In the connected IoT world we make there will be many Ethernet ports. It is inevitable that these vital network connections will be exposed to external environmental hazards. External sensors, cameras, WiFi hotspots, WiFi repeaters, and microcell backhaul are all examples of devices that may connect to a switch or router, exposing them to harmful overvoltage threats. In an effort to maximize the survivability of their equipment (and minimize warranty returns), manufacturers will most certainly require their switches and routers to meet specific surge/safety standards or guidelines.
What started as recommendations given by experts in the field to help maximize the likelihood of successful, robust operation when deployed in the real world, quickly becomes mandated requirements when your customer demands compliance to one or more of them. You won’t have a choice. You’ll find yourself wading through the murky waters of international standards. It can be intimidating. Figuring out which standards apply to your particular implementation is confusing. Ethernet standards exist in IEEE, IEC, Telcordia, and ITU documents. They’re all similar, yet they’re all different.
The major Ethernet standards for lightning and surge protection are summarized below. It is important to only consider the most recent versions since fast-moving technology, such as PoE, will quickly make older versions obsolete. This blog will focus on ITU-T since, at the moment, it embodies the most stringent tests.
| Document | Notes |
| GR-1089-CORE | Issue 7, December 2017 |
| IEC 61000-4-5 | Edition 3.0 |
| IEEE 802.3 | 2015, but currently being updated to include bt |
| ITU-T K.20 | 7/2017, for central offices |
| ITU-T K.21 | 7/2017, for customer premises |
| ITU-T K.45 | 7/2017, for access and trunk networks |
Table 1: Major Surge Protection Standards
The ITU-T Standards You Need to Know
ITU (International Telecommunications Union) is a specialized agency of the United Nations dealing with Information and Communications Technologies (ICT). It is further divided into sectors, one of which focusses on standardization. It used to be known as CCITT but became ITU-T in 1993. Within ITU-T, the K series discusses protection against interference, including environmental hazards.
ITU-T standards are well organized as long as you know how to navigate the system. The latest version of the standard can be freely downloaded from https://www.itu.int/rec/T-REC-K/en. Specific test requirements (e.g., voltage and current levels, impedances, test duration, pass/fail criteria) are listed in K.20, K.21, and K.45, depending on the application. Each of these uses the general test setups and test schematics defined in K.44.
The illustration below can be used to help you determine which standard is the one you need to be concerned about for your product. Please note that ITU-T does not distinguish between Ethernet Power Sourcing Equipment (PSE) and Powered Devices (PD). They both use the same standards. Residential areas typically do not have good grounding systems in place, making them more prone to environmental surges. Therefore, the K.21 tests are more difficult than the others. Generally, if you can pass K.21 then you can pass K.20 and K.45 as well.
Figure 1: Illustrated are the ITU-T standards for Central Office, Access Equipment and Customer Premises equipment.
ITU-T is fairly simple to follow since it explicitly lists Ethernet and PoE requirements in tests 2.1.7 through 2.1.11 for external ports and tests 7.4 through 7.7 for internal ports. Conveniently, external and internal ports both have the same requirements. All surge waveforms are generated from a standard 8/20 Combination Wave Surge Generator. These tests are summarized in the Table below:
ITU-T K.21 Resistibility of Telecommunications Equipment to Overvoltage and Overcurrent
K.21 Test |
Test Description |
K.44 Test Setup |
Coupling |
Basic test levels |
Enhanced test levels |
Notes |
2.1.10 & 7.6 |
UTP Impulse Voltage Test |
Figure A.6.7-3a |
R=5 Ω |
±2.5 kV,
|
±6 kV,
|
Must be 1st test |
2.1.10 & 7.6 |
Insulation Resistance Test |
Figure A.6.7-3 |
Ammeter |
500 VDC |
500 VDC |
Must be 2nd test. Pass if R>2MΩ. |
2.1.7 & 7.7 |
UTP Transverse Test |
Figure A.6.7-5 |
R1 = 10 Ω, R2 = 10 Ω |
±2.5 kV,
|
±6 kV,
|
|
2.1.11 & 7.5 |
PoE Mode A & B Transverse Test |
Figure A.6.7-2 |
R = 10 Ω,
|
±2.5 kV,
|
±6 kV,
|
|
2.1.8 |
STP/UTP Simultaneous Port to Earth |
Figure A.6.7-4 |
R=10, x8 |
±2.5 kV,
|
±6 kV,
|
This test is only for external cables. |
2.1.9 & 7.4 |
STP Simultaneous Port to Earth |
Figure A.6.7-6 |
R=5 Ω |
±2.5 kV,
|
±6 kV,
|
Same as 1st test but with shield added. |
Table 2: List of the most current July 2017 ITU-T K.21 Basic and Enhanced tests.
The biggest change from the previous ITU-T version is the increased Enhanced test from 4kV to 6kV, which more accurately models what is seen in the field.
Indoor equipment may only need to meet Basic level tests. For outdoor equipment it is tempting to design for less than the 6kV Enhanced tests by relying on shielded ethernet cables to reduce the environmental coupling of voltages and currents, but this won’t protect from conducted voltages and currents already on the wires. The end-customer will determine whether a design must meet Basic or Enhanced protection standards. As usual, this involves a cost versus performance trade-off.
Test Example
Figure 2: An example of a K.44 test set-up.
There are often three different grounds available in PoE designs: Line-side signal ground connected to the cable shield (if there is one) at the RJ45 connector, PHY-side signal ground on the other side of the transformer where the Ethernet ASICs and SOCs exist, and PoE power ground which is isolated from the signals. These are shown above as different shaded boxes.
In addition, all PoE tests (except the Insulation Resistance Test) are performed with power applied. This is shown above by the 54V bench supply. It is appropriate to add a diode with sufficient breakdown in series with it to prevent the surges from destroying the power supply. This is shown above. Be sure to increase the power supply voltage to compensate for the forward drop of the diode so the PoE controller is given the proper voltage.
K.44 generally specifies that all grounds be connected together. However, it is also expected that the equipment be tested in its normal operating mode which, for PoE, means the 54V power supply should remain floating. This is will be reconciled in future ITU-T amendments but for now, take care that a floating power supply is used for these tests. Otherwise you may need to run these tests with no power applied.
Surge Protection Considerations
It is quite challenging to find an optimal design to handle K.21 surges and test conditions shown in the Table above. The coupling resistors are small. The surge voltages are big. Large transient currents will find their way to ground somehow, somewhere. It behooves the designer to think through the various routs this current could take if the driving voltage was large enough. Since transformer inductance and circuit capacitance is present, transients may not behave the way you think they do. So, it may be useful to discuss some of the pertinent issues that should be considered.
Line-side considerations:
Surges induced on the line side can be effectively blocked using a signal transformer rated to handle surges up to 6kV for Enhanced and 2.5kV for Basic levels. However, “true” surge-rated transformers are rare in the market as it is difficult and costly to test 100% in production. However, transformers are routinely exposed to a High Potential (i.e., Hi-Pot) test. This is a test involving a slow voltage rise so its breakdown mechanisms may not correlate very well to surge breakdown. Still, transformers with high Hi-Pot ratings are generally more desirable than those with lower ratings.
Most RJ45 connectors do not tolerate 6kV surges very well. This can be solved by incorporating a clamping device (MOV or TVS) on the line-side center tap of the transformer, limiting the voltage seen on the Ethernet lines. This has a side benefit of protecting the transformer as well. However, the clamping device must not trip during the 500V Insulation Resistance Test (shown in the table # above) so be sure to pick parts with a clamping voltage above 500V, including process and temperature tolerances.
Phy-side considerations:
The transformer may saturate during a surge, greatly reducing the voltages seen on the Phy-side of the circuit. But it is still wise to include a clamping device across the signal lines to protect the Ethernet ASICs. For Enhanced tests, a more robust design would also include current limiters such as a Bourns TCS product.
PoE considerations:
The surge can also enter the PoE section of the design through the transformer center taps. The PoE Controller circuitry should be protected by clamping devices. There is often a series RF choke in the DC supply path as well, but its frequency response is high enough that it offers very little transient surge protection. The clamping devices should be located close to the device they are protecting and sized appropriately to handle the voltages and currents that reach the PoE circuitry.
Want To See More Detail?
Hungry for more? A second article will be written that describes in more detail the challenges in passing the various K.21 tests. It will include a design example, waveforms and measurements. Stay tuned for more insight and surge protection tips from Bourns
Posted by