Fracking for 5G: Reconfigurable RF and High-Efficiency Millimeter-wave Circuits to Find Elusive Spectrum

Transcription

1 Fracking for 5G: Reconfigurable RF and High-Efficiency Millimeter-wave Circuits to Find Elusive Spectrum Dr. James Buckwalter RF & Mixed-circuit Integrated Circuits Laboratory University of California Santa Barbara

2 Mixed-Signal and RF Integrated Circuits HSIC Lab at UCSB Fundamental investigation of RF and mixed-signal integrated circuit techniques that enable high-speed/high-frequency applications. 1 Post-doctoral Scholar 10 Ph.D. Graduate Students Alumni: 8 Ph.D. Alumni, 3 M.S. Alumni Silicon RF/ Microwave / Millimeter-wave High-Speed Mixed Signal Wireline/ Silicon Photonics 2

3 Why UCSB ECE? UCSB ECE MOVES DATA! Traditional strengths in Photonics: backhaul and intra/interchip communication High-speed, solid-state devices These strengths are complemented with an Emphasis in wireless communication. Emphasis on CMOS integrated circuits. Future integrated circuit technologies will leverage III-V and other material systems to enhance CMOS. UCSB is posed to develop leadership. 3

4 Is Spectrum Inexhaustible? Number of wireless devices vastly exceeds the world s population and is increasing rapidly!. New devices are coming such as wearable electronics, sensors, and radars. More devices in the same area. 1 EB = 1B Gigabytes Traffic is projected at 5 EB per month in /20/14: British Telecom demonstrates 1.4 Tb/s at 5.7 b/s/hz; estimates 35% growth in bandwidth per year. What drives increase in data demand? Cloud applications, Internet of Things 1) Data Center Traffic: Optical transport > 10 Tb/s 2) Mobile Content Delivery: Wireless > 1 Gb/s 4

5 Beyond 4G? Look back at history licensed operation 2G Cellular launched in 1991 Birth of GSM and CDMA in wireless communications Maximum data rate (EDGE) of 1 Mb/s. 3G Cellular launched in 2001 Movement towards integrated data services over cellular Maximum data rate (HSPA+) of 28 Mb/s. 4G Cellular launched in 2011 Transition to IP based communication Extensive use of MIMO and signal processing Peak data rates (currently) of > 100 Mb/s What s next for 2021? 5

6 4H? 6

7 Where will the CAPACITY come from? RF: Evolution towards more fluid use of available spectrum Carrier aggregation Simultaneous licensed/ unlicensed operation Millimeter-wave: Use of beamforming to exploit the spatial dimension for short wavelengths 50 x BW in mm-wave bands 5G 30 GHz bands 60/140 GHz for short range 70/80 GHz (E-band) for point-to-point links Market does not jump to higher frequency bands until everything is squeezed from the lower frequency bands. Is it time? Source: Qualcomm Source: R. Heath 7

8 Circuit Challenges for RF Hardware Modern Cellphone 4G (LTE) Multiantenna / Multistandard / Multiband solutions 1 W transmit, 20 MHz channels, OFDM (peak 300 Mb/s) 700/850/1700/1900/2100/2600 MHz WiFi, Bluetooth, Near-field How do you go 10X in existing bands? Most signal processing tricks have been exploited! MIMO è Massive/Multi-User MIMO (10x improvement possible) Small Cells è Offload data traffic onto broadband network Carrier Aggregation: Capacity expands with number of bands. More bands, more interference. e.g. Qualcomm announces LTE-U chip for seamless use of LTE-A and unlicensed 5 GHz bands. 3x20 carrier aggregation. Software-defined Radios ad hoc operation Cognitive Radio: Sense (Spectrum), Discriminate (Blockers), Mitigate (Interference), Learn (Patterns) 8

9 Trends in IC Design Everything starts with devices RF (< 6 GHz) circuit design is blending with mixed-signal design techniques to produce reconfigurable circuitry/ software defined radios. Golden age for RF-CMOS Design! CMOS is pushing to millimeter-wave bands. Mm-wave/ Terahertz systems are not necessarily limited solely by IC cost. Performance demands InP or GaN devices. Blending of heterogeneous IC processes: InP on CMOS or GaN on CMOS for phased arrays.

10 Conclusions Wireless access is reaching a tipping point. Improvements in wireless capacity will require digging deeper from within the congested RF spectrum for available spectrum. Opportunities to get plenty of spectrum at mmwave bands at great cost to industry. At the same time, CMOS is reaching a point of diminishing return for scaling. Enhancing CMOS through materials and packaging approaches is a critical part of future mm-wave systems. 10

11 Questions?