A near-end crosstalk (NEXT) canceller using a fine-grain pipelined arc
hitecture is presented. Performance. of the proposed pipelined NEXT ca
nceller is demonstrated in the 125 Mb/s twisted-pair distributed data
interface and 155.52 Mb/s asynchronous transfer mode local area networ
k applications. In addition, we analyze the computational complexity o
f the proposed pipelined NEXT canceller. It is shown that this archite
cture can be clocked at a rate that is 107 times faster than the seria
l architecture with a maximum loss of 2.0 dB in signal-to-noise ratio
(SNR).