DIAGNOSING ETHERNET CABLES USING TIME DOMAIN REFLECTOMETRY IN ETHERNET CIRCUITS

Abstract: Ethernet cables are a vital component in modern systems, providing reliable data transmission between devices. However, like any other physical component, these cables are subject to wear and tear. Depending on several factors, this wear and tear can happen at different rates, and unplanned stops in a system must be avoided. Having a method to detect when a cable is degrading is where the built-in diagnostics functions of the ethernet circuits will be utilized. A fast and straightforward cable diagnosis can be made using time domain reflectometry functions built into the ethernet circuits. These functions are already built-in to most ethernet circuits but are not enabled, and their performance is not entirely known. They are potentially providing a low-cost solution to a complex problem.  The goal of this thesis is to determine the functionality and performance of these functions. Several experiments have been carried out by creating a test platform that can access these diagnostics functions. The test platform combines time domain reflectometry functions with error monitoring to detect faulty cables. The diagnostics functions of the ethernet circuits provide good detection for hard faults and the possibility of localizing them, while detecting soft faults could be more reliable; some are detected, while some are left invisible.  These functions provide good fault detection and localization in longer cables. Every hard fault can be detected and localized to an accuracy of \textpm 1 m. However, soft faults are not always detected and require additional error detection, but when they are detected, they have the same accuracy as hard fault detection. The time domain reflectometry functions are not enough for detecting soft faults, and some error monitoring or other method is needed. The time domain reflectometry functions of the ethernet circuits have some promise in use cases where fault location is essential, and low cost and low complexity are needed.