Addressing Cellular Network Capacity Hotspots - Sectorization with Bi-Sector Arrays or Carrier Add?

September 14th,2011

Posted by Mark Cosgrove

For CDMA and W-CDMA systems adding carriers when spectrum is available seems the most logical way of addressing capacity hot spots. Adding carriers has two direct effects on the capacity of the system:

For the downlink the additional power of the new carrier provides a proportionate increase in capacity,
For the uplink users can be split between the new and existing carriers reducing the UL Noise Rise and gaining improved coverage improvements.

Obviously as the carrier count increases adding carriers becomes less and less effective as a way of adding capacity: going from 1 to 2 carriers would double the DL capacity, whereas going from 2 to 3 carriers is an increase of just 33%. In CDMA and W-CDMA systems soft handover mechanisms do not work between carriers on different frequencies and mobiles need to perform a “hard handover” when moving between adjacent cells or even layers within a cell. While making a hard handover the mobile makes measurements on the other frequencies. Since today’s mobiles do not have full dual receiver capability, the mobile has to stop sending/receiving while it makes measurements. This is called compressed mode. For various reasons a mobile in compressed mode has to increase its output power, which means that the interference levels in the cell will increase. Or, capacity will be reduced for a constant interference level. The effect on the mobile link budget is about 2dB; therefore operators need to consider very carefully how additional carriers are deployed. Ideally the new carrier should be deployed over an area consisting of several sites rather than just the one with capacity/coverage problems. Admission control and load control would then have to steer traffic to minimise handovers between carriers. This increases the effective cost of deploying the new carrier as each site chosen will need additional amplifiers for the new carrier even if base band equipment is pooled.

TenXc’s Bi-Sector Arrays have been extensively deployed in CDMA and W-CDMA systems and shown to offer capacity gains in the DL of between 2 and 2.3 times capacity of a standard 3-sector deployment. The gains from the Bi-Sector Array have been shown to be higher than even the ideal sectorization gains of 2 times: improved link budget, interference control and optimized softer-handover losses all contributing to a greater than 2 times gain. Sectorization in a CDMA type system requires basically the same Node-B/BTS changes as adding a carrier and offers a clear alternative to new carriers. Sectorization with the TenXc Bi-Sector Arrays will always create more capacity than carrier addition, particularly as sectorization gains are independent of the carrier counts.

The question becomes one of which should be performed first - Sectorization with the Bi-Sector Arrays or Carrier Adds? The big advantage of carrier add is in the ability to create layered services: carrying R99/HSPA+ on F1 and HSPA+ only on F2 for example; however this does need layering to be extended across contiguous areas and involves many sites being deployed with new carriers, where traffic demands do not justify it. Load and Access control require many changes to parameter settings and introduces a need for close management of capacity in layers. However the rewards to the operator in terms of network efficiency may warrant such approach. Sectorization with the Bi-Sector Arrays allows an operator to put capacity exactly where it is needed by just converting those sites or even sectors with capacity needs to Bi-Sector Array. Avoiding the issues of hard handover, the Bi-Sector Arrays allows capacity to be deployed ad-hoc as needed without a reduction in coverage. Following up with carrier addition at a later stage allows for the sites to continue to grow. Higher sectorization of a site requires some increased investment in terms of the Bi-Sector Arrays and additional feeders. For those operators prepared to accept the reduction of coverage from hard handover boundaries, and deploy new carriers on a sector-by-sector basis, carrier addition before sectorization is more cost effective. However, operators with plans to deploy new carriers as layers, in order to reduce performance impacts, Sectorization offers a clear financial and performance advantage.

Personally, I see that sectorization with the Bi-Sector Array followed by carrier addition will likely become the favoured option for most operators. The advantages of layered services are mostly negated by the increased difficulty in managing the traffic profiles and I suspect that any gains are small. As networks become loaded and users’ demands for quality increase the risks to KPI of uncontrolled inter-frequency handover boundaries becomes too great. Hence I believe that most operators have little choice but to deploy new carriers over large geographic areas simultaneously. In this scenario the economics of deploying the Bi-Sector Arrays and deferring carrier additions is unquestionably favourable and hard to ignore, when this is coupled with the additional gains from the Bi-Sector Arrays over new carriers the choice for operators is clear.