A network of 8 TV white space (TVWS) base stations (BS) has been deployed in Cambridge as a part of Microsoft’s Cambridge White Space Trial. The BBC has been participating in this work to understand the potential opportunities and challenges offered by the emerging uses of TV white space.

The Cambridge base stations operate under non-operational licences issued for test and development by Ofcom. A total of 10 TV channels have been allocated for the trial for low power operation (EIRP <4W). The base stations have been deployed mostly at a number of public houses. The site at the Lion & Lamb PH in Milton was selected for a field survey and the coverage and performance of this BS is the subject of this report. For the field measurements, an experimental TVWS radio from Adaptrum was installed at the site, and a series of mobile and fixed measurements were made using a pair of BBC survey vehicles.

The objective of the field survey was to measure the usable TCP/IP throughput for both uplink and downlink connections using the prototype Adaptrum radios. The measurements were made at a number of test points at 10m and 1.5m height, in order to simulate fixed customer premises equipment (CPE) and mobile user equipment (UE) applications. TVWS spectrum is of variable quality and the measurements were repeated on CH59 and CH61 to explore how the coverage might vary as a function of DTT interference present from distant parts of the TV network. The test and development licences were chosen to avoid interference to TV services from Sandy Heath transmitter, but interfering signals from Tacolneston, Sudbury and Waltham are known to spill over into Cambridge at significant level. The work was carried out over 4 days commencing from 5th March 2012, in collaboration with Arqiva and Adaptrum.

The results show the 125mW base station achieved a coverage distance of up to 1.5km for fixed reception at 10m and around 400m for mobile reception at the coverage edge. Throughput analysis, revealed a useful TCP/IP bandwidth of between 7Mb/s and 1Mb/s for the uplink, dependent on the received signal levels. Coverage predictions were made using the UK planning model (UKPM); these showed good agreement with the CPE measurements. The Hata suburban model, which is appropriate for mobile coverage, showed good correlation with the 1.5m measurements and drive test logs.