Thierry Pollet(a), Paul S p r~y t '~) and Marc Moeneclaey(a) FFT (a) Communication Engineering Lab, University of Ghent Sint I'ietersnieuwstraat 41, 9000 Ghent, Belgium tel: i--32-(0)9264 34 12, fax: i-32-(019264 42 95Abstract -In fully digital receivers, carrier and timing information is derived from samples of the (anti-aliasing-filtered) received continuous-time signal. In case of synchronized sampling, this information is used to align the sampling clock of the receiver with the remote transmit clock. In nonsynchronized sampling systems, the sampling at the receiver is performed by means of a fixed free-running clock, and additional post-processing is necessary to perform timing correction in the digital domain. In this paper, we investigate the effect of non-synchronized sampling on the BER performance of OFDM systems. We calculate the BER degradation caused by a given frequency offset between receiver and transmitter clock, as compared with the case of ideal sampling. The obtained resulb arc compared with the performance of synchronized sampling systems.

No abstract

At present, Discrete Multi-Tone (DMT) transceivers receive much attention for the implementation of Asymmetric Digital Subscriber Lines (ADSL). This paper analyzes the effect of clipping a DMT-signal, i.e. limiting the signal's maximum amplitude. An exact expression is given for the signal-to-noise ratio ( S N . ) degradation due to clipping alone. By combining this expression with the well-known expression for the quantization noise in the AID-DIA converters, it is shown how clipping can reduce the number of bits of AID-DIA converters as well as the dynamical range of the line drivers while keeping the overall SNR the same as without clipping.

VDSL services are overviewed with an emphasis on the basic architecture, applications, and data rates, as well as the technological challenges of design. Discussions of the telephone line environment, radio-band interference ingress and egress, impulse noise, and symmetric and asymmetric multiplexing are included, along with brief descriptions of two popular implementations.

No abstract