Research into new frequency bands and techniques for the upcoming 5G standard is putting new demands on T&M, but these complex standards present problems during production tests.
5G wireless communications brings with it a number of challenges, not least for test and measurement. Not only is the upcoming standard likely to be a sprawling document that covers many different possible frequencies and modes but it will have much tighter requirements for system attributes above the RF interface such as round-trip latency. The tighter timings will be needed to support new applications such as autonomous driving and long-distance robotic control.
“The latency of 4G on average is 80ms. For car-to-car communications you need sub-1ms latencies,” says Rahim Tafazolli, professor of mobile and satellite communication systems at the University of Surrey.
The need for 5G to expand into new frequency ranges comes from a parallel requirement for higher data rates to support video and interactive gaming: “We have done a study on this and said we need at least 100Mbit/s of bandwidth [for individual users]. That requires 10Gbit/s of connectivity. So we will need a combination of low frequencies, below 1GHz; 1-6GHz and then the millimetre bands above 6GHz,” says Tafazolli. “5G will evolve. The first release is expected in 2020 and that will probably be for communications below 6GHz. It will then evolve to use the higher frequency bands.
The university aims to be a key part of the development process. Tafazolli adds: “We are deploying the first 5G testbed in the world over an area of 4km2, for connected cars as well as smartphones.”
Developing the protocols and suitable hardware, as well as compliance tests once the 5G standards are complete, will require major changes to instrumentation set-ups such as the one that will be needed by Tafazolli’s team. The big advantage of the millimetre bands above 6GHz lie in their comparatively large bandwidth, of the order of hundreds of megahertz compared to the tens available in established cellular ranges. With simple modulation schemes, they could support high data rates with the downsides that they require the radio beams to be steered towards the receiver and that water vapour absorbs their energy. Instruments to handle those bands need much higher bandwidth than those used for traditional cellular development.
Because of the technical challenges, the millimetre bands above 6GHz have up to now been largely the province of specialised and expensive circuitry built primarily for military and aerospace radar and point-to-?point wireless links between base stations.
“We need these technologies to be mass market and low cost. So we now want to pull designs into a lower-cost arena from today’s high-end, military projects. So, we see a lot of work on characterising devices that can handle these frequencies,” explains Jonathan Borrill, director of marketing at Anritsu, “Previously you could do very narrowband characterisation. Now people need to see hundreds of megahertz of bandwidth. That’s very wide bandwidth analysis.”
Meik Kottkamp, technology manager at Rohde & Schwartz, says: “Exploring a significantly different spectrum would require comprehensive testing, from channel measurements in order to characterise the channel to eventually create channel models up to the verification of physical layer, protocol and network performance.
“5G will address new IoT use cases covering both low power and a high number of connected devices as well as low-latency and high-reliability scenarios. Thus, it will be essential to provide test equipment that can efficiently cover the diversity of test cases including low latency as one important test criterion.”
Borrill adds: “The industry has some high-end generators and analysers but there is a need to make it more mainstream and affordable. It has to move from being a niche application to the mainstream. We have some new architecture that can make it into a mainstream technology that combines the RF architecture that we use with new analogue-to-digital conversion technologies. We are just hitting the point where the technology is available.”
To perform the measurements needed to experiment with frequencies that reach close to 100GHz and get a head start on development, Nokia assembled its own test rigs using modular instrumentation based on the PXI format from suppliers such as National Instruments.
At the 2014 NI Week, Nokia radio systems research lead Amitava Ghosh and colleagues unveiled a proof-of-concept system designed to handle a wide spread of frequencies. “We need a proof of concept design that works from 6GHz to 100GHz, mainly in the 70GHz to 80GHz range. There are large spectrum opportunities there. But designing a proof of concept system for this presents a lot of challenges,” he said.
Mark Cudak, principal research specialist at Nokia, adds: “One big challenge was finding a system that could process 1GHz in real time. We also had to build an access point for it with a transceiver and everything in between. We were able to use PXI for the baseband processing hardware of a 1GHz access point.”
According to Cudak, cards carrying field-programmable gate arrays provide real-time processing for modulation, demodulation and channel coding as these offer higher processing performance than software running on digital signal processors.
A further issue with millimetre-band testing, which will change the way products are compliance-checked in the future, comes from the need to steer the radio beams. The proposed systems will use antenna arrays to perform beamforming similar to that used in military phased-array radar systems to steer the RF radiation electronically.
Below the 6GHz divide, there are significant changes to test and measurement techniques. Anritsu’s Borrill says traditional tests that allow a system to be wired up to a harness using test points close to the antenna are impractical. The systems may use arrays with more than a hundred elements, an order of magnitude more than the multiple-in, multiple-out (MIMO) arrays used in LTE and Wi-Fi systems. “We pretty much have to do over-the-air testing,” says Borrill. “We haven’t standardised MIMO over-the-air testing and it will mean a real change in how to characterise systems when you only have over-the-air testing available.”
The use of such large antenna arrays will mean cellular equipment manufacturers picking up techniques from their military counterparts, Borrill says. “We have a lot to learn from the phased-array radar community with techniques such as sideband characterisation.”
Even below 6GHz, there are big changes coming to the air-interface protocols as the limitations of 4G become apparent. “One of the key things, and it’s starting to become the Achilles heel for LTE, is interference,” Borrill says.
The OFDM modulation scheme used by LTE is unlikely to survive the move to 5G as researchers such as Tafazolli look for techniques that are less prone to interference. Testing those schemes will require equipment that is capable of generating the potentially complex new signal types and process them at the receiver’s end in real time.
Although the air interface alone will involve many new techniques and technologies, 5G demands a holistic approach to protocol development. Rohde & Schwartz’s Kottkamp says: “Considering that applications running over the top will rely on fast system response times, application-layer testing, which includes the impact on the cellular network in terms of signalling load, for example, will have to be implemented.”
Borrill says the latency requirement will not only push more processing into the base stations, it will place stringent requirements on timing accuracy. “We need to look at the latency and the jitter to see whether we can run the apps that we want.”
The backhaul network, which itself may be based on millimetre-wave links, will play a key role. “The microwave link needs to have really stable end-to-end performance. When people put fibre down they can be confident of jitter and delay,” Borrill says, but radio links may switch coding schemes to take account of changing interference conditions that can affect round-trip times.
As a radical step in cellular communications, 5G is already leading to major steps in test and measurement and is likely to continue to do so on the way to the scheduled 2020 release for the first of the standards.