From: "Tony (remove \"_\" from email address)"
Subject: Re: Info on protecting UTP network / phone / AC mains so lightning on cable doesn't destroy everything (again)?
Date: Wed, 06 Nov 2002 09:58:15 +1000
X-Newsreader: Forte Agent 1.92/32.572
Maybe my intention wasn't clear to you. All the protection equipment I can find
(including the one I use) clamp the phone and AC mains. Such equipment is fairly
universally accepted as worthwhile for two reasons - it keeps the phone line and
AC mains near the protective ground, even if that's nowhere near real ground,
thus protecting equipment with no additional risk to people, and, it has MUCH
higher rated protective devices than any built-in protection.
But all the recent faults I am aware of have come in on the cable, NOT the AC
mains - maybe because the cable or network isn't clamped like the others?
Without effective clamp devices, the meagre isolation barrier in a NIC or cable
modem has virtually no effect (on both my cable modem and NIC it was obvious
that the energy just flashed right across the isolation transformers). Had there
been a decent shunt clamp somewhere, all the equipment downstream of the clamp
would have a much higher chance of survival (cable and other impedances limit
the current dramatically compared to your quoted example).
In previous work in the electricity industry I've seen a lot of lightning - in
one case we had to replace over 2000 TVSs installed over solid ground bars on
terminal strips without losing a single I/O channel. I suppose if you had my job
then, you would have relied on the nominal protection in the I/O cards
themselves, and been content to replace every card as nature took it's course
each time there was a lightning event.
I don't have a problem sacrificing equipment in the name of safety, but when a
single external event can take out other apparently distant equipment on my
network (because there is NO effective clamp anywhere), then it means that to be
sure I need to scrap every computer, printer and other peripheral on the
network, now, and each time this happens. However "ideal" that might be, it's
not practical. And it's certainly not practical to believe in the tooth fairy as
a future solution; - the "existing built-in lightning protection" and "arrestors
to clamp the fault currents to ground" are barely rated to cope with ESD
My intention is still to clamp everything to a common ground, adding network
cables to the conventional power / phone protection scheme. I know the
protective ground won't be anywhere near real ground potential during a
lightning event, but as long as everything electrical is at around the same
potential, I feel more confident the damage to both equipment and people will be
I noticed comp.dcom.cabling, but thought that electronics groups would be the
place to ask about components like GDTs and TVSs. Nevertheless, I'll take your
suggestion and air the question on that group too. But please forgive me if I
don't respond further here - I really only wanted discuss exactly what to do,
not WHETHER I should do anything.
On Tue, 5 Nov 2002 04:05:44 -0800, Watson 'Atto Parsec' Name
>In article , "Tony (remove
>\"_\" from email address)" talked about...
>> Hi all,
>> I'm just getting back on air after lightning somehow got into my cable and blew
>> out my cable modem, NIC, printer, CDRW, related ports etc (PC still goes except
>> for the lost ports, but who knows for how long). It seems dozens of neighbours
>> suffered the same or worse fate, too.
>> I therefore want to define a substantial ground plane, say at a network hub,
>> which all network lines, AC mains, phone and other I/O pass over and get clamped
>> to (electrically).
>> AC mains are easy - MOVs and some X/Y filter caps (maybe I'll add some common
>> mode and series mode chokes too, and maybe even a brownout relay to drop it all
>> out if the AC volts sag too much. And of course I'll make it safe for humans
>> Phone protection I assume is gas arrestors in parallel with TVSs; but what
>> ratings are reasonable? and do I need resistors, fuses or PTCs between them?
>> Network is harder - at 100Mb/s I don't feel I can put anything much across
>> without affecting the signal. Maybe shottky diodes clamping to a grunty +/-
>> Cable is harder still - I can't imagine any way to protect it, but I am content
>> to lose the cable modem if another strike occurs, as long as I can reduce the
>> probability of losing other equipment - and of course I know there's nothing I
>> can do about a direct hit (c'est la vie).
>> All suggestions would be welcome.
>You can't stop lightning. You could do everything to the PC and the
>extremely high currents of the lightning will induce enough current into
>some short cable such as the coiled cord of your keyboard, and this will
>burn out your computer.
>So make sure all your equipment is in good woerking order and that the
>already existing built-in lightning protection has not been compromised by
>the previous damage.
>You might consider adding an isolation transformer to the incoming AC line.
>Or else use a good UPS with built-in protection.
>Also, the proper place to post for the network cabling is
>comp.dcom.cabling. There you will get a stern lecture from the fanatic on
>lightning, telling you that the most important thing is proper grounding.
>He has a point, but I think he's a kook and a troll. But your hub and NIC
>already have protection with a transformer to isolate the signal, and
>arrestors to clamp the fault currents to ground. So you don't have to do
>anything. Same with the telephone cable. It should have proper protection
>and grounding, but make sure this is in working order. It could have been
>damaged by the previous strike.
>One more point. Your equipment has already served its purpose. Even tho
>it failed, it did so without causing a fire or loss of life. So be
>grateful that you came thru your previous experience with just a few burned
>out items. If you screw around with things, you may not be so fortunate
> Lightning struck a few hundred feet from my place, and I saw the
>damage it did. It hit a high voltage line on the top of the power pole.
>This cable consisted of 6 strands of #10 (2.6 mm) aluminum wire around a
>center strand of steel the same size. The bolt melted 5 to 10 yards of the
>aluminum wires into foot long pieces and these were strewn all over the
>ground around the strike point. It takes a *lot* of power to do that kind
>of damage. So don't mess with Mother Nature!
>> PS sorry for the cross post, but the subject seemd relevant to all 3 groups.