The 2003 ITU-T Telecommunication Equipment Resistibility Recommendations

Mick J. Maytum

Updates to the 2000 ITU-T equipment recommendations modify longitudinal testing and add internal and interport tests.

The International Telecommunication Union Telecommunication Standardization Sector (ITU-T) prepublished new versions of the K.20, K.21, K.44, and K.45 resistibility recommendations in July 2003.4 These recommendations are in force and replace the 2000 issued versions, covered by an earlier article5. This article is based on this prepublished material, but correcting some obvious errors (see Conclusions).

The increased coverage of the new recommendations withdraws four other ITU-T Recommendations. Recommendations K.156 on remote-feeding systems and line repeaters and K.177 on power-fed repeaters are both absorbed into K.45. Recommendation K.228 on ISDN T/S bus equipment is absorbed into K.21 and K.419 on internal interfaces at telecommunication centres is absorbed into K.20. Recommendation K.44 picks up the test circuits from the absorbed recommendations. This article provides an overview of testing ports connected to symmetric-pair cables.

The basic document structure remains the same. Recommendation K.44 is the test-circuit standard. Recommendation K.20 covers equipment located at a telecommunications center (central office), K.21 covers equip- ment located at customer premises (subscribers), and K.45 covers access and trunk equipment.

This article provides an overview of testing ports connected to symmetric-pair cables.

External-Port Test Circuits

Single-Port Test Circuits. There are three test circuit configurations for single-external-port testing: transverse (Figure 1), port-to-ground (Figure 2), and port-to-external-port (Figure 3). Powering or auxiliary equipment or terminations are included in the test circuits as required to ensure normal system operating levels. Appropriate coupling and decoupling elements are added to present low- and high-impedance paths to the test generator current. Any source or termination fed via a decoupling element can be ignored from the surge dynamics. Gas discharge tubes (GDTs), metal-oxide varistors (MOVs), and thyristors are suitable coupling elements that block under normal conditions and conduct during a voltage surge. When needed, special or agreed or appropriate primary protection is added to the equipment ports.

Figure 1. Transverse testing, single external port. R = resistor, G = ground terminal.

By comparison with those in the 2000 resistibility recommendations, the new test circuit figures provide more information on test setup. However, because of this, and because of the need to provide additional test options, the new test circuits appear more complex.

Transverse Test Circuit. Transverse testing is carried out with some untested ports of each type powered or terminated. The unused ports are left floating. Testing is carried out on both port terminals. Transverse testing uses a single generator output via a resistor.

Port-to-Ground Test Circuit. Port-to-ground and port-to-external-port testing applies the test voltage longitudinally via the two current-sharing resistors. The surge voltage is still applied longitudinally in the port-to-ground circuit, but with a subtle difference from previous longitudinal port testing. The difference appears at the right side of Figure 2, where an internal port is shown being coupled to ground. Previously, equipment, such as a modem, was not stressed between the external line port and an internal type of port, such as a universal serial bus. The new arrangement will test for external/internal port voltage withstand when the primary protector is not fitted.

The port-to-ground test takes two forms. The first is to test with all untested ports powered or terminated. The second is to test, in turn, with each type of internal port grounded via a coupling element.

Port-to-External-Port Test Circuit. This type of test is particularly suitable for power-fed repeaters, which are series devices in which the primary protector grounding may not be effective. All untested ports are powered or terminated. Each type of external port is coupled to ground in turn, including, if present, a port of the same type as the one being surged. Equipment that has only one external port type is tested between ports.

At the basic test level, the port-to-external-port test does not apply to K.45 and K.21 equipment designed to be always connected to ground. Port-to-external-port testing does not apply to K.20 equipment unless an enhanced requirement is specified for telecommunications center locations with fewer than 250 lines. In that case, such a test is applied using the K.20 enhanced test voltage levels.

Multiple-Port Test Circuits. Only port-to-ground and port-to-external-port configurations are used in multiple-external-port testing. Test circuits are those depicted in Figures 2 and 3 but with additional generator feed resistors to connect the multiple ports. For 10/700 testing (recommendation tests 2.1.3 and 2.1.4), the extra loading will reduce the tested-port stress below the level of single-port testing.

Single- and Multiple-Port 8/20 Current Testing

Port-to-ground and port-to-external-port tests for single (test 2.1.5) and multiple (test 2.1.6) ports are performed on equipment that contains high-current-carrying protectors that remove the need for primary protection. Despite modifications, flaws remain in these enigmatic tests.

The single-output 8/20 current generator does not promote current sharing, as the generator external-current-limiting resistor is zero. Any current sharing depends on the coupling element, the test lead length, and the equipment protection circuitry. Recommendation K.44 warns:

When applying the test to multiple wires, care should be taken to ensure that the current is divided equally between the wires. Particular care should be taken to ensure that the operation of one or more protectors does not prevent the operation of the other protectors.

However, practical solutions to promote current sharing are not given.

Intuition suggests that the test lead inductance would help with sharing. However, laboratory testing with 1-m lead lengths on a single port with GDT protectors resulted in only one instance of current sharing in 10 tests. In the remaining nine tests one or another of the wire protectors operated first and monopolized all the current. Failure of current coordination limits the total current-test current to that of a single wire.

The cause of this problem is the single-output 8/20 high-current generator. Such 8/20-waveshape generators are widely available and used to benchmark GDTs, MOVs, and primary protector overstress performance. Their association with the test term lightning current perpetuates a common myth: the 8/20 waveshape has nothing to do with the normal characteristics of real lightning.10

Recommendation K.44 mandates that primary-protector characteristics comply with Recommendations K.12 for GDTs11 or K.28 for semiconductor arresters.12 K.28 thyristors have life ratings of 100 applications of 100 A, 10/1000, and must fail low-resistance at 5 kA, 8/20. K.12 GDTs have life ratings of 300 applications of 100 A, 10/1000, and are supposed to survive 10 applications of 10 kA, 8/20. Obviously, these 8/20 current tests are created for K.12-type primaries and exclude K.28-type primaries. U.S. experience has been that 100-A, 10/1000 thyristor primary protectors are reliable in central offices, but that these protectors are likely to fail the basic-level 8/20 test of 1 kA. It is a pity that the testing of integrated GDT primary equipment is included before the completion of the generic test options.

To evaluate the overstress caused by near strikes of lightning, primary protector assemblies are tested at 8/20 wire current levels of 5 kA and 10 kA. Assemblies are mandated not to show external damage. The absence of a let-through measurement is a major omission in these tests. In practice, a single measurement cannot capture the different equipment sensitivities; the most secure approach is to test the combination of primary protector and equipment. Converting the high-current 8/20 equipment testing to evaluation of primary/equipment failure coordination would improve system integrity. The requirement is that damage not occur to equipment using an external primary. Damage can occur to equipment with an integrated primary, but hazards shall not be caused.

Internal-Port Test Circuits

Recommendation tests 7.1 and 7.2 cover only lightning-induced voltages, whereas the internal-port tests of GR-1089-CORE cover both lightning and ac power fault conditions.13 The test circuit to be used depends on the internal cabling. Figure 4 shows the test circuit for ports with unshielded cabling. Some untested ports of each type are powered or terminated during testing, with the unused ports left floating.

Figure 4. Unshielded-cable longitudinal test circuit, single internal port. R = resistor, G = ground terminal.

An 8/20,1.2/50 combination impulse generator is used for unshielded cable testing with series resistance of 10 (omega). The resistor changes the wire short-circuit current waveshape from 8/20 to 3.3/30 (see Figure 5) for a 5-(omega) load. Any equipment input resistance moves the current waveshape further toward the open-circuit voltage waveshape of 1.2/50.

Ports with shielded cabling are tested in a multiple-port configuration with cables and shield bonded at the generator end of the cable (see Figure 6). An 8/20,1.2/50 combination impulse generator is used for shielded cable testing with series resistance of zero. Internal-port tests do not apply to K.45 equipment.
 

Figure 5. Combination generator current waveshape variation with external load.

Coordination

The coordination requirement of primary (if fitted) conduction at the highest test-voltage level remains the same (recommendation test 2.1.2). Changes to Appendix 1 of K.44 have removed the coordination figure errors that have been reported.5 The line-card example now used has symmetrical protection voltage levels, which is not so informative as the previous asymmetrical case. In the rest of the discussion in this section, the primary and secondary protection is assumed to be a switching-type protector such as a GDT or thyristor.

Manufacturers of general-purpose equipment have had a major coordination-testing problem because they do not know what kind of primary protector will be used with their equipment. This issue of the resistibility recommendation as of now misses the opportunity to fix the problem by specifying a default primary protector as was done with GR-1089-CORE Issue 3.

Figure 6. Shielded-cable test circuit, multiple internal ports. R = resistor, G = ground terminal.

Adding the port-to-external-port test arrangement increases the lightning coordination tests from two to three. The coordination mechanism for the port-to-external-port test depends on the port impedance to the equipment ground terminal. For high-impedance ports, the only high-current (nonflashover) path is through the port primary protector. Because the port and external-port primary protectors are in series, the generator voltage for coordination is going to be above the combined primary-protection voltages. The generator coordination voltage may be reduced if there are capacitance differences between the port/protector combinations.

Figure 7 shows equipment in which the two ports connect via low-impedance coordination elements and conducting secondary protection. As the impulse voltage increases, if a section of protector P1 operates, it will cause a voltage jump in the equipment ground terminal. This voltage increase should cause conduction of a section of protector P2. Similarly, first conduction of a section of protector P2 will cause conduction of a section of protector P1. In either case, once conduction to ground occurs, the test becomes port-to-ground coordination

Note that only protector P1 needs to coordinate with the equipment. To have protector P2 coordinate with the equipment, a current-sharing resistor, as present on the protector P1 side in the figure, needs to be added to each ground return coupling element. Omission of these extra current-sharing resistors looks like a test oversight, as one would expect the circuit to be symmetrical about the equipment ground terminal. Lack of coordination means that the current rating of a protector P2 section can be exceeded, because it carries the total current conducted by both sections of protector P1.

Multiple-port testing (recommendation test 2.1.4b) further increases the return current. Including current-sharing resistors for protector P2 and using the same number of ground return ports as tested ports will prevent excessive section current.

The table in Figure 7 gives the minimum coordination-element resistance for 400-V primary protectors at three voltage levels. For the port-to-ground coordination, the coordination resistance has to develop 400 V across the nonconducting section of protector P1 after the first section conducts. For port-to-external-port coordination, the coordination resistance has to develop 800 V across the series-connected nonconducting primaries. These coordination resistance values converge at higher voltage levels.

AC Tests

Apart from the addition of the port-to-external-port test, which increases the number of configurations from two to three, the equipment resistibility recommendation updates make little change in the three test categories of induction, induction with fitted primary, and power contact. The port-to-external-port ac power fault testing does not apply to K.20 equipment. K.44 has the 320-(omega) power contact source resistance error corrected to the 300-(omega) value used in the equipment recommendations.

One anomaly is that the power contact source resistance values do not change for different values of ac mains voltage (test 2.3.1). On 230-V rms mains and with the lowest source resistance of 10 (omega), the prospective current is 23 A rms. On 120-V rms mains, the prospective current is down to 12 A rms, yet established testing for GR-1089-CORE and UL 60950 uses a maximum current of 25 A rms.13,14 There is a case, based on existing 120-V rms test practice and symmetric-pair wire gauge conformity, for expressing power cross testing in terms of a prospective fault current range of 0.25-25 A rms rather than a source resistance range of 10 (omega)-1 k(omega).

Test Levels

The designation of test levels as basic and enhanced remains. As before, Appendix 1 of K.44 discusses higher lightning test voltages for certain applications. K.21 equipment used in customer-premises locations with poor primary protection merits extension of the inherent testing voltage up to 5 kV, 10/700. To encompass field measurements that show that peak voltages over 7 kV exist, coordination testing up to 10 kV is allowed.

Test Tables

Tables I, II, and III summarize, respectively, the external-port lightning tests, external-port ac power fault tests, and internal-port lightning tests defined in the prepublished new recommendations for symmetric-pair cables. The central portion of each table shows basic and enhanced test levels for the K.20, K.45, and K.21 recommendations. Blocked together are tests with the same values, in order to display commonalities and differences between these three recommendations. Right-hand columns show the primary protection used and the acceptance criterion A (withstand without damage) or B (the equipment, although damaged, shall not cause a fire hazard). Left-hand columns indicate the number of ports tested and test waveshape, and include notes, test applications, the test reference number, and the test type.

Conclusion

These prepublications cover internal and external symmetric pair cable ports, dedicated power feeds, and mains power ports, leaving coverage of coaxial ports for future revisions. Integration of other ITU-T Recommendations increases the equipment and port coverage. With the increase in breadth comes complexity. The aim of this article is to give an understanding of the interrelationships of the recommendations and the test intent.

By far the biggest problem with the 2000 Recommendations was coordination. Missed, in this revision, was the opportunity to define a default primary protector to solve the problem. Table 7 in the K.20 and K.21 prepublications has the wrong K.44 Figure 6 references.

For tests 7.1, 7.2, 7.3, and 7.4, the correct figure references are A.6.5-1/K.44, A.6.5-2/K.44, A.6.3-2/K.44, and A.6.3-1a/K.44, respectively. Content change is likely before final publication and the most-up-to date information will be at http://itu.int/ITU-T.

References

01. ITU-T Recommendation K.20, "Resistibility of telecommunication equipment installed in a telecommunications centre to overvoltages and overcurrents" (Geneva: International Telecommunication Union, July 2003).

02. ITU-T Recommendation K.21, "Resistibility of telecommunication equipment installed in customer premises to overvoltages and overcurrents" (Geneva: International Telecommunication Union, July 2003).

03. ITU-T Recommendation K.44, "Resistibility tests for telecommunication equipment exposed to overvoltages and overcurrents--basic recommendation" (Geneva: International Telecommunication Union, July 2003).

04. ITU-T Recommendation K.45, "Resistibility of telecommunication equipment installed in the access and trunk networks to overvoltages and overcurrents" (Geneva: International Telecommunication Union, July 2003).

05. Mick J Maytum, "The New ITU-T Telecommunication Equipment Resistibility Recommendations," Compliance Engineering 19, no. 1 (2002): 30-37; available from Internet: www.ce-mag.com/archive/02/01/Maytum.html.

06. ITU-T Recommendation K.15, "Protection of remote-feeding systems and line repeaters against lightning and interference from neighbouring electricity lines" (Geneva: International Telecommunication Union, 1988).

07. ITU-T Recommendation K.17, "Tests on power-fed repeaters using solid-state devices in order to check the arrangements for protection from external interference" (Geneva: International Telecommunication Union, 1988).

08. ITU-T Recommendation K.22, "Overvoltage resistibility of equipment connected to an ISDN T/S bus" (Geneva: International Telecommunication Union, 1995).

09. ITU-T Recommendation K.41, "Resistibility of internal interfaces of telecommunication centres to surge overvoltages" (Geneva: International Telecommunication Union, 1998).

10. IEC 61024-1-1, "Protection of structures against lightning, Part 1: General principles, Section 1: Guide A--Selection of protection levels for lightning protection systems" (Geneva: International Electrotechnical Commission, 1993).

11. ITU-T Recommendation K.12, "Characteristics of gas discharge tubes for the protection of telecommunications installations" (Geneva: International Telecommunication Union, 2000).

12. ITU-T Recommendation K.28, "Characteristics of semiconductor arrester assemblies for the protection of telecommunications installations" (Geneva: International Telecommunication Union, 1993).

13. Telcordia Technologies Generic Requirements, GR-1089-CORE, Issue 3, "Electromagnetic Compatibility and Electrical Safety--Generic Criteria for Network Telecommunications Equipment" (October 2002).

14. UL 60950-1, "Information Technology Equipment--Safety--Part 1: General Requirements" (April 2003).

Mick J. Maytum is the telecommunications standards specialist at Bourns Ltd. (Bedford, UK). He can be contacted via e-mail at mick.maytum@bourns.com.