Impulse Noise and Narrowband PLC

TL;DR

This paper examines how impulse noise, both random and periodic, affects narrowband Power Line Communications, highlighting differences from theoretical models and analyzing capacity loss due to FFT processing.

Contribution

It provides a comparative analysis of measured impulse noise against theoretical models and investigates the impact of noise on OFDM-based narrowband PLC systems.

Findings

Measured impulse noise has memory, unlike theoretical models.

FFT operation transforms impulse noise into Gaussian noise with specific variance.

Periodic noise directly affects FFT output, influencing system performance.

Abstract

We discuss the influence of random- and periodic impulse noise on narrowband (< 500 kHz frequency band) Power Line Communications. We start with random impulse noise and compare the properties of the measured impulse noise with the common theoretical models like Middleton Class-A and Mixed Gaussian. The main difference is the fact that the measured impulse noise is noise with memory for the narrowband communication, whereas the theoretical models are memoryless. Since the FFT can be seen as a randomizing, operation, the impulse noise is assumed to appear as Gaussian noise after the FFT operation with a variance that is determined by the energy of the impulses. We investigate the problem of capacity loss due to this FFT operation. Another topic is that of periodical noise. Since periodic in the time domain means periodic in the frequency domain, this type of noise directly influences the output of the FFT for an OFDM based transmission. Randomization is necessary to avoid bursty- or dependent errors.