Power over Ethernet (PoE) The IEEE 802.3 Ethernet committee created a standard for powering network devices such as wireless access points, VoIP phones andsurveillance cameras off the copper pairs in a 4pair UTP cable. The standard was developed when it was realized that there were two unused pairs in the UTPcable at that time. Ethernet up to 100BaseTX, used only pairs 2 and 3, leaving pairs 1 and 4 available to provide power. Later versions of Ethernet used all 4pairs and the PoE standards were revised to also use all four pairs, allowing for higher power levels. PoE uses a 48 volt power supply and requires cable of Cat 5 rating or higher. Power may be delivered using what are called midspan devices, dedicated PoE power supplies that can be plugged into links or even patch panels, as well or endspan devices, typically switches designed to provide power as well as function as anEthernet switch. Power over Ethernet Standards
*PoE++ standards are expected to be approved in 2018 PoE has had a long development cycle toward standardization, so like many other systems, numerous “prestandard”system implementations exist, not all of which are compatible. In addition, by oversight the terms "PoE" and "Power over Ethernet" were not trademarked, allowing its use for manynonstandard applications. The term PoE is being used for applications like LED building lighting or powering devices as big ascomputers. Some devices and applications may use more power than available from standardized PoE systems (more than 100 watts), distances longer than structured cabling (100m) and some cables may notbe capable of carrying full power without excess heating. Writers of the US National Electric Code are now developing codes for the use of PoE because of the safety implications of higher power applications. Generally, PoE use should not affect the design or installation of standardized UTP cable as defined by structured cabling standards, but the cabling designer must be careful to understand what devices are being considered for use on the cabling and if the cabling does not conform to standards to ensure compatibility and safety. PoE Analysis From Ethernet Alliance (With Our Analysis)In late 2017, CI&M magazine ran an interesting article by two people from the Ethernet Alliance titled "Clarifying misperceptions about Power over Ethernet and cable losses" that is an attempted rebuttal of some recent articles talking about the dangers of running high current over the small conductors of UTP cable in PoE. One of their arguments is based on the size of the conductors in electrical cable vs. UTP communications cable. They show this graphic:And here is part of their argument: To be able to accurately compare the equivalent amount of copper in a network cable, the area of four conductors would need to be combined and displayed as if they were a single conductor at the same scale. Four conductors of a 24AWG network cable are equivalent to a 1mm copper conductor; 22 AWG is equivalent to 1.3mm. Our analysis: Looking at a table of wire gage specs, four conductors of a 24AWG network cable are equivalent to a 1mm2 copper conductor(18gage); 22 AWG is equivalent to 1.3mm2 (16 gage). As a guide, #14 wire (1.6mm) is designated for 15amp circuits, #12 wire (2.1mm) is designated for 20amp circuits, #10 wire (2.6mm) for 30 amps. Smaller wires are sometimes used for short extension cords, 16 gage for typical indoor cords used for lamps and 18 gage for low voltage lamps or short lamp cords.
So
their comparison is close  UTP is about equivalent to the
wire in a lightweight extension cord. Which
brings up another issue. Yes, 4 conductors at 22gage is
equivalent to 16gage, the wire used for short extension cords.
But PoE is looking at much longer runs  100m is a lot longer
than an electrical extension cord for a 100W lamp, the power
they want to run over UTP, so there is much higher resistance
and higher heating of the cable. And because the voltage of
PoE is lower (48 vs 120/240V), it needs much more current to
source the same amount of power. Since power dissipated in the
cable is calculated as Voltage (V) times current (I) or the
current (I) squared times the resistance (R), the power lost
in the UTP cable is significantly higher.
Let's calculate some circuits, it's easy: Powering 100W over UTP from 50V (rounding off) requires 2 amps. If we use 4 pair 22 gage wires, we have to conduct 0.5A per conductor in parallel. At 53 milliohms/m, at 300m the resistance is ~16 ohms and the voltage drop (I x R) on the wire will be about 8 volts and the power loss is 4 Watts (V x I). For 24 gage wire, the resistance is 84 milliohms/m, so at 300m, we have 28 ohms, a voltage drop of 14 volts and power loss of 7 Watts. By the way, a 14V voltage drop means the PoE powered device must be designed to operate at ~34V. Compare that to 100W on a 120V circuit on 14 gage electrical wire. At 120V, a conductor must carry 0.83 amps. On 100m, to make the comparison equal, the electrical cable has a resistance of 2.5 ohms for a voltage drop of 2.1V and a power loss of 1.7 Watts. The math is simple; try it yourself. And they say: A new version of the specification is being developed (IEEE P802.3bt), with an expected release date early 2018, and this new version will define operation over all four pairs of the network cable. This will mean that 100 percent of the copper in the network cable is used for power transfer in PoE. In the AC mains cable, 66 percent of the copper is used for power transfer. The reason that electrical cables use only two conductors for power is the third conductor is used for grounding connections for safety. PoE has no grounding requirements because of the lower voltage, 48V vs 120 or 240 for electrical power. The good news is that arguments for 20% power losses are hard to justify, but 7% is still significant. The concern is that UTP cables are often bundled by the hundreds in cable trays where heat is hard to dissipate, while electrical power cables are usually run on their own. Even if it's only 7 watts per cable, bundle hundreds of cables and the power is significant, likely to cause serious heating. Designers of electrical systems learn how to calculate losses when designing electrical power systems. One hopes the same happens with designers of UTP cabling systems. PoE And Counterfeit CableSuperior Essex Comparison Of Cable Efficiency This graph showed up in an interesting white paper on PoE from Superior Essex, LeGrand and Fluke. You might remember we did an analysis of PoE power efficiency in the November FOA Newsletter article "PoE Analysis From The Ethernet Alliance (With Our Analysis). This graph has some data we did not consider  the efficiency of counterfeit cable that uses copperclad aluminum conductors instead of pure copper. The inefficiency of this cable means that more than half the power will be lost in transmission  not only inefficient but dangerous as the power lost is converted to heat. Considering that the samples of counterfeit cable that we have tested use flammable insulation and jackets, that could be a very hazardous application! Recent conversations with people involved with adding PoE applications to the US National Electric Code bring up a further problem. The IEEE did not register "PoE" as a trademark so it's being used to describe applications like daisychained lighting and cables longer than the 100m limit of Ethernet. This makes the safety considerations even more important  and scary. Premises Cabling Website Contents Each page will open in a new window. Overview of Premises Cabling and StandardsTable of Contents: The FOA Reference Guide To Fiber Optics 
