrication and assembly of the proposed design and compare it to a set of design rules for circuit board fabrication and surface-mount assembly, reflow and testing. Compliance to the design rules ensures a build and assembly cycle with maximum yield and quality product."

"Sometimes, OEMs feel the pressure of time-to-market so intensely that they want to skip the DFX process and go directly to the production floor without consulting their EMS partner during the initial design phases," said Steiner. "In many cases, Rev. B is not far behind. Then a delay is introduced when it's least desirable, during ramp-up or full production. The effect of avoidable design changes is magnified when manufacturing is done offshore. When DFX is implemented and open discussions take place between the OEM designer and the NPI team at the EMS provider, best results are ensured. The key to success in implementing a DFX exchange without it adversely affecting the design cycle is to get involved early on. Much like day or concrete, the best time to make changes in a PCB-based product design is before it is firmly set."

One of the major selling propositions of the EMS industry is that EMS companies are manufacturing experts and that OEMs can benefit from the best practices evolving on the production floor of companies dedicated to manufacturing as their primary core competency. As the comments above indicate, in some cases the EMS provider is seen as the expert and trusted partner in product development. In other cases, the EMS provider is out of the loop until manufacturing begins. The key to building trusted partnerships seems to center around getting involved early in the process, having a well-enunciated design/redesign methodology, providing written guidelines for manufacturability/testability and being able to quantify savings in ways that allow OEMs to understand the return on investment.

Will doing these things ensure DFX recommendation adoption? Not always. Particularly in a competitive EMS market, some OEMs will always look to push the cost of bad design on their EMS provider by refusing to implement cost reduction recommendations, yet continuing to press for pricing reductions. Ultimately, the key to successful OEM/EMS partnerships is the ability to clearly define joint goals and work together to achieve them.

A Checklist for New Product Design

Often, third-party designers fail to consider manufacturability, testability or procurement issues in developing new products. Bob Hilliard, technical services manager of Sypris Electronics, has provided a simple checklist of issues to consider:

* Orient parts in the same direction so that process and optical inspection are optimized.

* Co-locate similar technology parts to minimize printing issues that stencils may not be able to overcome.

* Provide useful fiducials.

* Configure the entire assembly with thermal management in mind: isolate small, thermally sensitive parts from heavily heat-sinked components.

* Consider test planning as a priority in the design to minimize the non-recurring expense factor.

* A new bill of material will often specify obsolete parts. Look to the EMS provider for recommendations for additional component sources or alternate parts.

A Successful Design Cycle Formula

Scott Morey, president of the Morey Corp., outlines his company's formula for success in the design cycle:

* Developing a good product specification is critical. Without a good initial product spec, quantifying the impact of changes is difficult. The test specification also must be developed early in the cycle.

* A detailed program plan is also needed. This document should coverall key milestones for the development effort, including hardware and software design, field testing, certification and marketing. Identifying the critical path is important.

* Assign a strong program manager. A good program manager is focused on achieving planned goals, which can mean managing the customer's team.

* Make sure all engineering disciplines have early involvement and that communication is continued throughout the process.

* Have written design guidelines.

* Maintain open communication. Quantify the impact of proposed changes in terms of cost and time.

* Be honest. Tell the customer what you think will solve an issue. At the end of the day, the customer can either take the advice or not. If they don't take the advice, you have to do the best you can do.

* The philosophies espoused by W. Edward Deming and Eliyahu M. Goldratt are key and focus on strategies that eliminate waste, reduce process variation and address bottlenecks.

Susan Mucha is president of Powell-Mucha Consulting, Inc., El Paso, TX; email: [e-mail].