EuroWire May 2015

Technical article

Measured and simulated DC powering of data cables for power over Ethernet By Stephen W Simms, Brand-Rex Ltd

Abstract The increasing demand for higher power levels in Power over Ethernet (PoE) systems is evident, with a variety of non-standard products currently available on the market which provide power levels in excess of those stated in IEEE 802.3at. Higher power levels will allow PoE to be used in a wider range of applications. However, they will also increase perfor- mance risk. With this increase in demand for more power, and the fact that installations using PoE technology differ greatly in terms of their configuration and environment, it is beneficial to mitigate risk by using numerical simulation. The work presented here provides numerical simulation and experimental verification of the thermal properties of data cables under DC powering which is used in PoE applications. Introduction The supply of DC power to end devices along the same electrical path used for AC signal communication has been successfully employed for many years, eg in telephones and audio equipment. The technique used to provide this functionality is commonly known as ‘phantom powering’. In relation to Ethernet, this technique allows power from the Power Sourcing Equipment (PSE) to be delivered to the Powered Device (PD) on the same pair that is used for data. The DC power is applied to the centre tap of the signal coupling transformer and does not interfere with data transfer. This allows PoE to be deployed over 1000BASE-T systems, in which data is carried on all four pairs.

IEEE 802.3at standardisation in 2009 stated the system parameters required for Type 1 (PoE) and Type 2 (PoE+) [1] . The standard classifies nominal highest DC current values of 0.35A and 0.60A per pair, for Type 1 and Type 2, respectively. Some of the most common applications which use PoE technology include wireless LAN access points, VoIP telephones and network cameras. Applying electric current to a conductor releases heat energy, an effect known as Joule heating. In relation to Ethernet cables and components, this heating effect causes concern due to the rise in attenuation, which has a limiting effect on link length. This concern is heightened for cables with a higher resistance than standard cables, eg copper clad aluminium (CCA) [2] , and smaller diameter (26 AWG) solid copper conductor cables. In 2009, IEC subcommittee 46C put forward a test method (46C/906/NP) entitled ‘Proposal for measuring of heating of data cables by current’ [3] . In this paper, the aim is to achieve a strong correlation between simulation and the proposed measurement method regarding the DC powering of Ethernet cables for PoE applications. The paper also aims to compare temperature rise due to DC powering of CCA cable with cables which have solid copper conductors. Numerical modelling A 2D model was set up using COMSOL Multiphysics 4.4, a software package which utilises the Finite Element method [4] . The model was set up to replicate the proposed measurement method [3] , which allowed for a comparison between theory and practice.

Energised pairs

LSZH jacket

Air

Probe

Cu

Polyolefin

AI/PET tape

In order to achieve this, a five-cable linear configuration was set up with the intention of providing a good prediction of the thermal behaviour at the centre cable without the need for including additional cables in a model requiring higher computational resource. Heat capacity at constant pressure, density and thermal conductivity material properties were applied to represent the constituent parts of the Cat6A 26 AWG U/ FTP cable. These properties were applied to the copper (Cu) conductor, aluminium/ PET (Al/PET) tape, Low Smoke Zero Halogen (LSZH) jacket, and polyolefin insulation, see Figure 1 . Conduction, convection and radiation heat transfer mechanisms [5] were accounted for in the model. Simulated electric energy was applied to one pair of each cable in the model. A stationary solver was used to determine the thermal behaviour for (a), a point at the centre of one of the energised conductors (see probe position in Figure 1 ) and (b), a 2D temperature plot of the cross-section, Figure 2 . ▲ ▲ Figure 1 : Simulation setup in COMSOL Multiphysics

78

www.read-eurowire.com

May 2015