An Efficient mm-wave Integrated Circuit Synthesis Method with Accurate Scalable Passive Component Modeling

• Main Authors: Zhijian Pan, Wei Zhu, Qiang Yao, Di Li, Zuochang Ye, Yan Wang
• Format: Conference Proceeding
• Language: English
• Subjects: Computational modeling; Data models; Design automation; Integrated circuit modeling; low-noise amplifier; Mathematical model; mm-wave ICs; Optimization; scalable modeling; Semiconductor device modeling
• ISSN: 2375-0995
• DOI: 10.1109/RFIC.2018.8429004

With the operating frequency of radio-frequency (RF) integrated circuits (ICs) ascending gradually to milli-meter-wave (mm-wave) regime, the RF IC design automation methods encounter great challenges due to the complicated distributed effects and parasitic effects. In this work, a new synthesis framework for mm-wave ICs is presented, which is featured by two progressive stages: offline preparation and online synthesis. In the former stage, to cope with the difficulty of mm-wave IC synthesis caused by passive components, a scalable modeling method is proposed, in which geometric parameters are incorporated into rational functions to accurately model the S-parameters up to 120GHz. Benefited from the dedicated offline preparation, during the online synthesis, the circuit performance evaluation and optimization are carried out without the time-consuming EM simulation. High-quality solutions can be obtained by using evolutionary algorithms with enough iterations. We applied the proposed approach to the design of a four-stage differential wideband low-noise amplifier (LNA) covering various mm-wave applications. The synthesized LNA is implemented in 65nm CMOS technology and the measured results show that it achieves the highest bandwidth (34GHz) with other comparable performances to the similar state-of-the-art CMOS broadband LNAs. The synthesis only costs 26min, which is more than 50x time speedUP compared to the existing mm-wave synthesis methods.