Solving 5G’s Thorniest Issues

5G rollouts are beginning to hit the market, accompanied by a long list of unsolved technical and business issues surrounding this next-generation wireless technology. But progress is being made on some of the key challenges facing this technology, even though not all of those solutions will be in place at launch.

The real challenges are with millimeter-wave implementations of 5G, which operate at significantly higher RF frequencies than existing wireless communication. The first phase of rollouts will be sub-6GHz implementations, which behave much like 4G. Those signals still travel long distances with little disruption.

Millimeter wave is where things get much more interesting and difficult. The technology is capable of transmitting large quantities of data, such as streaming high-resolution video, extremely quickly. But there are some fundamental issues that need to be resolved before that can happen.

Among the top challenges:

• With mmWave technology, arrays of antennas will be required because of the limited reach of mmWave signals. Those antennas likely will be embedded into a package with no exposed leads to test. So far, there is no proven approach for high-volume production testing.
• The effectiveness of those antennas will depend on signal coverage, and so far it isn’t clear how many base stations and small cells will be required to supply adequate coverage in different environments. Numbers may vary greatly, depending on where the cells are located, how many people they need to serve, and how much interference there is between base stations, small cells and people. With mmWave technology, trees and weather can disrupt signals.
• There is no single mmWave spectrum or standard waveform. There are many of them, and they are not the same from one region to the next. That makes it difficult to build devices that will leverage 5G in multiple regions, and there is confusion about what happens with digital and analog technology.

Despite all of that, millimeter-wave technology itself is well understood. It has been used for military applications such as radar for decades. But applying it to billions of consumer devices, many of which are in motion at varying speeds, and maintaining connections is a nightmarish problem.

“It is a realistic technology, but it’s definitely a challenge,” said Neill Mullinger, product manager in the vertical market solutions group at Mentor, a Siemens Business. “The reason that it is worth taking the challenge up is that we have nowhere else to go. We’ve used up all the spectrum, and demand is still skyrocketing. If you look at any city in the world, you test out the performance of your cellular phone at 3 a.m. and it’s pretty good. You try it out at 6 p.m. and it’s lousy because of the sheer volume of users. It’s like being on a congested highway. You can drive freely at night, but you can’t at 6 p.m. It’s the exact same problems, so we have to go there. We have to go into the millimeter wave. But it comes with these really problematic attenuation problems—200 meters and the signal has faded away.”

And this is where things begin to get really complicated. Some of the solutions involve developments in other technology areas, including the edge, which is still a vague concept, and the addition of artificial intelligence, which is in a state of almost constant flux. All of the pieces need to work together seamlessly.

“There is quite a change in the architecture required for 5G,” said Ian Dennison, senior group director for Cadence’s customized IC and PCB Group. “We’ve got all these small cells distributed on lamp poles and in buildings. But we now have to centralize the baseband. And this is where edge compute comes in. You can share baseband processing across a large number of small cells and have a greater efficiency as a result, but the timing needs to be just right for us to be bringing in AI processing to the edge computer, as well, so that AI can assist with the baseband processing. So if I’ve got small cells in different places with different contexts, then AI can learn about how the small cell can best bounce the beams off of local buildings. It sends multiple beams by multiple pathways to the same user to improve the connectivity. If you’re in a strong signal area, it can provide multiple concurrent pathways to the same user via multiple beams. There are a variety of different ways that beamforming is going to have to be intelligent, and AI is arriving at just the right time. That will help a lot with some of the difficulty with 5G.”

Solving many of these issues involves a massive global effort, and it involves a lot of moving pieces— including many that are well beyond the scope of the people working on various pieces of the solution.

“When you have all those mini cells doing this, the 5G solution is not just wireless,” said Peter Zhang, R&D manager in Synopsys‘ Solutions Group. “The wireless portion of the 5G’s impact on the communication network is only one aspect of that. There is also the gigantic backbone land line networking systems that we’re constructing and working on. We all have to be working together congruently, each with our own 5G wireless portion, so you can access a data network everywhere with a high-performance type of experience.”

A key component of that will be the edge, but so far the edge remains more of a concept than reality.

“There is still a lot of uncertainty about the edge, but some characteristics are starting to come into focus,” said Mike Fitton, senior director of strategic planning at Achronix. “There is a wide variety of devices, there is an increase in the number of devices that are coming, and there is a variety of devices that produce very little data to sensors that produce significant amounts of data, with different characteristics, as well. Classically, we saw all of that data moving from edge devices to the cloud and back, but because of requirements from a latency point of view, we’re starting to push this closer to the edge or edge compute.”

Test and reliability
Millimeter wave will play a key role in that paradigm, but one of the big problems today with mmWave is the reliability of data communication. Beamforming, which directs signals from multiple cells around objects, is a proven approach to maintaining a connection. But how well an end device receives and processes that signal depends to a large extent on the antenna array, and testing of an array is a slow process using over-the-air testing.

“We’re starting to see millimeter wave base stations and early testing of the topology, and you’ll see more of those in a year or two,” said Adrian Kwan, senior business development manager at Advantest. “That will be followed by cell phones in 2021. It’s the same base station technology as today, but you need five to six times more pico cells. The wireless has to be in place first before you see customers developing devices to support millimeter wave.”

Kwan said that packaging and testing both have evolved to be able to support testing of the antennas over the air, but so far nothing has been done in high volume. But that’s not the only issue.

“What we’re seeing with multiple carriers is that the waveforms are different,” said Kwan. “There are new standards being defined to deal with this. That all has to be in place before 5G is ready.”

Fig. 1: Different wireless standards and where they fit. Source: Advantest

Others report similar challenges.

“To support these bands we have to significantly change the design and testing we do in these environments,” said Ben Thomas, director of technical marketing at Qorvo, in speech at the recent NIWeek. “Combine that with more carrier aggregation combinations, dual uplink, and more complex waveforms and modulation schemes, and quite frankly you’re looking at an exponential impact on the RF section.”

There are other issues on the test side, as well. “The challenge with 5G is that some of this stuff has to get pushed up to the system level,” said Doug Elder, vice president and general manager of the semiconductor business at OptimalPlus. “So you may need to do some of this stuff with system-level test, as opposed to functional test or earlier in the process. If you collect the data that comes out of that test process, you can look at that data and draw some conclusions that you can drive back into your process.”

In addition, just testing a 5G device in the fab or packaging house isn’t sufficient. Because this technology is so new, quality of service is expected to be spotty at first. The early generations of cell phones dropped calls regularly with far lower data rates.

“The current frequencies clearly aren’t going away, so we have to continue to test those in this way that we have done in the past,” said Mentor’s Mullinger. “But then you augment that with new solutions to deal with the expanded frequencies of 5G. The current family of testers allows you to do a lot of that virtually, which means we can set up a test environment for the lab or the field and run the exact same environment.”

That will be critical for improving coverage and quality of service.

“Part of this an expectation-management problem, and we should expect that our 5G systems start to improve over time as the AI systems start to kick in, start to develop better quality of service,” said Cadence’s Dennison. “But straight out the box, it may not be perfect. So testing as it were is going to be a continuous thing, and something that the network operators are going to be relying on to improve the quality of their service. So there is yet another sort of difficulty that is new to us as we move from 4G to 5G.”

Thinking differently
So far, there is no consistency in how companies propose to solve the signal attenuation problem for millimeter wave. Still, the solution may not just be blanketing the walls of buildings with small cells. At least part of the answer may be in thinking differently about how a 5G mmWave device behaves, or even what it looks like.

On the behavior side, one possible solution is to use different technology for the uplink and downlink.

“If a mobile device moves relative to the base station, the real problem is the uplink,” said David Hall, principal product marketing manager at National Instruments. “To use beamforming is difficult, so you can use LTE for the uplink and millimeter wave for the downlink. The big questions are about the infrastructure. Most of the end devices have full radio capability, but for the base stations, the choice is whether to use digital versus analog for beamforming. Analog circuits are less efficient, which creates more heat in the base station.”

The 5G standard allows for both and analog and digital, but the beams look different.

“If you using a digital beam, it involves a change of the waveform itself,” said Hall. “You need to adjust the phase to wave carriers.”

Another change involves the end device. The introduction of mmWave technology has much broader ramifications than being able to download more video.

“We shouldn’t get overly focused on 5G just being the next increment over 3G and 4G for the end user with the mobile phone,” said Cadence’s Dennison. “5G is laying claim to a lot of things, and one of them is a very low latency to enable robots, in particular. Robots today are in factories and they’re in cages in factories. If we can get the latency down in communicating with them—imagine you’ve got that massive AI on the edge and it’s talking to robots through 5G and that has a very low latency of communication over milliseconds of order—then you’ve got a chance that those robots can be responsive in real time and not harm human beings that are around them. So 5G opens up a whole new set of opportunities like that. Maybe robots will break out of the factories and the cages, and it could be 5G low latency might be the enabler for that. While these chipsets already are being developed for handsets, that is only one possible device that can take advantage of 5G.”

Along similar lines, there is a lot of concern about relying on 5G for linking cars to the cloud because connectivity is unreliable. But communication to other vehicles is a different story.

“With wireless network, the cars can talk to each other,” said Synopsys’ Zhang. “We can create a link so that they can they can communicate with that, with short latencies, with the highest possible car reliability because they don’t want cars to crash. Those are the things that make 5G more interesting.”

Conclusion
Investments in 5G are rising, and so is the industry reliance on this technology for a variety of applications, including fixed structures such as smart buildings and cell phone towers that serve rural areas as well as densely populated areas. There is simply more data being produced and a growing demand for information everywhere.

The global communications business is betting heavily on 5G, even it remains to be seen exactly how that technology will be used and what an 5G end device actually looks like and how it will be served. Nevertheless, progress is being made on some of the most difficult problems in 5G, even if there is a long way to go before all the kinks are worked out and it is fully functioning like 4G LTE.

—Kevin Fogarty contributed to this report.