At last week's IWPC conference, held in Austin Texas, USA, there was considerable discussion of lessons learnt from 5G trials, as well as 4G/5G multi-band mode devices. These are both exciting and important topics, which are at the heart of the latest news and announcements related to 5G NR front-runners, early adopters, and 5G NR network launches.
Although there were few surprises, here are some notes from this standout event, together with our thoughts and comments. First, let's address the results of initial trials.
As is by now very well known, 5G NR comes with deployments both in low and high frequencies: respectively sub 6GHZ (e.g. 600MHz, 700MHz, 3.5GHz) and mmWave (e.g. 15GHz, 24GHz, 28GHz, 39GHz). As expected, the trend among operators is to deploy 5G NR coverage in the sub 6GHz spectrum for delivering blanket mobile connectivity, while using mmWave spectrum in hotspots for high bandwidth, demanding use cases, such as eMBB for various applications (for example video delivery, healthcare vertical services offering remote diagnosis), as well as for low latency use cases (for example, URLLC AR/VR, healthcare with remote surgery, or other IIoT - industrial IoT applications).
The majority of operators plan to make their first 5G NR deployment using low frequency spectrum and only later with mmWave. One exception is AT&T, having announced 39GHz as its first 5G NR, enabling some slow mobility (“walking” mobility) and supporting the Puck device. Another is Verizon Wireless, having already launched 5G FWA. This, however, is based on 5GTF technology, itself based on an advanced version of LTE-Pro, rather than pure 5G NR access.
Despite this, all of the trials to date - detailed in presentations from AT&T Labs, T-Mobile, Nokia and, last but not least, Samsung - were actually mmWave (28GHZ) scenarios, which makes quite a lot of sense. This is because the mmWave frequencies really represent a novel experiment in the mobile wireless domain — and are thus very much the unknown. If you are asking yourself which devices have been used for these trials, or which testing tool was used for the evaluation of results (and which therefore already supports 5G NR access), then the answer is inverted base stations, both as device as well as a measurement tool. With regard to the latter, the first 5G NR scanners from major vendors such as PCTEL are now emerging.
Turning to lessons learnt, several are worth noting. The first relates to issues regarding mmWave propagation. Experiments have shown that the NLOS (non-line of sight) environment demonstrated better coverage than expected, with indoor giving benefits over outdoor. This indicates that reflections could be leveraged, but not diffractions. However, it remains to be seen how well devices will be able to pick up and benefit from mmWave reflections.
In addition, mmWave can penetrate ordinary glass (but not low E-glass) and even walls (often very effectively, depending on materials) but with significant loss. While weather, unless there is very heavy rain or snow, is not of concern at these ranges, vegetation - especially clutters — can be major issues. Therefore, delivering consistent coverage can be difficult, to say the least.
However, FWA scenarios can benefit from LOS (line of sight) deployments, but with antenna being mounted on the roof tops of multi-residence dwellings or offices (e.g. at about 40m height) rather than lower-level windows (e.g. at about 2m height). In tests, the coverage range increased by about 25% in the former case. Considering these findings, one can conclude that mmWave spectrum is feasible for residential areas, but much less so for enterprises. In addition, for fixed access, mounting the antenna outside can improve range and performance, mainly due to likely LOS environments.
The second set of findings relate to some mmWave mobility experiments, which although based on limited data, proved first, that mobility at low speeds is viable; second, that higher speeds can be feasible; and, third, that many more scenarios and configurations must be tested and evaluated. As a result, we should not expect to see real mmWave mobility before at least another year or more after the first 5G NR mmWave based network is launched.
Finally, the third set of lessons are concerned with performance achieved, throughput, and latency. It has been determined that, for bandwidth of 400MHz, Layer L1 throughput can reach 1Gbps, at distances of about 300m. However, with RAN latency below 10 msec, low latency applications will more likely need to leverage edge computing in order to deliver the expected performance.
Let's now take a look at various challenges that were discussed in regard to 4G/5G multi-band/mode devices. Although 3GPP test specifications and certification procedures are not fully finalized, it is very much expected that the first 5G NR device (AT&T's Puck router) will reach the market soon. This will support 39GHz bandwidth as well as all LTE bandwidths. That said, operators, chip vendors, antenna vendors, and network vendors are agreed on the challenges that have emerged, mainly from mmWave test scenarios which require OTA device testing.
In addition, the complexity of the tests, as well as the time taken to complete them, and the associated costs are significantly higher than in previous generations of mobile access. This is largely due to the support of multi-bands, lower and upper frequencies, and multi-modes, which demand validation of many frequency combinations.
It should also be noted that only one high frequency (mmWave) bandwidth is recommended to be supported. This is because of the device power consumption needed to transmit in high frequency bands, which could result in the overheating of RF components and, consequently, in performance degradation. Worse, this effect is increased with the size and data rate of the bandwidth. Dual connectivity mode (LTE-5G NR), particularly if 5G NR operates in mmWave spectrum, could exhibit the same effect.
The device multi-mode feature is expected to remain in the market for some time to come. This is due to two reasons. First, NSA solutions based on dual connectivity (EN-DC, E-UTRA-NR Dual Connectivity), which channel signaling through eNB and the data through gNB. Second, because the LTE macro layer is likely to be used for a while (depending on specific markets) as the coverage blanket between mmWave hot spots, until sub 6GHz 5G NR solutions are fully deployed and operational, and/or until mmWave mobility is achieved.
Some simulation results showed that specific multi-band RF front-end designs, which cover both low and high frequencies, are feasible, but it could be expensive to ensure the performance of RF filters. Finally, massive MIMO antenna implementation in devices, especially for mmWave scenarios, raises concerns, because of loss that may be encountered in the human body (for example, in the head and hand), as well as loss caused by the housing material used in devices (e.g. type of glass). Typical problems include but are not limited to: transmission and/or beamforming loss, antenna frequency shift, and impedance mismatch. In addition, device form factor is challenging for massive MIMO antenna placement, particularly if we consider the fact that devices are typically hand-held.
So, from a consideration of these issues, one can conclude that it is crucial to understand and control device performance for 5G NR access. Therefore, device-based measurements, which have always been important, will be even more so with 5G.
Find out more about IWPC discussions on both 5G trial results and 4G/5G devices at: https://www.iwpc.org/ResearchLibrary.aspx