TECHNOLOGY BRIEF

Why 5E Cable is Unsuitable for T1 Extensions

As a manufacturer of high end data cable Quabbin is often asked the question, “Can we use a Category cable to extend our T1 service from the telephone company demarc (demarcation) point to our own internal equipment?”

Well you can, but that doesn't make it right. The purpose of this paper is to resolve two items. The first goal is to clear up the confusion about which type of cabling should be used to extend T1 lines from the phone company demarc to the non-phone company equipment internal to a business. Many knowledgeable IT professionals will tell you that specialized cable should be used for the extension of a T1 circuit from the demark point and they're right. However, when pushed for reasons why, they may not be able to support their argument well enough to prevail over those who feel that a Category 5E cable will do the job. Consequently, the second goal of this paper is to tackle the reasoning behind the choice of a true T1 extension cable and present the reader with the information required to support their position.

Historical Background

First, a little history is in order. T1 is not a cable, although there are T1 cables. T1 is actually a high speed digital network developed by AT&T in 1957. This network scheme was implemented in the 1960's to support long haul digitized (digital) voice signals in an effort to upgrade and replace the full analog telephone system of the time. Just think about the number of phones that were coming online at the time and the rapid increase in the amount of phone traffic that needed to be handled by AT&T. As necessity is the mother of invention, we can certainly say that AT&T had the need. The T-carrier system they developed needed to be robust and it worked so well it's still in use today.

As we entered the 1980's, it became apparent the demand from subscribers for bandwidth greater than 56K was becoming a reality. A high speed digital network was needed and what better tool to use than one already field proven that didn't need to be re-invented? The extension of the T-carrier system into the subscriber network was a natural.

When T1 connectivity began to migrate outward from the central office, it became necessary to implement specifications to maintain the integrity of the T-carrier system. Rather than let the world run wild with their new friend the t-carrier system; ANSI accepted the documentation task and by the late 80's had developed and published a number of specifications relating to T1 interfaces. From the perspective of installation, ANSI T1.403 is of particular relevance.

The content of T1.403 is aimed primarily at the signal between the demarc point and the customer premise equipment. Before we get involved in what the specification says, let's examine what the telephone company would recommend for cabling if you asked them at the time. AT&T would recommend ABAM cables which are configured with heavily shielded pairs and usually heavier gauge copper conductors than a typical Ethernet cable. One problem with ABAM cables was the requirement for bulky connectors that weren't readily available. As more electronic equipment for the customer side of the demarc emerged, an industry push toward the familiar and easier to source modular connector began. Aside from the modular connectors ease of use there was another very important benefit - connector population density. The smaller modular design allowed far more connections in a given amount of patch panel `real estate' than alternative options.

At the time, the cables with the best electrical properties and the ability to fit a modular connector were computer network cables. So far, no one had developed a specific T1 cable for the modular connector. The market did the best they could to work with what was available and the T1.403 specification followed suit by referring readers to what at the time was ANSI/EIA/TIA 568 which covered LAN cables. Evidence of the imperfect

solution is found in T1.403 Annex G under cable characteristics. For starters, we know that the T carrier system was designed to use shielded cable and the telephone company would recommend heavily shielded ABAM cables. We also know from the T1.403 specification that 100 ohm cables were required for T1 extensions. Unfortunately, ANSI/EIA/TIA 568 was really a specification covering Ethernet cable not T1 cable and there were some obvious mismatches present. For instance, Annex G of T1.403 states that the 568 standard “contains specifications for backbone and horizontal cables. Within each category, the following metallic cables with their characteristic impedance are referenced:

  • 100 ohm unshielded twisted pair
  • 150 ohm shielded twisted pair
  • 50 ohm coaxial

The standard then immediately states: “Only 100 ohm nominal cable is suitable for the network to customer interface.” The specification's wording points the reader directly at the unshielded option since it's the only 100 ohm option. It's very possible this could have been the foundation that to this day leads people to believe an unshielded Category cable is suitable for use in a T1 extension. The fact is, at the time this was all the market really had to work with. Today there are far superior solutions.

Technical Brief

Obviously, since the T1.403 specification directed installers to the ANSI/EIA/TIA 568 spec for category LAN cable, some people must have used it for T1 extensions. In fact it may work OK for installations limited to a single extension out to a max distance of approximately 150 meters (500 ft). However, expect the run to be susceptible to incoming EMI and also radiate its own EMI around your building. We will expand on this later. First let's discuss a certain kind of interference called cross talk.

Cross Talk is the phenomenon seen within a cable when the signal from internal wires jump over to other internal wires and present as EMI. Cross talk is most prevalent at the end nearest the transmitter where the signal is the strongest. The official name for this type of cross talk is Near End Cross Talk (NEXT). NEXT interference affects system performance by creating link errors which lead to multiple re-transmissions which `steal' network bandwidth and lead to a degradation of network speed/throughput. The problem with massive re-transmissions is that you may never know its happening. Network users will notice the performance is sub par, but they may never realize why. The most common knee jerk reaction is to upgrade the network, but to no avail. Cross talk phenomenon is well known in the network world and UTP category cables have been designed to eliminate these problems so UTP (unshielded twisted pair) Category cables can successfully be used for Ethernet. The question that begs to be answered is why is this cable `OK' on an Ethernet network and `no good' for a T1 extension? Well, it comes down to an easy answer that most installers either don't realize or have forgotten….Power. The mistaken assumption is that the data rate for T1 is lower than Ethernet, so everything will be fine. Even though the data rate for T1 is lower than Ethernet, the voltage in a T1 circuit is 300% greater. Yes, where Ethernet operates at 1 volt peak to peak, a T1 signal gets banged out at 3 volts peak to peak. The reason the T-carrier signals use the higher voltage is because they were designed to drive the signal thousands of feet between poles, not hundreds of feet on an internal network. T1 circuit designers never meant for T1 signals to run over UTP cable because they knew it would leak emissions. Now add to the equation a cable that was designed and regulatory tested for emissions at only 1 volt and you can see the problem. Further, the `power spectrum' (power at each frequency) is different in a T1 circuit than it is in an Ethernet circuit. (see graphs below) In a T1 circuit there is more power available at high frequencies because of the higher voltage. Additionally, T1 switches can be very noisy. In fact, in the early days many T1 switches could not pass FCC testing for emissions without the use of shielded cables. With this information in mind it's easy to see why more emissions will radiate out of a category cable running a T1 signal than may have initially been expected.

As a side note, the radiated frequencies are in the 1 to 100 MHz range which is often being utilized to transmit other important signals that don't need the added competition, such as radio communication.

Now consider large campus type scenarios where multiple T1 bundles are present. Not only is internal cross talk an issue within single cables, but we now have to account for something called Alien Near End Cross Talk or ANEXT. Alien cross talk is exactly what it sounds like, cross talk from an alien (external) cable run nearby the victim cable. Very early in this paper the statement was made that a category cable may be OK for a T1 extension if it was a single run. It won't go as far and it will definitely radiate like a son of a gun, but as long as it's only your neighbor's computer that gets interfered with, I guess you could say it's OK. However, if you plan to run multiple T1 lines in a bundle, forget about using a category cable, even if it's shielded. In a bundle `loose' T1 signals from the bundled cables are going to be interfering with each other regularly and nobody wants that, especially after investing the money, time and resources in a high speed T1 connection. The end goal is to get all the performance you're paying for and bundling category cables is a good way to make sure you don't get your moneys worth.

So how do you save yourself from the doom and gloom picture painted above? A NASCAR crew chief would never put re-treads on a million dollar race car; it would only be the best purpose built tires for the job. Laying out a T1 extension should follow the same basic premise. Use cable specifically designed to do the job.

Quabbin Wire & Cable Co., Inc. has designed a family of T1 cables that will ensure you get the performance you paid for from your T1 connection. Let's just take a quick look at the highlights.

Copper size

Quabbin offers T1 cable in 22 AWG which is a heavier gauge than most category cables which are usually 24 AWG. Insertion loss on a 22 AWG cable is approximately 25% lower than that measured in a 24AWG cable. This leads to more distance. Quabbin's 9720, 22 AWG cable can meet the required T1 pulse mask at a full 200 meters rather than the approximate 150 meters from a 24 AWG cable.

Shielding

Quabbin T1 cables are shielded. Some category cables are also shielded, but there's a difference. First, most category cables, if shielded at all are shielded with an overall shield that wraps the two pairs together. This may help to keep ANEXT out of the cable, but it doesn't do a thing to suppress the internal pair to pair crosstalk. Quabbin's T1 cables individually shield each pair. This configuration not only shields the pairs from each other, but also from the outside world. Further, the Quabbin cable is shielded with a heavier shielding material (more metal) than is used in category cable. The thicker shielding is meant for the higher 3 volt signals transmitted on the T1 extension and is far superior to the thin LAN cable shielding meant to protect against EMI from 1 volt signals.

Original ABAM performance

Quabbin uses a special dual layer `foamed' primary insulation which allows the cable diameter to be reduced while retaining the characteristics of a larger cable. The technique allows the construction to be heavily shielded and retain the required 100 ohm impedance without growing to an unmanageable diameter. This results in a cable with the performance of the original `AT&T recommended' ABAM cables while still easily fitting in a standard modular connector. This is truly a winning scenario, high performance, low EMI and high patch panel density via the much loved modular connector. How can you go wrong? When laying out the money to bring a true T1 line into your business, ensure that you get true T1 performance. The difference between the right cable and the wrong cable is only a couple of pennies per foot. Resist the urge to be penny wise and cable foolish.