Abstract
In this paper, we propose a transversal adaptive equalizer using the least mean squares (LMS) algorithm for digital power line carrier systems and discuss about the optimal number of taps, the step size parameters, and the achievable bit error rate (BER) performance. First, we present two power line models: Model 1 with a line trap on the power line branch and Model 2 without a line trap on the branch. Then, we present theoretical analysis and the computer simulation results. In the case of Model 1, the sufficient number of taps is around 12, while in the case of Model 2, it increases to 21. This is because a line trap in the case of Model 1 can suppress the electric energy of the delay path. It is desirable to change the step size parameter µ in the training and tracking modes in order to achieve fast mean square error (MSE) convergence rate and a good BER performance; i.e., a relatively large value of µ=0.01 should be used in the training mode while a small value of µ=0.001 should be used in the tracking mode.
