TY - GEN
T1 - A low-power 64-point FFT/IFFT design for IEEE 802.11a WLAN application
AU - Lin, Chin Teng
AU - Yu, An Chu
AU - Van, Lan-Da
PY - 2006/12/1
Y1 - 2006/12/1
N2 - In this paper, we propose a cost-effective and low-power 64-point fast Fourier transform (FFT)/inverse FFT (IFFT) architecture and chip adopting the retrenched 8-point FFT/IFFT (R8-FFT) unit and an efficient data-swapping method based output buffer unit The whole chip systematic performance concerning about the area, power, latency and pending cycles for the application of IEEE 802.11a WLAN standard has been analyzed. The proposed R8-FFT unit utilizing the symmetry property of the matrix decomposition achieves half computation-complexity and less power consumption compared with the recently proposed FFT/IFFT designs. On the other hand, applying the proposed data-swapping method, a low-cost and low-power output buffer can be obtained. So as to further increase system performance, we propose one scheme: the multiplication-afler-write (MAW) method. Applying MA W method with R8-FFT unit, the resulting FFT/IFFT design not only leads to the balancing pending cycle, but also abbreviating computation latency to 8 clock cycles. Consequently, adopting the above proposed two units and one scheme, the whole chip consumes 22.36mW under 1.2V@20 MHz in TSMC 0.13 1P8M CMOS process.
AB - In this paper, we propose a cost-effective and low-power 64-point fast Fourier transform (FFT)/inverse FFT (IFFT) architecture and chip adopting the retrenched 8-point FFT/IFFT (R8-FFT) unit and an efficient data-swapping method based output buffer unit The whole chip systematic performance concerning about the area, power, latency and pending cycles for the application of IEEE 802.11a WLAN standard has been analyzed. The proposed R8-FFT unit utilizing the symmetry property of the matrix decomposition achieves half computation-complexity and less power consumption compared with the recently proposed FFT/IFFT designs. On the other hand, applying the proposed data-swapping method, a low-cost and low-power output buffer can be obtained. So as to further increase system performance, we propose one scheme: the multiplication-afler-write (MAW) method. Applying MA W method with R8-FFT unit, the resulting FFT/IFFT design not only leads to the balancing pending cycle, but also abbreviating computation latency to 8 clock cycles. Consequently, adopting the above proposed two units and one scheme, the whole chip consumes 22.36mW under 1.2V@20 MHz in TSMC 0.13 1P8M CMOS process.
UR - http://www.scopus.com/inward/record.url?scp=34547271863&partnerID=8YFLogxK
U2 - 10.1109/ISCAS.2006.1693635
DO - 10.1109/ISCAS.2006.1693635
M3 - Conference contribution
AN - SCOPUS:34547271863
SN - 0780393902
SN - 9780780393905
T3 - Proceedings - IEEE International Symposium on Circuits and Systems
SP - 4523
EP - 4526
BT - ISCAS 2006
T2 - ISCAS 2006: 2006 IEEE International Symposium on Circuits and Systems
Y2 - 21 May 2006 through 24 May 2006
ER -