10 Gbit/s Ethernet test in twisted pair cabling Environment

10GBASE-T field test and verification Overview

The transmission capability of the cabling link cannot be defined in isolation. First, the transmitter needs to encode the digital information and generate electronic signals that can be transmitted, and then send these signals through the wiring link. Secondly, the physical link must be able to transmit the transmitter signal to the receiver across the required distance, and ensure the signal quality at the receiver input end. Finally, the receiver must also meet a set of performance parameters to ensure that the signal arriving through the link meets the technical requirements, so as to ensure the success of its own functions. Therefore, it is necessary to carefully understand the requirements of the sender, the signal encoding method, the capabilities of the receiver, and the transmission characteristics of the link, and take them as a whole.

The Institute of Electrical and Electronics Engineers (IEEE) has developed and maintained the IEEE 802.3 Ethernet standard. The 10 GB/s Ethernet Working Group will establish the 802.3an standard. IEEE uses its expertise to design and determine the active electronic subsystem, and works with the telecommunications industry association (TIA) and the International Organization for Standardization (ISO) and other major cable standards organizations have jointly established performance specifications for cabling channels.

FlukeNetworks has been active in activities of TIA and ISO research committees since its establishment. Driven by Fluke Networks, a set of effective external crosstalk (AXTalk) measurement methods have been developed. External crosstalk is a key performance parameter for sending 10 GB/s Ethernet (10GbE) signaling over a twisted pair cabling link. We will explain the meaning of these parameters and test methods.

All test parameters specified in Cat 6 TIA/EIA-568-B file * are required when the 10GBASET twisted pair is verified on site. These parameters include insertion loss, ripple loss, line-to-line to-end crosstalk, integrated near-end crosstalk, line-to-line to-end crosstalk, integrated remote crosstalk, propagation delay, length, and latency time difference.

The limits for testing 10GBASE-T are the same as those for testing Cat 6 with a maximum frequency of 250 MHz, but to support 10 GB/s Ethernet technology for higher data transmission rates, the frequency range and performance requirements of these tests were extended to 500 MHz. In addition, on-site verification of 10 GB/s Ethernet must involve the external crosstalk test parameters discussed in this White Paper.

Measure the test taker's knowledge about external crosstalk and its impact on the performance of 10 Gb/s.

10GBASE-T signaling requires a wiring bandwidth of up to 500 MHz, which is significantly higher than the 100 MHz bandwidth of 1000BASE-T1 Gigabit/second Ethernet. These frequencies are very high, so an important set of Interference measurement parameters must be added to the wiring test parameters. This new set of test parameters is called external crosstalk.

Crosstalk measures signal coupling from one line to another in a twisted pair cabling link. This coupling will produce noise interference to the center online, so we don't need it. The effect of crosstalk is very similar to that of a transmission line affected by noise. The receiver cannot distinguish the signal sent from the transmitter at the other end of the link from the noise produced by crosstalk. Crosstalk is a key performance parameter in data communication activities. The degree of crosstalk deepens with the increase of the signal sending frequency, and the higher the signal frequency, the more severe the signal attenuation will become weaker when it is transmitted along the link ). It is precisely because of these two effects that we need to find such a frequency: the noise produced by crosstalk is equal to the signal sent by the transmitter. For the 100 mCat 5e channel, the frequency is usually around 120 MHz, and for the entire Cat 6 channel, it is usually around 240 MHz.

If an electronic device does not have advanced digital signal processing technology, it will not be able to achieve reliable signal transmission at this frequency or even higher. This phenomenon sets a limit for the bandwidth of twisted pair wires.

Nearby crosstalk or "NEXT") measures the crosstalk signal at the end of the wiring link that sends a test signal or interference signal. Describes the near-end crosstalk in a system with two-wire pairs. IF signal transmission is performed on multiple pairs at the same time, as in 1000 BASET and 10 GBASET, you must also consider and Test Remote crosstalk (FEXT), as shown in figure 2 ).

And the crosstalk in the description occurs between the wire pairs of the same cable in the same outer cable ). The external crosstalk is exactly the same as this phenomenon, except that the crosstalk coupling occurs between line pairs in different but adjacent two cabling links.

External crosstalk is an urgent challenge for unshielded twisted pair wires (UTP), because it is the main interference or noise source for 10 GbE applications using UTP cabling. Figure 3 shows the external crosstalk produced by a line pair in the adjacent cable.

The measurement of external crosstalk will include the external near-end crosstalk between line pairs and the external remote crosstalk. Since it is necessary to evaluate the overall impact of the numerous wires in the harness on the wires under test, which is often referred to as the victim wires, therefore, it is necessary to calculate and evaluate the comprehensive external near-end crosstalk (PSANEXT) and Integrated External remote crosstalk (PSAFEXT) in the harness ).

Test external Crosstalk

This section describes the measurement methods and testing strategies for external crosstalk. The measurement method describes the hardware and software configurations of some testing tools used to measure crosstalk between midline pairs of adjacent cables. The test strategy discusses how the cabling system is tested. In most cases, testing all possible threads is not economically feasible or unacceptable for the combination's external crosstalk.

DTX-1800CableAnayzer? The tester is mainly used to provide the bandwidth required to test the external crosstalk parameters. FlukeNetworks has developed some experimental measurement tools and techniques based on this tester. These technologies have been continuously developed and have been used by many cable manufacturers for laboratory and field tests to generate characteristics of external crosstalk parameters for them. The external crosstalk module will be provided with the DTX-1800 after the development cycle ends.

Measurement Method

When the test tester is used to test the near-end crosstalk performance of the cable, the main tester and the remote tester should be connected to the two ends of the same cable, respectively, and the two devices should use the tested link to synchronize their measurement processes. If an external near-end crosstalk is measured, use the DTX permanent link adapter to connect the DTX-1800 master device and the DTX-1800 remote device to a different cable, as shown in 4.

To allow the master and remote devices to perform the measurement process synchronously, a connection must be established between the two devices. Insert a special external crosstalk module to the back of the DTX-1800 device, where the insertion position is the same as the insertion position of the optical fiber loss test module such as the DTX-MFM. After both devices are mounted with an external crosstalk module, you can use a standard jumper to connect the two modules and establish the link required for synchronization. At this point, the remote end of the wiring link to be tested is not connected to the tester device. The open circuit at the end of the link will produce obvious reflection on the test signal. To avoid Remote launch and interfere with the measurement process and accuracy, a special end connector must be installed at the end of the two links. Figure 4 shows the connection between the DTX-1800 main tester device and the remote tester device configured for wire-to-External near-end crosstalk between the two cables in the test harness. There are 16 types of combinations of all the near-end crosstalk lines between the two cabling links. After being configured in the way shown in figure 4, The DTX-1800 will measure the external near-end Crosstalk of these 16 line pairs to the line pair combination at a frequency between 1 to 500 MHz in about 30 seconds.

The DTX tester is used to measure the connection between several cables in a harness to the wire to the external remote crosstalk. The two tester devices are now connected to different ends of the harness. The tester device must be equipped with the external crosstalk module we mentioned earlier to measure the external near-end crosstalk. You can create a synchronization path between the DTX-1800 master device and the DTX-1800 remote device using a secondary cabling link or link that is not needed during measurement. For the link involved in the test, the open end must be connected using the plug Used for testing the external near-end crosstalk. This method has achieved good results in field application. A very important performance parameter of the tester is the Noise Floor, which allows the tester to measure very accurately a very small line to the external crosstalk signal.

Comprehensive crosstalk Measurement

As mentioned above, the evaluation of the external Crosstalk of the cable bundle is not limited to the measurement of the external crosstalk coupling between individual cable pairs. During the entire network operation, all pairs in the cable bundle will transmit signals simultaneously in full duplex mode. Therefore, in the same harness, any line pair will be affected by several signal transmission lines around it. By calculating the parameters of Integrated near-end external crosstalk (PSANEXT) and Integrated Remote external crosstalk (PSAFEXT), we can know the overall impact of the surrounding cables.

The DTX-1800 master device must be connected to a laptop through a USB connection. As more and more cables in the harness are tested, software running on the laptop will import the measurement data of the wire pair to the wire pair for the near-end or remote crosstalk and calculate the test results of the comprehensive crosstalk.

External crosstalk Test Strategy

When you obtain the test results of comprehensive crosstalk for each line pair in the harness, the test will soon become very time-consuming and resource-consuming. As mentioned above, each time an additional cabling link is added to calculate the overall crosstalk, 16 line-to-line pairs are required for measurement.

If the cables shown in maintain the same relative position on the entire Link, they form a layout surrounded by six cables, which is not ideal for studying external crosstalk. This layout is regarded as the most unsatisfactory layout because it is almost impossible for seven cables in a harness to maintain such a tight layout throughout the cable bundle. In this layout, the central cabling link is defined as the affected cable. To calculate the overall external near-end Crosstalk of the four wire pairs in this affected cable, we need to measure the combination of 96 6x16 line pairs to the external near-end Crosstalk of the line pair. The total test time required to use the above method is about 180 seconds or 3 minutes, and of course it will take some time to re-connect the test wire and upload the results to the computer. This is not bad, but we only considered a broken cable in a very small harness. Most of the commonly used wiring harness not only contains seven wiring links.

On the one hand, industry standards have confirmed that external crosstalk is becoming a very important 10GBE signal transmission performance parameter in a twisted pair cabling environment. These performance parameters must be tested on site to ensure that the cabling system supports such high-speed network applications. On the other hand, we need a reasonable test time.

Guidelines for external crosstalk sampling

A set of practical external crosstalk testing strategies requires selecting representative link samples, including the most "suspicious" cabling link. The overall testing strategy consists of the following elements:

Only test the link that must meet the 10GBASE-T operation requirements, and the link must have enough length; the shorter link is less than 50 meters.) there is almost no problem. Note that the insertion loss of the link increases with the increase of length and signal frequency. This means that the signal arriving at the end of a long chain will be weaker, thus reducing the signal-to-External crosstalk interference intensity ratio. The 10 GB/s standard defines the insertion loss value, which is higher than this value. External crosstalk is a suspicious interference source.

When many links in a cable bundle need to meet the 10GBASE-T operation requirements, you can select a few long links as the affected cables and check whether they have similar PSAXtalk performance. When the cable bundle contains many cables, all the external crosstalk will become neutral to a certain extent. The result is that the PSAXtalk attributes of different affected cables tend to be roughly the same. If the PSAXTalk results of the selected damaged cable are basically consistent and the test limit value is passed, no additional test is required.

PSANEXT must be measured over a longer link of more than 50 meters or 160 feet. This is required for the six types of cables currently used. Use the interference link close to the affected link to start testing. As long as the PSANEXT results appear to be affected or changed, an interference link needs to be added.

PSAFEXT needs to be measured on a longer link bundled with a shorter link. Select a short interference link that is close to the affected link in the position of a large number of cable connections, that is, the distribution board. As long as the PSAFEXT results seem to be affected, an interference link needs to be added to the results. The frequency response of this test parameter is highly predictable. Therefore, you only need to observe a value at 100 MHz to make a decision.

Table 1-cable types that are expected to support 10 GBASE-T

Applicable cable type

IEEE for 10GBASET considers the cabling types listed in Table 1. The comments in Table 1 for certain cable types, the phrase "IEEE goals supported", are expressed as electronic device transmitters and receivers) specified requirements; these requirements, these electronic devices must be able to work properly with the specified cabling channel in a six-enclosed layout. As you can see, the 55-meter Cat 6 channel is a cabling link that fully complies with 10 GBASET requirements.

In other words, if you are designing a cabling architecture for a data center with no more than 55 meters for all links, the existing Cat 6 UTP cabling can meet your needs. Pay special attention to the installation and connection quality during installation. In addition, although the specification of the enhanced Cat 6 or Cat 6A component has not yet been determined or released, to minimize the external Crosstalk of the distribution board, it is still wise to specify and use the latest distribution board and connection hardware that is classified as Cat 6A components. Finally, it is very important to use the external crosstalk test strategy described earlier to verify the key wiring performance. Fluke DTX-1800 CableAnalyzer? Two draft 10 GBASET Cabling Standards are supported. Please refer to the Fluke Networks website to ensure that your tester runs the latest software and has the latest test standard specification Test Database ).

There are two draft standards documents that specify 10GBASET cabling performance requirements. The first document is the telecommunications system announcement (TSB) prepared by the telecommunications industry association (TIA), known as the TIA TSB-155. The second document is the "Appendix 10" (TIA/EIA-568-B.2) of the TIA/EIA-568-B.2-10 standard ). Why do two documents need to be used to define the wiring performance requirements? The first document TSB-155 is mainly for the link length limit of 10 GB/s Ethernet transmission requirements, to specify the requirements for the wiring link. For more information about how to verify the Cat 6 link in the 10 GBASET application, see this document. The second TIA document "Appendix 10") defines a new cabling class, namely, enhanced Cat 6 or Cat 6A, which performs better than Cat 6, and supports 100 M 10 GBASET channels. In terms of the performance of the external crosstalk parameters and the test parameters in the 250 MHz channel, enhanced Cat 6 provides a higher level of tolerance with the lowest 10 GB/s Ethernet requirements.

Suggestions

When using Cat6 or Cat 6A cables, you must carefully design the cabling system to minimize external crosstalk. Suggestions for reducing external crosstalk interference are mainly focused on cable bundling and cable management. Recall that the 10GBASE-T testing requirements have been extended to 500 MHz, and all on-site installation methods to provide high-quality cabling links must be carefully observed.

Only the cabling links in the same harness have an impact on the measurement of the integrated external crosstalk (PSAXtalk. Therefore, the fewer cabling links in a harness, the smaller the PSAXtalk. When the number of links in each harness is small, the test policy we discussed earlier is more effective. The maximum number of links in each harness, especially for Cat 6 cabling channels) is 12. Generally, it is recommended that there be no more than 24 links.

Do not bundle the cables in the harness too tightly; tie a bundle at about 2 feet) or longer intervals.

Most of the external near-end crosstalk occurs within the first 20 m from the test end on the link. The simulation and verification of the prototype indicate that, unless each cable runs through the entire harness in parallel, otherwise, the external near-end crosstalk produced by the cable after 20 meters from the test end will not actually have any impact on the overall PSANEXT. Therefore, the Jumper and distribution board layout of the link and the resulting rack cable management will have a serious impact on the degree of external crosstalk coupling between the links. The closer the distance, the higher the degree of external crosstalk coupling.

Many sampling suggestions or the test strategy suggestions discussed earlier clearly show that you will benefit from a proper understanding of the wiring Topology when testing the external crosstalk performance of the distribution board. It helps you identify which cables are bound to the same harness. You can design and adopt a cable naming scheme to identify the wiring harness for cabling links.

Conclusion

10 Gbit/s Ethernet is becoming a complex network technology in a twisted pair cabling environment. It requires a very good Cat 6 Cabling System and takes effective measures to reduce the impact of external crosstalk interference. In addition, it also requires an enhanced Cat 6 Cabling device. In either case, installation will play an important role in ensuring that the cabling system perfectly supports this ultra-high-speed new network technology. In-site verification channel and external crosstalk channel) is required to ensure that the wiring system supports 10 GB/s Ethernet.

Related Articles]

• 10 Gbit/S Ethernet test challenges twisted pair cabling Environment
  

• The effect of temperature on bandwidth in the Cabling System
  

• Cabling problems during Integrated Wiring Construction