Cabling systems are the bridges that transport information between various points in a network – from source to receiver. Ensuring that this network is functioning at its optimum levels means confirming that the cables are performing well.
To do that, a solid understanding of what to look for when measuring cabling performance is essential.
The fundamental measurement parameters a certifier tests are return loss and crosstalk. These two measurements are what separate certifiers from all other types of network cable testers, including qualifiers.
Crosstalk is the measurement of a signal coupling from the intended pair of cable to another. Ideally there is zero signal transfer from one pair on to another but the reality is that it cannot be avoided. Crosstalk is sensitive to frequency and at higher frequencies, even very small disruptions in the transmission line will result in high levels of crosstalk. Crosstalk occurs along the cable and is most pronounced at connectors where the physical construction (twist) of the cable is disturbed. Certifiers measure several types of crosstalk.
Near End Crosstalk (NEXT): Signal is injected and measured at the same side of the cable. This is measured at both ends of the cable between each of the four pairs resulting in Near End NEXT and Far End NEXT test results. Six results at each end of the cable. NEXT is the key measurement for determining the component quality in workmanship of the cabling.
Far End Crosstalk (FEXT): Signal is injected at one end of the cable and measured at the other end. FEXT is not reported as a direct test result, though the data is used in calculations for other measurements such as ACR-F (Attenuation to Crosstalk Ratio – Far end).
Powersum NEXT: A calculated value to simulate the combined crosstalk of any three pairs on the fourth pair of the cable.
Alien Crosstalk: Instead of measuring signal coupled from one pair onto another in the same cable, alien crosstalk measures signal from one cable coupled onto a different cable. Alien crosstalk becomes problematic at frequencies greater than 300 MHz but is nearly eliminated when shielded or screened cables are used.
In any of its various forms, crosstalk is undesirable because it creates interference between the channels of the Ethernet transceiver at either end of the cable. The transceivers are able to operate with some amount of crosstalk, but if it exceeds their tolerance levels, bits will be dropped and data throughput of the link will suffer.
The most common sources of excessive crosstalk are excessive untwisting of the pairs during termination, poor quality connectors and cable/connectors that are not rated to the frequency at which they are being tested.
Return loss is the measurement of the signal reflected from the cabling back into the transmitting device, like an echo. High levels of return loss can create strong echoes that interfere with the transmission of the signal in one direction but it can also reduce the effective length of a cabling link/channel. Any power that is reflected from the cabling is taken away from the intended signal meaning there will be less power available to travel down a long cable.
The most common sources of excessive return loss are plugs/jacks from vendors that are not compatible with each other, multiple connectors on a channel and poor contact between the cable conductor and contact of the connector (bad termination).
The measurements described above that a certifier performs are precisely defined in the ISO and TIA standards and are not performed by any other type of tester. Unfortunately, these measurements do not come cheap.
Certifiers are measuring signals with approximately 80dB of loss from the reference signal. Using crosstalk as an example, if 1 volt is injected on to the 1,2 pair and measured on the 3,6 pair; an 80 dB loss is equal to 0.001 volts. The equipment required to measure such signals from a frequency of 100 kHz to 1,000 MHz is very difficult to design and expensive to manufacture.
The laboratory instrument that cable certifiers mimic in the field is called a vector network analyser (VNA). A typical VNA costs about $50,000-60,000 and it will only be able to connect to two pairs of a LAN cable at a time. Therefore, an RF switch is needed at an additional $30,000-40,000 bring the total cost of a laboratory system for LAN cabling to $80,000-100,000. Further, testing a single cable with a VNA can take as long as 20 minutes! A handheld cable certifier is performing the same tests as a $100,000 laboratory system and does it in seconds, not minutes for a fraction of the cost.