 
|
  
   |
| About
CE-Mag Free Subscriptions Current Issue Article Archives ESD Help Mr. Static Web Gallery Staff Info Contact us
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Mick
J. Maytum GR-1089-CORE, Issue 3, came into force in November 2002. It applies to service provider network equipment and covers electromagnetic compatibility (EMC) and electrical safety. Three years after the previous issue, this revision has greater harmonization with other U.S. and international standards. It includes electrical fast transient (EFT) criteria and more-rigorous testing. This article is written for those familiar with the previous version of GR-1089-CORE. It gives an overview of the section changes and looks at lightning testing in detail. A second article on page 109 covers ac power fault. Structure The new requirement is divided into the same nine sections as the previous issue: Introduction, System-Level ESD and EFT, Electromagnetic Interference, Lightning and AC Power Fault, Steady-State Power Induction, DC Potential Difference, Electrical Safety Criteria, Corrosion, and Bonding and Grounding. Aside from Section 2, which now includes EFT in the title, the section titles are unchanged. Even with the added content, at 236 pages the 2002 edition is only four pages longer than the 1999 edition.
System-Level ESD and EFT. This section is now harmonized with IEC Standard 61000-4-2, "Testing and Measurement Techniques—Electrostatic Discharge Immunity Test," Edition 1.2, 2001-04. The IEC standard is a normative reference, and specific clauses within it are called up in 1089. The IEC reference is a dated one (2001-04), and compatibility with any future revisions issued by IEC should be checked to see whether there is a variance from Edition 1.2. New for this revision is EFT. EFT testing is meant to correlate with the real-world equipment effects of ac system switching transients, such as contact arcing. Testing of telecommunications and power ports is done by capacitively coupling the EFT signal either by individual capacitors or by a 1-m capacitive coupling clamp around the equipment cables. The normative reference for EFT testing is the IEC standard 61000-4-4, "Testing and Measurement Techniques—Electrical Fast Transient/Burst Immunity Test," 1995-01. Again, as this reference is dated, compatibility with any future revisions issued by IEC should be checked.
In crude terms, EFT pulses are like an ESD pulse on steroids; more energy and a higher repetition rate. Figure 1 shows a typical ESD current waveform at the 8-kV contact level. Figure 2 shows a typical EFT current waveform at the 2-kV test level. Into a 50-W load, an individual EFT pulse has a rise time of 5 nanoseconds with a duration of 50 nanoseconds. These pulses are applied in bursts at 5 kHz (every 200 microseconds) for test levels up to 1 kV, and 2.5 kHz (every 400 microseconds) for 2 kV. The burst duration is 15 milliseconds (75 or 37 pulses) and is repeated every 300 milliseconds (see Figure 3). Electromagnetic Interference. Several new technical requirement and measurement reference documents are now included: CISPR 22, IEC 61000-4-6, MIL-STD-461E, and TIA/EIA/IS-968. Standards are continually being updated, and the referenced TIA/EIA/IS-968 (July 2001) has now been upgraded and revised as TIA-968-A (October 2002). Objectives have been added for emissions limits for broadband leads and power ports (ac and dc) along with conducted immunity for broadband and voiceband leads. A conditional objective for conducted immunity of power ports is now included.
Lightning and AC Power Fault. This section is reviewed in detail in this article and in the article on page 109. Only a quick summary of the changes is given here. Harmonization moves have occurred in two areas. The ITU-T recommendation K.20, power induction i2t values and primary-equipment protection coordination, has been integrated into this section. In addition, the use of the UL 60950 Type MDL-2 A fuse has resulted in a new wiring simulator characteristic. The generator for the current-limiter testing has been modified to remove a test blind spot. Criteria for current limiters in digital subscriber line (DSL) applications are given through new conditional objectives and conditional requirements. Objective first-level power fault tests (6 through 9) have become requirements, which is a potential problem area for line feed resistors (LFRs). Clarifications have been made to the sample plan and first-level tests. First-level threshold testing of equipment current and voltage limiters is now done with the current limiter bypassed and the overvoltage limiter removed. Listing requirements, previously in this section, have been moved to Electrical Safety Criteria. Steady-State Power Induction. A requirement and an objective have been added for coaxial port immunity to longitudinal steady-state power induction. Corrosion and DC Potential Difference. These two sections are unchanged. Electrical Safety Criteria. This section has increased harmonization with international and U.S. safety standards. Clarification and guidance have been given in the criteria applied to network equipment, parts, and circuit packs. The short-circuit testing approach of ac and dc sources has been reviewed and improved.
Bonding and Grounding. Similar to the electrical safety section, clarification and guidance have been provided in the criteria applied to network equipment, parts, and circuit packs. The short-circuit testing approach of ac and dc sources has been reviewed and improved. General. This new issue of GR-1089-CORE tries to resolve industry comments on the previous version. The subsection structure has been changed to avoid repetition, separate out the different port types, and provide a logical flow. Harmonization with other standards, where possible, has been carried out. Overall, it is a worthy successor to the second issue. Lightning Generators With the exception of the IEEE C62.41 combination waveshape 1.2/50—8/20, all of the lightning waveforms were specified as maximum rise time, minimum duration time, and minimum amplitude. In this issue, Annex A contains waveform tolerance. For the first time, the minimum rise time, maximum duration time, and maximum amplitude are specified. However, this does not mean new impulse generators are required. The tolerance represents commercial generator values. Table I shows the waveform values. One change has been to modify the four-wire test condition A. A test has been added with all wires connected to the generator (see Table II). To allow test facilities to be in place, an Issues of List Report (ILR) is expected, which delays the introduction of this four-wire test. Ports and Number of Test Samples Generally, three samples of the equipment are to be tested. Equipment ports are either telecommunications, power, or coaxial. In the case of telecommunications ports, the requirement of three samples has been changed to three port samples. For single-port equipment, three pieces of equipment must be tested. For equipment of three or more ports, three ports of a single piece of equipment are required to be tested. Testing two-port equipment is awkward. The test requires two pieces of equipment with testing of two ports on one piece and testing of one port on the other. For first-level testing, there shall be no damage not only to the equipment tested, but also to any host system.
Telecommunications Ports First-Level Lightning Tests. The basic tests in this section are unchanged, but there are changes to the test notes and the voltage-limiter procedure. Test note 10 now contains a warning about the extra stress caused by the IEEE C62.41 combination generator (1.2/50—8/20) when it is used as an alternative to the 2/10 generator. Although the combination generator test current and voltage amplitude are adjusted to that of the 2/10 waveform, the current waveshape is no longer 8/20. Figure 4 is a plot of the current rise and duration variation with load resistance. Typically the combination generator would be testing the equipment with duration two to three times that of the 2/10 generator. Test note 5 now gives guidance on testing at the equipment voltage-limiter maximum operating threshold. The use of the word threshold is confusing. Normally it would be considered a voltage level beyond which the signal starts to be clipped. However, in this test note, the voltage-limiter maximum operating threshold means the maximum voltage limiting under surge conditions. The maximum limited value must be identified. The generator open-circuit voltage is then set to this level and applied to the port with the limiter removed. The advantage to this approach is that the equipment is stressed to the worst-case-condition voltage rather than the voltage of the fitted protector.
For example, a TISP4350H3BJ thyristor voltage limiter has a typical 10/1000 protection voltage of ±310 V and a maximum guaranteed value of ±350 V. Under test, this protector would be removed, and a 350-V generator impulse would be applied to the port. This test may apply up to 40 V more than the typical protector would allow. This additional voltage increases the confidence in the equipment withstand, in spite of the small number of samples used. Certain applications have asymmetrical limiting voltages, and the positive and negative test impulses will be different. For example, a plain old telephone service (POTS) subscriber line interface circuit (SLIC) might be tested with 5- and 70-V impulses. Gated thyristor voltage limiters are often used in a voltage-tracking mode, where the gate is connected to the protected electronics voltage supply. The limiting voltage of this arrangement will be the supply voltage plus the gate-trigger voltage of the thyristor. For example, a circuit consisting of an integrated circuit (IC) run from a 52 ± 2-V supply combined with a TISP61089B gated limiter having a maximum trigger voltage of 2.5 V would be tested with a 56.5-V impulse. Compared with 10/360 or 10/1000 testing, 2/10 testing has two problems: identifying the maximum limiting level and the possible circuit overstress. Voltage limiter manufacturers do not often quote the 2/10 voltage-limiting performance. Rather, they merely state that the limiter survives a certain 2/10 test condition. The faster-rising and higher-current 2/10 impulse causes the maximum limiting voltage to increase, often by 15% or more. The previously mentioned TISP61089B gate-trigger voltage rises from 2.5 V on 10/1000 to 12 V on 2/10, making the 2/10 test voltage 66 V.
For voltage-clamping devices (e.g., an MOV), applying a 2/10 impulse of the appropriate amplitude is acceptable because the 2/10 impulse has a smaller volt-second integral than the limiter voltage. For switching devices such as a thyristor, applying a 2/10 impulse equal to the maximum voltage limiting could cause IC failure because the 2/10 volt-second integral is larger than the worst-case-condition limiter voltage. The switching devices have a dynamic voltage overshoot condition that truncates the impulse voltage. The IC, therefore, is exposed to a much shorter voltage overstress than the overstress that occurs with a 2/10 voltage impulse of the same amplitude. Because the IC voltage withstand is time dependent, the abnormally long overstress from a full 2/10 impulse could cause failure.1 To follow the intent of note 5, the test impulse wave shape should be the similar to the limiter wave shape. Using a 1/2 wave shape for switching limiters (corresponding to asynchronous primary GDT sparkovers) would probably be more appropriate. However, a 1/2 generator is not standard. If the equipment does not survive testing with a 2/10 impulse, then further investigation is required, which could lead to an exemption request or alternative verification based on the above criteria. Second-Level Lightning Tests. The basic 2/10 test is unchanged. The warning on the use of the combination generator is repeated, and testing at the equipment voltage-limiter threshold can be skipped if this test was passed in the first level. Coordination. The 10/1000 primary-equipment coordination objective (desirable performance) is optional beginning January 1, 2005. It is required (must comply) beginning January 1, 2006. The 10/1000 generator open-circuit voltage is fixed at 1 kV for normal first-level equipment testing, but can be between 400 V and 2 kV for coordination testing.
Figures 5 and 6 show the coordination test configurations. If the primary protector has a series current-limiting element (e.g., a heat coil) the element impedance aids coordination. To comprehend this, an equivalent impedance is connected to the equipment terminals so that the measured system-voltage values incorporate the series impedance voltage. The actual primary-protector series element is not used for this test because it may fail. Bizarrely, the GR-974-CORE primary requirement only tests nonresetting and self-resetting current limiters to 25 A, 10/1000 and not 100 A, 10/1000 or a 10/1000 voltage wave shape with an amplitude equal to the primary limiting voltage. Besides the generator open-circuit voltage, the GR-1089-CORE coordination test circuits are the same as the equipment test circuits. To determine the coordination test level, the generator open-circuit voltage is set to the maximum voltage of the specified primary protector. GR-974-CORE, Issue 3 (June 2002), protectors have voltage levels of 400, 600, and 1000 V. The 1000-V level matches the GR-1089-CORE default carbon-block primary protector. The voltage and current levels at the system terminals are monitored when an impulse is applied. If the measurement at the initial generator voltage setting does not result in either the primary-protector voltage or a current flow of 100 A, the generator voltage is increased by 200 V, and the impulse applied again. This process is repeated until the system develops the required voltage, or conducts 100 A, or equipment failure occurs. Voltage interpolation is allowed in order to avoid overstressing the equipment. For example, if a generator voltage of 1000 V produced a current of 95 A peak, increasing the generator voltage by 200 V to 1200 V would result in a 114-A current, possibly causing equipment damage. This is not the intent of the test, and the generator should be set to an interpolated value of 1000 ´ 100/95 = 1050 V to produce 100 A.
When the coordination level has been determined, the equipment is further tested with impulses in each polarity in the longitudinal and transverse test configurations. To pass, the equipment must not fail during testing. The Telcordia coordination condition requires that the system is not damaged. It also requires that the system terminals develop specified GR-974 primary-protector voltage or conduct 100 A or both. This is different from the ITU-T coordination testing that requires the primary protector to operate.2 The Telcordia coordination rationale is that the normal GR-974-CORE primary protector can conduct a maximum of 100 A, 10/1000. GR-1089-CORE equipment can conduct a maximum of 100 A, 10/1000. Low-impedance equipment has the same current capability as the primary protector, and, hence, the coordination requirement is to withstand 100 A, 10/1000. Although the primary protector is prevented from operating, the equipment can withstand the expected current. High-impedance equipment develops more voltage and less current; therefore, the coordination requirement is to withstand the maximum primary-protector voltage. The GR-1089-CORE coordination criteria account for legacy equipment using fused current limiters. However, equipment using resistive current limiters may have problems. Assuming that the current-limiter resistance is the largest component of equipment input resistance under impulse conditions, the voltage across the limiter and the current through it will be the result of a resistive divider action with the 10/1000 generator 10-W source impedance. During 1000-V 10/1000 equipment testing, a 10-W current limiter will develop 500 V and conduct 50 A. During coordination testing for a 1000-V primary protector, the same limiter will develop 1000 V and 100 A, a normalized power increase of four times. For resistance values higher or lower than 10 W, the coordination test peak-power-increase factor will be less than 4. Figure 7 shows how the normalized peak power varies with resistance. Normal SLIC LFRs would need peak power capabilities about 50% higher than present. Low-resistance positive-temperature-coefficient (PTC) thermistors could require peak power capabilities from 1.5 to 4 times higher than present. The energy levels involved may even cause the PTC thermistor to trip into a high-impedance state. Tripping could cause the PTC thermistor to fail as a result of being exposed to the generator's nearly full open-circuit voltage.
One equipment builder achieved more than a 50:1 reduction in field failure rate when he changed from using low-resistance LFRs to fuses. Both equipment variants were 1089-compliant at that time. It has been shown that coordination requires a higher-capability LFR, and lack of coordination may have been part of the failure problem. Nonlinear current limiters, such as the GR-974-CORE fast current limiter (circuit voltage or current activated), cannot be emulated by a fixed value of series impedance. Therefore, the actual current limiter must be connected to the equipment terminals (see Figure 8). The failure criteria should then include the fast current limiter as well. Limiter failure can occur if the equipment terminals don't develop enough voltage for the limiter to operate correctly. GR-974-CORE uses a fast current-limiter test circuit (Figure 6-4 in the standard) with a specified equipment load resistance of 1 W. Intrabuilding Lightning Surge Tests. The basic 2/10 tests are unchanged. It also includes the familiar note changes concerning the combination generator warning and equipment voltage-limiter threshold test method. Lightning Surge There are no test changes in the subsection AC Power Port—First and Second Level Lightning Surge Tests. There are also no test changes in the subsection Coaxial Cable Ports and Broadband Communications Equipment—First and Second Level Lightning Surge Tests. Conclusion The changes to the previous lightning tests are incremental and reasonable. When the coordination test comes into force it could have a major influence on resistive current limiters. References 1. MJ Maytum et al., "Coordination of Overvoltage Protection and SLIC Capability," in The 2000 International IC China Conference Proceedings (Beijing: Global Sources Ltd., March 2000), 87–92. 2. MJ Maytum, "The New ITU-T Telecommunication Equipment Resistibility Recommendations," Compliance Engineering 19, no. 1 (2002): 30–37. Mick J. Maytum is applications director for Bourns Ltd. (Bedford, UK). He can be contacted at mick.maytum@bourns.com.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
