Mobile WiMAX has emerged as a prime contender for the de-facto ITU’s 4G standard. It provides high data rate with large coverage and vehicular mobility support. It has been, and is being, widely deployed in diverse applications like video streaming, broadcasting and data service. The FCC (Federal Communications Commission) issues licenses to operate Mobile WiMAX services in several spectrums, among which 2.5 GHz and 3.65 GHz are the most common. Because the 2.5 GHz spectrum suffers the least propagation losses, it has been widely applied commercially. For this reason, academic and industrial researchers have given it particular attention. However, in some cases, the 3.65 GHz is often a better choice, in spite of having a lower coverage, particularly due to its favorable licensing requirements. Due to limited amount of work published in the 3.65 GHz spectrum, the concerned parties do not have sufficient data to reliably select the spectrum. In this research work, a thorough quantitative analysis of the two Mobile WiMAX spectrums, 2.5 GHz and 3.65 GHz, are presented and compared. Actual physical testing of commercial equipment in real-world settings has been done to provide a generic overview of the performance of the two spectrums. The results presented serve multiple purposes. First, they provide reliable technical data for decision-making. Second, they can be used for link budget analysis. Finally, they can be utilized as benchmarks for future testing and quality control of equipment production. It is shown that the maximum achievable downlink throughput for the 2.5 GHz and the 3.65 GHz systems is around 22 Mbps and 21 Mbps respectively. Assuming an average user bandwidth demand of Mbps, they can both reliably serve 40 to 45 users within a coverage radius of 12 km and 8 km respectively. Other than the lower coverage, opting for the 3.65 GHz spectrum over the 2.5 GHz spectrum will cause no significant performance loss and should be preferred if the loss of coverage can be tolerated.
Advisor: Hamid R. Sharif-Kashani