You should test data cabling each time you install, move, or troubleshoot a LAN-attached workstation.
Testing is a critical part of datacom installations. However, this testing is much different than the electrical power tests you're used to performing. Datacom work requires you not only to test each individual cable, but also to document the results and furnish copies of this documentation to the customer. Testing datacom cabling takes a lot more time than testing power wiring.
Testing during the installation process helps catch problems before they escalate. Even if the cable and connectors you install meet a certain category specification, you must test them. If you don't install them properly, the overall system performance can be substantially below the minimum requirements of that category definition.
To make matters worse, the effects of poor installation work may not be immediately evident if the first user is only operating at a relatively low speed, such as 10 Mbps. Many problems will only become apparent when the user tries to switch up to 100 Mbps or higher, which is the worst time for such a problem to emerge. Even when you use high-quality cable and connectors, up to 20% of all installed cable runs can deliver subcategory performance if you're not using the proper techniques.
Testing detects interactions between cable and connectors, which are not detectable independently. Some brands of cable and connectors can perform poorly when installed together in short runs (under 60 ft). Field-testing is the only way to find marginal incompatibilities.
Where do you start? As a cable installer, you should assure the quality of your materials and work by certifying all installed cable runs at the end of each job using equipment designed for testing these data links. Digital multimeters alone are inadequate for this task. While a good multimeter can confirm the signal is getting from one end of the cable to another, it can't tell you anything about the quality of the signal.
Understanding decibels. In data cabling, we express most energy and power levels, losses, or attenuation, in decibels rather than in watts. The reason is simple: Transmission calculations and measurements are almost always made as comparisons against a reference (received power compared to emitted power, energy in versus energy out, etc). Generally, we express energy levels (emission, reception, etc.) in dBm. This signifies the reference level of 0 dBm corresponds to 1 mW of power. We express power losses or gains (attenuation in a cable, loss in a connector, etc.) in dB. The unit dBu is used when measuring very low levels.
So what does the decibel measurement tell us? A 3-dB gain in power means the optical power doubled. A 6-dB gain means the power doubled, and then doubled again (four times the original power). A 3-dB loss of power means the power has been cut in half. A 6-dB loss means the power has been cut in half, then cut in half again (1/4 of the original power).
A loss of 3 dB in power is equivalent to a 50% loss. For example, 1mW of power in, and .5mW of power out. A 6-dB loss equals a 75% loss (1mW in, .25mW out).
Copper cable test equipment. The most common testing tools for copper data cabling are:
• Digital voltmeter (DVM), which measures volts.
• Digital multimeter (DMM), which measures volts, ohms, and capacitance. Some models even measure frequency.
• Time domain reflectometer (TDR), which measures cable lengths, locates impedance mismatches, and pinpoints fault locations.
• Tone generator and inductive amplifier, which traces cable pairs in walls or ceilings. The tone generator typically puts a 2 kHz audio tone on the cable under test, and the inductive amp detects and plays this through a built-in speaker.
• Wiremap tester, which checks a cable for open or short circuits, reversed pairs, crossed pairs, and split pairs.
• Noise tester (10Base-T), which measures noise levels in several frequency ranges. Various handheld cable testers can perform these tests.
Testing UTP cables. Many of the problems encountered in UTP cable plants are a result of mis-wired patch cables, jacks, and cross-connects.
Horizontal and riser distribution cables and patch cables are wired straight through (end-to-end), so pin 1 at one end connects to pin 1 at the other. Note: Crossover patch cables are an exception to this rule. Normally, jacks and cross-connects are designed so you always punch down the cable pairs in a standard order from left to right (i.e. pair 1, blue, pair 2, orange, pair 3, green, and pair 4, brown). You usually punch down the white-striped lead first, then the solid color. The jack's internal wiring connects each pair to the correct pins, according to the relative assignment scheme (i.e. EIA-568A, 568B, USOC).
One common source of problems is an installation in which USOC jacks mix with EIA-568A or 568B. When you see an installation like this, the punch downs appear to be correct, but some cables will work and others will not.
Wiremap tests will check all lines in the cable for the following errors:
• Open: Lack of continuity between pins at both ends of the cable.
• Short: Two or more lines short-circuited together.
• Crossed pair: A pair connects to different pins at each end (i.e. pair 1 connects to pins 4 and 5 at one end and pins 1 and 2 at the other).
• Reversed pair: Two lines in a pair connect to opposite pins at each end of the cable. For example, the line on pin 1 connects to pin 2 at the other end; the line on pin 2 connects to line 1. We also call this a polarity reversal or tip-and-ring reversal.
• Split pair: One line from each of two pairs connects as if it were a pair. For example, the blue and white-orange lines connect to pins 4 and 5, white-blue and orange to pins 3 and 6. The result is excessive near end crosstalk (NEXT), which wastes 10Base-T bandwidth and usually prevents 16 Mb/s token-ring from working at all.
Typically, you check cable length by using a TDR, which transmits a pulse down the cable and measures the elapsed time; until it receives a reflection of the signal from the far end of the cable. Each type of cable transmits signals at something less than the speed of light. This factor is the nominal velocity of propagation (NVP), expressed as a decimal fraction of the speed of light. (UTP has an NVP of approximately 0.59 to 0.65). From the elapsed time and the NVP, the TDR calculates the cable's length. A TDR may be a stand-alone piece of equipment or built into a handheld cable tester.
The 10Base-T standard defines limits for the voltage and number of occurrences per minute of impulse noise acceptable over several frequency ranges. Many of the handheld cable testers include the capability to test for this.
To understand NEXT, imagine yourself speaking into a telephone. You can hear the person on the other end and also hear yourself through the handset. Imagine how it would sound if your voice was amplified louder than the other person's voice. Each time you spoke you could barely hear any sound coming from the other end, due to the contrasting levels of volume. A cable with inadequate immunity to NEXT couples so much of the signal being transmitted back onto the receiving pair (or pairs) that incoming signals are unintelligible.
Using poor practices when installing cable and connecting hardware can significantly reduce your NEXT performance.
A signal traveling on a cable becomes weaker the farther it travels. Each interconnection also reduces its strength. At some point, the signal becomes too weak for the network hardware to interpret reliably. Particularly at higher frequencies (10MHz and up), UTP cable attenuates signals much sooner than does coaxial or shielded twisted-pair cable. Knowing the attenuation (and NEXT) of a link helps you determine whether it will function for a particular access method, and how much margin is available to accommodate increased losses.
Testing optical cable. After you install and terminate all fiber optic cable, you must test the whole system. We test to make sure light passes through the system with minimal signal loss. The three main types of optical testing in the field today are: Continuity testing. This is a simple visible light test. You perform this test with a simple device (basically a modified flashlight) and the naked eye, to make sure the fibers in your cables are continuous.
Power testing. You perform this test to accurately measure the quality of your optical fiber links. A calibrated light source puts infrared light into one end of the fiber, and a calibrated meter measures the light arriving at the other end. We measure this loss of light in the fiber in decibels. Most data links can handle a loss of 12 dB to 16 dB.
OTDR testing. An optical time domain reflectometer (OTDR) uses a light backscattering technique to analyze fibers. In essence, the OTDR takes a snapshot of the fiber's optical characteristics. It does this by sending a high-powered pulse into one end of the fiber and measuring the light scattered back toward the instrument. You use the OTDR to pinpoint breaks in the cable, splices, and connectors, as well as to measure light loss in the system.
In part 2 of this series, we introduced attenuation as it relates to copper voice/data cables. Attenuation in an optical fiber is a result of two factors: absorption and scattering.
The absorption of light and conversion to heat by molecules in the glass give us the term absorption. Primary absorbers are residual deposits of chemicals used in the manufacturing process to modify the characteristics of the glass. This absorption occurs at defined wavelengths (the wavelength of light signifies its color and place in the electromagnetic spectrum). This absorption is determined by the elements in the glass at wavelengths around 1000 nm (nanometers), 1400 nm, and above 1600 nm.
The largest cause of attenuation is scattering. This occurs when light collides with individual atoms in the glass. The light deflects and hits the core/cladding boundary at a high angle. It passes through the boundary, is lost, and enters the core of the fiber no more.
Chromatic dispersion and modal dispersion refer to the spreading of light pulses, until they overlap one another, and the data signal is distorted and lost. Chromatic refers to color, and modal primarily refers to the light's path. Thus, chromatic dispersion is signal distortion due to small color differences in the various rays of light, and modal dispersion is signal distortion due to the paths the various rays of light take as they pass through the fiber.
Don't set yourself up for these post-installation problems. Testing the system during installation and upon completion is simply good practice.