Every decision made in the design process of any PCB or system has an economic as well as an engineering effect. Some of these effects are more visible or tangible than others but if the “dollars and cents” elements are weighted along with the design decisions it is more likely that that the resulting product will be both economically efficient and well-engineered.
As with most everything in life, price and cost are not the same things. Successful PCB product realization is a carefully orchestrated balance of trade-offs between design considerations and overall cost of the product. Often, PCB suppliers—fabricators and assemblers–are chosen based on the price of the PCB delineated in a purchase order rather than on the ownership cost of that PCB. Beyond design, total PCB cost includes the cost of fabrication, assembly, rework, test, repair, scrap, replacement and field service costs. The lowest-priced PCB is often scrapped during final assembly due to poor PCB quality.
This article will address the various factors that contribute to the overall cost of a PCB and the trade-off decisions that go into achieving the right balance between price and cost while ensuring that all elements within a PCB are right the first time.
Board Economics—The Basics
Layer Count: Obviously, the greater the number of layers, the more expensive it is to design and produce a board. Most of the time, it is possible to implement RF and microwave designs in two or three layers. High-speed digital designs and more complex designs need more than two layers but it is often the case that the layer count is higher than it needs to be. Most often, this is due to poor layout and routing, incorrect component placement and incorrect assignment of power and ground layers. In some cases, the layer count can be too low which results in malfunctioning PCBs. The ultimate goal should always be to have enough layers to make the design electrically sound but also have it be the most cost-efficient from a manufacturing standpoint.
Laminate Selection: As noted in the references at the end of this blog, the key factors in laminate selection have to do with the type of product being designed and the key properties of the laminates that best suit that product design. These key properties include Tg (glass transition temperature; er (relative dielectric constant); Tan f (loss tangent); DBV (dielectric breakdown voltage) and Water Absorption (WA). But, material selection is also driven by some economic factors.
- Avoiding the use of “exotic” or more “expensive” board materials. If you select board materials that are not widely used, you may have to pay a premium for that material. Or, the supplier may require that you purchase far more material than you need for your particular design. In addition, the fabrication cycle may be extended while special order material is manufactured.
- Avoiding the use of sole-source materials. if you select a sole-source material, you may have to pay a premium for it. This is basic supply and demand economics—if only one supplier carries your selected material, that supplier can basically charge whatever they want because they have no competition in the market.
- Selecting materials that your fabricator uses on a regular basis. You want to select a fabricator that has a lot of experience with the material you have selected because they are more likely to have the chemistry used in their board fabrication process correct for that material, and they will also have experience with whatever subtle nuances may arise with the material. It is also important to make sure that your fabricator uses the laminate from a single supplier. Even if the laminate from different suppliers has the same key properties, there can be subtle differences in how the laminate was produced that can impact overall board quality. For instance, woven glass, which is one of the three components that comprise a laminate, always shrinks during the heat of lamination. Laminate from different suppliers shrinks by different amounts. When a fabricator uses laminate from more than one supplier in the production of a single PCB design, this difference in shrink rate often results in warped PCBs. Warped PCBs have to be scrapped and even if the fabricator assumes the responsibility for the warped PCBs, the schedule impact of having to reproduce those boards can affect overall time-to-market of the final end product and resulting product revenues.
Cheaper is Not Always Better
Often times, there is a tendency to go with the knee-jerk response to pick the cheapest board fabricator. With corporate mandates to cut costs wherever possible, it can be a real uphill battle to select a fabricator that charges more for its PCB production services. That’s why it’s important to pay the most attention to the overall impact of the choices that are economically driven. There are two critical factors to remember:
You get what you pay for.
It is rare that the lowest PCB bid turns out to be the lowest in overall cost.
Cost of the fabricator selected must also take into account the fabricators that keep the risk of failed prototypes and production PCBs to the absolute minimum. A failed PCB involves not only the cost of the board itself but the components mounted on it, the troubleshooting labor used to determine the cause of the failure and the cost of elapsed time while scrapped PCBs are replaced. Note: this particular cost can be much larger than all of the other costs combined. Costs associated with failures that occur during the prototyping process include the cost of project delay as well as the cost of delayed time-to-market.
As noted in the references at the end of this article, sales engineers who work for various fabricators often have a tendency to oversell customers on the capabilities of their facilities. In a robust economy, this over-selling is based on the desire to “move up” the scale in terms of the complexity of PCBs that a facility can manufacture. More often, it’s based on a “let’s get the contract first and we’ll worry about how to build the boards later” philosophy.
When you factor in today’s global manufacturing environment, it gets even more confusing. PCB fabrication facilities in China, Taiwan, and Korea have moved significantly up the scale in recent years in terms of the complexity of the boards that they can manufacture. But, they may not have the right equipment and processes in place and their engineers may not have the skill set needed to manufacture complex boards. The fabrication notes and stack-up drawings that are shipped with designs can be given very short shrift or ignored altogether. This results in production designs that don’t work. Right now one of the bigger challenges is offshore manufacturers who try to save their clients money by using a different laminate other than what is specified. This often occurs without the PCB owner’s knowledge or consent.
What all of the foregoing means is that the ownership of a PCB, from inception to final product assembly has to rest with the company developing the product. All too often the excuse is made that “we don’t have time to do a thorough investigation of PCB fabricators” or, “we can’t afford to send a team over to investigate a fabricator’s facility and its processes”. These same companies will agree that three or more respins of a PCB is the natural order of the universe.
This is no situation where investing the time and money to build a PCB right the first time is ever going to be more costly than having a failed product, having to do multiple respins of a board or having to switch fabricators during a production program. In fact, just the opposite is true. Designing, fabricating and assembling a board right the first time is the one way to ensure that you have designed a product that is not only functional from an engineering standpoint but also very cost-effective.
The More Subtle Side of PCB Cost
In addition to basic design and production costs, there are other appurtenant costs associated with a PCB that must be taken into account. These include EDA tool costs and knowing when to preserve a legacy design or move onto the next generation design.
The cost of ownership of an EDA tool involves not just the purchase price of the tool but also any license and maintenance costs that are imposed as part of the overall tool cost. And, as noted in the two references at the end of this blog, a robust PCB design process assures that a successful PCB implementation is based upon a virtual prototyping design flow that includes a variety of toolsets as depicted in Figure 1. There are some additional guidelines to follow relative to these tools:
If you go and purchase all the tools you may need at once and attempt to integrate them into your design environment, you can bring the efforts of your design team to a screeching halt.
Any EDA tool is only as good as the engineer operating it. No tool will ever take the place of an engineer and no engineer should relax his or her design discipline based on the availability of a particular tool.
Figure 1. PCB Design Flow with SI Engineering Tools
The places in the process where most failures occur with present logic technologies are signal integrity and the design of power delivery systems. Therefore, it is advisable to introduce the tools and methods that address these issues first. In the Virtual Prototyping Design Flow, this tool would be the signal integrity analyzer that includes the power plane simulators. As the skill level rises in these areas, the remaining tools can be added.
Still today there isn’t enough being invested in properly training engineers in either design practices or tool use. When there are economic pressures within a company, it is often the training budgets that are severely cut back or eliminated altogether. At the same time, engineers are under pressure to develop products that do work right the first time, meet critical time-to-market windows and are brought in under tightened product development budgets. Without all of the necessary weapons in a design arsenal, including the necessary training in design skillsets, it can be extremely difficult if not impossible to meet these goals.
Unfortunately, there is still a disconnect between how training is valued in the United States versus other countries. In the U.S., the continued training of engineers is often viewed as an expense with an ROI that is difficult to measure. In other countries, electrical engineers are regarded as professionals similar to lawyers and doctors and the cost associated with their continued training is viewed as an investment not just in the engineering staff but the overall longevity and profitability of the company for which they work. Overseas, electrical engineers are required to continue their training as part of their ongoing certification process.
Legacy vs. Next-Generation Designs
There are still a number of companies who want to move their legacy designs forward using newer component technologies. While they may understand that there are some design challenges—the use of components that are of a similar process technology; interface issues with other design elements—they don’t often do a critical analysis of the old design to determine if it is still really viable from both an engineering and economic perspective. Utilizing old components for some elements of a design, such as keeping a component in a QFP package because it has always worked in the package in the past, carries the risk of a design failing from the outset. And, assuming that integrating these new elements in the old package is perhaps a plug-and-play process with a few design twists is probably naïve at best from an engineering standpoint and not very cost-effective from an economic standpoint.
Sometimes, a long-standing design should be more realistically regarded as an outdated design. While conventional wisdom would say that if 50 percent or more of a design needs to be modified to bring it up-to-date, then it’s probably not cost-effective, there are other factors to be considered. The legacy vs. next-generation criterion should be applied when:
There are elements or components within the design that are still available but becoming difficult to obtain.
Updating the legacy design ties up significant design engineering resources for several man-months.
The overall market life of the redesigned legacy product is less than 18 months.
It’s a natural desire to try and save money wherever possible. But, what may at first blush seem like a “great deal” in terms of PCB design and fabrication can actually end up costing more money if a design has to be scrapped, reworked or is delayed significantly beyond original time-to-market windows. A thorough examination of the entire project that carefully takes into account both technical and economic factors and strikes a realistic balance between the two is the course that will most likely ensure that a product will be well-engineered, cost-effective and work right the first time.
Ritchey, Lee W. and Zasio, John J., “Right The First Time, A Practical Handbook on High-Speed PCB and System Design, Volume 1.”
Ritchey, Lee W. and Zasio, John J., “Right The First Time, A Practical Handbook on High-Speed PCB and System Design, Volume 2.”