Eric, I assume you are aware of the published clarification on this subject published by Telcordia in December 1999. I suspect this might have led to your questions, which are very good.
Rather than discuss your points specifically, I feel that simply covering the intent of this generic requirement might clarify the issue. I am also confident you will also find varying interpretations at the various test houses. The first point is that this requirement is found in Section 9 of GR-1089-CORE which is titled "Bonding and Grounding". Just as the title implies, the intent behind the requirements is to ensure there is a reliable path for fault currents to flow and for reliable equipotential grounding. This section was not intended for maintained functionality after a fault condition. This may be desired, but the short circuit tests that require continued functionality are in Section 4 and deal with Tip and Ring circuits. The second point is that R9-20 and R9-21 are "type" tests. That is the lab will short the output at the output terminals and again at the load (if it is remote to the EUT). The assumption is that during normal operation, a component fails and the output is shorted. The test is to determine if a safety hazard is possible if this situation occurs. The pass/fail criteria are simply to make sure that the bonding and grounding paths (fault current carrying paths) remain undamaged, and there is not a risk of a fire hazard or mechanical hazard. This is best accomplished by using fusing or other sacrificial component that limits the current or opens the current path before any damage to the bonding and grounding paths can occur. If a non-bonding or grounding trace opens benignly without catching the cheesecloth on fire, that would in my opinion be acceptable. However if a trace opens and in the process damages a frame ground trace, -48VR trace, or other trace that will provide electrical safety (even without fire) that would be a failure. The third point is that this test is not an in-depth investigation into what happens under every conceivable change of load impedance or overload condition. These types of investigations are normally part of a safety listing program performed by a NRTL "Nationally Recognized Testing Lab". That is why I strongly endorse the Listing of all equipment regardless of its location in the network. The testing program for a Listing would be far more thorough than what is in R9-20 or R9-21. Overload testing, component fault analysis, as well as short circuit testing would be conducted. Not only would this type of evaluation look for damaged fault return paths, and blatant fire hazards, it would look for components that would overheat and potentially cause a fire. A major omission to the requirements in R9-20 and R9-21 is that the length of time the output is shorted is not defined. Therefore technically for NEBS compliance a 1 ms short is as valid as a 30 minute or 10 hour short. The tests for a safety listing would require the current to stop flowing or require thermal equilibrium to be reached or 30 minutes before the faults are removed. On a second somewhat related issue, the status of safety listing standards for Network equipment is a mess right now. There is technically not a safety standard to which most network equipment should be allowed to be Listed. UL 1950, UL 1459 (which goes away April 1), and the new UL 60950 all require that the equipment listed by the safety standard be "intended to be installed per the National Electric Code". Network equipment that resides on the telco side of the demarcation point is exempt from the NEC per Article 90-2 section b which states that the following are not covered, "Installations of communications equipment under the exclusive control of communications utilities located outdoors or in building spaces used exclusively for such installations". Only the installation methodologies of the wire and cable are controlled by the NEC, the telco owned "equipment" is exempt. Therefore telecommunications equipment that is intended for deployment in C.O.'s, the outside plant, or inside the customer premise (if on the telco side of the demarc) cannot be listed to appropriate safety standard as one does not exist (although several NRTL's are inappropriately doing it to UL 1459 and UL 1950 due to NEBS requirements or customer requests). What I would propose is that industry (either through UL, T1E1.7 or TIA TR41.7) develop an ANSI safety standard for network telecommunications equipment that resides on the telco side of the demarc point or that makes up the demarc point. The new safety standard could be closely modeled after UL 1950, 1459, 1863, GR-63-CORE and GR-1089-CORE. It could incorporate several GR-63-CORE, and GR-1089-CORE items so that the original compliance as well as continued compliance could be verified by the end customer simply by seeing the listing mark on the product. This could also be used by CLEC's to demonstrate compliance with the FCC report and Order last spring stating they only need to comply with NEBS Level 1 as defined by SR 3580 if they want to co-locate equipment in a C.O. They could show the listing information and the FCC part 15 information to the ILEC and demonstrate complete NEBS Level 1 compliance. Examples of what could be incorporated into the new safety standard are: 1.) GR-1089-CORE section 4 - Second level power fault and lightning requirements 2.) GR-1089-CORE section 4 - Listing requirements 3.) GR-1089-CORE section 7 - Safety 4.) GR-1089-CORE section 8 - Corrosion 5.) GR-1089-CORE section 9 - Bonding and Grounding (which would clear up your original question) 6.) GR-63-CORE section 4.2 - materials analysis for flammability (use ANSI T1.307-1997) 7.) GR-63-CORE section 4.2 - burn test (use ANSI T1.319-2000 which is in progress now, assuming it includes pass/fail criteria as has been proposed. Otherwise use its methodology and GR-63-CORE pass/fail criteria) 8.) UL 1950, UL 1459, 1863 - use the basic principles with the understanding that this equipment is exempt from the NEC and certain concepts such as voltage limits/classifications in GR-1089-CORE are in direct violation of the NEC (which only allows up to 60 VDC on customer premise telco wiring and 150V on any telecommunications wiring, GR-1089-CORE theoretically allows 400VDC with GFI). Replace TNV1, TNV2, TNV3, and RFT with Class A1, Class A2, and Class A3, AB, and B. Eliminate the European concept of creepage and clearance distances in telco circuits as the infrastructure of the North American network has proved itself safe with dielectric tests only. The creepage and clearance distances specified in UL 60950 for circuits above 60 VDC are not obtainable with the current infrastructure. With UL 60950 ILEC's will need to do voltage management for the Listings to be valid rather than service management which is not logistically feasible. Other industry standard items that are incompatible with UL 60950 are wire wrap pins and in some cases champ style connectors and even the twisted pair wiring which in many cases does not have thick enough insulation and in many cases is not Listed (as required by the NEC). Also "frame" the standard more from a perspective of a service person or craftsperson as they are the only ones who are intended to be using this equipment. 9.) If the equipment is a PBX, computer, router, or other type of equipment that can have a dual role as customer owned and operated or telco owned and operated, it would require a listing to UL 60950 (or meeting the requirements of UL 60950) as well as items 1 through 7 above. I believe this would be greatly beneficial to the end users, equipment vendors and the test labs. Of course questions and issues regarding specifics of sections of GR-63-CORE and GR-1089-CORE covered by a new ANSI safety standard could easily be addressed by clarification and changes in the standard. This cannot be done with the Bellcore NEBS documents. Jim Jim Wiese NEBS Project Manager/Compliance Engineer ADTRAN, INC. 901 Explorer Blvd. P.O. Box 140000 Huntsville, AL 35814-4000 256-963-8431 256-963-8250 fax jim.wi...@adtran.com > ---------- > From: Meunier, Éric[SMTP:emeun...@teknor.com] > Sent: Friday, January 14, 2000 5:44 PM > To: nebs@world.std.com > Subject: DC Power Short-ciruit test GR-1089 Section 9.8.1 > > Hi, > I am trying to find out how to interpret the short-circuit protection > requirements of GR-1089 Section 9.8.1 R9-20 amended by Issue ID:1089-9. > > 1. Are the very short traces connecting IC's and discrete components pins > to > high current power planes required to survive a short at the components > pins > end? > > 2. If adequate current limiting can not be provided at the backplane > level, > is it acceptable to place the current limiting device on the circuit pack > when the connector power pin or the trace connecting the power pin to the > current limiting device can not handle the backplane current? > > 3. Similarly to the above question, is it acceptable that a few components > associated with the active current limiting circuit of a circuit pack be > powered by connector pins and traces that would not handle the backplane > current? > > 4. If multiple low current connector pins are used in parallel to connect > the power between a backplane and a circuit pack, is it acceptable that > each > pin be not able to handle the backplane current individually? > > Thanks in advance for your collaboration, > Éric > > =================================== > > Éric Meunier > Architecte de conformité / Conformity Architect > > E-mail: mailto:emeun...@teknor.com > > Teknor Applicom Inc. > 616, rue Curé-Boivin > Boisbriand, Québec > Canada, J7G 2A7 > > Tel: 1-450-437-4661 ext. 419 > Fax: 1-450-437-8053 > > Web: http://www.teknor.com > >