material gains
High-Speed Telecom Requires ‘Dialed-In’ Materials
Stronger, faster and more robust networks will be one of the positive legacies of the pandemic.
Hopefully, we will soon be able to start living our post-Covid-19 lives. Going forward, some of us want fundamental changes. Others are keen to return to the way things were. Although we will be pleased to put this situation behind us, some things are here to stay. One, obviously, is the lethal group of coronaviruses that will surely continue to take lives after lockdown (we hope at a greatly reduced rate). Another, I believe, is the tendency for many of us to continue working from home (WFH) to a much greater extent than before.

WFH has been one of the headline trends of this crisis. Although clearly not to everyone’s taste, it could turn into a revolution founded on the internet technologies that allow us to meet with colleagues online, access data and tools remotely, and benefit from high connection speeds wherever we are – wired or wireless. That so many can do meaningful work this way also reflects the soft nature of many tasks associated with getting things done in developed economies. These soft deliverables liberate us from location and will be critical to our economic survival of this pandemic.

Many are keen to recover the social dimension to our working lives. While physically working together in the same space and time to achieve shared goals is a powerful part of team building and cohesion, we can also take advantage of the flexibility to ease some of the more stressful aspects, such as traveling and being away from loved ones.

WFH has also powered a rise in IT equipment sales as lockdown periods have been enforced across the world. Internet use has surged 30% in the past months, challenging the networks to support this explosion of connectivity. In addition, service providers and end-users are more aware of security vulnerabilities and the importance of protecting people and assets against cyberattacks. As if securing connected devices within a restricted campus is not difficult enough, widespread WFH significantly extends the attack surface available to hackers.

Among the techniques put forward to protect and manage telecom networks, AI is perhaps one of the most powerful and exciting. The ability to predict the usual and identify the unusual in its midst is the key to AI’s power in these roles. By anticipating usage patterns, AI can help network operators provision resources optimally to ensure resilience and maintain uptime. At the same time, the technology has become sophisticated enough to distinguish the many small aberrations that characterize human behavior from the more seriously unusual events likely associated with fraudulent activity and cybercrime.

Underlying everything, of course, are demands for greater speed and capacity driven by the uptick in reliance on the internet for work and social connections. This involves committing to rolling out the new generations of high-performance hardware needed to handle the extra subscribers, extra traffic density, and extra speeds. In the last telecom boom, just before the 2001 dotcom bust, I worked with leading hardware vendors involved in building incredibly large multilayer telecom switch backplanes, up to 60″ x 48″, with as many as 40 layers. Getting these boards to operate at the speeds required, while also achieving the thermal performance needed to ensure reliability, proved extremely challenging.

While the physics of the situation have not changed in the intervening years, the materials at our disposal have. So, too, has our understanding of the interdependencies among parameters like signal speed, trace dimensions, energy loss, and dielectric properties. Getting the best tradeoff is critical to meet the world’s connectivity needs going forward.

Skin effect is the greatest limiting factor in data transmission through PCB traces and depends on trace dimensions, particularly trace width. As signal frequency increases, the region at the edges of the trace where current, and therefore data, flows becomes thinner. Widening the trace can combat the restriction: the wider the trace, the more “skin” for current to flow, and hence more data can be transmitted. More skin from wider traces also helps reduce skew and lower trace resistance, which, in turn, lowers the system power requirement.

PCB geometry places a burden on layer usage that could be relieved if more layers are added in the same or smaller physical structure. However, as dictated by the dielectric constant (Dk) of the material, thinner spacing between layers requires smaller, narrower traces as the trace gets closer to the reference plane. This clashes with the drive to increase trace widths for increased data speeds and lower power. Smaller or narrower traces also increase pressure on PCB manufacturing processes. Larger traces are easier to manufacture – at every stage of the process, from imaging to etching – and permit better yields.

We have pointed out dissipation factor (Df) and Dk are important parameters to consider when choosing a dielectric material for high-performance, multilayer PCB fabrication. With a lower Dk, the trace width can be wider, thereby increasing its effective data carrying capacity, while at the same time easing manufacture, reducing pressure on end-of-line yield. Historically, however, lowering Df tends to drive up Dk. The latest PTFE materials offer a new tradeoff between these important parameters, permitting low-loss materials with Df in the region of 0.0015 to have Dk about 2.6. These properties offer a much better tradeoff between trace width and layer thickness than previous low-loss materials, enabling large, high-speed multilayer boards that are affordable and reliable.

Currently, we regard the latest PTFE low-loss, low-Dk materials as exotic. With familiarity, they will undoubtedly be standard, especially when the next generation arrives to deliver even greater performance.

Alun Morgan smiling
Alun Morgan
is technology ambassador at Ventec International Group (;