Just because 5G is rapidly rolling out doesn’t mean everything works and it’s time to move on. Indeed, the design challenges are just beginning, as this two-part series from IMS 2020’s 5G Summit shows.
5G technical summits have changed over the years. What started with theories has evolved into implementation and optimization. While the business world talks mostly about apps and “monetization,” engineers know that now that the technology works, it needs improvement. That’s the takeaway from the 2020 (virtual) International Microwave Symposium.
Moderator Walid Ali-Ahmad, Vice President, Samsung Electronics USA moderated the panel. Panelists, in order of appearance, consisted of:
- Naveen Yanduru, Vice President and General Manager, Renesas Electronics. Talk title: “Sub-6GHz and mmWave RFICs for 5G Wireless Infrastructure RF Front Ends.”
- Shahriar Shahramian, Director, Bell Labs. Talk title: “The 5G Quest: System, Deployment & Application Challenges.”
- Ir. Michael Peeters, Program Director Connectivity, IMEC. Talk title: “FR 1,2,3,4,… PA and FEM technology approaches for 5G and beyond.”
- Bami Bastani, Senior Vice President, RF Business Unit, Globalfoundries. Talk title: “Differentiated end to end silicon solutions for the new 5G reality.”
- David Pehlke, Senior Director of Systems Engineering, Skyworks. Talk title: “RF Front-End Evolution from 4G to 5G.”
- James Chen, Associate Vice President, MediaTek. Talk title: “5G – Evolution or Revolution.”
- Chih-Lin I, China Mobile Chief Scientist, Wireless Technology, China Mobile. Talk title: “The Myths and Facts of O-RAN Whitebox.”
5GTW will cover talks from Bastani, Pehlke, Chen, and I in Part 2, which will appear during the week of August 24.
Each presenter took a different of 5G. Looking at ICs for 5G, Yanduru discussed power, efficiency, beamforming, and cost, a theme that carried through the other talks. “LTE was all sub 6 GHz,” he said. “5G added 3 GHz to 5 GHz plus mmWave (FR2) frequencies.” See Figure 1. “Frequencies are being added around the world.”
“Digital beamforming is coming,” added Yanduru. Digital beamforming has suffered from power and heat issues because each antenna in a phased array needs its own ADC and DAC. As a result, today’s beamforming system use a hybrid analog/digital architecture. The challenge with digital beamforming is power, and with power issues come heat issues. Figure 2 shows an RF chain for a digital beaming system. In this case, showing two of 32 or 64 antennas. Each uses a dedicated DAC for transmit and a dedicated ADC for receive.
Yanduru noted that power amplifier (PA) efficiency is still a problem. “We need innovation in PA design to reduce power consumption.” That’s because PSa have to operate in the linear region to minimize clipping and other forms of distortion. As Fig. 2 shows, each antenna element consumes 4 W to 8 W and that’s too much, according to Yanduru. Another issue comes from the high cost of antenna elements in base stations. “We need more integration to reduce costs.”
Shahriar Shahramian from Nokia Bell Labs cited exponential growth in network traffic, particularly resulting from COVID-19, as the impetus for developing a new semiconductor process for radios and antenna elements: glass. “Wireless keeps getting closer to fiber in speed,” he noted. As a result, we could see more fixed-wireless access (FWA) services (Figure 3) because with wireless service providers don’t need to bring fiber or cable to the customer, but deployments needs to be optimized so that networks can provide the needed coverage while minimizing costs.
Bringing FWA to homes and businesses first requires studying mmWave signal propagation, then designing mmWave networks that deliver the services with minimal network installations. Shahramian and others designed a 384-element phased array using RFICs developed at Bell Labs based on a concept called radio-on-glass (RoG), which could result in 100 Gbit/s wireless backhaul using a D-band (130 GHz to 174.8 GHz) transceiver. Figure 4 shows the module mounted on a breakout board.
Figure 5 Shows a block diagram of the radio and how the circuits and antenna arrays integrate into a module.
Michael Peeters from IMEC looked beyond 5G’s current FR1 (sub 6 GHz) and FR2 (24.25 GHz to 52.6 GHz) bands. FR4 covers 52.6 GHz to 71 GHz and is expected to find use in autonomous vehicles and vehicular Radar. FR5 covers 95 GHz to 325 GHz. Peeters described FR3 (10 GHz to 20 GHz) as “no man’s land.” He went on to describe the challenges at each frequency band. “By 2026 and 2027, we will see IoT devices that look nothing like the low data rate devices of today.”
The challenges of getting to FR and FR5 (and beyond) also revolve around spectrum efficiency and power efficiency. “GaN technology is maturing,” he said. “It’s time to look at it. User equipment that uses GaN will depend on how the process is integrated with silicon.”
Peeters said that beamforming has advantages over massive MIMO because of fewer antennas in a device. Doherty amplifiers need to get more efficient, but that is happening.
“FR5 is a completely different beast compared to FR4. We have challenges everywhere from a UE perspective.” Such challenges span from the antenna to the baseband signals and 3D integration (Figure 6), plus calibration and beamforming issues. “We will need die-to-wafer and wafer-to-wafer stacking.” On the plus side, Peeters noted that at 140 GHz, we can get “really high power efficiency per bit numbers.” As for the semiconductor processes needed to operate at these frequencies, bulk CMOS and InP look promising.
Continue to part 2.