Copper cabling has a future, and the ARJ45 connector is a big part of it.
By Yakov Belopolsky, Stewart Connector
Some time ago a panel at a fiber-optic meeting discussed the topic, “How to finally kill copper in data centers.” The question is: Why kill copper? One can argue that because both the fiber-optic and copper cabling (and, for that matter, wireless) compete in the area of signal transmission, there should be winners and losers. However, a trip to your local supermarket should illustrate the fallacy of such an argument; if the meat and potatoes compete for a space on your dinner table, one of them should have disappeared from the supermarket shelves.
In many a discussion on the cabling infrastructure, the valid technical and economic facts have been replaced by myths, and common sense replaced by emotional outbursts. In this author’s memory, no copper-centric meeting ever discussed “killing optical fiber.”
In every single case, the benefits of each type of cabling need to be examined using technical information and common sense. The decision must be suitable for a specific application. The networks of the 21st century contain now, and will contain in the future, a multitude of means of connecting billions of users to the internet. The fiber-optic and copper cabling shall find their proper places. This article examines the benefits of copper cabling in comparison to optical-fiber cabling, and discusses the new copper technologies for high-speedcommunications.
The ARJ45 jack shown here has the same form factor as the RJ45.
The copper-vs.-fiber myth
A popular myth goes, “Fiber optics are faster than copper because photons are faster than electrons.” The optical fiber can deliver greater bandwidth, it is true; but the bandwidth is a different characteristic than speed. The technical term describing the signal speed is the velocity of propagation. The velocity of propagation of electromagnetic signal through twisted-pair copper cabling is about 0.66 of the speed of light in a vacuum (C); coaxial cabling allows more than 0.8C. The best singlemode glass fiber transmits photonic signals at about 0.71C. Two messages sent from San Francisco—one over an optical-fiber network and the other by electromagnetic pulse over copper cabling—will arrive in New York at approximately the same time.
The aforementioned panel could not find a sure way to kill copper, because the cost of a copper link is less than that of fiber optics. The cost of adding transceivers at both ends of the channel is the penalty to pay for other advantages of optical fiber. The competitive cost of copper connections within data centers, assuming 10GBase-T switches and Category 6A cabling, is about $0.50/Gbit/sec per link. An installation cost equation includes the cost of connectors and cables proper. The copper and fiber cables are approximately at par, with fiber possibly being lower. The copper connectors are less expensive; in particular, taking into account that a duplex fiber channel needs four connectors versus two needed for a copper link. However, the copper is even less costly if one considers the lifetime maintenance and repair. In one case, the replacement of a failed copper connector (RJ45 plug) takes 2 minutes and less than $3. A repair of a failed fiber-optic link could require the replacement of a link, at a cost from $50 to $250. An optical transceiver could add about 20 percent to the cost of a laptop computer. If you have not seen many laptops with a fiber data port lately, the cost is the reason.
In the fiber-optic transmission, the great source of power consumption is the dual transformation of the power from the electronic to the photonic signal. Per Encyclopedia of Laser Physics and Technology, such efficiency is typically about 50 percent. The dual transformation would quadruple the power consumption in comparison to the pure electronic transmission. The fiber cabling has lower attenuation as a function of frequency. The optical fiber is energy-efficient when one needs to transport huge blocks of data over long distances. However, a common-sense approach indicates that if both the optical receiver and transmitter are in the same building—and if you have to pay for both transceivers’ power consumption (and the building’s cooling)—you may want to have another look.
PoE, IoT, EMI
A huge advantage of copper connectivity is the ability to support Power over Ethernet (PoE), where the same cabling can transmit power as well as signals. The majority of the network switches on the market have built-in PoE options. The PoE also helps to dispel another myth that the copper is being replaced by wireless. In fact, each wireless transmitter, commonly referred to as wireless access point or WAP—is connected to a copper cable. Considering the places where WAPs are installed (ceilings, attics, rafters, towers) it is often difficult and expensive to route both power cabling and network cabling to such places. Many WAPs have the same cable supplying the power and signals. The copper cabling PoE provides the most economical solution.
A number of processes within the industrial and automated sensors generate an electrical signal. The Ethernet protocol dominates local area networks (LANs), and copper-based Ethernet over twisted wiring dominates the premises wiring market. Industry 4.0 is envisioned as an integral part of the Internet of Things. Industry 4.0 grows from the present industrial automation by creating an integral network to the office automation. The PC on an engineer’s desk becomes a piece of equipment used in the production process. That process is naturally accommodated by the copper-based Ethernet protocol.
This is near-end crosstalk (NEXT) test data for a Category 8.1 channel that includes an RJ45-ARJ45-RJ45 architecture. The two connector types use different approaches to controlling differential NEXT. The RJ45 uses compensation; the ARJ45 uses isolation.
The fiber has wider bandwidth and longer reach, and the optical cable is not susceptible to electromagnetic interference (EMI). (Fiber-optic transceivers are not immune to EMI, by the way.)
The latest penetrations of copper cabling Categories 7A and 8 use shielded cabling. Many installations even of earlier cable generations, 6 and 6A, used shielded cabling. The preferred cable construction is PiMF—pairs in metal foil. According to current market surveys of new installations in the European countries of Germany, Switzerland, Austria and France, PiMF is becoming the norm, approaching 90 percent of all newly installed copper cabling.
The fiber reach advantage, if examined in detail, may not be so important in some applications. Consider cases in which the cabling is in the confines of a given network such as a data center or an office building. When the distance from your equipment to the next piece of equipment is about 20 to 40 meters, the ability to transmit signal over a singlemode fiber at a 500-meter distance became moot, and it will be a waste of resources to use fiber in the application.
The same Category 8.1 channel exhibits return loss that significantly outperforms a Category 6A channel. The Category 8 standard specifies 16.0 dB return loss at 100 MHz, while Category 6A specifies 12.0 dB. At 500 MHz, Category 8 specifies 10.7 dB return loss, while Category 6A specifies 6.0 dB.
New standards for copper
The networking industry recognizes the tremendous potential of copper connectivity and continues working on development of new standards for high-speed copper interconnect. The ubiquitous presence and continued rapid growth of Ethernet is the main driver.
IEEE 802.3bq, a new internet standard published by the Institute of Electrical and Electronics Engineers, specifies the networking protocols for 25 and 40 Gbit/sec Ethernet. The standard is a response to the industry demand for higher data density traffic in data centers.
A Category 8.1 channel with an RJ45-ARJ45-RJ45 construction can connect all the existing RJ45 ports and support applications from 1- to 40-Gbit Ethernet.
The IEEE 802.3bt standard expands PoE by using all four wire pairs for transmission of power; it introduces Type 3 (60 Watt) and Type 4 (100 Watt) power sourcing equipment.
An international standard ISO/IEC 11801-99 (currently in the final stages of development) introduces Class II copper cabling up to 2000 MHz. It covers Category 8 and allows the use of alternative connectors in addition to the RJ45.
The IEC 61076-3-110 standard covers ARJ45 – Augmented RJ45. This is the first standard that expands the copper premises wiring connector to 3000 MHz.
Category 8—In comparison to Category 6A, Category 8 increases the bandwidth of copper connectivity by a factor of 4, from 500 MHz to 2000 MHz. The Category 8 standard specifies improved channel return loss performance. It is useful to note that there are two Category 8s: Category 8.1 and Category 8.2. In practice, both use the same copper cables. Category 8.2 allows other connector types in addition to the RJ45.
Data centers, where a high density of interconnects results in a large number of short transmission lines, is a natural environment for Category 8 applications. In this environment the ease of maintenance and repair, short distances between transmitter and receiver, in combination with a wider bandwidth of Category 8 make copper cabling very attractive.
RJ45 and ARJ45—Two copper connectors used for Category 8 twisted-pair channels are RJ45 and ARJ45 (Augmented RJ45). The RJ45 connector is one of the most popular connectors in the world. Network appliances, computer and data terminals have billions of RJ45 ports. The ARJ45 is a relatively new connector. Augmented RJ45 was derived from the RJ45. Presently, at 3000 MHz it is the fastest standard connector for twisted-pair cabling. ARJ45 is also used with twinaxial cables. RJ45 and ARJ45 are robust, inexpensive to make, and user-friendly connectors.
To avoid misconnections, an ARJ45 plug cannot be inserted into an RJ45 jack.
The RJ45 and ARJ45 use different approaches to control differential near-end crosstalk. RJ45 uses compensation to reduce the differential NEXT. Compensation is a method of creating the crosstalk equal in amplitude, but opposite in phase, by adding capacitive and inductive elements. ARJ45 uses isolation to avoid differential NEXT. A Faraday cage is built around each differential pair. As a result, ARJ45 is better suited for high-frequencyapplications.
Category 8 for 25- and 40-Gbit/sec Ethernet. Customers demand the connectivity that can be plugged into the existing 1-, 2.5-, 5-, and 10-Gbit/sec equipment and will support 25- and 40-Gbit/sec switches and peripherals. Such connectivity will deliver the robust transmission performance and support the all-importantautonegotiation.
Category 8 copper cabling is such connectivity. The best way to connect RJ45 ports and take advantage of ARJ45 performance is to combine RJ45 and ARJ45. That is commonly done by using a patch cord with an ARJ45 plug on one and an RJ45 plug on the other end. The IEC standard design provides means to prevent plugging an ARJ45 plug into an RJ45 port.
As depicted in the nearby figure, a Category 8.1 channel can combine the advantages of both—a common RJ45 and the high-performance ARJ45, delivering robust performance. The RJ45 plugs at the channel ends allow the use of existing equipment. The RJ45/ARJ45 channels fully supportautonegotiation.
Make no mistake, copper cabling will continue to compete with and complement fiber-optic cabling. Both optical-fiber and copper cabling networks will continue to coexist. Rumors of copper cabling’s death are at least somewhat premature. Copper cabling is not dead, and is not going to die anytime soon. Category 8 and the ARJ45 will extend the life of copper connectivity for years to come.
Yakov Belopolsky is manager of research and development with Stewart Connector. He is a member of the U.S. Technical Advisory Group (TAG) to the ISO and a member of the ISO/IEC committees on connectors and cabling. He has been awarded 81 U.S. patents and has published more than 30 technical papers.