The explosive growth of the global communications industry has been enabled by the amazing evolution of CMOS integrated circuit (IC) technology. The relentless scaling of CMOS IC technology has made the increasingly sophisticated communication protocols, digital signal processing techniques, and microprocessors applied to communication systems practical. As a result, today’s communication systems are staggeringly complex systems composed largely of digital circuitry, and the trend toward increased digital complexity is likely to continue into the foreseeable future.
For example, by 2025, next-generation (5G) mobile communication systems are expected to handle an astonishing 1000-fold increase in data traffic, a 100-fold increase in subscribers including “Internet-of-everything” applications, and a several orders of magnitude reduction in energy consumption relative to current-generation (4G) systems. They will incorporate massive MIMO technology, extensive carrier aggregation, and operate at higher maximum frequencies and over far wider bandwidths than existing systems. These advances will be enabled by significantly increased digital signal processing (DSP) complexity made possible by the extremely high-density and low-power digital circuitry now feasible with advanced CMOS IC technology.
Yet the world remains an analog place. Communication systems inevitably require components based on analog circuitry to interface the digital circuitry with the real world, such as amplifiers, mixers, data converters (ADCs and DACs), and phase-locked loops (PLLs). Furthermore, the accuracy required of the analog circuitry tends to increase as communication standards evolve toward higher data rates. Unfortunately, the very phenomenon that enables complicated digital circuitry actually works against conventional high-accuracy analog circuitry. Each IC technology iteration inevitably involves trade-offs that favor digital circuitry over conventional analog circuitry. Nevertheless, the communication industry’s razor-thin margins and emphasis on miniaturization dictate that as much as possible of the analog circuitry in each communication device be integrated in the same IC technology as the digital circuitry.
Thus, there is a fundamental disconnect between the requirements of present and future high-performance communications systems and the evolution of IC technology. Our research addresses this disconnect. It involves the invention, detailed analysis, and development of techniques that overcome fundamental analog circuit limitations, and the demonstration of the techniques as enabling components in ICs with record-setting performance.
Most of our research contributions to date fall into three general categories: 1) statistical digital background calibration of mixed-signal ICs, 2) dynamic element matching (DEM) techniques, and 3) mixed-signal circuit techniques based on edge-time signal processing. The contributions are new techniques that represent significant departures from the corresponding prior art techniques and we support their introduction with deep theoretical analyses and industry-quality proof-of-principle ICs enabled by the techniques.
As can be seen from our papers, our research is both theoretical and experimental. We develop and analyze highly-mathematical techniques and then validate them in state-of-the-art ICs. While the development of enabling new techniques and supporting analyses are the primary objectives of the research, a goal of each research project is to apply the new techniques to implement ICs that achieve record-setting performance relative to all other comparable ICs. This enables fair evaluations of the new techniques; whenever record-setting performance can be achieved in a mainstream application using the new techniques with otherwise conventional methods, it stands to reason that the new techniques have significant value. Moreover, the IC design field is very mature and competitive, so claims of improved techniques must be backed up with experimental verification. Such experimental results also encourage rapid adoption of our new techniques by industry. Many of the techniques have been licensed by companies including Analog Devices, Cambridge Silicon Radio, Maxlinear, Microchip, Motorola, NXP Semiconductors, Qualcomm, RFMD, and Texas Instruments, and are now widely used in consumer electronics products including mobile telephone handsets, Bluetooth transceivers, and digital television tuners.