Two Port Memory-based FFT Architecture for DVB-T
| Autor: | Yung-Ming Chang, 張雍明 |
|---|---|
| Rok vydání: | 2014 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 102 Orthogonal frequency division multiplexer (OFDM) is used widely in communications systems now and modulation technique is also be used usually, such as ADSL, DVB, 802.11a etc. To achieve the purpose of modulation and demodulation, the Fast Fourier Transform (FFT) and Inverse Fast Fourier Transform (IFFT) are implemented into a circuit. The FFT plays significant role in applying digital signal processing, radar, medical electronics, and other fields. With the expansion of application fields as well as the deepening of application requirements, FFT processor data points need to be addressed in different way standards for different systems have different points specification. It discusses to analysis several of FFT’s algorithm complexity and hardware architecture design algorithm in this paper. In traditional architecture, it can be divided into two architectures pipeline and memory-based. The advantages of pipeline have higher throughput rate and simple control unit, but it’s hardware larger more; otherwise memory-based has small area but low throughout rate compared with pipeline architecture. This paper proposes a two port memory-based FFT architecture. This processor is Radix-2 FFT processor and proposes a memory read and write method to make the two ports input data path mode operating. According to require of the computing process make its time and the value in the memory locations be all correct. In the process of the memory stored, because of the operation process for each stage produced of bit size had increase. Therefore, we do dynamic scaling to reduce request of memory access of bit size. In addition, we also research for twiddle factor, and make use of a function in math symmetrical, quantization and differential compensation to reduce ROM size. Finally, the architecture can support to DVB-T 2K/8K FFT point and can reach higher than 40dB in SNR. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|