On copper cable, data signals are represented by voltage levels that represent binary ones and zeros. The voltage levels are measured with respect to a reference level of zero volts at both the transmitter and the receiver. This reference level is called the signal ground. It is important that both transmitting and receiving devices refer to the same zero volt reference point. When they do, they are said to be properly grounded.
In order for the LAN to operate properly, the receiving device must be able to accurately interpret the binary ones and zeros transmitted as voltage levels. Since current Ethernet technology supports data rates of billions of bits per second, each bit must be recognized, even though duration of the bit is very small. The voltage level cannot be amplified at the receiver, nor can the bit duration be extended in order to recognize the data. This means that as much of the original signal strength must be retained, as the signal moves through the cable and passes through the connectors. In anticipation of ever-faster Ethernet protocols, new cable installations should be made with the best available cable, connectors, and interconnect devices such as punch-down blocks and patch panels.
There are two basic types of copper cable: shielded and unshielded. In shielded cable, shielding material protects the data signal from external sources of noise and from noise generated by electrical signals within the cable.
Coaxial cable is a type of shielded cable. It consists of a solid copper conductor surrounded by insulating material, and then braided conductive shielding. In LAN applications, the braided shielding is electrically grounded to protect the inner conductor from external electrical noise. The shielding also helps eliminate signal loss by keeping the transmitted signal confined to the cable. This helps make coaxial cable less noisy than other types of copper cabling, but also makes it more expensive. The need to ground the shielding and the bulky size of coaxial cable make it more difficult to install than other copper cabling.
There are two types of twisted-pair cable: shielded twisted-pair (STP) and unshielded twisted pair (UTP).
STP cable contains an outer conductive shield that is electrically grounded to insulate the signals from external electrical noise. STP also uses inner foil shields to protect each wire pair from noise generated by the other pairs. STP cable is sometimes called screened twisted pair (ScTP). STP cable is more expensive, more difficult to install, and less frequently used than UTP. UTP contains no shielding and is more susceptible to external noise but is the most frequently used because it is inexpensive and easier to install.
Fiber optic cable is used to transmit data signals by increasing and decreasing the intensity of light to represent binary ones and zeros. The strength of a light signal does not diminish like the strength of an electrical signal does over an identical run length. Optical signals are not affected by electrical noise, and optical fiber does not need to be grounded. Therefore, optical fiber is often used between buildings and between floors within the building. As costs decrease and demand for speed increases, optical fiber may become a more commonly used LAN media.
Attenuation and insertion loss on copper media
| Attenuation is the decrease in signal amplitude over the length of a link. Long cable lengths and high signal frequencies contribute to greater signal attenuation. For this reason, attenuation on a cable is measured by a cable tester using the highest frequencies that the cable is rated to support. Attenuation is expressed in decibels (dB) using negative numbers. Smaller negative dB values are an indication of better link performance. There are several factors that contribute to attenuation. The resistance of the copper cable converts some of the electrical energy of the signal to heat. Signal energy is also lost when it leaks through the insulation of the cable and by impedance caused by defective connectors. Impedance is a measurement of the resistance of the cable to alternating current (AC) and is measured in ohms. The normal, or characteristic, impedance of a Cat5 cable is 100 ohms. If a connector is improperly installed on Cat5, it will have a different impedance value than the cable. This is called an impedance discontinuity or an impedance mismatch. Impedance discontinuities cause attenuation because a portion of a transmitted signal will be reflected back to the transmitting device rather than continuing to the receiver, much like an echo. This effect is compounded if there are multiple discontinuities causing additional portions of the remaining signal to be reflected back to the transmitter. When this returning reflection strikes the first discontinuity, some of the signal rebounds in the direction of the original signal, creating multiple echo effects. The echoes strike the receiver at different intervals making it difficult for the receiver to accurately detect data values on the signal. This is called jitter and results in data errors. The combination of the effects of signal attenuation and impedance discontinuities on a communications link is called insertion loss. Proper network operation depends on constant characteristic impedance in all cables and connectors, with no impedance discontinuities in the entire cable system. |
Crosstalk involves the transmission of signals from one wire to a nearby wire. When voltages change on a wire, electromagnetic energy is generated. This energy radiates outward from the transmitting wire like a radio signal from a transmitter. Adjacent wires in the cable act like antennas, receiving the transmitted energy, which interferes with data on those wires. Crosstalk can also be caused by signals on separate, nearby cables. When crosstalk is caused by a signal on another cable, it is called alien crosstalk. Crosstalk is more destructive at higher transmission frequencies.
Cable testing instruments measure crosstalk by applying a test signal to one wire pair. The cable tester then measures the amplitude of the unwanted crosstalk signals induced on the other wire pairs in the cable.
Twisted-pair cable is designed to take advantage of the effects of crosstalk in order to minimize noise. In twisted-pair cable, a pair of wires is used to transmit one signal. The wire pair is twisted so that each wire experiences similar crosstalk. Because a noise signal on one wire will appear identically on the other wire, this noise be easily detected and filtered at the receiver.
Twisting one pair of wires in a cable also helps to reduce crosstalk of data or noise signals from an adjacent wire pair. Higher categories of UTP require more twists on each wire pair in the cable to minimize crosstalk at high transmission frequencies. When attaching connectors to the ends of UTP cable, untwisting of wire pairs must be kept to an absolute minimum to ensure reliable LAN communications.
Types of crosstalk
There are three distinct types of crosstalk:
Near-end crosstalk (NEXT) is computed as the ratio of voltage amplitude between the test signal and the crosstalk signal when measured from the same end of the link. This difference is expressed in a negative value of decibels (dB). Low negative numbers indicate more noise, just as low negative temperatures indicate more heat. By tradition, cable testers do not show the minus sign indicating the negative NEXT values. A NEXT reading of 30 dB (which actually indicates -30 dB) indicates less NEXT noise and a better cable than does a NEXT reading of 10 dB. NEXT needs to be measured from each pair to each other pair in a UTP link, and from both ends of the link. To shorten test times, some cable test instruments allow the user to test the NEXT performance of a link by using larger frequency step sizes than specified by the TIA/EIA standard. The resulting measurements may not comply with TIA/EIA-568-B, and may overlook link faults. To verify proper link performance, NEXT should be measured from both ends of the link with a high-quality test instrument. This is also a requirement for complete compliance with high-speed cable specifications. Due to attenuation, crosstalk occurring further away from the transmitter creates less noise on a cable than NEXT. Power Sum NEXT (PSNEXT) measures the cumulative effect of NEXT from all wire pairs in the cable. Some Ethernet standards such as 10BASE-T and 100BASE-TX receive data from only one wire pair in each direction. However, for newer technologies such as 1000BASE-T that receive data simultaneously from multiple pairs in the same direction, power sum measurements are very important tests. |
This is called far-end crosstalk, or FEXT. The noise caused by FEXT still travels back to the source, but it is attenuated as it returns. Thus, FEXT is not as significant a problem as NEXT. 
PSNEXT is computed for each wire pair based on the NEXT effects of the other three pairs. The combined effect of crosstalk from multiple simultaneous transmission sources can be very detrimental to the signal. TIA/EIA-568-B certification now requires this PSNEXT test. 
| The TIA/EIA-568-B standard specifies ten tests that a copper cable must pass if it will be used for modern, high-speed Ethernet LANs. All cable links should be tested to the maximum rating that applies for the category of cable being installed. The ten primary test parameters that must be verified for a cable link to meet TIA/EIA standards are:
The Ethernet standard specifies that each of the pins on an RJ-45 connector have a particular purpose. A NIC transmits signals on pins 1 and 2, and it receives signals on pins 3 and 6. The wires in UTP cable must be connected to the proper pins at each end of a cable. The wire map test insures that no open or short circuits exist on the cable. An open circuit occurs if the wire does not attach properly at the connector. A short circuit occurs if two wires are connected to each other. The wire map test also verifies that all eight wires are connected to the correct pins on both ends of the cable. There are several different wiring faults that the wire map test can detect. The reversed-pair fault occurs when a wire pair is correctly installed on one connector, but reversed on the other connector. If the orange striped wire is on pin 1 and the orange wire on pin 2 at one end, but reversed at the other end, then the cable has a reversed-pair fault. This example is shown in the graphic. A split-pair wiring fault occurs when two wires from different wire pairs are connected to the wrong pins on both ends of the cable. Look carefully at the pin numbers in the graphic to detect the wiring fault. A split pair creates two transmit or receive pairs each with two wires that are not twisted together. Transposed-pair wiring faults occur when a wire pair is connected to completely different pins at both ends. Contrast this with a reversed-pair, where the same pair of pins is used at both ends. Transposed pairs also occur when two different color codes on punchdown blocks, representing T568-A and T568-B, are used at different locations on the same link. |
| The combination of the effects of signal attenuation and impedance discontinuities on a communications link is called insertion loss. Insertion loss is measured in decibels at the far end of the cable. The TIA/EIA standard requires that a cable and its connectors pass an insertion loss test before the cable can be used as a communications link in a LAN. Crosstalk is measured in four separate tests. A cable tester measures NEXT by applying a test signal to one cable pair and measuring the amplitude of the crosstalk signals received by the other cable pairs. The NEXT value, expressed in decibels, is computed as the difference in amplitude between the test signal and the crosstalk signal measured at the same end of the cable. Remember, because the number of decibels that the tester displays is a negative number, the larger the number, the lower the NEXT on the wire pair. As previously mentioned, the PSNEXT test is actually a calculation based on combined NEXT effects. The equal-level far-end crosstalk (ELFEXT) test measures FEXT. Pair-to-pair ELFEXT is expressed in dB as the difference between the measured FEXT and the insertion loss of the wire pair whose signal is disturbed by the FEXT. ELFEXT is an important measurement in Ethernet networks using 1000BASE-T technologies. Power sum equal-level far-end crosstalk (PSELFEXT) is the combined effect of ELFEXT from all wire pairs. Return loss is a measure in decibels of reflections that are caused by the impedance discontinuities at all locations along the link. Recall that the main impact of return loss is not on loss of signal strength. The significant problem is that signal echoes caused by the reflections from the impedance discontinuities will strike the receiver at different intervals causing signal jitter. |
Propagation delay is a simple measurement of how long it takes for a signal to travel along the cable being tested. The delay in a wire pair depends on its length, twist rate, and electrical properties. Delays are measured in hundredths of nanoseconds. One nanosecond is one-billionth of a second, or 0.000000001 second. The TIA/EIA-568-B standard sets a limit for propagation delay for the various categories of UTP.
Propagation delay measurements are the basis of the cable length measurement. TIA/EIA-568-B-1 specifies that the physical length of the link shall be calculated using the wire pair with the shortest electrical delay. Testers measure the length of the wire based on the electrical delay as measured by a Time Domain Reflectometry (TDR) test, not by the physical length of the cable jacket. Since the wires inside the cable are twisted, signals actually travel farther than the physical length of the cable. When a cable tester makes a TDR measurement, it sends a pulse signal down a wire pair and measures the amount of time required for the pulse to return on the same wire pair.
The TDR test is used not only to determine length, but also to identify the distance to wiring faults such as shorts and opens. When the pulse encounters an open, short, or poor connection, all or part of the pulse energy is reflected back to the tester. This can calculate the approximate distance to the wiring fault. The approximate distance can be helpful in locating a faulty connection point along a cable run, such as a wall jack.
The propagation delays of different wire pairs in a single cable can differ slightly because of differences in the number of twists and electrical properties of each wire pair. The delay difference between pairs is called delay skew. Delay skew is a critical parameter for high-speed networks in which data is simultaneously transmitted over multiple wire pairs, such as 1000BASE-T Ethernet. If the delay skew between the pairs is too great, the bits arrive at different times and the data cannot be properly reassembled. Even though a cable link may not be intended for this type of data transmission, testing for delay skew helps ensure that the link will support future upgrades to high-speed networks.
All cable links in a LAN must pass all of the tests previously mentioned as specified in the TIA/EIA-568-B standard. These tests ensure that the cable links will function reliably at high speeds and frequencies. Cable tests should be performed when the cable is installed and afterward on a regular basis to ensure that LAN cabling meets industry standards. High quality cable test instruments should be correctly used to ensure that the tests are accurate. Test results should also be carefully documented.
Testing optical fiber
| A fiber link consists of two separate glass fibers functioning as independent data pathways. One fiber carries transmitted signals in one direction, while the second carries signals in the opposite direction. Each glass fiber is surrounded by a sheath that light cannot pass through, so there are no crosstalk problems on fiber optic cable. External electromagnetic interference or noise has no affect on fiber cabling. Attenuation does occur on fiber links, but to a lesser extent than on copper cabling. Fiber links are subject to the optical equivalent of UTP impedance discontinuities. Because noise is not an issue when transmitting on optical fiber, the main concern with a fiber link is the strength of the light signal that arrives at the receiver. If attenuation weakens the light signal at the receiver, then data errors will result. Testing fiber optic cable primarily involves shining a light down the fiber and measuring whether a sufficient amount of light reaches the receiver. On a fiber optic link, the acceptable amount of signal power loss that can occur without dropping below the requirements of the receiver must be calculated. This calculation is referred to as the optical link loss budget. A fiber test instrument checks whether the optical link loss budget has been exceeded. |
When light encounters an optical discontinuity, some of the light signal is reflected back in the opposite direction with only a fraction of the original light signal continuing down the fiber towards the receiver. This results in a reduced amount of light energy arriving at the receiver, making signal recognition difficult. Just as with UTP cable, improperly installed connectors are the main cause of light reflection and signal strength loss in optical fiber. 
On June 20, 2002, the Category 6 (or Cat 6) addition to the TIA-568 standard was published. The official title of the standard is ANSI/TIA/EIA-568-B.2-1. This new standard specifies the original set of performance parameters that need to be tested for Ethernet cabling as well as the passing scores for each of these tests. Cables certified as Cat 6 cable must pass all ten tests.
Although the Cat 6 tests are essentially the same as those specified by the Cat 5 standard, Cat 6 cable must pass the tests with higher scores to be certified. Cat6 cable must be capable of carrying frequencies up to 250 MHz and must have lower levels of crosstalk and return loss.
A quality cable tester similar to the Fluke DSP-4000 series or Fluke OMNIScanner2 can perform all the test measurements required for Cat 5, Cat 5e, and Cat 6 cable certifications of both permanent links and channel links. Figure shows the Fluke DSP-LIA013 Channel/Traffic Adapter for Cat 5e.
Cisco Systems, Inc.
0 comments:
Post a Comment