Richard, You mention several safety standards but not the pertinent EMC standards. One is the EN 61000-4-4 or IEC 1000-4-4 which is called the EFT/Burst test and it has a test for AC mains and DC mains included. For my equipment at overvoltage category III the +/- 2kV burst is appropriate. In addition there is a voluntary test for Surge EN 61000-4-5 or IEC 1000-4-5 which puts a 1.5/50uS open circuit voltage pulse on the mains. Surge is for lightning immunity and is similar to the impulse test in the safety standards. If you wish to learn more about the testing for the various points on the distribution network check the surge standard however the definitions there do not directly correspond to categories I, II, III, IV. They are what are called installation classifications 0 through 5.
Generally 5 is for overhead power lines, 4 is for outdoor cables, 3 is for a power installations and cables run in parallel, 2 is where cables are well separated and there usually is a transformer used for first level isolation, 1 is for circuits with primary overvoltage protection and essentially not influenced by lightning, and 0 is for well protected areas. Please do not these remarks alone for making decisions as Annex A and B in the standard use several pages to cover the information. It is possible to build your own test equipment using the schematics shown in the standard however calibration can be an issue and there a few manufacturers producing these today. You can find these advertised in most trade magazines. By the way, the "alternative" method for achieving clearance is as I understand, requires 100% routine testing, whenever you design to homogenous field spacings and not the inhomogenous field numbers. If you go this route you must have one of these impulse testers in your production area for testing every production unit. Typically I have used the industrial safety standard prEN 50178 but on occasion use other IEC 664 derivatives such as EN 60950 / IEC 950 and EN 61010-1 or IEC 1010-1. All of these mention impulse testing under certain circumstances. I don't believe that you can just throw away any regard for any spacings but you can get significant reductions in space. Concerning the use of MOV's, the EMC and Safety technical committees are at odds on this one. Safety people hate these and EMC people love them. Unfortunately I have to wear both hats. Fortunately there are a few VDE approved MOV's on the market, unfortunately they are all limited to circuit protected to 16 Amps (I need up to 200A). Gas suppressor tubes have a limited life and transorbs are semiconductor material (another discussion). One possibility is to build the MOV circuit with a thermal fuse attached to the body to prevent overheating and sputtering of metal. I think there is a company in Japan that makes these. The problem here is that if the fuse blows you now have an undetected fault and have lost any surge protection. For a price monitoring circuits can be added. Placing MOV's in an enclosure or plastic box only creates an explosion problem by trying to contain the expansion. At my company we have been able to use MOV's in the "electronic" portion of the equipment with success by using the test guidelines of UL 1449 (new version is available) and by performing abnormal fault tests in the product application. Pass criteria: no fire, no products of combustion, no ejecting of material, no melting insulating material, etc., etc. This is all done at "worst case normal operation" meaning in most cases high line and minimum loading of the mains. One other possibility that I have not tried yet is the Phoenix Contact "Trabtech" products, catalog #7. These are surge protection products that are designed with spark arrestors and do not rely on special gases. Made in several configurations including power strips, outlets, din rails, etc. Call 717-944-1300 (US) or 905-890-2820 (Canada) I do not have a European number. Once again, comments are mine... Doug Powell Advanced Energy Industries, Inc. [email protected] ---------- From: Richard Hughes To: [email protected] Subject: Data on mains transients requested List-Post: [email protected] Date: Thursday, July 24, 1997 9:48AM Dear Safety & EMC folks, I am interested in obtaining data on the number, duration and amplitude of mains transients at different parts of the mains distribution network. It would be great if someone could point me to information published by the IEE, IEEE or whoever. IEC 664-1 classifies different parts of the mains distribution network as being Installation Category II, III, IV etc. IEC 950 re-states IEC 664-1 in saying that equipment plugged into a normal socket outlet in an office or home is Installation Category II: hence the clearances in Clause 2.9.2 of IEC 950 are based on 2.5 kV transients for a 230 V supply. I am also aware that various EMC (well, immunity) standards require equipment to be hit with transients, but not being an EMC engineer I cannot quote standard numbers or values at you. It is one thing knowing the assumed maximum transient voltage, but that is not sufficient to work out the amount of energy to be absorbed by an MOV or other such device during its working life. I have seen MOVs of various sizes used in power supplies, but just how big is big enough I dont know (but Id like to find out). For those of you that dont keep up with standards let me say that the draft 3rd edition of IEC 950 has an Annex that provides an alternative method for determining minimum clearance distances that includes a test for injecting the assumed transient at the mains input and then measuring what the resultant transient at the insulation under consideration. Unfortunately there is also a requirement that "surge suppressors in PRIMARY CIRCUITS are disconnected". This restriction may be due to the nasty habits that MOVs can get as they get older (i.e hot and leaky) - but then maybe devices that exhibit this behavior were not adequately rated in the first place?! Information and constructive comments please. Thanks (I hope!) Richard Hughes Nortel plc (UK) (By the way, this is all the signature I think is needed, but no doubt someone will correct me).
