IEC61000-4-5 is also known as the surge immunity test. This mainly tests the immunity of electrical equipment to a unidirectional surge caused by overvoltage from switching or transients. Unlike ESD and EFT tests, Surge test involves an injection of high energy or power pulses into the system. These pulses are in the millisecond range, unlike the nanosecond pulse range for ESD and EFT tests. The surge test is defined by the ability of electrical equipment to withstand high energy pulses without any damage in the device or system application. The system needs to recover automatically without any power cycle or hardware reset after the surge injection. Unlike other EMI tests, where no packet errors or loss is allowed in the system, surge tests mainly look for Class B performance. For this, packet errors or loss is allowed, but the PHY must automatically recover by itself.

Within IEC61000 4-5 Surge test, there are different tests defined for: Open circuit, Short circuit, and Telecom ports. Both Open and Short circuit testing is mainly used for power surge tests. There are Section 5 tests that can improve the power circuitry performance. However, power EMC tests depend on power ICs other than Ethernet. Therefore, we can mainly focus on the signal ports' surge test.

In signal or telecom port surge tests, there are different test setups for Shielded and Unshielded cables. Here are the differences between the two test setups:

• For unshielded cable applications, surge test involves a direct noise injection through the coupling decoupling network (CDN) to the cable. Since unshielded cable does not have ground shield cover around the cable, there are more possibilities for differential noise injected into the system. As a result, both differential mode noise injection line to line and common mode noise injection line to ground surge test need to be perform.
• For shielded cable applications, noise is directly injected to the cable shield or earth ground. Therefore, only "line to ground" need to be tested since most of the noise is acting as common mode noise. Here are common ways the surge pulse is injected for shielded cable applications:
  • Direct surge noise injection cable shield through clamp or direct contact
  • CDN has a solid connection to earth ground, with the Ethernet cable connected to each side. The noise is directly injected on the shield of the CDN

Both types of surge test setups noted above simulate real life scenarios. Compared to shielded cable surge tests, unshielded cable surge tests normally give worse performance. This is due to the noise directly injecting into the cable. Again, Surge test is a high power noise injection into the system. Having a potential path for high power noise to inject into the signal lines is not always recommended for the design. In industrial applications, most customers use shielded cables. The following section focus on shielded cables or line to ground surge tests.

Surge Test level:

• Level 1: ±0.5kV (line to line), ±1kV (line to ground)
• Level 2: ±1kV (line to line), ±2kV (line to ground)
• Level 3: ±2kV (line to line), ±4kV (line to ground)

Surge waveform:

For Signal and Telecom port surge tests, the pulse width is 1.4ms with 10us Front time. The pulse is a high-power pulse injected into the system:

Figure 4-8 Surge Test Waveform