Cat. 6 Channel Measurement Made Easy

Testing Cat. 6 equipment may be easy in theory, but without a clear understanding of how it's done, it can be a time-consuming process.

When it comes time to measure your Cat. 6 equipment, it's important to understand how to do it and do it smoothly. Three methods exist for Cat. 6 channel measurement, and choosing the right method for the job can save you time and money — two things we could all use more of.

Before we get into a discussion of how to make channel measurements, we should review exactly what a channel is. As defined by the Telecommunications Industry Association (TIA), a channel includes all the connecting hardware and cabling necessary to send signals from a server or hub in an equipment closet to terminal equipment in an office (typically a PC). Thus, the channel usually comprises an equipment cable (in an equipment closet) that runs from the hub to a patch panel, the patch panel, the horizontal cable to the telecommunications outlet (with an optional transition point), and the user patch cable connecting the telecom outlet to the PC.

However, the channel does not include the mated connection (plug and jack) at both ends of the link. The performance characteristics of this mated connection depends a great deal on the plug termination on the user patch cord. Therefore, although the mated connection is assigned to the terminal equipment, its characteristics depend greatly on the attached patch cable — which is part of the channel.

Channel measurement.

When you measure a channel, replace the terminating equipment with test equipment. Connect the plugs at the end of the equipment cables to jacks in the test equipment. This mated connection must not influence the measurement of the channel. So the test equipment must make this mated connection “invisible” during the test, while measuring all of the channel (from within 1 cm of the plug onward). The technique of measuring through a mated connection without having its performance affect the measurement of the subsequent cabling is called “connector cancellation.”

Take the following tests when making a Cat. 6 measurement:

• Wire map

• Length

• Delay and delay skew

• Attenuation

• Return loss

• NEXT

• PSNEXT

• FEXT

• ELFEXT

• PSELFEXT

Wire map, length, delay, and delay skew do not require any adjustments when tested in a channel configuration. Because the mated connection is a small and controlled variable, field test equipment can compensate for its attenuation.

PSNEXT, ELFEXT, and PSELFEXT are actually calculations, not measurements. NEXT, FEXT, and attenuation measurements ease the computation of these values. The measurements that pose a challenge for connection compensation are NEXT, FEXT, and return loss.

Channel connection cancellation.

There are three specific techniques that compensate for the connection in the NEXT, FEXT, and return-loss tests mentioned above. Let's take a closer look at each one.

Cut off the plug. When discussing channel measurement issues, it is important to look at what is done in the laboratory. The typical lab setup consists of a network analyzer connected to a set of precision baluns. This technique involves cutting the plug off the end of the cord and carefully connecting the bare wires to the balun. The process of separating the pairs from the jacket and terminating them in a balun often introduces additional crosstalk and return-loss components that were not present in the uncut cord.

It's possible to destroy the cord while testing it, so this approach is counterproductive. This is not a workable field solution, and there are clear difficulties in getting close agreement between this method and one that includes the plug in the measurement. Recent research has demonstrated significant performance differences between patch cords in the same batch, so a destructive test on one cord does not necessarily prove a batch of similar cords will perform the same.

Time-gating. Time-gating eliminates the channel connection via time domain subtraction through the following steps:

• Link-measurement of the time domain.

• Mathematical exclusion of a short section at the start of the link, including the channel connection and part of the patch cord.

• Conversion of the leftover link segment to the frequency domain.

• The performance of the leftover link segment is determined.

As stated before, the channel connection can be eliminated without getting rid of any of the attached user patch cord. It is mandatory to measure the cord for standards compliance, and recent TIA contributions have proved this cord has a significant impact on the performance of the link. For the time-gating method to meet the standards' requirements, it must subtract only the connection and not any of the attached cord.

Time-gating resolution is an inverse function of bandwidth. As the graph in Fig. 1 on page 62 (of original article) indicates, subtracting only the connection (a length of 3 cm) requires an available bandwidth of 15 GHz. If you consider the current bandwidth of today's field testers (250 MHz to 350 MHz), the corresponding time resolution is 1 m or more. This means that if a field tester uses time gating to subtract the connection, it will create a 1-m dead zone of unmeasured performance in the patch cord. This dead zone may have a significant impact on overall channel performance, so throwing it away will affect the channel performance test result. In addition, this method is not compliant with standards requirements because it does not allow measurement of the full channel.

Adaptive vector cancellation. A third alternative for making field channel measurements is frequency domain adaptive vector cancellation (AVC). With AVC, you can dynamically measure the link's response and create a customized digital vector cancellation algorithm to eliminate the effect of the connector by determining the link response and then add signals of opposite phase and magnitude at the right time for all frequency points in the spectrum.

Advantages of AVC include:

• Complete cancellation of the NEXT in the channel connection.

• Retention of the NEXT and return-loss characteristics of the attached patch cord.

• Complete compliance with the channel model.

• Easy identification of plug termination performance.

• The Cat. 6 channel adapter becomes effectively invisible to the measurement.

Remember, the jack/plug interface may not be part of the channel, but you should still care about its performance. Improper plug terminations cause most problems in patch-cord connections. Your test method must be able to determine if the crosstalk in the plug is expected for a compliant connection.

Summary.

True channel measurement at Cat. 6 levels requires a very sophisticated cancellation methodology. Even though it's acceptable in the lab, cutting off the plug is not a practical field solution. Time-gating techniques can work, given sufficient bandwidth, but they do not comply with standards at today's field-tester bandwidths. On the other hand, AVC completely and accurately cancels the NEXT and return loss of a Cat. 6 connection while leaving the patch cord as part of the measured link, regardless of the performance of the attached cabling.