Waiting For 5G Technology

New wireless standard will significantly speed up communication, but dealing with mmWave technology isn’t going to be simple.


For some time, carriers, equipment OEMs and chipmakers have been gearing up for the next-generation wireless standard called 5th generation mobile networks, or 5G.

5G is the follow-on to the current wireless standard known as 4G, or long-term evolution (LTE). It will enable data transmission rates of more than 10Gbps, or 100 times the throughput of LTE. But the big question is whether 5G will disrupt the landscape or fall short of its promises.

Regardless, the 5G market is heating up. Anokiwave, Broadcom, Intel, Qorvo, Qualcomm, Samsung and a growing list of others are developing 5G chips.

But there are a multitude of challenges to deploy a 5G wireless network. For example, although OEMs and chipmakers are developing 5G products, the standards are still not set.

Today’s LTE networks operate from 700 MHz to 3.5 GHz. In comparison, 5G will not only co-exist with LTE, but will also operate in the unlicensed or millimeter wave bands. This involves the band of spectrum between 30 GHz and 300 GHz, which in turn enables more wireless data capabilities.

Initially, 5G will likely operate at 28 GHz (United States) and 39 GHz (Europe). Over time, 5G could involve other spectrum, such as 60 GHz, 71 GHz to 86 GHz, or even 300 GHz.

For this, mobile systems and base stations will require new and faster application processors, basebands and RF devices. In fact, 5G will require a new class of RF chips, including a mmWave technology called phased-array antennas. Already used in aerospace/defense applications, mmWave devices are moving into the automotive radar, 60-GHz WiFi, and eventually 5G.

Bringing mmWave technology from the aerospace/defense sector into the commercial markets is no simple task. “Millimeter wave technology presents some challenges,” said Peter Rabbeni, director of RF marketing at GlobalFoundries. “The designs, the operations of those designs, and testing at millimeter wave can be challenging. This is mainly because of the frequencies that are associated with those bands.”

Designing these chips is difficult, but testing them is even more challenging. “We, as an industry, have done mmWave measurements for a long time, but that’s primarily been in aerospace/defense applications,” said Eric Starkloff, executive vice president of sales and marketing for National Instruments (NI). “And that’s been done at very high cost points. We are going to have to scale test costs down significantly at mmWave in order for it to be viable.”

Despite the challenges for 5G, Verizon hopes to roll out some 5G services in the United States by 2017 or so. In addition, Korea Telecom, along with Samsung, plans to offer 5G services during the 2018 Winter Olympics in Korea.

Source: Samsung

Source: Samsung

But the mass deployment of 5G isn’t expected to occur until 2020 and beyond. “I’m very skeptical about 5G as a generally available cellular service, even by 2020,” said Will Strauss, president of Forward Concepts. “Sure, there will be network trials by 2018 for higher speed cellular operation, but the likelihood of consumers being able to buy 5G cellular handsets is quite small by then.”

Still, the industry needs to keep a close eye on 5G. To help vendors get ahead of the curve, Semiconductor Engineering has taken a look at the following parts of 5G—the status of the technology; the chips and associated processes for 5G handsets and base stations; and the test and packaging flow.

What is 5G?
So what exactly is 5G and why do we need it? Today, many carriers have deployed a 4G wireless standard called LTE Advanced (Release 10). Within that framework, carriers have deployed Category 4 and Category 6 LTE-A mobile networks, which enable data downlink speeds of up to 150Mbps and 300Mbps, respectively.

Over time, carriers will launch LTE Advanced Pro (Release 13), which is considered 4.5G technology. LTE Advanced Pro is a stepping stone towards 5G. It will consist of up to 32 component carriers (from 5 in Release 10), massive MIMO, and LTE in the unlicensed spectrum, according to NI. Many of those same elements are part of 5G, but 5G will take it a step further by operating in the mmWave spectrum. For LTE Advanced Pro, carriers will deploy Category 10, which supports data speeds of 450Mbps. Some are even talking about Category 16, which provides 1Gbps.

LTE-A might be good enough for now. The problem is that mobile data traffic is expected to rise at an annual rate of 45% from 2015 to 2021, according to Ericsson. The data traffic per smartphone is projected to reach 8.9-GB per month in 2021, compared to 1.4-GB per month in 2015, according to Ericsson.

For that reason and others, the world may require 5G. Each carrier will likely have a different set of 5G services, but basically, the technology consists of three separate elements—enhanced mobile broadband; the Internet of Things (IoT); and machine-type communications.

Enhanced mobile broadband involves mmWave technologies that will enable data transmission rates of more than 10Gbps. Compared to 4G, 5G is expected to provide 1,000 times more capacity and one-tenth the latency.

The IoT, the second element of 5G, involves WiFi-based technology. For this, the industry has recognized a narrowband, wireless standard called NarrowBand IOT (NB-IOT).

Meanwhile, the machine-type communication part of 5G involves a separate machine-to-machine protocol. For this, the industry has recognized a M2M wireless standard called LTE-M.

“Some of the requirements for 5G are for lower power and a longer battery life,” said David Hall, principal marketing manager at NI. “Both (NB-IOT and LTE-M) are modifications of mobile communication networks. They are designed for machine-to-machine. As a result, the radios are simpler.”

The problem? Other WiFi wireless technologies could also find themselves moving under the 5G umbrella, creating more complexity, uncertainty and chaos in the market. For example, 5G may include a 60 GHz WiFi technology known as WiGig. Other wireless technologies are also emerging, such as LoRa and Sigfox.

But it’s unlikely, if not impossible, to build a RF radio chip that can support all of these standards in every country. “Can you try and satisfy all of these requirements simultaneously? No,” said Thomas Cameron, chief technology officer for the Communications Infrastructure unit at Analog Devices.

So over time, carriers will support some but not all of the proposed standards for 5G. “The goal (among carriers) is to make a flexible network that can be sliced and directed towards vertical markets,” Cameron said.

There are other issues. As the mmWave frequencies go up in 5G, there is a reduction in the cell radius. The transmission losses are due to absorption and other factors. In 5G, the cell radius might be roughly 200 meters. To handle the traffic within that radius, 5G will require a system called massive MIMO, which multiplies the capacity using multiple antennas. That’s just the tip of the iceberg in what is expected to be a complex network.

Inside the 5G phone
Today’s 4G smartphone includes an application processor and modem, which represent the digital side of the system. A 4G phone also has an RF front-end, which includes the power amplifier or radio and switches. The power amp, which amplifies RF signals in the phone, is generally based on gallium arsenide (GaAs) heterojunction bipolar transistor (HBT) technology.

5G smartphones also will have applications processors and modems. But unlike 4G systems, 5G phones also will incorporate phased-array antennas. Phased-array devices consist of an array of antennas with individual radiation elements. Basically, a phased-array antenna can electrically steer a beam in multiple directions using beamforming techniques.

A 5G-based smartphone may need to up 16 elements. “The antenna elements can have separate PAs and phase shifters and connect to a single transceiver that covers an entire wide band,” said Chris Taylor, an analyst with Strategy Analytics. “Ideally, you put the antenna on top of or as part of the transceiver. So, you have a transceiver with multiple emitters, which are basically small PAs. All the modification to the signals going to or from the antenna is done in the analog domain.”

Designing a system with mmWave devices is challenging. “A lot of customers have different views on not only the architectures, but also what technology could be used to address that,” GlobalFoundries’ Rabbeni said. “A lot of that depends on how much you are going to integrate and where are you going to partition the system.

“In addition, millimeter wave circuits are very sensitive to layouts,” Rabbeni said. “Things have to be compacted pretty close together in order to minimize losses. Interfacing to circuits that operate at those frequencies can be challenging.”

Generally, phased-array devices are made using various processes, but many of today’s devices are based on standard CMOS and silicon-germanium (SiGe). “SiGe technology has already been proven for mmWave phased-array/active antenna applications,” said Amol Kalburge, senior director of strategic marketing at TowerJazz, a specialty foundry vendor.

“Additionally, SiGe enables integration with advanced CMOS and on-chip passives, thus providing area efficient system-on-chip integration capability and a cost/performance tradeoff,” Kalburge said. “We believe that SiGe will play a key enabling role in 5G front-end ICs, and will coexist with other III-V technologies.

“While SOI switches will continue to play a pivotal role at <6 GHz frequencies, their role and challenges are not so clear at mmWave frequencies. With beamforming antennas supporting different Rx and Tx paths, it is possible that the two paths may be completely isolated eliminating the need for an antenna switch at mmWave. If a switch were needed for mmWave frequencies, current SOI based switches will likely have too high an insertion loss and will become unusable. This limitation will create opportunities for switches based on MEMS or other novel technologies," Kalburge said.

Meanwhile, SiGe makes use of a standard CMOS manufacturing flow in 200mm fabs. In addition, foundries continue to improve SiGe. For example, GlobalFoundries recently introduced a new 130nm version of its SiGe process. It has an fMAX up to 340 GHz, a 25% increase over the previous process. In addition, TowerJazz recently announced a new 130nm version of its SiGe process.

Like 4G phones, a 5G mobile system will require a power amp. “For the millimeter wave radio, the power amp will be the dominate factor in terms of power consumption,” said Jeffery Curtis, a staff research engineer at Samsung Research America. “There are already some products available, but the requirements for those front-ends are quite a bit different than we need in mmWave for mobile communication systems.”

For 5G, Samsung has developed a 28 GHz power amp with an integrated low-noise amplifier and switch. The device is based on a 0.15-micron GaAs process. “By implementing application specific designs of the PA and LNA, we can reduce the power consumption by more than 65%,” Curtis said. “The integration of these components is a critical step to implementing them in a handset form factor.”

Besides GaAs, the industry is looking at other III-V technologies as well as SiGe for the power amp. “Compared to some of the other MMIC technologies for PAs, the performance capability makes GaAs a better choice for efficiency, linearity and frequency ranges,” said Eric Higham, an analyst with Strategy Analytics. “The disadvantages of GaAs relate to the cost of a device and the relatively limited integration capabilities, versus a silicon-based process.”

Generally, GaAs foundries are making devices using 4-inch wafers today, although many are migrating towards 6-inch wafers as a means to reduce cost, Higham said.

At the lower frequency ranges, the gate lengths for GaAs HBTs are typically in the 0.25- to 0.5-micron range. “To get to mmWave frequencies, most device manufacturers use a process with a gate length in the 0.1- to 0.15-micron range,” he said. “Qorvo has released a 90nm process, but that is about the current lower limit for production GaAs.”

The infrastructure
Basically, a handheld device with phased-array antennas would send signals to a multitude of base stations and/or micro cells in the field. Base stations and micro cells would also incorporate phased-array antennas.

For this, there are several challenges for mmWave. For example, weather conditions can impact the signal path. “When you go to mmWave frequencies, you get more path loss due to oxygen and absorption,” said Robert Donahue, chief executive of Anokiwave. “So you counteract that by building a radio that has beamforming capabilities.

Anokiwave recently introduced what it calls a 5G Quad Core IC, which is a 28 GHz, phased-array device. Based on SiGe, this device could be incorporated in a micro cell or other system, which is situated on a house or a utility pole.

In theory, this type of chip could communicate to a base station. Unlike a 4G system, a 4.5G and 5G base station requires massive MIMO technology. Generally, base stations use RF power transistors based on laterally diffused metal oxide semiconductor (LDMOS) technology. But even today, LDMOS is gradually being displaced by RF gallium-nitride (GaN) technology.

“As with LTE-A, the 5G infrastructure will move to higher frequency bands to take advantage of greater available bandwidth,” said David Danzilio, senior vice president at Win Semiconductors, a GaAs/GaN foundry vendor. “GaN technology has begun to take significant share as LTE moves to higher frequency bands.”

Today, the majority of GaN is produced on 3- and 4-inch wafers. Qorvo is transitioning its GaN production to 6-inch wafers by year’s end, according to Strategy Analytics. GaN is migrating from 0.25- to 0.5-micron geometries to 0.15-micron, with some going as low as 60nm.

“GaN is what’s known as a wide bandgap material,” Strategy Analytics’ Higham said. “This means it can withstand higher electric fields. And this means higher power densities capable of withstanding higher operating temperatures. As a result, GaN devices can handle more power than other high frequency technologies like GaAs and InP, with better frequency performance characteristics than other power technologies like LDMOS and SiC.”

In the future, GaN might even be used as the power amp for a 5G phone. “GaN will also be added, particularly at higher frequencies,” said Sumit Tomar, general manager of the Wireless Infrastructure Products Group at Qorvo.

Today, GaN is used for handheld systems in military applications. But it will take time before GaN moves into a smartphone. For this, the industry must make some breakthroughs in terms of low power processes for GaN.

Test challenges
Meanwhile, the test and measurement steps are arguably the most difficult part of the 5G manufacturing flow. These steps are different for mmWave, as compared to today’s RF chips.

“Today, nearly all RFICs are tested by cabling from the test equipment to the RFIC,” NI’s Hall said. “Cabling, known as conducted measurements, is generally preferred because you don’t have to deal with uncertainties such as path loss.”

But for Bluetooth and some other RF chips, engineers also perform radiated measurements over an antenna. Meanwhile, for production test, the industry uses various automatic test equipment (ATE) and instruments for today’s RF chips.

It’s a different story for mmWave devices, however. For example, the phased-array antenna may get bonded on the RF front-end device. “The package will actually include the antenna,” said Mike Millhaem, a 5G technical architect at Keysight. “So you will have no RF connectors and terminals on the semiconductor.”

So, in simple terms, the traditional cable-based testing methods for mmWave won’t work. So, how do you test and measure mmWave devices?

Each vendor has a different solution, but it could involve the use of several and expensive instruments in a rack to do the job.

“Right now, part of the challenge with mmWave is the wider bandwidth of many signals at those frequencies,” NI’s Hall said. “There are production test methods for mmWave devices, but not with modulated measurements. Today, for example, engineers can buy mmWave vector network analyzers (VNAs) up to 100-GHz and above, but these are only good for S-parameter measurements.”

VNAs can test mmWave components like filters, couplers and some power amps. “However, VNAs do not have the ability to test modulation quality, which is important for RFIC and radios designed to handle communication signals,” Hall said.

Testing 28 GHz parts are possible, however. “When 28 GHZ 5G comes out, and the spec requires 500 MHz of bandwidth, that’s pretty doable,” he said.

But there is a gap for testing 60 GHz parts. “Several vendors are working on 802.11ad test solutions, but I don’t believe there is any released solution on the market for WiGig test right now,” he said. “In the absence of such a solution, engineers rely on the ‘golden DUT’ method. We check to see that a WiGig RFIC can appropriately connect to a radio that is known to be good. This method is very unreliable. And it is part of the reason why we see so many quality issues with the WiGig products in the market today.”

Today’s military-based mmWave devices are packaged using ceramic or metallic packages. Generally, these packages are reliable, but they can be expensive.

As a result, some are migrating towards QFN packaging and multi-chip modules. Still others are looking at advanced packages for mmWave. “People are also trying out fan-out and embedded,” said Harrison Chang, vice president of corporate R&D at Advanced Semiconductor Engineering (ASE).

Indeed, packaging engineers must look at several options, and design considerations, in mmWave packaging. “The RF front-end is a lot more complicated,” Chang said. “We need to make sure those structures, such as the connection lines, pads or vias, work with and not against the RF design of the die.”

Related Stories
How Will 5G Work?
RF chip giant Qorvo explains how 5G will get deployed.
Inside The 5G Smartphone
Video and overall demand for higher bandwidth and throughput are creating new opportunities in the smartphone market, sparking new technologies, approaches and materials.
Mobile Edge Computing For The IoE
The network edge will become one of the most critical metrics of the IoE and 5G, but it will require new technologies and solutions.
Plotting The Next Semiconductor Road Map
Industry leaders examine enablers, implications, and perspectives for a changing technology ecosystem.