155 megahertz of spectrum was aggregated in the tests on T-Mo’s network in California
Three component-carrier TDD-FDD carrier aggregation (3CCA) results in better spectral efficiency in 5G, but there are some interesting anomalies in play in regards to devices and spectrum, according to recent testing by Signals Research Group.
SRG tested 3CCA in a 17-site test cluster for T-Mobile US in the southern California city of Torrence. The cluster covers about 24 square kilometers of urban and residential neighborhoods. The three component carriers aggregated were 15 megahertz of TDD spectrum at 600 MHz, 100 megahertz of 2.5 TDD and another 40 megahertz of TDD 2.5 GHz.
Using test tools from Accuver and Spirent Communications, SRG tested and analyzed data from two Samsung devices: a Galaxy S21 that supported 2 component carriers, and a Galaxy S22 with a special software build that supported 3CC. Both phones supports 5G Standalone operations, but for testing purposes SRG sometimes locked the S21 to NSA operation.
SRG said that it looked at 3CCA in ideal radio conditions on an empty network; differences in performance between the two smartphones; and relative differences in performance between 600 MHz and 2.5 GHz at a distance from the cell site, as well as comparing FDD-TDD carrier aggregation as opposed to TDD-TDD.
SRG also took note of T-Mobile US’ recent achievement of 3 Gbps download speeds in sub-6 GHz 5G. That testing was done in a commercial network (probably during off-peak hours or a maintenance window, the report theorized) with the same Qualcomm Snapdragon X65 modem in the S22 device that SRG was using—but “ample” 5G spectrum is necessary to reach that level of speed and despite SRG’s S22 being allocated “virtually all possible network resources,” it still didn’t see 3 Gbps speeds. “Depending on the market, that much spectrum may not be available today because the operator still needs to support its legacy LTE traffic,” SRG said.
In T-Mo’s Torrence cluster, the carrier has quite a bit of spectrum available, the analyst firm pointed out—but it’s being used by customers on LTE. It’ll take more customer migration to 5G phones before more of the available spectrum can be used for 5G services.
SRG said that it saw double-digit spectral efficiency in the midband spectrum—where “5G really shines” compared to low-band frequencies, it said. And it has generally found that operators usually have a dense enough site grid in outdoor urban environments that they can deploy midband spectrum without much fear of a coverage penalty.
SRG said that the differences between the S22 and S21 were the most interesting thing to turn up in this particular testing. There were scheduling anomalies, but more intriguingly, the two devices (which use two different Qualcomm chipsets) delivered very similar throughput performance in very different ways, through how they used 256QAM, multiple-input-multiple-output and Modulation and Coding Scheme (MSC) allocations. There were also, SRG found, unexplained and inconsistent differences between the two midband channels, with one of the Band 41 channels “[performing] much, much better than the other band for a given RF condition.”
Read more from SRG here.
