Last month, in continuing to report on the Wet Link Phenomena, I heavily referenced different parts of the National Electrical Code (NEC). One notation I made at the very end of my reporting, which did not make it into the final cut of the printed column due to space limitations, was the following: If the NEC represents minimum safety requirements, and TIA standards represent minimum performance requirements, how can we get the NEC code panels’ attention?
That theme is particularly noteworthy with respect to a topic I have covered several times over the past few years-Power over Ethernet. Currently, I am researching the dynamics of PoE and the NEC. Essentially, I’m interested in knowing (and I bet many of you also will be interested in knowing) what, if any, role the NEC code-making panels plan to play in setting requirements for the distribution of power via PoE.
I hope to have more detailed information on the nuts-and-bolts of the NEC/PoE dynamic next month. For now, here is some background information on PoE that comes from research I have conducted over the past couple years.
PoE: How and why
Almost all networked devices require both data connectivity and a power supply. Basic analog telephones are powered from the central office through the same twisted pair that carries the voice.
By using Power over Ethernet (PoE), the same thing is possible for Ethernet services, such as:
VoIP (Voice over Internet Protocol) telephones;Wireless LAN access points;Bluetooth access points;Web cameras.In response to a growing need to supply power over the Ethernet cables, IEEE began work on 802.3af in 1999.On June 12, 2003, the Power over Ethernet Standard was approved by the IEEE Standards Board to be published as: IEEE 802.3af-2003, Data Terminal Equipment (DTE) Power via Media Dependent Interface (MDI).The IEEE 802.3af standard was developed to provide a 10Base-T, 100Base-TX, or 1000Base-T device with a single interface to both the data it requires and the power to process the data. Besides the obvious-“You do not have to plug the device into a local power receptacle”-there are other advantages.PoE allows use of a centralized uninterruptible power supply (UPS), guaranteeing power to the devices during power failures. This is becoming very important as more security, access control, and building automation services are moving to Ethernet. And my personal favorite: devices can be shut down or reset remotely-no need for a site visit to press the reset button or toggle the power switch. More control with fewer truck rolls.IEEE 802.3af uses ISO/IEC IS 11801:2002 Generic Cabling for Customer Premises Standard cabling. The minimum cabling channel requirement in ISO/IEC IS 11801:2002 is Category 5e; however, use of lesser-category cable is also allowed by IEEE 802.3af. IEEE always wants to use the installed base of cabling for their new applications. (Let me fill you in on a little secret: that’s because it leaves more money in our IT budgets for their new equipment.)Getting power onto cablingPower is injected into the cabling by power sourcing equipment (PSE), which is located at the data terminal equipment (DTE) at either end of the channel or at a cross-connection point along the way (midspan). The power is used by a powered device (PD) located at the end of the channel.A PSE can provide power to the PD in one of two wiring configurations:Alternative A, sometimes called “phantom mode powering,” supplies the power over pairs 2 and 3.Alternative B supplies power over pairs 1 and 4.Endpoint PSEs can be compatible with 10Base-T, 100Base-TX, and/or 1000Base-T systems, and can use Alternative A, Alternative B, or both.Midspan PSEs can be used only with 10Base-T and 100Base-TX, and can only use Alternative B. Why? Because using a midspan PSE contributes an additional connection in the signal path (and puts three connectors close together in the telecommunications room), which can affect the channel’s near-end crosstalk, power-sum near-end crosstalk, equal-level far-end crosstalk, power-sum equal-level far-end crosstalk, and return-loss performance. Trading line powering of the device for a channel that cannot support the application would not be a good idea.How much power?IEEE 802.3af allows delivery of a maximum of 12.95 watts of power over two pairs to a PD at one end of the cabling. While there are several power classifications in the standard, Class 0, which covers the entire range, is the default. Class 0 allows a maximum power level of 15.4 watts at the output of PSE, and power levels ranging from 0.44 to 12.95 watts at the input of PD.The maximum continuous output current from the PSE under normal operation is 350 milliamps direct current (DC) per pair, or 175 milliamps DC per conductor. This was based on the calculation of the maximum capacity of 26 AWG stranded conductor cables at their maximum environment operating temperature with the maximum allowable temperature rise. But now, consider that at startup, peak current can be as high as 5 amps for the first 1 millisecond, and 450 milliamps DC for up to 50 milliseconds.Power is only applied when a PSE detects a device that requires power and is capable of accepting it. This ensures that there is no electrical arcing when inserting the plug into the jack, and protects ordinary DTE that would likely be destroyed if power was injected onto the signal conductors.Three-plus years after IEEE approval of the 802.3af standard, a “higher-Power over Ethernet” standard is in the works. The IEEE 802.3at specifications, commonly referred to as “PoE Plus,” are currently being constructed with an anticipated publication date in the second quarter of 2008. Under the forthcoming specifications, PoE Plus most likely will permit at least 30 watts of power to each PD, and perhaps more.Meanwhile, back in TIA, a task group within the TR-42.7 Subcommittee is developing a set of PoE Plus Performance and Test criteria. One of the task group’s primary objectives is to develop a matrix of component and channel performance criteria (resistance unbalance, temperature rise, insertion loss at elevated temperature, etc.).Why are such examinations of the effects of PoE and PoE Plus necessary? Think of structured cabling as an interstate highway system. All sizes and shapes of vehicles travel on the system. But no vehicle is allowed on the highway that cannot meet the minimum speed requirements or would damage the highway so that other vehicles could not use the same highway in the future. Size, speed, and weight were all considered by the civil and structural engineers before they designed an interstate highway system. The same is true of a structured cabling system, only the engineers tend to be electrical and mechanical.For further information on Power over Ethernet, see: www.PowerOverEthernet.com. To download a free copy of IEEE 802.3af, see: standards.ieee.org/getieee802. For further information on work in TIA TR-42, see: www.tiaonline.org.