Articles
  • Register

Structured Cabling Installation Practices - Part Four Certification and Testing

Here in part four and the final article of our series, we cover the required Performance Testing for Twisted Pair Cabling within a Structured Cabling System.

What is a 'certified' cable? This term is used by vendors or testers and it indicates that the cabling has been tested to the minimum specifications of ANSI/TIA standards and should work with any network designed to operate on a Cat 5e/6/6a link. As we covered in Part One of our Installation series, ANSI (American National Standards Institute) and TIA (Telecommunications Industry Association) have joined forces, and in the latest generic Telecommunications Cabling Standard - ANSI/TIA-568-D - they provide a set of Standards that enable the planning and installation of a Structured Cabling System for commercial buildings.


Wire-Mapping 

​One testing parameter used is called 'wire-mapping' which simply means that each wire is hooked up correctly and that the two ends have been terminated pin for pin, i.e. that pin 1 at the patch panel goes to pin 1 at the outlet, pin 2 goes to pin 2 etc. The wire-map also checks for continuity, shorts, crossed pairs, reversed pairs and split pairs. Each pair must be connected to the correct pins at the plugs and jacks, with good contacts in the terminations.

Most of the failures are simple enough to understand, like reversed wires in a pair, crossed pairs, opens or shorts. One possible failure, crossed pairs, is caused when both wires of a pair are crossed at one termination. The usual cause of a crossed pair is a 568A termination on one end and a 568B on the other. The most difficult wire-map problem is a split pair, when one wire on each pair is reversed on both ends. It causes the signal to be sent on one wire each of two pairs. The usual DC wire-map will pass, but crosstalk will fail. It takes a more sophisticated wire-mapper or Cat 5e/6 tester to find a split pair, as some wire-mappers which use only DC tests do not check crosstalk. In our experience, a split pair is usually caused by someone using punchdown color codes on jacks which splits the pairs.

 Typical Wire-Mapping Errors


Cable Length 

Since 568 cables must be less than 90 meters (296 feet) in the link or 100 meters in the channel (328 feet), length must be tested. This is done with a "Time Domain Reflectometer" which is a fancy term for cable "radar". The tester sends out a pulse, waits for an "echo" from the far end and measures the time it took for the trip. Knowing the speed in the cable, it calculates the length. All cable certification testers include a TDR to measure length. If you have a short or open, the TDR will also tell you where the problem is, making it a great tool for troubleshooting problems. Length is measured by a cable tester using Velocity of Propagation (Vp), which is the speed that a signal travels through a transmission medium measured as a percent of the speed of light (186,000 miles/sec.). 


The Decibel (dB) 

The decibel (dB) is considered the baseline by which all telecommunications designers look to compare cabling system performance. But, what is a dB? And, what performance advantage does a margin of a few decibels really offer? The answer can be found by looking at the origins of the terminology.

First used for measuring the intensity of sound, the decibel was named after Alexander Graham Bell. A decibel is a convenient way for engineers to describe the input to output ratios of either power or voltages. Better crosstalk loss (NEXT, FEXT, and ELFEXT) and return loss performance is specified by a larger performance limit (in decibels) because less signal voltage is coupled or reflected. Better attenuation performance (described below) is specified by a smaller performance limit (in decibels) because less signal voltage is lost or attenuated.

The simplest way to examine how the decibel function operates is to assume a reference voltage of 1. Substituting 1 volt into the decibel function and solving for the corresponding decibel that relates to half of the signal strength (0.5 volts) demonstrates that an improvement in performance by 6 dB results in a reduction of signal strength by one-half. This means that:

 
Insertion Loss (Attenuation)

Signal transmissions over long distances are subject to attenuation, which is a loss of signal strength or amplitude. Attenuation is also caused by broken or damaged cables. Attenuation is the main reason why networks have various cable-length restrictions. If a signal becomes too weak, the receiving equipment will interpret it incorrectly or not at all. This causes errors, which require re-transmission, and loss of performance.

The illustration to the right shows the weakening of signal due to attenuation. This test requires a tester at each end of the cable, one to send and one to receive, then one of them will calculate the loss and record it. As described earlier, attenuation is measured in dB (decibels) of signal loss. 

A typical readout from a Fluke tester is shown in the below illustration. Note that the cable is tested at increasing frequencies. The margin is the difference between maximum attenuation allowed and the tested amount.

The insertion loss in a cable is largely dependent upon the gauge of wire used in constructing the pairs. 24 gauge wires will have less insertion loss than the same length 26 gauge (thinner) wires. Also, stranded cabling will have 20-50% more insertion loss than solid copper conductors. Field test equipment will report the worst value of insertion loss and margin, where the margin is the difference between the measured insertion loss and the maximum insertion loss permitted by the standard selected. Hence a margin of 4 dB is better than 1 dB.

​Troubleshooting Recommendations: Excessive length is the most common reason for failing insertion loss. Fixing links that have failed insertion loss normally involves reducing the length of the cabling by removing any slack in the cable run.

Crosstalk 

Whereas attenuation weakens the signal, capacitive and inductive reactance distort and corrupt signals. This corruption causes an unwanted signal called crosstalk. In structured cabling, crosstalk refers to electromagnetic interference from one unshielded twisted pair to another twisted pair, normally running in parallel. This interference can be minimised by the twists in the cable, with different twist rates causing each pair to be antennas, sensitive to different frequencies and hopefully not picking up the signals from its neighbouring pairs. 

Near End Crosstalk (NEXT) 

NEXT is a measure of the ability of a cable to reject crosstalk, so the higher the NEXT value, the greater the rejection of crosstalk at the local connection. It is referred to as Near End because the interference between the two signals in the cable is measured at the same end of the cable as the interfering transmitter. The NEXT value for a given cable type is expressed in decibels per feet or decibels per 1000 feet (as described above) and varies with the frequency of transmission. Cat 5e /6 testers measure crosstalk from one pair to all three other pairs and then compares it to the 568 specs, giving a pass/fail result. NEXT should be performed at both ends - the first 15 meters are the most important as the signal is strongest and there's more likelihood of crosstalk.

It is very important to keep twists as close to the terminations as possible to minimize crosstalk. ​Other reasons for a NEXT failure include: Cable ties are too tight; over-pulled cable; external noise; cable length too short, split pair, or a short cable with a far end connector problem.

Since NEXT is a measure of difference in signal strength between a disturbing pair and a disturbed pair, a larger number (less crosstalk) is more desirable than a smaller number (more crosstalk). Because NEXT varies significantly with frequency, it is important to measure it across a range of frequencies, typically 1 100 MHz. If you look at the NEXT on a 50-meter segment of twisted pair cabling, it has a characteristic "roller coaster going uphill" shape. That is, it varies up and down significantly, while generally increasing in magnitude. 

This is because twisted pair coupling becomes less effective for higher frequencies.

​Troubleshooting Recommendations: Excessive crosstalk is mainly due to poorly twisted terminations at connection points. All connections should be twisted to within 13 mm of the point of termination according to ANSI/TIA/EIA 568-B. An additional note common to all standards is that the amount of untwist should be kept to a minimum. Experience has shown that 13mm does not guarantee a PASS when field testing.


Power Sum Near End Crosstalk (PSNEXT) 

​PSNEXT is a NEXT measurement which includes the sum of crosstalk contributions from all adjacent pairs as an algebraic sum of the NEXT of the three wire pairs as they affect the fourth pair in a four-pair cable.

PSNEXT is calculated rather than measured. Cabling bandwidths in excess of 100 MHz (Cat5 cable bandwidth) make consideration of PSNEXT more important as Gigabit Ethernet through Cat6 uses all four wire pairs simultaneously and bidirectionally. The additional wire pair usage and growing bandwidth increases the need to keep NEXT in check.

​Since PSNEXT is a measure of difference in signal strength between disturbing pairs and a disturbed pair, a larger number (less crosstalk) is more desirable than a smaller number (more crosstalk). Because PSNEXT varies significantly with frequency, it is important to measure it across a range of frequencies, typically 1 – 100 MHz. 

If you look at the PSNEXT on a 50 meter segment of twisted pair cabling, it has a characteristic "roller coaster" shape. That is, it varies up and down significantly, while generally increasing in magnitude. This is because twisted pair coupling becomes less effective for higher frequencies. Typically, PSNEXT results are around 3 dB lower than the worst-case NEXT result at each end of the link.

​Troubleshooting Recommendations: PSNEXT is a calculation based on NEXT measurements. Troubleshooting for PSNEXT failures can be achieved by troubleshooting for NEXT problems.

 
Attenuation to Crosstalk Ratio - Near End (ACR-N)

Attenuation to Crosstalk ratio (ACR) expresses the difference between the NEXT and attenuation for the pair in the link under test, in effect telling you how much bigger the signal is than the background noise. Many cabling vendors sell systems based on ACR-N performance. It is an expression of 'headroom', and a key differentiator between cabling systems.



ACR = NEXT - Attenuation / 

PSACR = PSNEXT - Attenuation

High ACR and PSACR measurements
mean better performance.

​ACR is an important figure of merit for twisted pair links. 

It provides a measure of how much 'headroom' is available, or how much stronger the signal is than the background noise. Thus the greater the ACR the better.

​Troubleshooting Recommendations: Improving either NEXT or attenuation performance will improve ACR performance. This usually means troubleshooting for NEXT because the only way to significantly improve attenuation is to shorten the length of the cable.


Far End Crosstalk (FEXT)
FEXT measures the interference between two pairs of a cable measured at the far end of the cable with respect to the interfering transmitter. Surprisingly, FEXT is never reported. Why? Because FEXT is directly affected by the length of the link. Due to attenuation, FEXT on longer cables is less than FEXT on shorter cables of the same type, therefore FEXT is proportional to the received signal strength. So to make a meaningful measurement on all types of links, we need to compensate for the affects of attenuation.
Attenuation to Crosstalk Ratio - Far End (ACR-F) 

Previously called Equal Level Far End Crosstalk (ELFEXT) this is the computed ratio of the measured FEXT loss and the measured attenuation, so the equation is ACR-F = FEXT - Attenuation. Since ACR-F does not depend on length, it is used instead of FEXT to evaluate cable performance. Power Sum ACR-F is the combined crosstalk of the far end transmitter.

​Compare the results of measurements made from both ends of the link to the appropriate ISO or TIA limits. There are 12 ACR-F measurements made from each end, for a total of 23. This is because the attenuation can vary slightly depending upon which pair is energized. 

So as an example, the field tester will energize Pair 1 and listen on Pair 2 at the far end. Then it will energize Pair 2 and listen on Pair 1 at the far end.

​Troubleshooting Recommendations: The same factors that contribute to NEXT problems contribute to FEXT problems. Troubleshooting for ACR-F therefore means troubleshooting NEXT and attenuation problems, just as you would for ACR problems.


Return Loss 

Return Loss is a measure of the reflections from the cable due to variations in the impedance. These reflections can cause signal degradation, especially if the pairs are used in a full-duplex (bidirectional) mode. With 1000Base-T Gigabit Ethernet transmitting in both directions on each pair, return loss can cause big problems. 

​Some causes of Return Loss:

  • Poor installations practices like splices or taps in cable;
  • Over tight cable ties;
  • A poor bend radii. 

​All standards define the formulae to calculate the allowable return loss for each cabling link model (Channel and Permanent Link) over the frequency range. You will note that part of the limit is grey. Any measurements made in this frequency range are ignored under something called the 3 dB rule. Any measurements are ignored if insertion loss is less than 3 dB. On short links, the entire limit line can be grey, since insertion loss never reaches 3 dB. In this case, the return loss measurement is recorded as Information Only.

​Troubleshooting Recommendations: Installation practices are more important on Category 5e and 6 than they were for Category 5. Additional unnecessary untwist in terminations can add several dB of return loss in some cases.

Propogation Delay 

Propagation Delay is the length of time it takes for a signal to travel to its destination. This matters, because application standards need to know the maximum time it takes to send a signal across the furthest points of the network. By knowing the maximum time, they can design their signal collision sensing and detection circuitry to optimize network speed and performance.

When testing, one propagation delay measurement is made per pair.

Propogation Delay Skew 

​Propagation Delay Skew is a measure of the maximum difference in propagation delay across the 4 pairs in the cable. Signals must arrive at roughly the same time or the receiver cannot reconstruct and as cable constructions have different signal speeds (Nominal Velocity of Propagation, NVP) on each pair, problems arise leading to re-transmissions and network slowdown. When testing, Delay Skew must be under 50 nanoseconds for 100 meters of Cat 6 cabling. Only one result is reported.

Delay measurements are relatively straightforward. Most structured wiring standards expect a maximum horizontal delay of 570nS. If design specifications allow, higher delay can be acceptable. Since each pair in the cable has its own unique twist ratio, the delay will vary in each pair. This variance (delay skew, discussed in the next section) should not exceed 50nS on any link segment up to 100 meters. Standards require all pairs to meet the requirement. It is possible to report just the worst case pair. This will be the pair with the highest propagation delay.

​Troubleshooting Recommendations: Excessive propagation delay is caused by the cable being too long. If propagation delay fails, check to ensure that the pass/fail criteria match the design specifications. If so, the cable is too long.

Well-constructed and properly installed structured cabling should have a skew less than 50 nanoseconds (nSec) over a 100-meter link. Lower skew is better. Anything under 25 nSec is excellent. Skew between 45 and 50 nanoseconds is marginally acceptable.

​Troubleshooting Recommendations: If the skew is high, provided the intended application is a 2-pair application such as 10Base-T or token ring, the application should still perform. If one pair is much higher or lower in delay than the others, very high skew may result. 

Examine the delay results for each pair. If one pair exhibits uncharacteristically high or low delay, re-examine the installation.

Alien Crosstalk 

Alien Crosstalk is interference caused by other cables routed close to the cable of interest as opposed to signals contained in the same cable. As it is a measurement between cables, it is not impacted by outside interference eg. noise from motors, transformers, or florescent lights) in the environment.

According to the latest industry standards, ANSI/TIA-568-D.2 (Cat6A), ANSI/TIA-568-C.2-1 (Cat8) and IEC 61156-9/-10 (Cat8.1 and Cat8.2), the balanced twisted-pair telecom-unication cables have to go through different electrical tests. Among others, Alien Crosstalk is a critical test that deserves special attention. For this specific test, the cables have to be arranged in a 6-around-1 bundle configuration. The construction is defined as follows: one 4-pair disturbed cable in the centre (#1) surrounded by 6 additional 4-pair disturbing cables (#2 to #7). The "Alien Crosstalk" test measures the perturbation created by the 6 disturbing cables on the centered disturbed cable. It implies that the crosstalk between the 6 "external" cables (24 pairs) and each pair of the central cable (4 pairs) has to be measured. These measurements have to be performed from both the near and far end of the cabling under test.



Stay Up To Date

Thanks for coming to use the services provided by logging in to the DINTEK website. Press the login button.

Once you are logged into the DINTEK Website. You will have access to additional content and services depending on the level of access that has been assigned to you.

Register