Fibre cabling systems need precise testing processes
Wed 4 Apr 2018 | Michael Akinla
Cabling infrastructure has become a critical component in today’s data centre environments. Performance and reliability are important, but now the margins for error are even finer. When deploying high speed optical networks such as 100Gb/s Ethernet and 128Gb/s Fibre Channel, the network stakeholders must have accurate knowledge of the fibre cable performance against the application standards.
As network performance has increased, the specification on components in the fibre cable plant have become more rigorous. Since the standardisation of 10Gbs Ethernet in 2002, the 2.55dB total Channel Insertion Loss (CIL) for 50/125μm OM3 multimode fibre has been reduced to 1.9dB for 40GBase-SR4 (and 100GBASE-SR10). Today, a maximum connector loss of 1.0dB is required for a 150m OM4 channel that may contain multiple connector interfaces.
Current plug and play multimode fibre structured cabling systems assembled around LC and MTP connector systems have minimal insertion loss compared to the required cabling and component standards. Current cabling standards and application standards (Ethernet and Fibre Channel) require that no mated connector pair exceed 0.75dB Insertion Loss (IL). The latest multimode LC connectors offer average losses of less than 0.1dB and leading vendors offer ultra performance MPO connectors that show no more than 0.25dB Insertion Loss.
After cabling installation, the permanent links are qualified against cabling standards with nearly ideal patch cords, that yield excellent loss measurement repeatability and reproducibility. This type of measurement with “reference grade” patch cords assures that the chance of costly measurement errors are minimised and the links are tested and commissioned in a timely fashion.
Testing is a serious business, and the use of legacy test equipment (designed for SC cable plant) can produce anomalous results, as the test gear must be adapted to the LC-style cable plant via the use of ‘hybrid’ jumpers (SC to LC). This introduces extra connector interfaces into the measurement which can have a significant impact on the efficacy of field testing of links with LC or MPO connectors. The potential to produce false fail results and false passing results increases markedly with in-field testing.
To reliably measure the loss of a 30m OM3 pre-terminated permanent link in-field to the relevant standards requirements, where the expected total loss is a little over 1.6dB, the measurement system repeatability and reproducibility needs to be a fraction of 1.6dB. In fact, the industry recommendation is 20-30% of the test limit. This is further constrained when customers intend to measure to even lower “engineered limits”. At some point, the test equipment and operator start to make significant quantities of measurement errors for these low loss test limits.
Moving forward into the era of high performance cabling such as OM5 and permanent links installed with low-loss multimode fibre (MMF), we will increasingly see the need for ultra low loss connectors that support the higher speed protocols. These cabling systems require compliance with tight customer and industry specifications and highly accurate and reliable insertion-loss measurement processes.