Data communication circuit with equalization control

Abstract

An adaptive equalizer comprises an adjustable equalizer circuit (10) that allows to enhance the frequency dependence of contents of the transmitted signals which suffer from losses in the connected transmission channel. A blind equalization tuning procedure is proposed that operates without knowledge about the characteristic of transmission channel. Phase positions of transitions in the equalized signal are detected through a transition phase detector (44). A digital post-processing circuit evaluates a measure for spread of the detected phase positions of transitions, accumulated over a plurality of the symbol periods. The digital post-processing circuit (14) controls the adjustable equalizer, setting the adjustable equalizer to a setting wherein the detected spread is minimized.

Description

technical field [0001] The present invention relates to data communication circuits. More specifically, the present invention relates to automatically adjusting channel equalization to compensate for frequency dependent distortion of signals transmitted over communication channels. Background technique [0002] The need to equalize signals in high-speed communications has been described in J. Liu et al., entitled "Equalization in High-Speed ​​Communication System", published in IEEE Circuits and Systems Magazine, pp. 4-17, 2004. [0003] A wired communication system conceptually consists of three distinct building blocks: a transmitter (TX), a channel (eg, cable or optical fiber), and a receiver (RX). Due to non-ideal channel characteristics such as limited bandwidth and crosstalk noise, the RX input signal is degraded such that data recovery on the receiver side stalls due to an unreasonable bit error rate (BER). This problem is exacerbated by the need to increase system ...

AI Technical Summary

Problems solved by technology

Here, data dependent errors cause equalizer tuning misalignment, since received power also depends on the sequence / order of transmitted bits

Second, ROM / RAM memory is required

[0012] Also, most prior art approaches require a lot of analog hardware like filters, peak detectors, amplifiers, etc., so these techniques consume a lot of power and silicon area

[0013] Therefore, prior art equalizer tuning algorithms are either limited to certain applications or require considerable power and area for consuming analog circuits

Images