What does it take to design PCB interconnects with good analysis-to-measurement correlation up to 40 GHz? Is it doable with typical low-cost PCB materials and fabrication process, typical trace width, via back-drilling and the shortage of space to place the stitching vias? This paper reports lessons learned from validation projects with the goal to build a formal procedure for systematic prediction of interconnect behavior up to 40 GHz. Topics include: selection of test structures, connectors and measurement equipment, and analysis uncertainties.
