Tips and Trends: Meet the requirements of multi-carrier, multi-protocol picocells and relay designs

Capacity and coverage upgrades of existing mobile access networks are necessary in order to bring high mobile data rates to subscribers with smart phones, web tablets, and laptops. This capacity upgrade is taking place in the middle of the transition from 3G-to-4G service access technology, with the vast majority of today’s subscribers connected to 3G networks.

The products that enable operators to deploy this capacity will be silicon-based small cells – system-on-chip (SoC)-based picocells and relays that can easily be deployed in indoor and outdoor locations. In order to bridge current and future voice and data services, these picocells need to operate concurrently in multiple spectrum bands utilizing multiple carriers for 3G and 4G service access.

Figure 1. Reference design for multi-carrier picocells and relays based on the DAN3400 SoC platform

In order to facilitate the design of compact, passively cooled picocells, these SoCs need to ensure very low system-level power consumption. At the same time, these SoC platforms need to integrate all processing layers for each supported protocol to support the design of cost-effective, easily mountable, single-chip products. Concurrent operation of W-CDMA and LTE requires sufficient processing capacity to support high-capacity mobile data services.

Initially, LTE data services will be introduced in 10, 2x10 or 20 MHz channel configurations in the context of LTE MIMO and evolving to LTE Advanced MU-MIMO deployments. This requires the support of 2x2, 2x4 and ultimately 4x4 RF configurations with SoC-embedded multi-channel antenna array and multi-carrier digital RF front-end functionality. As W-CDMA networks are struggling to meet mobile data demand, operators are resorting to deploy 2x5 MHz channel configurations with MIMO. This requires concurrent support of a 2x2 RF configuration operating in W-CDMA bands. Altogether, the SoC-embedded digital RF front-end should support “single-RAN” multi-carrier, multi-mode operation to avoid the cost and power consumption of a larger external FPGA.

Operators wouldn’t want to experience any cross-service impact, such as a 4G-service interruption in the context of 3G software updates. Nor should 4G software problems, such as outages, have any impact on 3G-service performance. So operating several protocols in a single-SoC picocell isn’t really an embedded software integration task.

Underlying SoC platforms must provide “hardened processing islands” in a scalable multi-layer, multi-core architecture. Sufficient independent processing and hardware acceleration cores must be available on all processing layers, enabling the isolated operation of all protocol layers for each supported service access technology. In-field software evolution is a mandatory requirement, as operators wouldn’t deploy a larger number of in- and outdoor picocells just to return to the many installation sites for future hardware upgrades or even picocell replacements. SoCs must operate with sufficient spare capacity on each processing layer, and individually for each service access technology.  To facilitate future in-field upgrades, e.g. in support of spectrum harmonization on the path to a single LTE Advanced network, the embedded digital RF front-end layer must support LTE Advanced features such as multi-carrier spectrum aggregation.

Not all operators have access to, or chose to deploy outdoor infrastructure on the same type of locations, and not all metropolitan areas support cost-effective connection of many small cells to fiber optic facilities.  In essence, the deployment site dictates many of the configuration parameters of a picocell product as a consequence of cell size and demography surrounding the location, and as a result of physical and mounting restrictions for any specific installation site.

Cell size and demography drive requirements for simultaneously registered and active subscribers, as well as required RF output power to cover the service area. However, operators aren’t going to deploy a plethora of different products and expect the picocell line-up to cover all requirements with identical software content for each required form factor to limit UE certification efforts.

This requires the support of a range of subscriber configurations from as little as 16 or 32, all the way up to 64, 128, or even more subscribers independently for each service access technology. It’s fair to assume that vendors will need to offer a variety of RF and antenna form factors in their all-in-one picocell product line-up, ranging from low-power indoor access points (~23dBm) and outdoor picocells (~27dBm) to high-power microcells (>30dBm). Hence, embedded PA linearization becomes a critical requirement for picocell SoCs, minimizing system power consumption and heat dissipation as the determining criteria for the selection of enclosure materials, related product size and weight, and ultimately cost.

Finally, the actual mounting location will determine new the requirements for cell-site backhaul, specifically in the context of high-density small cell deployments in outdoor locations, such as lampposts, traffic lights, or side walls of buildings. While wire line fiber-optic or copper access is preferred, it is not readily available in these locations, and building permits, as well as construction time and costs, quickly erase operators’ business case. Hence, wireless backhaul is critical for deployment of picocells in such locations.

Low-cost wireless backhaul must be integrated with the picocell product, as many of these locations do not permit systematic deployment of two pieces of equipment. Hence, it is essential that the SoCs used for picocell products support suitable wireless self-backhaul (sub-6GHz, or E-band) operating concurrently with W-CDMA and LTE service access. This can be facilitated by means of dedicated self-backhaul technology, e.g. in W-CDMA picocell deployments, or utilizing standard LTE Advanced relay technology. Each option requires integrated sub-6 GHz NLOS or E-band LOS RF subsystems.

As the market for small cells matures, operators are no longer seeking to merely extend femtocell solutions to outdoor locations. Instead, they are looking for highly capable, infrastructure-grade picocell products that warrant the investment to deploy high-density small cell layers. If cost-effective enough, these products will, in turn, serve the bulk of indoor infrastructure deployments as well.

About the Author

Joachim Hallwachs joined DesignArt Networks in 2007, serving as Vice President of Marketing and Business Development. Joachim has over 20 years of experience in telecommunications and networking infrastructure markets, with executive assignments in marketing, business development, product management and M&A for leading start-ups and major infrastructure vendors.