Slashdot Mirror


How to Protect a Network Against Lightning?

RichiH asks: "The monsoon, started about a month early in India this year. While it is not sure if that is due to global warming or not, there are more pressing issues for the IT world at hand. Until about the end of July, there will be major thunderstorms in this area. How do you protect a network that is spread over 100 square kilometres in a land where the concept of a lightening arrestor is next to unknown? The network in question consists of about 2500 boxes of various kinds which are connected using 10BASE2 (aka BNC), 10BASE-T (aka RJ45) and 10BASE5 (aka thicknet), where only the last one may be new to some readers. The big question is: how can you protect yourself against these storms in a way that is both fast to implement and does not require laying of new lines?"

3 of 120 comments (clear)

  1. Cheap Hardware by Micro$will · · Score: 5, Interesting

    A few years ago my friend and I decided to set up a network between our houses. We ran about 200 feet of 10BASE2 along a fence and used an old 486 DX33 box on each end as transparent bridges between the cable and out LAN segments. Every once in a while we'd get a close lightning strike in the summer and it would fry one of the combo cards we used. Fortunately, they were old Linksys NE2000 compatable ISA cards I picked up used for about $2 each. I'd go through about 1 or 2 every year. I tried using a spark gap type arrestor, but it wasn't enough, besides a few bucks a year was worth it.

  2. Re:Lighting tips by tzanger · · Score: 3, Interesting

    We use substation-class arrestors to protect or industrial motion controls from direct strikes. Not cheap, you're right, but they do work.

    Something else to keep in mind is that unless you're talking about spark-gap or gas discharge type arrestors (i.e. anything like that will be SPECIFICALLY mentioned on the box), you're dealing with Metal Oxide Varistors (MOVs) and the protection should be REPLACED after every major storm since you cannot practically test if the MOVs will clamp properly again. The only way to test them is to hit them with enough voltage to cause them to clamp, but you just hit them and now don't know if they'll clamp again next time. :-)

    Oh, and MOVs, when hit with sufficient Joule energy will turn into a beautiful plasma cloud. Plasma is conductive. We used to get units back that would have survived a nearby strike if the plasma cloud hasn't bridged two phases and caused a line-to-line short which blew the shit out of the unit since the short happened before the fusing. :-)

  3. Re:WiFi by rcw-home · · Score: 2, Interesting
    Lightning acts like a very large spark-gap transmitter. As we all know, spark-gap transmitters throw out nasty RF over several freqencies

    s/several/all/

    Picture a lightning bolt. It's white, right? White is the sum of all colors. White (RF) noise is the sum of all frequencies.