I’m not sure how many circuit turn ups I have been a part of that just fell through because the on site technician couldn’t get a CD light on the WIC card. Moreover, I think the number of turn ups that I have been a part of that went completely as planned are fewer than 5%. The reason? Wiring. Almost 99% of the time if there is an issue with a T1, its because of the wiring. I’m going to start by explaining some of the basics of T1 wiring, discuss some relevant terminology, and pepper in some of my own experience throughout that might help your troubleshooting.
2 wire or 4 wire?
I can’t tell you how many people I have heard complaining about having a ‘2 wire’ T1. They wonder why they would pay T1 prices for what they equate to a ADSL circuit. They have a point; if it’s 2 wire, it certainly sounds like a DSL circuit. The fact of the matter is that both 2 wire (single pair) and 4 wire (dual pair) T1s do exist. The interesting part of this discussion is that regardless of whether it’s a single or dual pair circuit you’ll ALWAYS have dual pair coming out of demarc and going into your WIC card. T1 circuits traditionally use 2 pair. However, with the emergence of better standards and technologies the telco is now, in most cases, able to deliver the T1 to demarc over a single pair. How do they do this? In most cases they use a version of HDSL. HDSL stands for high bit rate digital subscriber line. Now, you might be wondering why the telco is selling you a DSL circuit and magically converting it into a T1 at demarc. The concept is, for the most part, fairly simple but it took me a very long time to fully grasp the entire picture. The basics of it is that the only single pair part of the connection is between HTU’s. HTU stands for host terminal unit and there is usually a HTUC (HTU Central Office) and a HTUR (HTU Remote Location). The span of copper between HTUs when using HDSL is single pair. Without going into too much detail, the HTUs do the conversion for us back to dual pair standard T1. Worried about some of the normal downsides of DSL like overloading or over-subscription? No worries, HDSL is just like T1s. They are either in sync (full 1.544 up and down) or out of sync (not working). Next time you have a T1 installed, ask the telco tech if you can watch him do his cross connects. He’ll probably install a T1 card in an available cabinet and then run a single pair of wire off ‘his’ (the telco’s) blocks into the cabinet. Then coming off of the T1 cabinet he’ll run dual pair over to, what is in most cases, a 66 block where he’ll loop and tag (more to come on what that means below). The point that I want you to get out of this section is that T1’s can be delivered 2 wire or 4 wire but any wiring you’ll be involved in (unless you are the telco) is going to be standard 4 wire T1 cabling.
Now that I have thoroughly confused you, let’s go over some of the terminology for wiring and T1’s
Demarc (Demarcation Point) – Think of it as the end of the telco’s responsibility. That’s where they terminate the circuit and hand it off to you. If you have issues with the circuit the only thing they can do is test from demarc back towards them. The rest of the wiring (between demarc and your WIC) is your problem.
CO (Central Office) – This is where all of the telco’s equipment is located. The circuits that are terminated at your end in your buildings demarc are terminated at the telco’s end in the CO. COs are sometimes also called ‘local exchanges’. Think of it as a building where lots of connections occur.
TLM (The Last Mile) – Regardless of the fact that it isn’t usually a mile any more, the term refers to the connection between the CO and the client’s demarc. Most of the time it’s significantly longer than a mile and with optics it can be much longer. Sometimes it’s referred to as the first mile or local loop.
ILEC (Incumbent Local Exchange Carriers) and CLEC (Competitive Local Exchange Carriers) – I only bring up ILEC and CLEC because you’ll most likely hear about them somewhere along the line. There’s a very LONG story here about AT&T being the only phone company in the US and during the 80’s and 90’s the government made AT&T basically become seven smaller companies. These companies are called ILECs. Basically, the big seven own the majority of the copper laid all over the US. That being said, it’s hard for smaller ISP’s and circuit providers to get a hold in the market when they don’t have any of the lines. Long story short is that the government forces ILECs to let CLECs rent equipment in their local exchanges so that they can provide services as well. Hence, sometimes you buy a circuit from a smaller company but still need to deal with AT&T (the ILEC) during provisioning. ILEC in most cases is the big company who owns the equipment and CLEC is in most cases more a less a smaller sort of reseller.
UTP (Unshielded Twisted Pair) – You should know this one but for the sake of being clear I’ll define it again. Think standard Ethernet cable. The only shield is the rubber wrapped around all 4 pairs in a standard Cat5 cable.
STP (Shielded Twisted Pair) – The real difference here is that there is either a cable or pair shield in the cable. Some say that true STP is only the cable that has each pair shielded independently while others will say that STP just has to have a single shield around all of the cables.
ISTP (Independent Shielded Twisted Pair) – ISTP cable that has shielding around each pair. Sometimes referred to as ‘T1 wire’. I’ve see it sold in 2 and 4 wire configurations. As mentioned in the STP definition, some people will call ISTP STP. I make a point of saying ISTP when I’m referring to a cable that has each pair shielded separately. I think in some cables you can still get grounding wires on each pairs shield as well. If you do, only ground on side of the run.
HTUC (HDSL Terminal Unit Central Office) – As defined above, the equipment on the telco’s side of a T1 that does the HDSL magic.
HTUR (HDSL Terminal Unit Remote Location) – As define above, the equipment on your side of the T1. Sometimes also referred to as a T1 card or Line card. After the HTUR, you’ll have dual pair standard T1 wiring to work with.
Ring and Tip – If you actually understand the reference and are under the age of 30 I’ll buy you a beer. Ring and tip comes from the old days where telephone connections were made at switchboards by plugging a cord with larger head phone-like jacks between two callers to make the connection. The ‘tip’ would be the wire that was on the tip of jack and the ‘ring’ would be the wire connected to the ring around the jack. If I recall correctly, the ring was used to sync the call and actually sent the signal to make the other phone ring (funny huh?). The tip connection was actually used for the phone call. Ironically, these terms are still around today and every T1 cabling guy I have ever met still uses them.
Loopback – Much like the loopback on a network interface a circuit loopback crosses the receive and transmit wires to cause a physical loop. You can make a loop back jack yourself (we’ll talk about that more below) for testing purposes or most WIC cards let you set loopback in software for testing. Most T1 testers prefer the physical loop back over the software.
RJ48X (Smart Jack) – A T1 should be terminated in a RJ48 junction box. Never heard of RJ48? Thought I meant to type RJ45? Wrong! RJ stands for ‘registered jack’ and while some people think that there is a physical difference between RJ numbers there really isn’t. Most of this confusion comes from the difference between RJ11 (having a 6 position connector) and RJ45 (having a 8 position connector). So why does RJ48 use a different type of connector? Because RJ more or less defines the type of wire used and their associated pinouts within the connector. For instance, RJ45 specifies UTP and a send/receive pins of 1236. RJ48 specifies STP and send/receive pins of 1245. Sometimes RJ48 connectors have an extra side groove so they won’t fit in a standard RJ45 jack. That being said, I still see tons of people who use RJ45 jacks and connectors to do T1s. You just need to properly understand the wiring patterns to make it work. Back to what I was initially talking about though, an RJ48x jack has what are called grounding (or shorting) bars in it. When you plug a connector into the jack the bars move out of the way allowing the connection to flow through the wire. When you remove the connector the bars come back into place which ground pin 1 to 4 and 2 to 5 causing a loopback. Loopbacks are a big part of T1 testing so its a good thing to have a jack that has this characteristic.
CPE (Customer Premise Equipment) – Basically the router with the WIC that you are terminating the T1 in.
WIC (WAN Interface Card) – The card you terminate the T1 in. Usually has a built in CSU/DSU
CD (Carrier Detect) – On a Cisco WIC card its the green light that comes on to tell you that the WIC has sync. You need a CD light on the WIC before you proceed with turning up a circuit.
Extending the Dmarc
Now that we have some definitions lets talk about the actual wiring. Since we are past the single and dual pair debate we can talk about straight T1 wiring which is ALWAYS dual pair. In most cases when you order a T1 circuit a telco tech will come out to your building demarc and connect the circuit up to some point of demarcation. That being said demarcation can mean a lot of different things. Lots of demarc rooms have telco equipment on one side of the room and then a wall of blocks and biscuit jacks that have all of the circuits on them. A typical T1 install involves installing the T1 card and patching the connection to it. Then they’ll patch the circuit from the T1 card either to a 66 block or a RJ48x jack. Most of the time in a large building you’ll see the circuit being terminated on 66 blocks where the tech will ‘loop and tag’. Since there isn’t anything in line to automatically loop the circuit for testing the telco tech will usually put a pair of wire between the send and receive pairs on the 66 block to physically loop the circuit. They’ll usually leave excess wire on the loop and tie the circuit ID tag onto the loop wire. Then when you come into the picture you can pull off the loop wire and extended the circuit from the 66 block to house pairs going to your suite where you can terminate in a RJ48x jack. Now is an appropriate time to talk about the actual wire used when extending the demarc. If you don’t already have house pairs running to where you want to terminate the circuit and you aren’t low voltage certified call a wiring contractor. There are all kinds of things to take into account when extending a circuit such as wire type, extension length, interference, binder groups, etc.. If you decide to do it yourself, I suggest ISTP (T1 wire) and if you can’t get that, run two Cat5 cables with one pair in each wire. I’m not even going to touch on the subject of how long the extension can be. I’ve heard any where from 10 to 100 feet with Cat5 alone.
T1 Pinouts and Wiring
It’s pretty straight forward
Pin – Description
1 – RX Ring
2 – RX Tip
4 – TX Ring
5 – TX Tip
Now what isn’t as straightforward, for some reason, is the big picture view of this. For some reason the first thing that any on-site tech does when he doesn’t get a CD light on the WIC card during turn up is blame the wiring guy. What you SHOULD do is verify the wiring into the jack and the cable you are using to connect the WIC to the jack. 95% of the time the wiring guy got it right, he just didn’t tell you how to finish the wiring. The most common issue is cross over. It seems to me that some cabling technicians assume you’ll be using a T1 cross over cable and others assume a straight through will be used. If a straight through is going to be used, you can simply use a normal Cat5 patch cable as the pinouts on both sides of the wire will be identical. If you need to cross over before you get to the WIC a standard Cat5 cross over cable wont work. Why wont it work? Because a T1 uses pins 1,2,4,5 and a Cat5 cable uses pins 1,2,3,6. Below I show you what the connectors for each type of cross over would look like is using the full 4 pairs. The concept here is that you simply need to cross receive and send pairs so that send on one end goes to receive on the other.
Cat5 Cross over Cable
T1 Cross over Cable
Do you see what the issue would be if we were using a standard Cat5 cross over cable? Pins 1 and 2 cross correctly but 4 and 5 don’t cross at all in the Cat5 cross over. Just keep this in mind when you are troubleshooting the wiring. This what the end result of the T1 wiring has to be.
Telco TX Ring – CPE RX ring
Telco Rx ring – CPE TX ring
Telco TX tip – CPE RX tip
Telco RX tip – CPE TX tip
I seem to be throwing out a lot of percentages so I might as well stick with it. I think about 60% of the time when I don’t get a CD light on a WIC card I can use the same cable and simply swap the pairs in the RJ48X jack. If you don’t like messing with jacks you can simply carry both standard T1 patch cables as well as T1 cross over cables in your bag. One of them ought to make the connection work.
T1 Loop Back Jacks
Another area of disagreement between cabling professionals is what pairs need to be looped to cause a loopback. The loop back jack I carry in my bag loops pin 1 to pin 4 and pin 2 to pin 5. That sort of configuration loops both pairs. Others have told me that you only need to loop pins 1 and 4 since those are the only ones used for syncing the connection. I think I’ll stick with looping both pairs to be safe. You can use any old RJ45 jack and some cross connect to make the loop back jack.
I hope this was more helpful than confusing. I tried not to go into so much detail about the telco side of things as most of you will only be concerned with the T1 after demarcation. And I realize that I didn’t get into as much detail about some items. Unfortunately there are entire books written about T1s so a blog entry can really only graze the surface. If you read this and disagree with anything I’ve said, I’d love to hear from you. All of this is based purely on my experiences with T1 circuits. Any additional input would be appreciated!