By Patrick McLaughlin
The 2017 edition of the National Electrical Code includes several new articles that directly relate to the use of twisted-pair communications cables to carry direct current (DC) to power networked devices. An article we published last month (“Data/comm cables and the 2017 National Electrical Code,” October 2017, page 22), authored by Stanley Kaufman, PhD of CableSafe, addressed a number of changes made in the 2017 NEC. Dr. Kaufman is a member of the National Fire Protection Association’s NEC Code-Making Panels 12 and 16 as well as being a member of the NFPA Technical Committee on Electronic Computer Systems.
This month’s article focuses on just a couple of the topics covered in Dr. Kaufman’s article. It is based on a web-based seminar delivered by Cabling Installation & Maintenance on September 29, 2016. That seminar will be available for on-demand viewing until March 2017 at cablinginstall.com/webcasts.
Many in the industry use the term “Power over Ethernet” to refer to any type of remote powering-technology that permits the carriage of DC over the conductors of a communications cable. In a vast majority of cases, the products and systems that provide this ability do so in accordance with the IEEE’s 802.3af or 802.3at specifications. Officially this set of specifications is called Data Terminal Equipment Power via Media Dependent Interface, or DTE Power via MDI. Nowhere in the standard’s official title will you find the term Power over Ethernet or PoE. Nonetheless, references to these specifications almost always include the term PoE.
The IEEE does not “own” the term PoE. And just because a product says “PoE” or “Power over Ethernet” on it, does not indicate that it complies with either the existing or forthcoming specifications that we very frequently refer to as PoE.
Under the original IEEE standard, 802.3af, the power sourcing equipment (PSE) injects between 44 and 57 volts, with 48 being typical, at 350 to 400 mA. In total the power sourcing equipment emits 15.4 Watts and by the time it reaches its destination of the powered device, 12.95 Watts are available to the PD. 802.3af uses two pairs of a four-pair cable.
The “at” standard, commonly referred to as Power over Ethernet Plus, injects 50 to 57 volts (50 is typical) at up to 600 milliamps. It’s worthwhile to point out that the IEEE established that 600-milliamp limit for its 802.3at standard based on a 50-degree Celsius ambient temperature. Generally cables are rated to 60 degrees Celsius, meaning they can operate up to that temperature without degradation to their performance characteristics. Through collaboration between a group in the TIA’s TR-42 cabling standards committee, and the IEEE, it was determined that sending power over twisted-pair cables at 600 mA per pair could account for as much as a 10-degree Celsius rise in temperature.
A couple points to note from these facts. 1) The notion that sending power at certain amperages over twisted-pair cables will cause temperature rise is not new. The 802.3at specification was finalized in 2009 and it accounted for the fact that there would be temperature rise in some cables that were carrying this current. 2) The IEEE and the TIA collaborated extensively as the PoE Plus standard was developed, for exactly reasons like this. That collaboration has continued over years.
Reason for caution
But as mentioned earlier, the term “PoE” does not necessarily equate to compliance with IEEE 802.3af or 802.3at specifications. In our October 2014 issue we published an article authored by Steve Carlson, representing the Ethernet Alliance, addressing this issue. In that opinion article he observed, “Unfortunately, the success of PoE led to products that took shortcuts with the standard, or simply ignored it. Most of the products did not come from mainstream networking vendors who followed the standard. Rather, they frequently came from companies that were not in the networking business, or ‘no-name’ … producers. It’s too bad that the IEEE didn’t trademark ‘Power over Ethernet.’ It might have saved a lot of future trouble. Many cheap ‘PoE injectors’ appeared on the market. These units did not have any of the IEEE standardized features, and were frequently a power supply inside a box that interrupted the ‘idle’ pairs and placed permanent voltage on them. In many cases this voltage was not the correct IEEE standardized voltage, or the power supply could be switched between different output voltages. The current supplied frequently far exceeded those specified in the standard, leading to possible damage of remote devices or the network cabling. These non-standards-based devices were frequently marketed as ‘Power over Ethernet,’ causing potential market confusion regarding the integrity of the IEEE standards-based solutions.”
Carlson authored that article in 2014-11 years after the IEEE 802.3af standard was published and 5 years after 802.3at was. That long after the completion of the IEEE’s standards, rogue devices continue to exist in the market.
While justification exists to use caution when considering deploying power injectors that do not comply with 802.3af or 802.3at, some non-IEEE powering technologies can be deployed safely, without concern about harming network electronics or cabling. One example is the Power over HDBase-T (PoH) specification, developed by the HDBase-T Alliance. The PoH specification is backward-compatible with IEEE specifications. PoH calls for 95 to 100 watts of power over twisted-pair cabling. In a technical paper titled “Introduction to Power over HDBase-T,” the HDBase-T Alliance notes, “PoH enables the PD [powered device] to identify the cable length/resistance and draw more power, as long as the overall power consumption does not exceed 100W. PoH is fully backwards-compatible with the IEEE 802.3at-2009 specification, including the section 33.7.1 mandate that all power sourcing equipment (PSE) conform to IEC 60950-1:2001 and be classified as a Limited Power Source (LPS) carrying no more than 100 volt-ampere (VA) per port without the need for special over-current protection devices. PoH also does not infringe on any of the mandated PoE safety requirements.”
Additionally, Cisco introduced proprietary technology that it calls UPOE-Universal Power over Ethernet-in 2011. UPOE is a 60-watt technology that has been successfully and safely deployed in the years since its introduction. Also, as a historical note, UPOE is not Cisco’s first remote-powering technology. Before there was an IEEE 802.3af specification, Cisco served the market with a working remote-powering system that primarily was used to power Voice over IP phones. As it turned out the 802.3af specifications did not precisely match Cisco’s technology, but nonetheless, safe and effective remote powering predates IEEE 802.3af’s publication in 2003.
It’s possible that sometime in 2017 the IEEE’s next-generation remote powering specification, 802.3bt, will be published. The standard will specify two different remote-powering methods, referred to as Type 3 and Type 4 (802.3af technology is referred to as Type 1 and 802.3at technology is referred to as Type 2). Both powering types specified in 802.3bt will send power down all four pairs in a twisted-pair cable. Based on the standard’s current draft, Type 3 will specify wattage levels up to 60 and Type 4, wattage levels up to 100. Type 3 will employ 600 mA, like 802.3at (Type 2) does. Type 4 will employ 960 mA.
There is general acknowledgement that for 802.3bt Type 4, heat dissipation-the heat generated by the cable carrying power at the specified current of 960 mA-requires attention. In that vein, the Telecommunications Industry Association’s TR-42.7 subcommittee, which deals with twisted-pair communications cabling systems, initiated work on an “A” revision of its TSB-148 specification. The original TSB-184 document is titled Guidelines for Supporting Power Delivery Over Balanced Twisted-Pair Cabling. It was published in 2009-the same year as 802.3at Type 2/PoE Plus remote powering specifications. Work on TSB-184-A began in 2014 and, through collaboration with the IEEE, the document has progressed in parallel with 802.3bt.
The 2017 NEC
As previously mentioned, an article in our October issue addressed several revisions made to the 2017 National Electrical Code that relate to remote powering over communications cables. The National Electrical Code is published by the National Fire Protection Association (NFPA). Its document number is NFPA 70. On its website, the NFPA explains the NEC “is the benchmark for safe electrical design, installation and inspection to protect people and property from electrical hazards.” The association also states, “The NEC addresses the installation of electrical conductors, equipment, and raceways; signaling and communications conductors, equipment and raceways; and optical fiber cables and raceways in commercial, residential and industrial occupancies.”
As is often the case, the 2017 edition of the Code includes cross-references that span multiple articles and sections of the publication. To some extent that is the case with cable temperature ratings.
Most communications cables are rated to 60 degrees Celsius. As of the 2017 Code, temperature rating requirements are consistent for all communications and data cables. All cables have to be rated to at least 60 degrees C. Any cable with a rating that exceeds 60 degrees C must have that rating marked on the cable.
Section 310.15(A)(3) of the 2017 NEC requires that conductors will be installed and operated so they don’t exceed their temperature limits. In some previous editions of the Code there was some gray area about certain cable types and whether or not this requirement applied to them. Those gaps or loopholes were closed in this edition of the Code. So every cable has to have a temperature rating of at least 60; if the cable is rated above 60, that rating must be marked on the cable; and in every case, a cable has to be installed and operated so as not to exceed that rating.
Section 725.144, titled Transmission of Power and Data, is new in the Code’s 2017 edition. It is like other sections of the Code in that it addresses conductor heating through an ampacity table. As Dr. Kaufman explained in his recent article, the term “ampacity” is referenced in the NECStyle Manual as the current-carrying capacity of conductors only. It doesn’t include the current limit of the 8P8C connectors used with cables. But Section 725.144 has a requirement that the current in a power circuit does not exceed the connectors’ current limitations. So the maximum current that can be carried by each conductor in a LAN cable often will be determined by the connector’s current limit.
And speaking of conductors, the conductor sizes in twisted-pair cables represented new territory for the NEC. The ampacity tables that have existed in the Code over many revisions consider conductors that are 18 AWG or larger. (The smaller the AWG number, the larger the size of the conductor.) With twisted-pair cables, 22 AWG represents the large end of the spectrum, down to pairs as small as 26 AWG. So the code-making panels were in new territory here. To determine ampacity tables for conductors in this size range, they turned to information that came from a fact-finding investigation carried out by Underwriters Laboratories and commissioned by SPI, the Plastics Industry Association. The ampacity table developed as a result of the investigation, and included in Section 725.144, is included in this article, on the opposite page.
The table is a matrix of conductor sizes, bundle sizes, and temperature ratings. Based on those three variables, the table dictates the cable’s ampacity-the maximum current that a conductor can carry continuously under conditions of use without exceeding its temperature rating.
Section 725.144 introduces a cable classification called LP, which stands for Limited Power. Underwriters Laboratories offers the LP certification program, through which cables can achieve the LP rating for specific ampacities. The 2017 NEC permits an unlimited number of LP cables to be installed in bundles and carry the amount of current to which the cable is certified.
Article 840 of the NEC is titled Premise-Powered Broadband Communications Systems. It has a new part in the 2017 edition, Part VI: Premises Powering of Communications Equipment over Communications Cables. A new section, 840.160 Powering Circuits, requires compliance with Article 725.144 when the power supplied to a circuit is greater than 60W.
Appeal denied, but …
After the 2017 NEC was finalized at the NFPA annual meeting, an appeal was filed to the NFPA Standards Board specifically related to Section 840.160. The appeal said there was insufficient specificity in that section’s reference to more than 60 watts. The basis for the appeal was that Section 840.160 did not specify an ampere limit. By not specifying an amp limit, the appeal contended, it does not prohibit the types of rogue and potentially unsafe powering devices that have existed in the market for years. The appeal requested that specific language, to include amperage, replace the existing language in 840.160.
The appeal was heard in August and was denied. The language in Section 840.160 is not going to change as a result. But in its decision, the NFPA Standards Council wrote that it “acknowledges there may be value garnering additional input on Ethernet communications to inform future changes to NEC.” The Standards Council also directed a correlating committee to establish a task group that “should specifically include representation of those with knowledge and experience in telecommunications and Ethernet communications.”
As of the time this article was written, this author was unaware of any activity related to the task group, such as its formation, makeup, objectives, or timeline for activity. We will follow this activity and continue to report on it.
These changes to the NEC have been a topic of great interest and much conversation among professionals in the structured cabling industry. This author attended BICSI’s Fall Conference in mid-September, where several formal presentations and many private conversations explored the subject in detail. An overriding issue is how to ensure that a new cabling installation complies with the Code - which is law in the jurisdictions in which it is adopted. And an overriding best-practice recommendation has been to get in touch with the authority having jurisdiction (AHJ) as early in the process as possible in an effort to ensure compliance.
The section of the NEC that addresses enforcement states, “The authority having jurisdiction for enforcement of the Code has the responsibility for making interpretations of the rules, for deciding on the approval of equipment and materials, and for granting the special permission contemplated in a number of the rules. By special permission, the authority having jurisdiction may waive specific requirements in this Code or permit alternative methods where it is assured that equivalent objectives can be achieved by establishing and maintaining effective safety.”
The AHJ can be any number of entities, depending on the jurisdiction. It may be a building inspector, fire marshal, municipal or other government agency, or any number of other individuals or agencies.
As stated by the NFPA and quoted earlier in this article, the NEC concerns itself with safety and installation. Sections 725.144 and 840.160 of the 2017 NEC came into being because of safety concerns about the heating of communications cables that carry power. As a benchmark for safety, the NEC does not consider whether or not a power source complies with any IEEE specification, Power over HDBase-T, UPOE, or any other similar specifications. Hand in hand with that, the NEC does not consider whether or not a twisted-pair communications cable is going to be able to successfully carry a signal from a transmitter to a receiver.
Despite that fact, professionals in the structured cabling industry must consider all those possibilities. Network end-users count on their internal IT departments and/or their cabling contractors to ensure that a physical layer cabling infrastructure will successfully support the travel of a signal from its transmitter to its receiver.
In that regard, when the TIA’s TSB-184-A specification ultimately is completed, it will provide guidelines for that type of assurance. As of the early October meeting of TIA TR-42.7, TSB-184-A remains a work in progress. As mentioned previously, work began on the document approximately two years ago. In the meantime, the 2017 NEC-particularly including the ampacity table in Section 725.144-has taken hold. TR-42.7 is addressing that reality, and some of the group’s current work on TSB-184-A is related to bundle sizes, conductor gauge sizes, and cable temperature ratings. While there is no official projection for the completion of TSB-184-A, the industry looks toward the finalization of that document as another piece in this puzzle.
Patrick McLaughlin is our chief editor.