Estimation Method for Statistical Eye Diagram in a Nonlinear Digital Channel

An estimation method for statistical eye diagram in a digital channel with nonlinear circuitry is proposed, which is based on the superposition of multiple bit pattern responses method. The proposed method captures the nonlinear effects in the channel by using 2^m bit pattern responses as bases for doing superposition. By selecting a larger m, the nonlinear performance of the channel can be captured more accurately. Besides the statistical eye diagram, a corresponding bathtub curve computing method is also proposed, which not only gives the total bit error rate (BER) distribution, but also the BER when transmitting a logic “1” signal and the BER of transmitting “0.” The proposed method is verified using two different channel topologies: the push-pull output driver and the open-drain output driver. The predicted results are in agreement with the result of PRBS, while the CPU time of this method is much less than PRBS.