EE4: Figures-of-merit on trial

Abstract

Mixed-signal/RF circuits are characterized by a wide variety of performance parameters and diverse functionality. Afigure of merit (FOM) provides a unique, simple and objective metric that allows normalizing and comparing circuits and systems of the same class. On the other hand, does the minimalistic simplicity of any single metric sacrifice more than it offers? Doesn't engineering practice intrinsically require designing and judging a far more complex reality than the monochromatic reductionism that an FOM can provide? For instance, in the case of analog-to-digital converters, the ability to drive the ADC's input, to clock it, to integrate it or interface it with other processing units, to supply power to it, are just a few real-life examples of factors that can make or break a converter architecture and the signal chain embedding it. These factors are not considered in any FOM, with potentially catastrophic consequences. Enough already with the cult of FOMs? Open the doors to a new age of purely human subjective calls? You, the audience, be the judge. This panel will probe the weaknesses and strengths of popular analog FOMs in an entertaining and educational way: To this end, the room will become a tribunal with the moderator as judge. For each FOM on trial, two panelists will officiate, one becoming the defending advocate of the FOM, and the other the prosecutor, while the audience will become the jury, that will decide which of the two contestants will win.