Space Coding Applied to High-speed Chip-to-chip Interconnects

This dissertation presents new signaling schemes and circuit architectures for reducing the power and cost of high-speed chip-to-chip links. After an overview on chip-to-chip interconnects and its building blocks, a new signaling scheme is proposed that can provide many advantages of a fully-differential signaling scheme while employing as few as N + 1 signal paths for communicating N differential signals. Next, power-efficient signaling schemes that use channel coding to achieve appreciable coding gain are proposed. One discussed method is to use 6-PAM signaling instead of 4-PAM to transmit two bits of information per channel. The proposed low-complexity architecture of this scheme makes its high-speed implementation feasible. A realistic model for a typical interconnect channel is used for simulations. We then introduce a coding scheme which employs a convolutional encoder in space to achieve 3-5 dB coding gain without expanding the modulation from 4-PAM to 6-PAM. The functionality and performance of the proposed scheme is verified by experimental results obtained from a fabricated chip based on this method. Finally, a novel power-efficient architecture for multi-level PAM drivers is presented. In addition, a data-look-ahead technique, used for high-speed implementation of this method, eliminates the need for a pre-driver to further reduce the driver power. Based on this architecture, a 4-PAM transmitter is designed in 0.18μm digital CMOS technology. The transmitter achieves 10 Gb/s with a 2-V supply and it occupies an area of 0.16 mm. The output driver and the entire transmitter consume only 20 mW and 121 mW at 10 Gb/s, respectively, which are the lowest reported powers at this speed. i Acknowledgements During the course of this research and the writing of the dissertation, I have been very fortunate to be surrounded by teachers, family, friends, and colleagues who have continually offered me support, encouragement, and aid. I would like to express my deepest gratitude to my advisor, Professor David Johns, for his continual support, guidance, and friendship. This work would not have been possible without his excellent teaching and great insight. I really feel that I could not have asked and hoped for a better advisor. I would also like to thank Professor Kschischang, Professor Loeliger, Professor Konrad, Professor Ng, and Professor Genov for taking the time to serve on my committee and for reviewing my thesis. Also, I like to express my appreciation to Professor Kschischang for his great comments throughout this research. I feel very proud to be a part of such a talented research group: Bahram Zand, Amir Hadji-Abdolhamid, Prof. Tony Chan-Carusone, Samira Naraghi, Sherif Abdalla. I would like to thank them for their friendship and for offering unconditional assistance throughout this project. I am also grateful to my colleagues, Prof. Shahriar Mirabbasi, Prof. Anas Hamoui, Afshin Rezaee, Mohammad Hadji-Rostam, Afshin Haft-baradaran, Farshid Rezaee, Kasra Ardalan, Mehrdad Ramezani, Sotoudeh Hamedi-Hagh, Saman Sadr, Jorge Verona, Masoud Ardakani, Mehrdad Shamsi, Mehrdad Eslami, Yadi Eslami, Takis Zourntos, Sebastian Magierowski, Tooraj Esmailian, Reza Jafari, and many others for being there for me and for all the good time that we had together. Going through the difficulties of the graduate years would have not been possible without the joy, fun and happiness that have been brought to my life by great friends such as Kayvan, Farzad, Amir, Samira, AfshinHF, Afshin, Farshid, and Bahram. It is certainly impossible to find the appropriate words to thank my wife, Arezoo, for her love, support, and patience over the years. Her presence gave me the will and strength to go through the difficulties in my graduate years. Also, I would not have been able to carry myself all the way through the Ph.D. degree without my parents, sister, and brother. ii Growing up, I never had to worry about anything but my education. Thanks to my parents for all their sacrifices and all they have done for me. Thanks to my sister, Nasrin, and my brother, Kayvan, who cheerfully took over all my responsibilities at home during the term of this research. Finally, I would like to dedicate this thesis to my wife, who endured this long process with me, and to my mom and dad, for being the best parents in the world.