10M/100M/1000M Ethernet interface measurement

    Ethernet is widely used in the fields of PC and data communication. Since the introduction of the 10Base-T standard in 1990, the technology of Ethernet has developed rapidly. Currently, 10M/100M/1000M Ethernet based on twisted pair media is popular, and 10G Ethernet technology is also gradually being applied. In order to ensure the interoperability between different Ethernet devices, it is necessary to perform corresponding consistency tests according to the specifications. The standards based on the test are mainly the corresponding chapters in IEEE802.3 and ANSI X3.263-1995. According to different signal rates and rise times, the bandwidth of the oscilloscope and probe required is also different. For the test of 10Base-T/100Base-Tx/1000Base-T, a bandwidth of 1GHz is required. For the test of 10G Ethernet, there are many standards, such as 10GBase-CX, 10GBase-T, 10GBase-S, etc. Some are electrical interfaces, some are optical interfaces, and the signal rates of different interfaces are also different. The testing of 10GBase-CX, XAUI, and 10GBase-T requires a real-time oscilloscope with at least 8G bandwidth. The testing of optical interfaces such as 10GBase-S requires a sampling oscilloscope with corresponding bandwidth according to different rates.

10M/100M/1000M Ethernet interface testing

    For 10M/100M/1000M Ethernet signal testing, you can choose Agilent's 9000 or 90000 series oscilloscope. To test the Ethernet interface, an oscilloscope alone is not enough. In order to conveniently analyze the Ethernet signal, a test fixture and test software are also required. The purpose of the test fixture is to lead out the Ethernet signal and provide a standard test interface to facilitate testing. The model of the test fixture is N5395B, as shown in Figure 1.

Figure 1 N5395B Ethernet test fixture

    The N5395B test fixture is divided into different areas for 10Base-T/100Base-Tx/1000Base-T measurements. There is also a dedicated area for connecting a network analyzer to measure return loss. The short cable that comes with the fixture can connect the fixture and the device under test, and the small board that comes with it is used to calibrate the network analyzer when measuring return loss.

    IEEE802.3 specifies many parameters of Ethernet signals. For the electrical parameters of 10Base-T/100Base-Tx/1000Base-T, you can refer to Sections 14, 25, and 40 of the IEEE802.3 specification. Without the help of corresponding software, it is a very tedious, time-consuming and labor-intensive task to measure these parameters completely manually. In order to facilitate users to complete the measurement of Ethernet signals, Agilent provides Ethernet consistency test software N5392A on the 8000/90000 series Infiniium series oscilloscopes. Figure 2 shows the test items provided by the N5392A Ethernet consistency test software. Figure 2 N5392A Ethernet consistency test software

    The software is also very easy to use. Users only need to select the test rate and test items in order and connect according to the prompts, and then run the test software. After the software is running, the oscilloscope will automatically set the time base, vertical gain, trigger and other parameters and perform measurements. The measurement results will be summarized into an HTML format test report, which lists the test items, whether they pass, spec requirements, measured values, margins, etc., as shown in Figure 3.

Figure 3 Test results of Ethernet conformance test software

    It is worth noting that Ethernet signal testing is not done by placing a probe on a normally working cable, because Ethernet signals are high-speed signals, and the test must be measured at the termination resistor at the end of the signal, which is the purpose of using a test fixture. During the test, the test software will prompt the user to set the device under test to different test modes to complete the test of different items. For example, Gigabit Ethernet specifies 4 test modes for different tests, so users must be able to set their own device under test to the corresponding test mode.

    For the jitter test of Gigabit Ethernet, the specification requires a cable consisting of five sections with different impedances to simulate the transmission channel, so the jitter test of Gigabit Ethernet also requires the use of the N5396A jitter test cable. Figure 4 shows the test cable and its requirements.

Figure 4 Gigabit Ethernet jitter test requirements for test cables

    In addition, for Ethernet testing, it is also necessary to test the return loss of the DUT, that is, the return loss, to consider the impedance matching of the DUT. Excessive return loss will cause signal reflection, distortion, crosstalk, etc., especially for Gigabit Ethernet, because it is a 4-pair cable working in both directions at the same time, so the return loss requirements are higher. To measure Return Loss, it is not enough to rely on an oscilloscope alone, and a network analyzer is also needed. However, the use of a network analyzer may not be familiar to engineers working on digital circuits, so the N5392A test software provides a control function for the network analyzer. The oscilloscope host can be used to control the network analyzer through the GPIB interface to complete the Return Loss test, and the test data can be analyzed and calculated, and the test results can be attached to the test report. Figure 5 is an instrument connection diagram for Return Loss measurement.

Figure 5 Return loss test instrument connection diagram