Power line communication (PLC) is the use of power lines for the purpose of electronic data transmission. The presence of additive noise, namely, background noise and impulsive noise, significantly affects the performance of a PLC system. While the background noise is modeled by Nakagami-m distribution, the impulsive noise is modeled using Middleton class A distribution. In this paper, we study the performance of a PLC system under the combined effect of Nakagami-m background noise and Middleton class A impulsive noise assuming binary phase shift keying signalling. The probability density function of decision variable under the influence of additive noise (sum of background noise and impulsive noise) is derived. We also derive an analytical expression for the average bit error rate of the considered PLC system. The analytical expressions are validated by close matching to the simulation results. The analysis presented in this paper closely predicts the behavior of the PLC system under the combined effect of background and impulsive noises.Keywords -Background noise, binary phase shift keying, bit error rate, impulsive noise, Middleton class A distribution, Nakagami-m distribution, power line communications.I. INTRODUCTION In the past few years, power line communications (PLC) has been gaining the attention of the researchers due to the ubiquitous nature of power lines and widespread availability. PLC is an emerging field of communication for the home area network of smart grid. PLC, with its huge coverage area and existing power line infrastructure, has the potential to cater to the ever-increasing demand of high speed access to video and data [1]. However, PLC offers a difficult environment for communication due to the random fluctuation of its key parameters, such as, noise, impedance, and attenuation with time, frequency, and distance [2].There are two broad categories of additive noise in PLC, namely, background noise and impulsive noise. The longterm measurements of noise spectrum in PLC conducted in [3] over a frequency range of 1 MHz to 30 MHz suggest that time domain amplitude of the background noise follows Nakagami-m distribution. Using this noise model for the background noise, a maximum likelihood detector for binary and quaternary transmitted signals is derived in [4] and [5], respectively, and the corresponding average bit error rate (BER) is obtained. The performance of a PLC system under the influence of Nakagami-m additive background noise over Rayleigh fading is evaluated in terms of BER and outage probability in [6]. In [7], the outage probability analysis of a PLC system with Rician channel under Nakagami-m background noise is performed. However, the effect of impulsive noise has been ignored in the BER analysis in [3]- [7]. Although, the Bernoulli-Gaussian distribution is used to represent the