And a plan for resolving the disconnect.
In any industry, standards provide a vital catalyst for market development by supporting a variety of assurances that are needed by product suppliers and buyers. Over time, however, the pace of technical development can outstrip standards-making processes, and a change of approach is needed.
Specifications for flammability or electrical safety such as those maintained by UL are unequivocal. Those that influence general performance, on the other hand, provide limits for parameters like dielectric constant (Dk) and dissipation factor (Df). These limits are now often simply too wide for designers to predict PCB behavior based on conformance to industry standards alone.
Today, we find that end-user demands across markets from consumer mobile and automotive to industrial automation and wireless infrastructure are pushing signal frequencies higher and higher in search of ever-faster data rates and lower latency. So, even some of the most humdrum devices we use every day must be designed within exacting parameters to meet demands for functionality, performance form factor and cost. When an accurate assessment is needed, such as knowing exactly how much signal attenuation to expect, simply knowing the substrate material conforms to a category of materials based on a broad definition of its chemistry will not provide the answers.
Tightening published limits only addresses part of the issue, when the specifications are defined according to chemistry and construction. A more performance-centric approach is needed to help designers get the right material properties for their application.
From a technical standpoint, the industry will always be ahead of the standards. Moreover, many industry standards are established through consensus of technical committees comprising experts from various stakeholders, and the goalposts tend to be wide. The resulting standards can make it difficult to be sure of performance based on compliance alone.
Could the industry benefit by changing the standards-making processes to deliver new specifications with higher minimum-performance requirements more quickly? This is probably impractical, and, in any case, proper consideration has to be given to setting workable specifications that meet a variety of industry requirements. That takes time. Elsewhere, we often see the results of rushed and poorly considered legislation. It’s to the credit of the technical committees that they take care to avoid making mistakes the industry will regret later.
Another approach is to introduce multiple performance categories within a specification. The problem here is the choices can become confusing and force designers into over-specifying materials for their application, as the tightest requirements available becomes the de facto expectation over time, regardless of cost or actual need.
Alternatively, companies that need strict performance assurances to address specific applications, such as 5G infrastructure equipment or a commercial communications satellite, can establish their own proprietary minimum specifications and require suppliers to deliver against these. We can look to the automotive industry for an example of this type of approach. However, it’s labor-intensive and negates one of the key advantages of having industry standards, which is to avoid such duplication of effort.
An alternative is for trade groups to come together to formulate specifications for materials, looking ahead to the forthcoming demands of their own industries. We see one example of this in the work done by the High-Density Packaging Users’ Group (HDPUG), which has recently started a sixth-generation project for testing materials aimed largely at datacom applications. Candidate materials are evaluated using the notoriously challenging MRT-6 test board developed by Nokia to produce a shortlist of products that have known performance to meet the advanced requirements of the datacom industry. This looks like a workable and efficient model that could be adopted by other industry groups.
The aerospace sector is another example of an industry that sets its own standards based on performance criteria to ensure suitable levels can be achieved. As performance is critical in aerospace applications, these documents are not created by industry consensus but rather are shaped by the demands of the specific application areas. I would like to see many more industries move toward this model, to show more clearly how any given material will perform in a specific application to help designers make informed choices.
More dialogue and greater understanding provide the basis for a way forward. In this case, board designers need to ask materials suppliers for more information about their process capabilities. The materials suppliers, for their part, should have the statistical process data to be able to provide practical distribution curve data for key parameters, such as those influencing Dk. In this example, board designers could then put the distribution curve data from each supplier on their shortlist into impedance-solving tools to determine which has the process capability to meet the performance needs of their application.
This provides a reliable way for designers to be sure the boards they receive will satisfy minimum performance requirements. It’s certainly faster and more efficient than individually screening all boards upon delivery to eliminate those that are outside requirements.