From @delta.eecs.nwu.edu:telecom@eecs.nwu.edu Sun Aug 18 04:49:48 1991 Received: from MINTAKA.LCS.MIT.EDU by gaak.LCS.MIT.EDU via TCP with SMTP id AA19131; Sun, 18 Aug 91 04:48:56 EDT Resent-Message-Id: <9108180848.AA19131@gaak.LCS.MIT.EDU> Received: from delta.eecs.nwu.edu by mintaka.lcs.mit.edu id aa22807; 18 Aug 91 4:48 EDT Received: from [147.28.0.33] by delta.eecs.nwu.edu id aa32447; 17 Aug 91 2:47 CDT Received: by m2xenix.psg.com (/\==/\ Smail3.1.22.1 #22.2) id ; Sat, 17 Aug 91 00:35 PDT Received: by puddle.fidonet.org (mailout1.26); Sat, 17 Aug 91 00:03:12 PDT Date: Fri, 16 Aug 91 19:36:00 PDT Message-Id: <7752.28ACCA30@puddle.fidonet.org> From: Pat Verner Subject: Lightening Surge Protection To: telecom@eecs.nwu.edu X-Mailer: mailout v1.26 released Resent-Date: Sun, 18 Aug 91 3:51:11 CDT Resent-From: telecom@eecs.nwu.edu Resent-To: ptownson@gaak.LCS.MIT.EDU Status: RO Hi Pat, The South African Highveld has one of the highest lightening occurences in the world, and to operate equipment without protection can be very costly. The following article was written a short while ago, and I am forwarding it to you with the authors permission after seeing several complaints in your newsgroup! It is a bit long, and I have shortened it a little, but you may still find it of general interest! --------------- > Cut here <------------------- TELEPHONE LINE SURGE ARRESTOR (C) 1990 Tim Jackson P.O. Box 199 Cresta, RSA 2118 Tel: +27-11-476-1346 Fido: 5:7101/1.17 What follows is a bit on why most lightning protection schemes for modems and other such devices (cordless phones, answering machines, fax machines etc.) fail at what they were intended to do and how you can build a circuit yourself that will work as well as any circuit for the purpose can. Firstly let me say that the circuit itself is not unique at all. There are a number of people such as the manufacturers of PABX equipment and others who use either the same circuit or a variation thereof. The important point is that this circuit WORKS! There are others, some very similar, that don't work nearly as well. Also it is critically important to understand HOW lightning causes the damage it does. Once you get this you'll see why most protection schemes fail and you'll have a good idea of how to avoid damage. There are schemes for protecting your equipment such as Ken Burke's patented Spark Gap (R) system. This is a system whereby the modem is unplugged from the phone line and the modem's telephone jack is left lying on the desk or floor thus creating a spark gap between itself and the wall socket of about half a metre. There is no doubt that this system is very effective although it won't give much protection against a raving lunatic with a fourteen pound hammer. Neither will my circuit but I'm working on it........ The problem with the Spark Gap (R) system is that it is manually activated and thus subject to Murphy's law. Murphy's law of Spark Gap (R) systems is that you'll remember to unplug your modem ninety nine times when there is actually no need to and on the hundredth time, when you rush out in a helluva hurry and forget to unplug it lightning will zap the blasted thing. There was another system written about on this board a while back which was designed to automatically disconnect your modem from the line by means of a relay powered by your PC's PSU so that when you switched off your PC your modem would be automatically disconnected. This is a good system for those who do switch their PC's off when not in use provided you follow the instructions carefully regarding earthing and use a relay with well-spaced contacts. The details are available on Golden City Opus (and other boards, I'm sure) under the name of TRANZORB.ZIP. The problem is that while it does have surge arrestors in it, these would only give limited protection. I have three phone lines (I work from home) which between them have a fax machine, a modem, an answering machine, a recording unit and several "fancy" phones connected to them. Clearly, it would be tedious to unplug this lot every time Annie hid in the cupboard. Annie is one of my cats who is scared to death of thunder and hides away when she detects a rumble five hundred kays away. A couple of years ago I stayed in Kyalami where the telephone lines are strung up on poles for miles and miles just inviting lightning to zap them. Every time there was a storm I would quickly set up my stuff and experiment on various methods of "taming" the phone line. I can honestly say from practical experience and as a result of many incinerated attempts that the circuit presented here works and works well. Before we get to the circuit itself let me squash a few myths and establish a few facts: I assume that anyone reading this is at least slightly technical. There is no difference, for practical purposes, between internal and external modems when it comes to lightning damage. Most people are under the impression that if your modem is external the lightning is somehow contained. Bongggg. You provided a wonderful path right into your PC when you connected your external modem to your serial port. The fact that part of the package (the modem) is outside your PC and part (serial card) is inside means zilch. As far as lightning is concerned there is no difference. It is true that SOME external modems have better lightning protection circuitry than their internal counterparts but that is incidental. Now for the part that most people (and I'm talking about suppliers here too) don't realise. Your phone line consists of a pair of wires across which a fairly wide voltage range appears depending on exactly what's happening. The voltage will typically be between about 10 Volts (when you're using the line) and a maximum peak voltage of around 200 Volts (when it's ringing). The voltage ACROSS the line is one thing. The voltage between the line and earth is another thing altogether. The problem is that most people provide protection for a voltage surge across the line, such as the gas arrestor found in SAPO plugs, but almost no-one worries about the voltage between the line and earth and this is what does the real damage. In Telcospeak the two wires of a phone line are referred to as the "A" and "B" wire. If you are running a bog standard phone (that doesn't mean a bathroom phone) then protection against surges across A and B is adequate. Hence the gas arrestor in SAPO telephone plugs. Problem is, as soon as you connect a device to the phone system that is also connected to your electrical mains you are introducing a third "line" and that is EARTH. Devices in this category include fax machines, modems (internal modems are connected to mains by virtue of being in your PC), answering machines, some fancy telephones, cordless phones etc. What happens is that although the modem (I'm not going to list all the other devices each time) is not connected directly to 220 Volts it is fed power via a transformer which IS connected to mains. Now the transformer has a certain insulation between the primary and secondary windings but if you force the secondary to a potential a few thousand Volts higher than earth potential then the insulation will break down and suddenly your modem IS connected directly to 220 Volts, even if it's just briefly. Although your phone line is at earth potential at the exchange end it is ripped up to several thousand volts at your end if lightning strikes nearby. This high voltage, although not necessarily between the A and B wires, is certainly felt between A and B as a unit, and earth. In other words a high voltage is induced into the line pair pulling the line way above earth potential. What happens is that this voltage finds its way through your modem to earth by one of many possible routes often with catastrophic results. The way to prevent this happening is to provide a path from your phoneline to earth for the surge to follow without going via your modem. The circuit at the end of this blurb does just that. It uses a chip from Texas Instruments which was designed for just that purpose. It is essentially transparent when the line is at normal working voltages (up to about 200 Volts) but above that voltage it shorts your line to earth for as long as the surge lasts. The trick is to install the unit in the line between the telephone jack and your modem (ie: not too far from the modem, like in another room) and connect the earth lead from the circuit to the earth pin on the SAME PLUG that feeds your PC. You can connect it to the chassis of your PC if you find that easier. Just check that the chassis is connected to earth. I've yet to find a PC where it wasn't but just check anyway. ================================================================== 10 ohm 5 Watt >----------+-----/\/\/\/------------------+-------------> Phone Line A | | <=>300V Gas Arrester ____|_____ To Modem | | A | | |TISP2290| |--------+--------------------------|C | Earth | | B | | ----+----- <=>300V Gas Arrester | | | | | To Modem Phone Line B | | >----------+-----/\/\/\/------------------+-------------> 10 ohm 5 Watt Circuit drawn as best as possible with ASCII by Pat Verner. ===================================================================== The telephone line enters the circuit on the left of the diagram and the feed to your modem is on the right of the diagram. What happens in between is this: The phone line has a gas arrestor from each leg to earth. In other words, two gas arrestors. One from A to earth and one from B to earth. The line then has a resistor in series with each leg (A and B) before being connected to the TISP2290 (the Texas Instruments chip mentioned earlier). This chip has three pins. The outer two (A and B in the diagram) are connected to the resistors while the centre one (C) is connected to earth. The metal tag of this component is internally connected to the earth pin (C), just for the record. The modem is fed from the outer two pins of the TISP2290. The way the circuit works is as follows. The bulk of the energy involved in a surge is dissipated by trusted (and slow as treacle) gas arrestors. The TISP2290 absorbs the high speed spike that the gas arrestors miss and is itself protected by the two resistors which provide a little current limiting. The modem, being fed from the same point as the TISP2290 is protected by the whole circuit. For those who have to know, the TISP2290 works in a manner similar to a zener array connected between the A and B wire and earth so as to limit the voltage between any of three points to about 200 Volts. As you know this is not entirely effective and so if the voltage rises to 290 Volts (hence TISP2*290*) then this crafty critter cuts in triacs to crowbar the offending points to earth until the surge has passed. This all happens superfast so as to afford the best possible protection. Just a bit about the individual parts. The Gas arrestors SHOULD be available from most electronics hobbyist shops. Anything in the range 250 Volts to 400 Volts will do. The resistors are 5 Watt wire wound jobbies of 10 Ohms although any power rating will do (be prepared to change them every lightning strike if you use quarter watt jobs :-) I recommend you use 5 Watt wire wounds) and the value of 10 Ohms can actually be anything from 10 to 47 Ohms. The TISP2290 may be hard to find. It comes in a TO220 package (looks like a 78 series voltage regulator) in case you have to explain it to anyone. The agents in JHB are Multikomponent (Tel: (011) 974-1521) and I'm sure they will send you one COD (I'm not sure about this) or tell you where you can get one if your local electronics shop dude goes blank when you ask him. I buy directly from them but then I'm in JHB and you might not be. Assembling the unit is piece of cake. I have supplied the artwork for a printed circuit board and, while not essential that you use it, it makes life easier. ********************************************************************** --------------------> End Cut <-------------------- -- uucp: uunet!m2xenix!puddle!5!7101!22.6!Pat.Verner Internet: Pat.Verner@p6.f22.n7101.z5.fidonet.org