Lightning protection
L. M. Rappaport
local telco serviceman what they do to protect the CO. He said that
in addition to gas tubes, they use something called "sneak current
devices" or something like that. In essence it's a fast acting
circuit fuse which monitors line current and blows if it exceeds some
amount.
Does anyone know what these are and where I might be able to buy some
along with whatever is necessary to contain them? I have four lines
here at the moment.
Btw, besides the gas tubes the telco provides, I have two sets
individual protectors, but my modems seem to be unusually sensitive to
induced strikes - I use Courier v.everythings, btw.
Larry
--
ra...@lmr.com
Macy Hallock
+---------------
> After losing yet another modem to induced lightning strikes, ...
> [snip]
> Btw, besides the gas tubes the telco provides, I have two sets
> individual protectors, but my modems seem to be unusually sensitive to
> induced strikes - I use Courier v.everythings, btw.
Don't know if your hits are coming in on the phone line or power line.
Could even be a combination of the two, especially if one or the other
is not properly grounded. It's entirely possible you have more than
one deficient situation at your location that needs to be solved.
Protection is only as good as its grounding.
Have the telco check your protector ground with a "megger".
A water pipe is not a good ground for this purpose. A good ground
rod (stake) is. Should be a minimum of 6', better to be 10'
and should be 3 feet from the foundation (common mistake).
The telco installer will complain about this, but will do it
if you insist, especially if you tell them you've not only had
computer/fax/answering machine damage, but also been shocked
by the phone during a storm. It's really not a good idea to
bury the top of the ground rod, the top connection should be
visible so it can be inspected.
Ask the phone tech to use a 3 element type gas tube protector,
if they have them available. Most, but not all NID's use
these today. Many older models were not 3 element.
Be sure you have a good power ground, also can be checked with
a megger. The telco and power should not share the same ground rod.
National Electrical Code says your panel must be grounded to
a grounding rod (stake) and the inside water line (if it's not PVC).
Consider installing a secondary power protector in your power
panel main breaker. Joslyn, TII and SquareD make good ones
(These are most commonly used to protect submersible well pumps
and are readily available at electrical supply houses.)
Get a secondary lightning protector with a real toroid type inductors
in it. TrippLite and APC's best high end units have these (they are
much heavier than the less expensive units.) Ask if you are not sure,
and don't trust a discount or computer store clerk, talk to the mfr.
first and know the models you want.
A few of these high end models have fairly good phone line protection
in them, but I'd still use a separate unit, like the APC in line
ProtectNet series. Again, check with the mfr.
All major protection vendors have websites with 800 numbers. Tell
them you have a persistent and recurring problem. Don't hesitate
to ask for an experienced engineer since you have a persistent issue.
For the best protection, run a separate ground wire directly
from your power protection ground rod connection at your power panel,
using the correct type of grounding connection clamp. (Some
severe cases use a completely separate ground rod.) Use solid
12 or 10 AWG wire, run it in a straight line, with a minimum of
bends, and no sharp 90' bends if possible. Connect it directly to
the grounding connection of your protector (most really good ones
will have a ground screw connection, or use a case screw on a metal
unit like the TrippLite.)
EFI and DiTech make even better lightning protection for phone
and power, but they are a little harder to find.
FYI: Quality UPS's have good lightning protection, cheap models do not.
--
Macy M. Hallock, Jr. N8OBG 216.241.7166 fax:216.241.7522 ma...@apk.net
APK Net, Inc. 1621 Euclid Ave. Suite 1230 Cleveland, OH 44115 USA
Rich Campbell
Rich Campbell
"Macy Hallock" <ma...@apk.net> wrote in message
news:9i4d31$k...@newstaff.apk.net...
w_tom
to earth ground, either directly or through a surge protector, before it
enters the building, then effective surge protection is missing. But this
is important. The phone line ground and the electric line ground must be
same - as required by the National Electrical Code and also by UL 497(?).
Both ground wires must be separate until they meet at central earth
ground.
Fuses don't protect from surges. Fuses only blow after damage is
created and only so that damage does not create a fire. Actually, your
local serviceman is talking about obsolete technology. For over the past
15 years, telco protection has been semiconductor based and, as stated
elsewhere, is only as good as its earth ground.
To learn from previous modem damage - to avoid all future modem
failures: trace the incoming and outgoing surge paths. The incoming path
exists because a direct lightning strike was not shunted to earth ground
when it entered the building. AC electric lines are the most common
source of incoming, modem destructive surges. Also any surge protectors
adjacent to the computer may contributed to surge damage.
Why are modems being damaged? What is your central earth ground. How
good is it? What kind of soil is it installed in? How long are every
connection to central earth ground? Is that transformer on the nearest
telephone pole still connected to earth ground - the wire to ground still
intact? What other earth grounds exist - ie water well? These are the
first questions you must ask. These are the first of many reasons why
modems can suffer damage.
L. M. Rappaport
Thanks for the advice, but perhaps I should have mentioned a bit
more. In NH, at least using Verizon, the telco ground is connected to
the power line ground at the service entrance of both. That is
Verizon's requirement (and as you said, also UL and the NEC). My
ground is about as good as it gets. The power line ground is further
connected (with no loops) to at least 6 other ground rods (an 8' rod
buried 2' below surface) by #2 or larger copper conductor. Those
grounds include a 70' tower with three sets of guy wires each
separately grounded. Soil is pretty good - not a lot of rock or sand
and several of the rods are in wet, loamy soil.
The house has multiple layers of surge protection and I've had no
problems with any of the stuff within. Those layers start with a
fairly massive device at the service entrance and go down to simple
outlet style surge protection at various outlets.
Since the only device in my home office which seems to suffer is
the computer modems which sit on one line of the four incoming.
Besides the gas tubes installed by Verizon, I have a Cutler Hammer
telephone line protection device on each line.
The telephone guy said that the NID is the only one they have. He
was the one who suggested that I get these "sneak current" protectors.
Yes, they are some kind of fuse and they do have to be replaced. He
said there are boxes of them in the CO and each time a workman goes in
there, he is supposed to check a panel and replace those which have
blown. He said this is the secondary protection system for the switch
and recommended I get them and use them as well.
Now, here is my take on what's happening: The only modem I use
are Courier v.everythings as they perform the best up here. (I've
tried just about every one made, btw.) We're in the country - it
doesn't take a direct hit to a telephone line, induced voltages occur
simply from lightning occurring - even if it's a purely cloud to cloud
occurrence. (Trust me on this, I've monitored the lines during
thunderstorms) The gas tubes have a flashover of something like 3 -
500 volts according to the Verizon serviceman. I believe that is too
high to effective protect my equipment. My interior protection looks
ok, but I don't have a megger to test it - maybe I better get one.
I guess I'd like to know specific brands of protection which you
folks use. I'm not familiar with telephone equipment, but I'm an EE
with experience protecting radio equipment (I've had several direct
strikes hit my tower with no equipment damage) so I understand the
issues, just not the products out there for telephony.
Thanks
Larry
--
ra...@lmr.com
George Driskill
get some info regarding the "bubble of protection." They have some devices
that will protect both your phone lines and your power lines and they'll
insure your installation. Again, however, there's nothing you can do against
a close lightening strike, except to disconnect all your equipment.
2. I'm only a low-voltage contractor, but I am pretty sure you can have an
electrician drive the 10-foot ground rod, like someone else suggested, and
connect it to your power service entrance ground. This will really ground
your power entrance and the GTE (veri-zone) protector.
3. There's no law against installing your own protector in-line after the
telco protector.
4. The heavy-duty power protection company is Porta Systems, 575 Underhill
Bl., Syosset, NY 11791, 516-364-9300, or 1-800-93PORTA. I think they're the
guys who make the stuff for the telco's.
George Driskill
Driskill Communications
geo...@driskill.com - EMail
32320...@pagenetmessage.net - EMail Pager
415.892.2659 Office
707.795.6674 Same Office
707.664.8282 Fax
707.316.3020 Pager
707.321.3775 Cel
"L. M. Rappaport" <ra...@lmr.com> wrote in message
news:3p3ekt8ppe3ccnalu...@4ax.com...
Macy Hallock
+---------------
> 1. Call/visit Panamax (http://www.panamax.com)
They make very good protectors. Their high end solid state stuff may help.
> 4. The heavy-duty power protection company is Porta Systems
They make good primary and secondary protectors. They have a line of
modules that combine gas tube and solid state technology that work
quite well.
Another vendor for retrofit secondary solid state protection is ITW-Linx.
All these items are available from Graybar and other major telecom
supply houses.
One more suggestion...for a really hard to solve lightning problem,
Poly Phaser is the company many pro's turn to. They have literally made
a busienss out of playing with lightning. They make some incredible
stuff, mostly for RF coax but they have some telecom items as well.
Some good info and papers on their web site http://www.polyphaser.com
w_tom
technology installed in those Verizon NIDs. GDTs degrade with use and leave
no indication of their degradation. It is one reason why NIDs use
semiconductor based surge protectors that trigger at about 320 volts. Each
line should shunt any 320+V surges to earth ground if the Verizon NID is
connected less than 10 feet to the central earth ground, a green wire not
passing through any conduits, no sharp bends, etc - all the other good
installation procedures used in RF installations. Your ground system is
substantial - should more than meets the definition of adequate.
With that NID and with your ground network, then the surge is not entering
on the telephone line. Furthermore most modem surges don't come from the
telephone line anyway. Modems are most often damaged by surges that seek that
phone line earth ground.
The "fairly massive device at the service entrance" is what? Is it also
connected less than 10 foot, direct, not through conduit or metal pipes, no
splices, sharp bends, etc to the same central earth ground? The
effectiveness of that service entrance unit is, like the NID, a direct
function of how it connects to earth ground.
A problem with your installation are all those 'layers' of surge
protection. Layering only works when each layer also connects less than 10
feet to earth ground. Layers of plug-in surge protection can instead create
massive ground loops - the induced currents so often discussed in papers. If
connected direct to earth ground, then layering works. If simply connected to
other wires (such as the safety ground wires which should not be considered
connected to earth ground - too long), then those 'layered' surge protectors
can create massive ground loop problems.
A surge protector only shunts a surge from one wire to all others. Networks
of such devices simply make a big interconnected network of bypass and ground
loops that create havoc to surge protection. Montandon and Rubinstein discuss
this in their paper "Some Observations of the Protection of Buildings Against
the Induced Effects of Lightning". Their conclusion point three notes, "Do
not establish equipotentialization by multiple bonding of sensitive power of
data cables to different potential reference points within the structure."
Those layers of surge protector may be doing just that - shunting surges to
different potential reference points.
Lets look closer at your modem and computer for example. We can assume (as
long as the Cutler Hammer unit is also grounded less than 10 feet to earth
ground) that the phone line is not the incoming surge path. It has surge
protection connected short to a superior earth ground. So what could be the
other incoming surge paths? Green wire ground wire is one. Connected to many
other devices that might shunt the surge into the green wire safety ground
network, then a surge may simply choose the best path to earth ground, through
computer and modem to that excellent phone line earth ground.
Computer network. Does it exist? What is it connected to? IOW in another
newsgroup, lightning crashed the guys computer (without damage) because it
entered via powered off computers elsewhere on the network, entered his
computer via the NIC, passed through the motherboard ground plane (creating
enough noise to crash the computer), then out via phone line or another
connection to earth ground.
Computer network could be one incoming surge source. What else does this
computer connect to? Montandon and Rubinstein also make conclusion point 4,
"Follow within the building a bonding and routing concept to interconnect
different equipment by power and data cables according to the principle shown
in Figure 12". Figure twelve shows all cables interconnect only in one
bundle (not a lot of separate wires) so that the interconnecting connection
first contacts a voltage reference point AND that voltage reference point
connects directly, a dedicated connection, to earth ground.
IOW we are not just looking for a incoming surge source on some specific
service such as CATV. We seek that possible incoming source due to transient
voltage variations between different equipment in the building.
What about a satellite dish. Does it also enter the building first
connected to that central earth ground window - your power line and phone line
earth ground system? Again, I am seeking other incoming sources that would
use those layers of surge protectors, etc to enter your computer.
I would reassess all those 'layers' of surge protection that could actually
create more problems because surge protectors only shunt a surge from one wire
to all others. IOW a common mode surge not shunted direct to earth ground
creates induced surge problems as cited in that IEEE Transaction on
Electromagnetic Compatibility article of 4 Nov 1998.
POTS modems are most susceptible to surge damage. Although they have a
2000+V isolation transformer, they also have bypass circuits. For example,
the off-hook relay is typically rated at only 500 V which is one common path
used by surges from motherboard to phone line. IOW search for the incoming
source of that surge; the outgoing path being the phone line. Where can that
surge income from? What are all other wires that could carry an incoming
surge?
John Dearing
:
: With that NID and with your ground network, then the surge is not entering
: on the telephone line. Furthermore most modem surges don't come from the
: telephone line anyway. Modems are most often damaged by surges that seek that
: phone line earth ground.
This is a hard thing to explain to most folks. That the modem connected to
the phone line was blown out by a voltage surge entering through the AC
power supply but *exiting* through the phone line.
At the frequencies involved in lightning strikes, the telephone circuit
looks like a fairly low impedance "ground" to the higher vottages induced
onto the AC line.
Since the higher voltage is always looking for a way to ground (now from
*inside* the modem) the phone line is as good a place as any to go. Then
you have a dead modem.
This was expalined to me at a presentation by APC sometime ago at one of
our PC User Group meetings.
Very interesting presentation they had. Most of the presentation was
general in nature, educational. Only the *very* end could be considered a
"sales pitch" for their products.
John
--
John Dearing : Phila Area Computer Society http://www.pacsnet.org
Email : jdearing "at" netaxs "dot" com
L. M. Rappaport
> If the Verizon stuff is still using GDTs, then it is not the current
>technology installed in those Verizon NIDs. GDTs degrade with use and leave
>no indication of their degradation. It is one reason why NIDs use
>semiconductor based surge protectors that trigger at about 320 volts. Each
>line should shunt any 320+V surges to earth ground if the Verizon NID is
>connected less than 10 feet to the central earth ground, a green wire not
>passing through any conduits, no sharp bends, etc - all the other good
>installation procedures used in RF installations. Your ground system is
>substantial - should more than meets the definition of adequate.
The NID is a modular unit which accepts 6 incoming lines. Since I
only use 4, I moved the offending line to a new protector, so
hopefully, if the old one was degraded, it's no longer a factor. Btw,
the technician used two different instruments to check the old one and
both said it was good. Was one a megger? I don't know, both were
proprietary and I have never seen either one.
> With that NID and with your ground network, then the surge is not entering
>on the telephone line. Furthermore most modem surges don't come from the
>telephone line anyway. Modems are most often damaged by surges that seek that
>phone line earth ground.
Phone line and power line grounds are the same - they are bonded
together and that bond has not corroded - at least not that I can see.
>
> The "fairly massive device at the service entrance" is what?
There are two of them. The first is a set of 3 300 Joule MOV's. They
are installed at the disconnect where neutral and ground are bonded,
less than 6 feet from the ground system. They are connected each leg
to ground and leg to leg (a higher voltage MOV). The second is at the
distribution panel. I believe it is a 400 Joule device connected
through a double breaker. That panel, however is 25 ft. from the
service entrance.
>Is it also
>connected less than 10 foot, direct, not through conduit or metal pipes, no
>splices, sharp bends, etc to the same central earth ground? The
>effectiveness of that service entrance unit is, like the NID, a direct
>function of how it connects to earth ground.
Yes, see above.
> A problem with your installation are all those 'layers' of surge
>protection. Layering only works when each layer also connects less than 10
>feet to earth ground. Layers of plug-in surge protection can instead create
>massive ground loops - the induced currents so often discussed in papers. If
>connected direct to earth ground, then layering works. If simply connected to
>other wires (such as the safety ground wires which should not be considered
>connected to earth ground - too long), then those 'layered' surge protectors
>can create massive ground loop problems.
I understand what you are saying, but I don't think it applies in this
case. What happened in this latest storm is that there was one and
only one strike. It was cloud to cloud (I actually saw the damn
thing). Further, judging from the following thunderclap, it was
several miles away (guessing about 10 seconds x 1100 fps = 11,000 or
about 2 miles). IAC, I am not sure I completely agree with what you
say above. (no flame here, I'm trying to learn). In radio work, we
would say that the problem is one of impedance of the ground system.
That is why we coil incoming cables (to raise the impedance of an
incoming strike) and is due to the fact (that you probably know) that
lightning voltage is basically AC as the charge keeps getting
transferred over the ionized path until it has dissipated. If we are
talking about a substantially reduced voltage - or at least clamped to
300 volts or so, wouldn't the increased impedance apply to both signal
and ground wires, making the use of "layers" at least effective at the
site of the device?
> A surge protector only shunts a surge from one wire to all others. Networks
>of such devices simply make a big interconnected network of bypass and ground
>loops that create havoc to surge protection. Montandon and Rubinstein discuss
>this in their paper "Some Observations of the Protection of Buildings Against
>the Induced Effects of Lightning". Their conclusion point three notes, "Do
>not establish equipotentialization by multiple bonding of sensitive power of
>data cables to different potential reference points within the structure."
>Those layers of surge protector may be doing just that - shunting surges to
>different potential reference points.
I guess I better read this paper. Do you know where I might?
>
> Lets look closer at your modem and computer for example. We can assume (as
>long as the Cutler Hammer unit is also grounded less than 10 feet to earth
>ground) that the phone line is not the incoming surge path. It has surge
>protection connected short to a superior earth ground. So what could be the
>other incoming surge paths? Green wire ground wire is one. Connected to many
>other devices that might shunt the surge into the green wire safety ground
>network, then a surge may simply choose the best path to earth ground, through
>computer and modem to that excellent phone line earth ground.
>
> Computer network. Does it exist? What is it connected to? IOW in another
>newsgroup, lightning crashed the guys computer (without damage) because it
>entered via powered off computers elsewhere on the network, entered his
>computer via the NIC, passed through the motherboard ground plane (creating
>enough noise to crash the computer), then out via phone line or another
>connection to earth ground.
Yes, I have a computer network. There are currently five machines on
it. A laptop was turned off. Another workstation was off. The Win
2k server was on. The file server was on and my workstation (with the
modem) was on. There are no outside connections except for the modem
in my workstation. Another point - all machines are within 25 feet of
each other.
>
> Computer network could be one incoming surge source. What else does this
>computer connect to? Montandon and Rubinstein also make conclusion point 4,
>"Follow within the building a bonding and routing concept to interconnect
>different equipment by power and data cables according to the principle shown
>in Figure 12". Figure twelve shows all cables interconnect only in one
>bundle (not a lot of separate wires) so that the interconnecting connection
>first contacts a voltage reference point AND that voltage reference point
>connects directly, a dedicated connection, to earth ground.
>
>IOW we are not just looking for a incoming surge source on some specific
>service such as CATV. We seek that possible incoming source due to transient
>voltage variations between different equipment in the building.
No CATV.
>
> What about a satellite dish. Does it also enter the building first
>connected to that central earth ground window - your power line and phone line
>earth ground system? Again, I am seeking other incoming sources that would
>use those layers of surge protectors, etc to enter your computer.
Yes, there is a satellite dish located 150 feet from the building. It
enters through the building via some protection devices (MOV's, I'd
guess) and a dedicated grounding rod along with a cable which bonds to
the main ground system. (yes, I know, a potential ground loop, but
that's what the mfg. said) However, that is not connected in any way
to anything other than the sat receiver, etc.; i.e., no connection to
the computer network.
>
> I would reassess all those 'layers' of surge protection that could actually
>create more problems because surge protectors only shunt a surge from one wire
>to all others. IOW a common mode surge not shunted direct to earth ground
>creates induced surge problems as cited in that IEEE Transaction on
>Electromagnetic Compatibility article of 4 Nov 1998.
>
> POTS modems are most susceptible to surge damage. Although they have a
>2000+V isolation transformer, they also have bypass circuits. For example,
>the off-hook relay is typically rated at only 500 V which is one common path
>used by surges from motherboard to phone line. IOW search for the incoming
>source of that surge; the outgoing path being the phone line. Where can that
>surge income from? What are all other wires that could carry an incoming
>surge?
Where can they come from? That's what I'm trying to figure out and
eliminate!!! My best guess is that the surge comes from the telephone
lines where it was clamped to say 300 - 500 volts (telephone lineman
said it could be that high even with gas tubes) and that level was too
high for the Courier. Note: both modems which have died were Courier.
Another cheaper modem on a free-standing (non-networked) computer
which is on the same telephone line has survived both strikes. Based
on your post, I took a look at the power protection strip on that
machine, but there are no obvious physical or electrical problems.
Thanks for your help.
Larry
--
ra...@lmr.com
L. M. Rappaport
>w_tom (w_t...@usa.net) wrote:
>:
>: With that NID and with your ground network, then the surge is not entering
>: on the telephone line. Furthermore most modem surges don't come from the
>: telephone line anyway. Modems are most often damaged by surges that seek that
>: phone line earth ground.
>
>This is a hard thing to explain to most folks. That the modem connected to
>the phone line was blown out by a voltage surge entering through the AC
>power supply but *exiting* through the phone line.
>
>At the frequencies involved in lightning strikes, the telephone circuit
>looks like a fairly low impedance "ground" to the higher vottages induced
>onto the AC line.
>
>Since the higher voltage is always looking for a way to ground (now from
>*inside* the modem) the phone line is as good a place as any to go. Then
>you have a dead modem.
...snip
Actually I understand it quite well. It is a simple application of
ohm's law except that both resistance and reactance have to be
considered. However, one would think that the AC protection at the
service entrance should rule that out.
The problem is further compounded by the fact that there is no
physical damage indicated at any of the protection involved and even
the modem "looks" ok.
Thanks,
Larry
--
ra...@lmr.com
Roy McCammon
> The telephone guy said that the NID is the only one they have. He
> was the one who suggested that I get these "sneak current" protectors.
> Yes, they are some kind of fuse and they do have to be replaced. He
> said there are boxes of them in the CO and each time a workman goes in
> there, he is supposed to check a panel and replace those which have
> blown. He said this is the secondary protection system for the switch
> and recommended I get them and use them as well.
sounds a lot like something called a "heat coil".
--
If you are one in a million; there are 6000 people just like
you.
Opinions expressed herein are my own and may not represent those of my employer.
L. M. Rappaport
current fuses there. The more common use for them appears to be in
multi-story buildings, at least that was mentioned in the application
notes...
Larry
--
ra...@lmr.com
rbmcc...@mmm.com (Roy McCammon) wrote (with possible editing):
w_tom
joules protection. That line to line surge protector really does nothing. Minimum
AC line surge protection is 1000 joules. That is minimal. Most decent 'whole
house' surge protectors are about 2000 joules. But this measures a surge
protector's life expectancy IOW if your 600 joule surge protection degrades after
4 strikes, then a 1000 joule surge protector would degrade to the same point in
about 35 strikes. The 2000 joule surge protection would last less than 300
strikes.
Properly sized surge protectors never blacken or vaporize. Their trigger voltage
just increases with age. This degradation is often charted by manufacturers based
upon a 10% trigger voltage increase. However I rather doubt having only 300 joules
per leg is a problem. Again, I suspect you have some path to earth ground superior
to your rather extensive ground network. Therefore even a tower strike would be
looking for a path through the building for this better earth ground. Ie. I asked
if you had a water well connected to the building. It is why water wells are
routed to enter the building first grounded to the central earth ground so that a
surge need not seek that ground destructively across the building.
In another town, some homes are built across a change in geology. No matter how
good the ground to a service entrance on the left side, surges still search
connections to the right side of the house that is sitting is significantly
superior conductive soil (graphite).
POTS modems are typically more susceptible to surge damage. Most other modems
contain a 2000+V isolation transformer (as in the POT modem). But POTS modems have
bypasses such as the off-hook relay that is only good to 500 V. Once that 500 V
isolation is circumvented, then the surge has an even lower voltage to maintain a
destructive current through the modem.
The problem with a plug-in surge protector is not the incoming impedance, but the
outgoing impedance back to earth ground. For example, a 100 amp surge shunted by a
plug-in surge protector leaves surge protector and adjacent appliance at maybe
15,000 volts relative to earth ground (50' of 12 AWG). Clearly that voltage is too
high. Therefore that 15,000 volt surge voltage will be reduced because other
destructive paths to earth ground are located. For example, one path may be
through modem, to earth ground on the phone line, or to earth ground through a wire
from modem to linoleum tile. IOW a surge protector adjacent to an appliance only
shunts a surge to many other paths into an appliance providing a surge many new and
possibly destructive paths to earth ground.
However if the same surge protector is located at the other end of that 12 AWG
wire, then a nearby earth ground receives most of a surge and other 'sneaky'
grounds at the appliance become an additional 150 ohms distant. IOW with the surge
protector 50' from the appliance, then increased impedance from a coiled wire only
improves surge protection. If the surge protector is located at the appliance end
of the 50', then the coiled wire only makes the surge protector assist in creating
more appliance damage.
Another also suggested: the coax wire down the tower should bond the shield to
the tower in various spots on the tower. If not, a resulting potential difference
can enter the building via this coax even if the shield is grounded to the central
earth ground before coax enters the building. I have no experience with towers so
I can only pass on his advise.
Computer network. It makes little difference whether the computer was on or
off. In one case, I have a three computer network - all computers off. Two were
plugged into surge protectors which simply shunted a surge into motherboard
grounds. The driver/receiver chip grounds shorted the surge into the network where
this surge entered a third computer. The surge left the third computer (also
powered off) via modem. All three network cards and modem were damaged. Surge
passes through ground plane of all three motherboards but no motherboards were
damaged.
Montandon and Rubinstien in 4 Nov 1998 IEEE Trans on Electromagnetic
Compatibility. I get it from college libraries. Otherwise, the IEEE wants $10 for
a reprint.