It's hard to ignore the fact that the worlds of PCB (largely digital) and high-frequency (RF and microwave) design have converged.

The wireless revolution has driven GHz functionality onto higher and higher layer-count, and more dense boards. In consumer electronics, it has resulted in an environment of smart phones with amazing functionality and the nearly wireless nature in both the office and at home of everything from telephones to printers.

In telecoms, the large racks of equipment found in cell phone base stations and central offices are now operating at speeds well into the microwave regime, while simultaneously ensuring that each bit is preserved and power is conserved.

Yes, RF and PCB have converged and there is no going back.

What is truly amazing to me, having worked in both the RF and PCB communities during my electrical engineering career, is that the wireless revolution is happening without these two worlds really figuring out (or attempting to figure out) how to work together. Back in my early days of designing, the PC was in its infancy and the whole notion of using EDA software was reserved for the "digital guys" developing ICs who were laying out thousands of gates, or for the "analog guys" who had to use SPICE to simulate hundreds of transistors. RF design was a fairly self-contained discipline with tools customized for the demands of frequencies near, above, and beyond one GHz.

Meanwhile, on the PCB side, designers were concerned with physical tangibles - managing more parts, more layers, and streamlining the combination of both to achieve manufacturability. Suffice it to say, it was possible to work in PCB or RF and not even know that guys on the "other side" existed.

Time marches on; so does Moore's law and the world's demand for wireless everything. Eventually the notion of integrating RF right onto the PCB happened - perhaps spawned from the desire to simply have everything on a single board. Since the digital circuitry is needed, in part, to control the RF, and the RF feeds the analog baseband signal for some sort of processing or display, it all seems quite natural in retrospect to have the whole darn thing on one board. As a result, a file-based translation system emerged, which works well for easily distinguishable, partitioned PCB functionality. However, when the design grows more complex and coupling takes on a bigger role (whether intentionally or not), a simple file-based translation system is not enough.

At this stage of the game, cross-pollination is necessary.

This is where we find ourselves today. The two worlds have converged. The PCB has begun to look more like RF's territory - the PCB environment has risen in frequency, driven by faster clock rates for digital chips and higher data rates for the baseband design. This is evidenced by the use of finer geometries and materials other than the venerable FR4 in complex, high layer-count boards.