Chipset makers, network infrastructure companies, device OEMs and test and measurement companies have been heavily involved in the research and development process and the standards work for 5G for years at this point, and that work is coming to fruition as 5G devices begin to hit the market. Both consumer smartphones and a variety of other devices, from 5G-backed routers and mobile hot spots to IoT devices, are emerging this year.
But that is just the first wave of 5G chipsets — the development of 5G is still in its early days. For a more detailed look at what’s happening on the 5G chipset front, RCR Wireless News reached out to Qorvo for an email Q&A with Ben Thomas, the company’s director of mobile 5G business development. Qorvo announced in late February that its integrated front-end modules for 5G were in high-volume production in support of the device and infrastructure ecosystem that is ramping up this year. The following Q&A has been lightly edited and condensed.
RCR: How would you describe the general state of development of 5G chipsets at this point?
Thomas: It’s progressing well, in terms of mobile handset-focused chipsets. The first set of chipsets, which were really based on pre-Release 15 assumptions, are helping handset trials through the end of this year. Then in 2020, we will have good, Release 15-compliant chipsets from multiple vendors. I would see three primary drivers in our industry: Qualcomm, Samsung and HiSilicon. That is, at least until we see the rise of mid-tier in 2021. As that volume begins to take hold, other mobile chipsets will come.
RCR: What are some of the challenges in 5G semiconductor development and production, both at millimeter wave and at sub-6 GHz?
Thomas: For chipset-related challenges, I’ll only touch lightly on what I am aware of as we work through RF front end architectures with our customers. Those include:
-The transceiver impact is really about 100 megahertz of carrier bandwidth, as well as good throughput at the required waveforms, both DFT-S and CP-OFDM, especially at 256QAM. The dual uplink requirements are the biggest change compared to LTE.
-For the modem, handling the EN-DC LTE anchor and New Radio (NR) is also a challenge. This is the big change moving from 2019 to 2020. In 2019, we will see what are termed Type 2 UEs — that is, those handsets driven from chipsets where the LTE effectively does not know what the NR side is doing. That has its own set of limitations. This really is solved in the Type 1 chipsets coming in 2020. Further, the coordination in managing handoffs to and from LTE and EN-DC 5G might not be as smooth in this early going, but will improve in 2020.
-Lastly, [sub-6 GHz] FR1 and [millimeter wave]FR2 are running on completely separate RF chains, all the way back through the TCVR. It’s a big challenge just to enable EN-DC FR1, but enabling FR2 through the same modem — and all of it working well — will be a battle through 2020. …
-Other challenges include new RF components, wider bandwidths, and again, the exponential increase in waveforms to test. An extra design consideration in most cases certainly, but the biggest challenge is verifying the massive number of combinations across the handset. Then, once compliant, choosing the place to optimize: full power PC2 for coverage or shorter range, CP 256QAM for high data-rate MIMO applications. These are very much driven by both carrier and OEM requirements, so the variation in design in itself drives complexity.
-One area where Qorvo has helped historically is the use of ETICs: envelope tracking power management for the [power amplifiers]inside these RF front ends. Envelope tracking has become the standard way to handle the varying needs of RF front ends on the transmit side, based on different use cases and points. However, we run into a roadblock in early 5G. Most companies’ ET, and the chipset integration of that ET, for the wide 100 megahertz bandwidth signals isn’t ready for prime time. That will take through 2020 to really have all the bugs worked out, but a return to ET will be important to help efficiently manage all these use cases and transmit power levels. So, what are they using now? A blend of average power tracking (APT) and ET. Which leaves a more complicated, sub-optimized PA and RF front end module.
-Lastly, there is mmW. As we move up to millimeter-wave 5G bands like 28 and 39 GHz, we should recognize this is driven by a very different use case versus sub-6. Because of challenges like propagation and shorter range, one of the goals of millimeter-wave spectrum is to add network capacity in high-density environments. From an RF perspective, this is a whole new animal in handsets; very much a system solution, rather than buying discrete components as in early days of a sub-6 standard, like LTE, mmW demands close collaboration of the front-end vendors and the chipset suppliers. A great many things to juggle.
-Then there is the size challenge of fitting mmW into what is already a space-constrained form factor. Thus far, it leads these mobile handset designers to sacrifice best performance (larger array sizes and placements) for size-conscious methods. Which in turn leads to the need for very high-performance mmW front end solutions. Quite frankly, NOT what the industry initially thought of when they started down this path. Large arrays of silicon-based solutions simply won’t work in handsets like they do in [customer premise equipment]or base stations (which are now also seeing more traditional GaAs or GaN based solutions as advantageous).
RCR: Where is Qorvo at in terms of generations of 5G RF front ends? How quickly do you think the move from first-generation to second-generation 5G will happen and what will drive that? What are some of the features of 5G specifications that have yet to be implemented and what opportunities do they open up?
Thomas: Qorvo was the first to MP with our 5G Band n77 RF front end module. And we have a full portfolio of RF front ends and components right after it. We have been preparing for this for a while. It was challenging to release a 100 megahertz-carrier bandwidth PA up at the 3.3-4.2 GHz frequency range, while meeting 5G’s stringent RF requirements. We learned a lot from that solution, and we will be seeing our second round in 2020 that aims to improve up on our first understanding and efforts to support 5G trials in 2019. Those 2020 products will take us through until, say, mid-2021, when we will see mid-tier-oriented 5G front end modules. Perhaps [there will be]a slight tilt toward being more cost-conscious, [but]these mid-tier 5G front ends will still be high performance. It’s not so much a choice by Qorvo to do so … the 5G standard demands it!
Additionally in 2021, we could see an update related to the 3GPP Release 16 ratification in 2020. Since Release 16 is focused heavily on expanding to other non-mobile markets, we won’t see many fundamental changes in specs for handset-related products. However, we will see even more EN-DC and even NE-DC band combinations to deal with within the RF front end’s overall architecture. I really don’t see the RF front end for 5G stabilizing until 2022 at the earliest. It will be at this point that, as always, an eye can be focused on optimization for either performance improvements or cost reductions, depending on the handset segment you’re dealing with.
Throughout this timeline, Qorvo’s opportunities are multifold. Our high-power, super-linear GaAs HBT PA processes will be benchmarks for those bands needing PC2 high power operation. It will be a differentiator for us. Because of this capability, the UHB’s (3.5 and 4.5 GHz) are ideal product candidates. The heavy [carrier aggregation]requirements lend themselves well in our mid/high band S-PADs, which use high-order multiplexers to handle these complex combinations in a low-loss, high-rejection fashion. The technologies we have been working on even before the merger creating Qorvo lend themselves well to the 5G challenges.
[In addition], DRx, or receive diversity — which in the mid band through UHB is mandatory as 4DL — certainly has promise to provide opportunities. In LTE, Rx diversity or even MIMO was seen more as a “check the box” capability. However, in 5G, I see a real press from mobile operators/carriers to demand high performance and multi-band CA receive diversity at the same time. That is right in line with Qorvo’s capability to differentiate.
Lastly, with the need for high performance mmW, Qorvo perhaps has an opportunity to provide RF front end solutions. A little further out in time, as mmW in high volume is a little further out in time, but the opportunity is shaping up nicely.
RCR: What does it take to get to a reference design for 5G? Can you tell us about the basic process and timeline that it takes to get a 5G chipset to market and any lessons learned that Qorvo has gleaned along the way?
Thomas: Our customers are the ones assembling these designs today, but with Qorvo’s help. Qorvo has been known throughout the industry for some time (especially to our customers) as the “go-to” company for helping to consult on complete system design for new RF front end architectures. Why? Because that is where we start when we specify our RF front ends. Not just piece by piece specs. We take a holistic view.
You can see the 5G complexity — now consider that the timeline for these handset OEMs hasn’t changed! It’s still on constant, 12 month innovation cycles. What’s worse is this expectation to release 5G trial handsets this year, when the spec itself wasn’t locked until late last year.
The pace is not changing, but the complexity is exponentially more difficult. Without starting at a system view, a customer or even a supplier really can’t hope to release products on-time, meet specs AND provide the high performance our customers and their users demand.
As we move forward with a little breathing room, we are trying to start earlier — say, 18 months before launch, on overall architectures — which means we are already looking at 2021 handset needs. But as I stated, due to Release 16’s lock not being until 2020, we simply don’t see 5G stabilizing anytime soon such that we can truly work far in advance. We will need to continue to be nimble, despite our desire to work ahead. Responsiveness to changes in spec, carrier needs and OEM implementations will be critical moving forward.
So, there is still a long road ahead to seamlessly bring 5G to the masses. But, despite all those challenges the future for 5G is a bright one!
The post Q&A with Qorvo: What’s going on in 5G semiconductor development? appeared first on RCR Wireless News.