If the physical layer isn’t right, the network will have big problems. In this video, you’ll learn about fiber mismatches, cable categories, crosstalk, EMI, attenuation, and proper termination.
When you’re working with fiber optics, you have to make sure that you’re using exactly the right kind of fiber optic connection. This is because the light works very differently in one type of fiber versus the other. For example, with multimode fiber, the fiber can use multiple modes to be able to find its way to the other end of the connection. This means the light will use different paths to find its way from one end of the fiber to the other. This is very different than using single mode fiber.
With single mode fiber, the light source uses a single path to go all the way through the fiber. And you can see with single mode fiber that it is a much smaller fiber diameter than it is with multimode fiber. And it’s difficult to tell the size of the fiber core and cladding unless you look very carefully, and even then, it becomes very difficult. For example, in multimode fiber, you often find two different sizes. One that is a 50 micron and one that’s a little bit larger that is a 62 and 1/2 micron.
And there’s also cladding around that that makes the total size of the fiber 125 microns, which is identical if you were to hold them in your hand. Single mode fiber is very different. You can see it’s around 9 microns, but still, we have cladding around the outside that brings it up to 125 microns. So you could be holding one piece of fiber, but that fiber could be one of three different standards that are commonly used on your network.
If you do mix fibers where you plug in a single mode fiber where multimode should be or multimode fiber where single mode should be, you should see signal errors and have a difficult time communicating across that link. Often, the fiber will have different colors to help designate what size it happens to be, but you cannot always rely on those colors. So make sure you document all of the fibers that you receive. There will often be writing on the outside of the fiber that will tell you what type of fiber optic is inside and what sizes it happens to be.
This will help when you’re troubleshooting to verify exactly what type of fiber you might be using. Not only can we mix fiber, but we can also mix copper. There are many different types of copper cables, and they all look very similar from the outside. The way that these cables are constructed is very standardized by the Telecommunications Industry Association or TIA. The TIA sets standards for how certain types of cables should be manufactured, and they have tests that confirm that the cable that was manufactured meets their strict guidelines.
Once cables are manufactured and tested, they will qualify as a certain category of cable. And that’s how we reference different types of cable and different constructions of those cable types. So if you were to look at category 5 cable versus category 6 versus category 7, you will not only see that they are constructed very differently, but they’re able to support different types of signals across that copper. So if you look at the IEEE standards for ethernet networks, it will specify the minimum cable type that could be used for that particular standard.
So if you’re reading through the 1000BASE-T standard, it states that the minimum cable you can use for this is a category 5 cable. But if you read through the 10GBASE-T standard, you’ll see that the minimum cable is a category 6 and a category 6A depending on how far you need to transmit that signal. These IEEE standards will often specify the speed of the signal moving through that copper cable, so you’ll see a bandwidth rating or speed rating associated with that.
This is the theoretical maximum data rate that we can get across that particular piece of copper, and it’s often measured in a number of bits per second. Sometimes you’ll hear bandwidth and throughput mentioned in similar ways, but throughput is different than bandwidth. Bandwidth is our theoretical maximum data rate over that link, whereas throughput is the total amount of data that we’re putting over that link in a particular time frame. This is often measured also in bits per second, but sometimes it’s also measured in bytes per second.
So you may have to do a bit of a calculation to be able to translate between both of those. And with all of these cable standards, there is a maximum length that we’re able to extend that cable and still maintain the standard amount of signal. We need to know the maximum distances, and they are usually part of the IEEE standard. For example, the IEEE standard for 1000-BASE-T specifies a minimum category type of category 5, and you can extend that particular link to 100 meters of distance.
If you’re using unshielded twisted pair and you are running 10GBASE-T signals over that cable, you need to use category 6 and that unshielded cable will be able to run 55 meters and still comply with the standard. If you’re installing cable into an existing building or you’re building out a new network run, it’s always useful to test the cable that you’re installing. You need to validate the cable that it is not only installed properly, but that it can support the signals that you plan to put over that particular link.
It’s common to use a cable tester that will perform tests that are very similar to the ones that are used to qualify a particular category of cable. So you can install your cable and run your tests, and your tester will tell you what category of cable best matches the cable that you’ve recently installed. If you use the wrong category of cable, you’ll see errors on the link and that will certainly cause a slowdown. These physical errors are often caused by signal attenuation.
You might have a loss of signal completely, or you may be receiving CRC errors. UTP stands for unshielded twisted pair cable. This is a good example of an unshielded twisted pair cable. You can see that there are four pairs of wires inside of this cable and none of those wires have any type of shield around them. You can also see that each of these pairs is twisted, and that’s how we get the term unshielded twisted pair. There’s also a type of cable that is more resistant to electrical interference.
This is a shielded twisted pair, or STP. Shielded twisted pair has a shield either around the entire cable or around individual pairs inside of the cable to help prevent interference. And you’ll also see that shielded twisted pair has a grounding wire so you can ground that shield. Here’s a shielded twisted pair cable that not only has a braid around the entire set of cables, but each individual pair of the cables is shielded. You would need to pull those back to be able to put it into an RJ45 connector or to punch it down to a punch down block.
The wires inside of those cables are very close to each other, and sending an electrical signal through one pair of wires could affect the signal that’s being transmitted or received on another pair of wires. We refer to that interference as crosstalk or XT. This crosstalk value describes how much of that signal is leaking into other wires, and we can often measure exactly how much of this signal is being transferred over from one wire to another. One of these crosstalk measurements is known as NEXT.
This is near-end crosstalk. This means that we are measuring how much of this signal is leaking into those other wires at the closest end next to the testing device. This is usually the strongest part of the signal, and often that is the first crosstalk value we’d like to see. But we might also want to see crosstalk values at the far end of the cable. We refer to that as far end crosstalk. These two measurements can give you an idea of just how much signal happens to be shared between those wires as it’s going from one end of the cable to the other.
You might also find that there’s crosstalk in your cable that’s being transmitted from another cable. We refer to this as alien crosstalk because that’s how you would describe interference that’s coming from other cables. And then you have the attenuation to crosstalk ratio, or ACR. This allows us to compare two different values. One of those values is attenuation, or it’s our signal loss that we see when we plug into a connection. The other value is near-end crosstalk.
So we’re able to compare how much signal we’re losing versus how much signal is leaked between the different wires. We can often see this as a signal to noise ratio, where we have a total amount of signal that we’re putting on the wire, and then we have interference that’s being caused by the near and crosstalk. We can then see how much good signal is getting through the wire versus how much is noise that may be disrupting that communication. We would obviously like that ratio to be very large. If we had a 10 to 1 signal to noise ratio, our signal would be 10 times stronger than the noise.
But if the ratio was a 1 to 1, your signal and your noise would be exactly the same signal strength on that wire. If you’re seeing a lot of cross talk, it may be related to the punchdown or the connectors that you’re putting on the end of the cable. It might be worth reconnecting that particular cable. You also want to be sure to maintain as many twists as you can in this cable. You obviously need to untwist a little bit from the end to either put it into a punchdown block or into an RJ45 connector, but you want to maintain as much of that twist as possible.
Category 6A has increased the cable size, and many category 6A cables will have the spacer in between that physically separates the different wires in order to minimize the amount of crosstalk. And of course, if you’re wondering if a cable is working the way it should or you’re installing new cables into a facility, it always makes sense to perform tests on that cable to see if it passes a particular level of cable category. Earlier we described electrical interference and how a shielded twisted pair cable would resist that type of interference.
It’s always useful to minimize the amount of EMI or electromagnetic interference that we would have both on an unshielded and a shielded cable. You always want to be careful to protect the wire that is inside of these cables, especially if it is a shielded cable, and maintain that shield all the way through that particular cable run. If there is a break in that shield, there’s a possibility that electromagnetic interference will get into those wires.
There’s also a minimum bend radius that you can use, and you’ll want to check what the documentation for your cable to see just how much of a bend you can make to that cable without causing any damage. And of course, you never want to use staples on these cables, and you want to limit the number of plastic cable ties. A good best practice is to use some type of removable cable tie, such as Velcro, so that you don’t crimp any of those wires inside of the cable.
It’s also useful if you avoid any of these external sources of interference, such as power cords, fluorescent lights, any type of electrical system or generator, and any fire prevention components. All of these can put out electromagnetic interference, and that interference can find its way into your network cables. And if you’re wondering if your cable is coming close to one of those sources of interference, you could always run a cable test and see what your signal to noise ratio might be over that link.
Attenuation describes a loss of signal as it is sent over a particular distance. Attenuation is something you’ll find both in copper cables and in fiber optic connections. As you are moving down that fiber or moving through that copper, we are losing some of the strength of that signal as we are moving farther and farther away from the source. You’ve probably seen this on wireless networks. As you move farther and farther away from the antennas, the attenuation of the signal becomes stronger and you start to lose the signal as you move farther away.
This is one of the reasons that we see distance limitations when we start referencing IEEE standards and the different categories of cable that we can use. If you’ve ever worked with a cable plant or you’ve been terminating cables into a punch down block or putting RJ45 connectors on, you know how challenging it is to have the perfect type of cable termination. And often you can make or break a network configuration just by the terminations that you’re doing with the cables themselves.
Part of the problem is that many of these connectors look very similar to each other, and yet there may be different specifications depending on the different connector type. You also may find that different cable installers may have a different quality to the punchdowns that they’re doing. And it’s true that connectors that are made by one individual may work better than connectors that were installed by a different individual. It’s always a good idea to find somebody who is well-versed in the standards for installing network cables versus other types of cables. In most cable runs, we’re trying to match the pins on each side of the cable.
For example, pin 1 on one end of the cable should be matched to pin 1 on the other side of the cable. Pin 2 should be matched to pin 2, pin 3 to pin 3, and so on. If you have a mismatch between these different pins or you miss connecting the pins together, you may find that the cable speed drops to a different standard. For example, you may be expecting to get a 1 gigabit connection and it’s only able to support a 100 megabit connection. It’s very easy to make one little mistake where you might connect pin 1 to pin 2 and pin 2 to pin 1, and when you plug in that ethernet cable, you get no signal at all.
Sometimes you may find that a cable has crossed pairs, where pin 1 might connect to pin 3 and pin 3 might connect to pin 1. This is a half crossed view where you have four of the wires are crossed and the other four wires are straight through. This is sometimes a mistake that can be made at the ends of the cable when you’re crimping on the RJ45. It could also be a problem at the punchdown block, so you’ll want to check every connection along that path to see where this cross might have occurred. This will be something that’s very easy to see when you plug in your cable tester, which is another good reason to always test a cable that you’ve installed.
And it will show you that instead of pin 1 connecting to pin 1 and pin 2 connecting to pin 2, you have a very different cross set of cables. Some ethernet chipsets, however, will recognize when a cable cross has occurred and it will reverse that cross in the electronics of the ethernet card itself, effectively turning it back into a straight through cable. This is something that’s built around the Auto-MDIX capability, but it’s not always turned on and it’s not something that will always work across every ethernet chipset. So it’s always a best practice to install a properly terminated cable rather than relying on Auto-MDIX.