A phased array feed (PAF) provides a contiguous, electronically synthesized wide field of view for large-dish astronomical observatories. Significant progress has been made in recent years in improving the sensitivity of PAF receivers though optimizing the design of the antenna array, cryogenic cooling of the front end, and implementation of real-time correlation and beamforming in digital signal processing. FLAG is a 19 dual-polarized element phased array with cryogenic LNAs, direct digitization of RF signals at the front end, digital signal transport over fiber, and a real time signal processing back end with up to 150 MHz bandwidth. The digital back end includes multiple processing modes, including real-time beamforming, real-time correlation, and a separate real-time beamformer for commensal radio transient searches. Following a polyphase filterbank operation performed in field programmable gate arrays (FPGAs), beamforming, correlation, and integration are implemented on graphical processing units (GPUs) that perform parallelized operations. Parallelization greatly increases processing speed and allows for real-time signal processing. During a recent test/commissioning of FLAG, Tsys/efficiency of approximately 28 K was measured across the PAF field of view and operating bandwidth, corresponding to a system temperature below 20 K. To demonstrate the astronomical capability of the receiver, a pulsar (PSR B1937+21) was detected with the real-time beamformer. This thesis provides details on the development of the FLAG digital back end, the real-time beamformer, and reports on the commissioning tests of the FLAG PAF receiver developed by the National Radio Astronomy Observatory (NRAO), Green Bank Observatory (GBO), West Virginia University (WVU), and Brigham Young University for the Green Bank Telescope (GBT).



College and Department

Ira A. Fulton College of Engineering and Technology; Electrical and Computer Engineering



Date Submitted


Document Type





phased array feed, correlation, beamforming, digital signal processing