burning up NICs

[Milleron]

I've lost two NICs in the last two months.
1 -- there was an electrical storm, and cable modem and router were
interposed between outside and NIC but the next day Linksys NIC was
dead and two ports of the four-port router were dead. I don't know
when the router ports went out -- could have been ages ago. The cable
modem, which would have been FIRST in the path of a surge over the
cable, was perfectly fine, and the television was perfectly fine, so
I'm inclined to believe that this was not the result of a lightning
strike.
2 -- yesterday, experienced a shutdown problem (forgot to close an old
16-bit program before shutdown) and the computer ran for 20 hours in
shutdown limbo. Today, the new Netgear NIC is dead.

Is this coincidence, or could there be something wrong with my
old-reliable A7M266's PCI slot?
Ron

 

[Paul]

I've lost two NICs in the last two months.
1 -- there was an electrical storm, and cable modem and router were
interposed between outside and NIC but the next day Linksys NIC was
dead and two ports of the four-port router were dead. I don't know
when the router ports went out -- could have been ages ago. The cable
modem, which would have been FIRST in the path of a surge over the
cable, was perfectly fine, and the television was perfectly fine, so
I'm inclined to believe that this was not the result of a lightning
strike.
2 -- yesterday, experienced a shutdown problem (forgot to close an old
16-bit program before shutdown) and the computer ran for 20 hours in
shutdown limbo. Today, the new Netgear NIC is dead.

Is this coincidence, or could there be something wrong with my
old-reliable A7M266's PCI slot?
Ron

When you say the Netgear NIC is dead, is it the PCI side
which is dead, or the PHY side ? If it is the PCI side,
then your OS might fail to recognize it or be able to
enumerate it. If it is the PHY, then all the control
panels and other software will think the NIC is fine,
except you won't be able to send or receive packets.

Have you tried moving the Netgear card to another slot ?

The thing is, there are probably a few PCI devices on the
motherboard itself (CMI8738 sound chip?), and since PCI is a
shared bus, then why aren't they dead too ? Unless a pin is
bent in the PCI socket, and, say, a power pin has bridged to
some other signal ?

Maybe something is up with VIO ? Have you been modifying
the header lately ?

Paul

 

[rstlne]

I've lost two NICs in the last two months.
1 -- there was an electrical storm, and cable modem and router were
interposed between outside and NIC but the next day Linksys NIC was
dead and two ports of the four-port router were dead. I don't know
when the router ports went out -- could have been ages ago. The cable
modem, which would have been FIRST in the path of a surge over the
cable, was perfectly fine, and the television was perfectly fine, so
I'm inclined to believe that this was not the result of a lightning
strike.
2 -- yesterday, experienced a shutdown problem (forgot to close an old
16-bit program before shutdown) and the computer ran for 20 hours in
shutdown limbo. Today, the new Netgear NIC is dead.

Is this coincidence, or could there be something wrong with my
old-reliable A7M266's PCI slot?
Ron

Have you tried new cables..
a NIC going out by static discharges in the air is something I would def put
down as "possible" when there is just basic cat5 cable put outside.
It's also possible that if there was "Surges" going through the housing
electrical system that there could be a problem with the PSU itself..

So the first thing I would do is try is a new cable as that's really easy to
check.. After that it's all going to be just lots of trial and error but I
wouldnt fully rule the PSU being damaged. windows shutdown really isnt
going to do anything to the NIC even if it hangs.. It is possible that the
NIC was just a lemon anyhow..

 

[w_tom]

First review primary school science. Electricity that does
not flow in a complete circuit does not exist. It does not
flow like an ocean wave. Electricity flows through everything
in that circuit simultaneously - or it does not flow at all.
IOW things don't get damaged as the ocean wave passes
through. A surge flows through everything in the circuit. And
only after the flow exists, does something in that circuit
finally fail.

So where did the surge enter? Is cable properly installed?
That means cable drops down to make a less than 10 foot
connection to an earth ground rod before rising back up to
enter building. It properly earthed, then the surge never
entered via that path.

Look on utility poles. What are the highest wires that will
be struck before cable wires? AC electric. Now we have a
complete circuit from cloud to earth. Incoming on AC
electric, through computer (completely bypassing power
supply), through NIC, through router, to earth ground via
cable. What could be damaged? Well in that path, what
electronics paths could conduct? Typically the transceiver
and isolation transformer on NICs and same parts inside
router.

Today the NIC will fail. But other electronics parts have
been overstressed. So the router fails tomorrow. And the
most common reason for cable router and computer modem failure
- incoming surge on AC electric.

NICs and routers contain substantial internal protector
circuits. Anything that would protect same at the end of a
power cord is already inside/on those devices. For example,
they contain isolation transformers good to 2000+ volts. But
this internal protection assumes you and your building have
earthed the incoming transient BEFORE wires enter the
building. If not, then internal appliance protection is
overwhelmed.
IOW, there is no plug-in protector solution. No need for a
protector on cable. All protection systems must connect to a
surge protection - earth ground. Some do this using a surge
protector - ie telephone and AC electric. Others don't
require a surge protector. Instead they make a direct
hardwire connection to surge protection. Surge protectors are
only temporary connectors to the surge protection - earth
ground. So ineffective plug-in protector avoid all discussion
about being effective.

That cable should already have effective protection - a less
than 10 foot connection to the same earth ground rod that all
other incoming utilities connect to. Details such as single
point earth ground, no sharp bends in that earthing wire, and
earthing wire must be routed separate from all other wires -
all these details are essential and trivial to install.

'Whole house' protectors are so effective that the telco
installs one in your premise interface box - the NID. Your
phone lines should already have effective surge protection -
assuming that installation was performed correctly with a less
than 10 foot wire to the same SPG - single point ground.

We still don't build new homes as if the transistor exists.
The public is still too naive about effective protection to
demand it be installed routine. Therefore you (or your
electrician) must install a 'whole house' protector on AC
electric at the service entrance. And again, you must verify
or enhance the SPG so that even AC electric 'whole house'
protector completes a less than 10 foot connection to SPG.

Again, don't be fooled by plug-in protectors that could even
contribute to NIC damage. How to identify ineffective
protectors: 1) no dedicated and less than 10 foot connection
to SPG and 2) no mention of critical earthing - the SPG.
Plug-in protectors hope you believe electricity flows like an
ocean wave. Will that silly little protector stop, block, or
absorb what three miles of sky could not stop? Yes, they
would have you believe that to sell a protector that costs on
the order of 10 or 50 times more money per protected
appliance.

Forget the nonsense about surges entering like ocean waves.
Damage to NICs that were in a complete electric circuit from
cloud to earth. Damage probably because the building did not
have a properly earthed protection system on every incoming
utility wire.

Is the PCI slot overstressed? Typically not. But was the
motherboard electrically connect to chassis via only one
conductive standoff - or many? Again, factors that determine
what may and may not be overstressed. What was and was not in
a circuit from cloud to earth.

 

[Milleron]

When you say the Netgear NIC is dead, is it the PCI side
which is dead, or the PHY side ? If it is the PCI side,
then your OS might fail to recognize it or be able to
enumerate it. If it is the PHY, then all the control
panels and other software will think the NIC is fine,
except you won't be able to send or receive packets.

I'm not sure what you mean by PHY side, but Windows XP does detect the
NIC. Device manager says it's "working normally." However, neither
of its LEDs on the back panel illuminate and it transmits NO data. (I
believe that the sudden failure of the back-panel LEDs is a reliable
indicator that the card died.) The port LED on the switch to which
it's attached doesn't illuminate, but both that port and the cable can
be shown to be good by connecting through a USB Ethernet adapter I had
lying around.

Have you tried moving the Netgear card to another slot ?

That's difficult. The only slot left is #1, adjacent to the AGP card,
and I've found that on this MB, I can't run an NIC from that shared
slot , but I could at least move it over there to see if the
back-panel LEDs light up.

The thing is, there are probably a few PCI devices on the
motherboard itself (CMI8738 sound chip?), and since PCI is a
shared bus, then why aren't they dead too ? Unless a pin is
bent in the PCI socket, and, say, a power pin has bridged to
some other signal ? Everything else works fine.

Maybe something is up with VIO ? Have you been modifying
the header lately ?

Never. Haven't even cracked the case since replacing the last dead
NIC. That one really was dead because the new NIC was placed into the
same PCI slot and worked fine . . . for two months.

Paul

Ron

 

[Milleron]

Have you tried new cables..

First, thanks very much for responding.
Well, I proved that this cable is good by making a successful
connection with it attached to a USB Ethernet adapter I had lying
around (I'm sending this message to you over the original cable right
now).

a NIC going out by static discharges in the air is something I would def put
down as "possible" when there is just basic cat5 cable put outside.
It's also possible that if there was "Surges" going through the housing
electrical system that there could be a problem with the PSU itself..

I just spent $365 on a whole-house surge suppressor through which both
the AC and the cable run. The computer is connected through a UPS
that should clamp any surges that got through the whole-house
suppressor or originated from within the house. The cable goes
serially through the whole-house suppressor, a cable modem, a Vonage
VOIP phone adapter, a Linksys router, through about 50 feet of
Ethernet cable, and then another Linksys switch before getting to the
NIC in this machine.
A problem with the PSU cannot be ruled out. It is a high-quality
Antec 400W PSU, however, that's never given any other sign of trouble.

So the first thing I would do is try is a new cable as that's really easy to
check.. After that it's all going to be just lots of trial and error but I
wouldnt fully rule the PSU being damaged. windows shutdown really isnt
going to do anything to the NIC even if it hangs.. It is possible that the
NIC was just a lemon anyhow..

The cable has been proven good. I'm thinking that this was probably
just a lemon. Since NICs are now inexpensive, I think the next thing
I'll do is replace the NIC and see if the next one survives longer
than a couple of months.
Ron

 

[Milleron]

Thanks for responding. Most of what you're explaining here is over my
head, but I think that my whole-house surge suppressor and
single-point ground are done correctly. See info below.

First review primary school science. Electricity that does
not flow in a complete circuit does not exist. It does not
flow like an ocean wave. Electricity flows through everything
in that circuit simultaneously - or it does not flow at all.
IOW things don't get damaged as the ocean wave passes
through. A surge flows through everything in the circuit. And
only after the flow exists, does something in that circuit
finally fail.

So where did the surge enter? Is cable properly installed?
That means cable drops down to make a less than 10 foot
connection to an earth ground rod before rising back up to
enter building. It properly earthed, then the surge never
entered via that path.

Right. I don't think it entered from outside. I know that the cable
is connected directly to the ground rod before entering the building
(my condo is new construction and I had the cable installer explain to
me what he was doing as he brought it into the unit). The AC goes
through a state-of-the-art whole-house surge suppressor and then
through an APC UPS. AC to the cable modem's transformer goes through
a cheapy surge-suppressor power strip in addition to the whole-house
suppressor..
The cable follows this route:
whole-house suppressor --> cable modem --> VOIP phone adapter -->
Linksys router --> ethernet from basement to den --> Linksys switch
--> NIC.

Look on utility poles. What are the highest wires that will
be struck before cable wires? AC electric. Now we have a
complete circuit from cloud to earth. Incoming on AC
electric, through computer (completely bypassing power
supply), through NIC, through router, to earth ground via
cable. What could be damaged? Well in that path, what
electronics paths could conduct? Typically the transceiver
and isolation transformer on NICs and same parts inside
router.

Today the NIC will fail. But other electronics parts have
been overstressed. So the router fails tomorrow. And the
most common reason for cable router and computer modem failure
- incoming surge on AC electric. Seems unlikely. See above

NICs and routers contain substantial internal protector
circuits. Anything that would protect same at the end of a
power cord is already inside/on those devices. For example,
they contain isolation transformers good to 2000+ volts. But
this internal protection assumes you and your building have
earthed the incoming transient BEFORE wires enter the
building. If not, then internal appliance protection is
overwhelmed.
IOW, there is no plug-in protector solution. No need for a
protector on cable. All protection systems must connect to a
surge protection - earth ground. Some do this using a surge
protector - ie telephone and AC electric. Others don't
require a surge protector. Instead they make a direct
hardwire connection to surge protection. Surge protectors are
only temporary connectors to the surge protection - earth
ground. So ineffective plug-in protector avoid all discussion
about being effective.

That cable should already have effective protection - a less
than 10 foot connection to the same earth ground rod that all
other incoming utilities connect to. Details such as single
point earth ground, no sharp bends in that earthing wire, and
earthing wire must be routed separate from all other wires -
all these details are essential and trivial to install.

'Whole house' protectors are so effective that the telco
installs one in your premise interface box - the NID. Your
phone lines should already have effective surge protection -
assuming that installation was performed correctly with a less
than 10 foot wire to the same SPG - single point ground.

We still don't build new homes as if the transistor exists.
The public is still too naive about effective protection to
demand it be installed routine. Therefore you (or your
electrician) must install a 'whole house' protector on AC
electric at the service entrance. And again, you must verify
or enhance the SPG so that even AC electric 'whole house'
protector completes a less than 10 foot connection to SPG.

Done correctly, I think, but I have to trust my electrician. My
up-to-date whole house suppressor was installed on August 23.

Again, don't be fooled by plug-in protectors that could even
contribute to NIC damage. How to identify ineffective
protectors: 1) no dedicated and less than 10 foot connection
to SPG and 2) no mention of critical earthing - the SPG.
Plug-in protectors hope you believe electricity flows like an
ocean wave. Will that silly little protector stop, block, or
absorb what three miles of sky could not stop? Yes, they
would have you believe that to sell a protector that costs on
the order of 10 or 50 times more money per protected
appliance.

Forget the nonsense about surges entering like ocean waves.
Damage to NICs that were in a complete electric circuit from
cloud to earth. Damage probably because the building did not
have a properly earthed protection system on every incoming
utility wire.

Is the PCI slot overstressed? Typically not. But was the
motherboard electrically connect to chassis via only one
conductive standoff - or many? Again, factors that determine
what may and may not be overstressed. What was and was not in
a circuit from cloud to earth. NINE conductive standoffs.

Ron

 

[Paul]

I'm not sure what you mean by PHY side, but Windows XP does detect the
NIC. Device manager says it's "working normally." However, neither
of its LEDs on the back panel illuminate and it transmits NO data. (I
believe that the sudden failure of the back-panel LEDs is a reliable
indicator that the card died.) The port LED on the switch to which
it's attached doesn't illuminate, but both that port and the cable can
be shown to be good by connecting through a USB Ethernet adapter I had
lying around.
That's difficult. The only slot left is #1, adjacent to the AGP card,
and I've found that on this MB, I can't run an NIC from that shared
slot , but I could at least move it over there to see if the
back-panel LEDs light up.
Never. Haven't even cracked the case since replacing the last dead
NIC. That one really was dead because the new NIC was placed into the
same PCI slot and worked fine . . . for two months.

Ron

What that tells me, is the physical layer (PHY) has died.

What you should realize about the modern Ethernet interface,
is the link is transformer coupled. One of the purposes of
transformer coupling, is to prevent a DC path from forming
between interconnected devices. If it wasn't for this, you would
have the situation that exists on RS-232, where you can draw
sparks when connecting the ground of one computer to the ground
on another. (That used to happen at work, in my office building.
Building used false floors, and AC wiring spanned long distances,
and AC grounds were all over the map with respect to one another.)

Dev #1 ------ || ------------------- || ------ Dev #2
Ethernet \ || / Cable floats, \ || / Ethernet
/ || \ and has no DC / || \
\ || / path. \ || /
------/ || \-------------------/ || \------

The Ethernet cable functions like a big antenna. I've actually
connected an oscilloscope to the cable, and there are 10's of
volts of AC noise on the cable. The reason this doesn't affect
operation, is the same noise exists on both wires, and the noises
subtract from one another.

There is a sample Ethernet transformer spec here. Click "H303"
and then "Continue":

http://www.pulseeng.com/products/dataSheets.aspx?Category_Id=18

Now, if lightning strikes in the neighbourhood, the way your wire
is routed and where it is routed, could influence how much potential
builds on the wire. The transformer has a 1500 volt breakdown
voltage, and if you look at page 3 of the datasheet, there is a
0.001uF capacitor, with a 2000 volt breakdown voltage used as
part of a termination network. These voltages are selected to help
protect against accidental contact with AC power voltages, but
won't be enough to protect against lightning induced damage.

So, in the case of the NIC that lasted two months before dying,
I would look carefully at how your stuff is wired. Maybe there is
a long cable run along a wall of the building ? If this recurs,
you may want to go optical between the two devices. Generally,
optical is so expensive, going wireless is a much easier
alternative.

http://startech.com/ststore/itemdetail.cfm?tab=b&ProductID=ET9003ST&topbar=topbarn.htm

Here is another option. This device protects against high
voltage developing on an Ethernet cable. On a long cable
run, place one on either end of the cable run. Again, with
the expense of devices like this, wireless could be a lot
cheaper.

http://www.hyperlinktech.com/web/hgln_cat5_hp.php

HTH,
Paul

 

[Milleron]

What that tells me, is the physical layer (PHY) has died.

What you should realize about the modern Ethernet interface,
is the link is transformer coupled. One of the purposes of
transformer coupling, is to prevent a DC path from forming
between interconnected devices. If it wasn't for this, you would
have the situation that exists on RS-232, where you can draw
sparks when connecting the ground of one computer to the ground
on another. (That used to happen at work, in my office building.
Building used false floors, and AC wiring spanned long distances,
and AC grounds were all over the map with respect to one another.)

Dev #1 ------ || ------------------- || ------ Dev #2
Ethernet \ || / Cable floats, \ || / Ethernet
/ || \ and has no DC / || \
\ || / path. \ || /
------/ || \-------------------/ || \------

The Ethernet cable functions like a big antenna. I've actually
connected an oscilloscope to the cable, and there are 10's of
volts of AC noise on the cable. The reason this doesn't affect
operation, is the same noise exists on both wires, and the noises
subtract from one another.

There is a sample Ethernet transformer spec here. Click "H303"
and then "Continue":

http://www.pulseeng.com/products/dataSheets.aspx?Category_Id=18

Now, if lightning strikes in the neighbourhood, the way your wire
is routed and where it is routed, could influence how much potential
builds on the wire. The transformer has a 1500 volt breakdown
voltage, and if you look at page 3 of the datasheet, there is a
0.001uF capacitor, with a 2000 volt breakdown voltage used as
part of a termination network. These voltages are selected to help
protect against accidental contact with AC power voltages, but
won't be enough to protect against lightning induced damage.

So, in the case of the NIC that lasted two months before dying,
I would look carefully at how your stuff is wired. Maybe there is
a long cable run along a wall of the building ? If this recurs,
you may want to go optical between the two devices. Generally,
optical is so expensive, going wireless is a much easier
alternative.

First, thanks for this wonderfully thorough explanation.
1 -- There is a long cable run to the room housing this computer, but
it's NOT along an exterior wall.
2 -- Incoming AC and cable are routed through a whole-house
suppressor. I know that it's recommended that I have standard
[power-strip] surge suppressors at the points of use in addition to
this whole-house suppressor, and those secondary suppressors are in
place for my computers and entertainment system.
3 -- In the room where this computer is, the cable runs first to a
switch/print server. Nothing has happened to the switch or to the
printer or the NIC in the other computer connected to this switch.

Are you saying that an interior Ethernet wire along an exterior wall
could pick up current from somewhere in spite of the whole-house
suppressor?

http://startech.com/ststore/itemdetail.cfm?tab=b&ProductID=ET9003ST&topbar=topbarn.htm

I think optical is out of the question because the condo was prewired
for Ethernet, but those wires are no longer accessible now that
construction is complete.

Here is another option. This device protects against high
voltage developing on an Ethernet cable. On a long cable
run, place one on either end of the cable run. Again, with
the expense of devices like this, wireless could be a lot
cheaper.

http://www.hyperlinktech.com/web/hgln_cat5_hp.php

This device is described as a "lightning protector." Is it really
necessary when the cable coming into the house has been routed through
the whole-house suppressor?

HTH,

Yes, I think this helps. Thanks again.

Paul

Ron

 

[Paul]

First, thanks for this wonderfully thorough explanation.
1 -- There is a long cable run to the room housing this computer, but
it's NOT along an exterior wall.
2 -- Incoming AC and cable are routed through a whole-house
suppressor. I know that it's recommended that I have standard
[power-strip] surge suppressors at the points of use in addition to
this whole-house suppressor, and those secondary suppressors are in
place for my computers and entertainment system.
3 -- In the room where this computer is, the cable runs first to a
switch/print server. Nothing has happened to the switch or to the
printer or the NIC in the other computer connected to this switch.

Are you saying that an interior Ethernet wire along an exterior wall
could pick up current from somewhere in spite of the whole-house
suppressor?
http://startech.com/ststore/itemdetail.cfm?tab=b&ProductID=ET9003ST&topbar=topbarn.htm

I think optical is out of the question because the condo was prewired
for Ethernet, but those wires are no longer accessible now that
construction is complete.

Here is another option. This device protects against high
voltage developing on an Ethernet cable. On a long cable
run, place one on either end of the cable run. Again, with
the expense of devices like this, wireless could be a lot
cheaper.

http://www.hyperlinktech.com/web/hgln_cat5_hp.php

This device is described as a "lightning protector." Is it really
necessary when the cable coming into the house has been routed through
the whole-house suppressor?

HTH,

Yes, I think this helps. Thanks again.

Paul

Ron

The whole house protector is just that. It protects the entry to the
house. At least that is what the product offered in my area does. As
I described in my post, the Ethernet wires float, and if a potential
of more than 2KV is applied to a wire on the Ethernet cable,
the capacitor in the terminator network should arc over. Otherwise,
the wires happily float and have identical electrical noise on them.

The device on the Hyperlinktech page was created for a reason, and
my _guess_ is that it is used for illegal installs, such as running
an Ethernet wire between buildings. 100baseT is not intended for
exterior applications, and it almost looks like the Hyperlinktech
device is intended to "harden" a line for external use. If going
outside, an optical connection should be used (100baseFX), as you can
have all the lightning you want and nothing will happen.

In the interests of not wasting money, your first option would be
to replace the NIC with a new one, and wait. See if the PHY is
hosed on the new one. If a new one fails, proceed with the second
option.

Your second option is to inspect the house wiring. For example, I
have a three pronged gadget from the hardware store, that plugs into
a three prong outlet, and can tell whether it is wired properly.
(Hot, neutral, safety ground where they are supposed to be.) I would
then take my voltmeter, set it on 1000VAC scale. Unplug the Ethernet
cable from the back of the affected computer (leave the computer
plugged in, so the computer chassis remains grounded), and measure
from each of the eight wires in the cable, to ground via one of the
nuts on the I/O connectors on the back of the computer. Most voltmeters
will show either AC or DC when set to the AC scale, and touching the
wires on the Ethernet patch cord one at a time with the probe tip,
will show whether there is anything "funny" on the line. (You should
unplug the load at the other end of the house, if this is the router
for example, so the line terminator in the router doesn't absorb
whatever is on the Ethernet wires. I.e. You want the wire in the walls
all by itself, with no loads attached.)

If you don't want to test the wiring, another option is to run a
separate ethernet cable from the affected computer to the router.
That way, you know it isn't touching anything, and you know exactly
where it is going.

As for the house wiring, the more stuff is wired to it, in terms of
protection devices, the more chances for mistakes. And, the
inspectors who inspect for code violations, don't necessarily check
that the actual wiring is correct. I mentioned the Hyperlinktech page,
purely to show that protection devices exist (I thought it wasn't
possible to do one, but it didn't take long with a search engine to
find one). Ethernet inside a house should be able to work without
protection, and only the most extreme environment would need two of
these installed (like running Cat5 between the house and a separate
garage building, for example).

BTW - That particular Hyperlinktech device is intended for 100baseT
and protects pins 1,2,3,6 with the gas tubes. That one is not
designed for GbE (Gigabit Ethernet), as in GbE all eight wires carry
signals, and the Hyperlinktech device actually shorts some wires
together. In 100baseT, only four wires carry signals, a TX pair
and a RX pair.

Perhaps the other four wires in the Ethernet cable have been hijacked
for some other purpose ? Perhaps you should talk to the contractor
and see how the data jacks have been set up. I.e. Separate wire
for phone. Separate eight wire cable for Ethernet. Etc.

Another bit of trivial for you. The pins 1,2,3,6 of the Ethernet
cable float so nicely, that some genius decided to extend the Ethernet
standard, to carrying DC power down the Ethernet pairs. I personally
am not in favor of this change, but I guess it makes Cisco happy.
This is something else to watch for, and is another reason for
disconnecting Ethernet devices at either end of the cable before
making measurements. You wouldn't want to mistake a POE power source
for a problem with the wiring I don't know how widespread this
change is, and whether home routers do this or not. As my home router
uses an external 5VDC power brick, it would be tricky to get 30-60V
cheaply from it.

http://www.industrialnetworking.co.uk/mag/v9-3/f_poe.html

HTH,
Paul

 

[w_tom]

For some reason, implied is that a 'whole house' protector
is the magic protection device. Protector is nothing more
than an electrical switch or an electrical wire. Protector is
trivial. Your attention has been diverted to the one thing
you can see. Protection - a different device - is important
and forgotten because it is not visible. Earth ground (into
the protectors) is the 'state of art' protection. You
suffered damage. Therefore something is wrong with your
'state of art' protection.

You suffered damage. Somehow a transient got inside the
house. It should not. Therefore your protection system has a
serious defect.

The 'whole house' protector is only secondary protection.
Start by inspecting your primary protection, if possible.
Pictures that demonstrate a defective primary protection
system:
http://www.tvtower.com/fpl.html

Now we move on to the secondary protection system. Does
every incoming utility make a 'less than ten foot' connection
to that SPG? If not (if any incoming utility wire connects to
some other earth ground), then you don't have effective
protection. For example, plug-in protectors (power strip and
UPS) make no 'less than 10 foot' connection from AC electric
to earth ground. Therefore they are ineffective AND may even
contribute to appliance surge damage. A protector is not
protection. Unfortunately you imply otherwise. A protector
is effective when it shunts (diverts, connects, acts like a
wire) the transient from each utility wire, less than 10 feet,
to the single point earth ground (SPG).

Your posted chart of cable connections implies you don't
understand how cable is surge protected. Cable needs no
connection to a 'whole house' protector. Cable should be
earthed to the same earth ground before even entering the
building. This even required by National Electrical Code
(NEC). A ground block that costs maybe $2 in Home Depot
connects that earthing wire to incoming cable. Cable company
should have installed it. The protector for cable is simply a
copper wire from ground block to earth ground rod. No 'whole
house' protector required for cable. What does a surge
protector becomes during a transient? An electrical wire.
Protection is defined by the earth ground rod - the SPG.
'State of the art' is not a protector. 'State of the art' is
defined by the quality of that SPG and how wires connect to
that SPG.

In the meantime, previously defined: those plug-in power
strip and UPS protectors are ineffective. They enrich the
manufacturer. Obviously. How far away is the SPG? More than
10 feet? So manufacturer forgets to mention they don't
protect from the destructive type of surge, such as you have
apparently suffered. That's correct. They only claim to
protect from a type of surge that does not typically exist and
that is made redundant by internal appliance protection. The
protector is only as effective as its earth ground. So
instead, the manufacturer remains dumb.

Is the computer protected? Then your diagram must show the
computer in relation to each single point earth ground. SPG
is the protection; not a power strip or APC UPS. Moreso,
those plug-in protectors can even contribute to damage of the
adjacent electronics. Your protection system has a problem.
So we suspect a problem with where protection is provided -
the SPG or connections to that SPG.

Earthing wire from protector to earth ground must be
separated from all other wires. For example, some think a
plug-in protector is earthed through the wall receptacle's
safety ground. How does that connection get made? A safety
ground wire is bundled with other wires. It can induce
transients onto those other wires. It is bundled with other
wires AND is too distant from SPG. Therefore the plug-in
protector is not earthed - does nothing effective.

Same might be true of your 'whole house' protector. How
does the wire route from protector to a single point earth
ground? Distance? Sharp bends? Through any conduit or other
metallic container? Bundled with what? No splices? Each
earthing wire must no be shared until all earthing wires
connect to the SPG. These are important questions answered by
a visual inspection. You suffered damage. Discover why the
system was compromised. How does each incoming utility wire
connect to that SPG?

What type of soil? A quality SPG (and not the protector) is
important. Many now install two earth ground rods - one being
insufficient - due to non-conductive soils. What makes a Ben
Franklin lightning rod effective? Not pointed rod verse blunt
rod. That was hype. Quality of the earth ground and how that
lighting rod connects to earth. Surge protectors are doing
same. A surge protector is only as effective as its earth
ground. So how good is your earthing soil?

If damage exists, then your earthing system may have
problems. Things often overlooked because the installer did
not understand, for example, the "less than 10 feet" concept.
Or because one incoming utility is using a separate earth
ground - very bad. Or because something geological has
compromised your single point earth ground. Your secondary
protection system (which defines the 'whole house' protector)
apparently has been compromised.

Nothing here should be 'over your head'. The complicated
stuff has been removed. An introduction to concepts was
posted previously at "Pull the wall plug or not?" in
nz.comp on 7 Sept 2004 at
http://tinyurl.com/5ttwl

Effective surge protection does same thing as the Ben
Franklin air terminal as taught in elementary school science.
Its not complicated other than following certain important
rules such as only one earth ground - the single point earth
ground. "Less than ten foot" from each incoming utility to
that SPG. But this we know even from your damage. The surge
did enter from outside. Your damage is typical of an
electrical circuit from cloud to earth. For some reason,
electricity had to find earth ground, destructively via your
cable network. Your damage is a classic example of surges
entering from outside.

 

[Milleron]

For some reason, implied is that a 'whole house' protector
is the magic protection device. Protector is nothing more
than an electrical switch or an electrical wire. Protector is
trivial. Your attention has been diverted to the one thing
you can see. Protection - a different device - is important
and forgotten because it is not visible. Earth ground (into
the protectors) is the 'state of art' protection. You
suffered damage. Therefore something is wrong with your
'state of art' protection.

You suffered damage. Somehow a transient got inside the
house. It should not. Therefore your protection system has a
serious defect.

Maybe I didn't make this clear. I just moved. The whole-house
suppressor is at the new house, and there was NO electrical storm here
when the second NIC went out. I think this device failed from some
reason OTHER than a line surge

The 'whole house' protector is only secondary protection.
Start by inspecting your primary protection, if possible.
Pictures that demonstrate a defective primary protection
system:
http://www.tvtower.com/fpl.html

Now we move on to the secondary protection system. Does
every incoming utility make a 'less than ten foot' connection
to that SPG? If not (if any incoming utility wire connects to
some other earth ground), then you don't have effective
protection. For example, plug-in protectors (power strip and
UPS) make no 'less than 10 foot' connection from AC electric
to earth ground. Therefore they are ineffective AND may even
contribute to appliance surge damage. A protector is not
protection. Unfortunately you imply otherwise. A protector
is effective when it shunts (diverts, connects, acts like a
wire) the transient from each utility wire, less than 10 feet,
to the single point earth ground (SPG).

Your posted chart of cable connections implies you don't
understand how cable is surge protected. Cable needs no
connection to a 'whole house' protector. Cable should be
earthed to the same earth ground before even entering the
building. This even required by National Electrical Code
(NEC). A ground block that costs maybe $2 in Home Depot
connects that earthing wire to incoming cable. Cable company
should have installed it. The protector for cable is simply a
copper wire from ground block to earth ground rod. No 'whole
house' protector required for cable. What does a surge
protector becomes during a transient? An electrical wire.
Protection is defined by the earth ground rod - the SPG.
'State of the art' is not a protector. 'State of the art' is
defined by the quality of that SPG and how wires connect to
that SPG.

As mentioned clearly in my other posts, the incoming cable is grounded
before entry to the house. And, no, I don't understand how cable is
surge protected. God knows I'm sorry, but I'm a physician, not an
electrician or computer engineer.

In the meantime, previously defined: those plug-in power
strip and UPS protectors are ineffective. They enrich the
manufacturer. Obviously. How far away is the SPG? More than
10 feet? So manufacturer forgets to mention they don't
protect from the destructive type of surge, such as you have
apparently suffered. That's correct. They only claim to
protect from a type of surge that does not typically exist and
that is made redundant by internal appliance protection. The
protector is only as effective as its earth ground. So
instead, the manufacturer remains dumb.

The Web sites explaining whole-house surge suppression are all very
clear that the store-bought surge suppressors SHOULD be used IN
ADDITION to the whole-house suppressor.

Is the computer protected? Then your diagram must show the
computer in relation to each single point earth ground. SPG
is the protection; not a power strip or APC UPS. Moreso,
those plug-in protectors can even contribute to damage of the
adjacent electronics. Your protection system has a problem.
So we suspect a problem with where protection is provided -
the SPG or connections to that SPG.

I cannot reconcile this with what I've gathered about whole-house
suppression as presented for laymen on several Web sites devoted to
the subject. Your undoubtedly correct, but you're beyond my
understanding at this point.

Earthing wire from protector to earth ground must be
separated from all other wires. For example, some think a
plug-in protector is earthed through the wall receptacle's
safety ground. How does that connection get made? A safety
ground wire is bundled with other wires. It can induce
transients onto those other wires. It is bundled with other
wires AND is too distant from SPG. Therefore the plug-in
protector is not earthed - does nothing effective.

Same might be true of your 'whole house' protector. How
does the wire route from protector to a single point earth
ground? Distance? Sharp bends? Through any conduit or other
metallic container? Bundled with what? No splices? Each
earthing wire must no be shared until all earthing wires
connect to the SPG. These are important questions answered by
a visual inspection. You suffered damage. Discover why the
system was compromised. How does each incoming utility wire
connect to that SPG?

What type of soil? A quality SPG (and not the protector) is
important. Many now install two earth ground rods - one being
insufficient - due to non-conductive soils. What makes a Ben
Franklin lightning rod effective? Not pointed rod verse blunt
rod. That was hype. Quality of the earth ground and how that
lighting rod connects to earth. Surge protectors are doing
same. A surge protector is only as effective as its earth
ground. So how good is your earthing soil?

If damage exists, then your earthing system may have
problems. Things often overlooked because the installer did
not understand, for example, the "less than 10 feet" concept.
Or because one incoming utility is using a separate earth
ground - very bad. Or because something geological has
compromised your single point earth ground. Your secondary
protection system (which defines the 'whole house' protector)
apparently has been compromised.

Nothing here should be 'over your head'. The complicated
stuff has been removed. An introduction to concepts was
posted previously at "Pull the wall plug or not?" in
nz.comp on 7 Sept 2004 at
http://tinyurl.com/5ttwl

Effective surge protection does same thing as the Ben
Franklin air terminal as taught in elementary school science.
Its not complicated other than following certain important
rules such as only one earth ground - the single point earth
ground. "Less than ten foot" from each incoming utility to
that SPG. But this we know even from your damage. The surge
did enter from outside. Your damage is typical of an
electrical circuit from cloud to earth. For some reason,
electricity had to find earth ground, destructively via your
cable network. Your damage is a classic example of surges
entering from outside.

Ron

 

[w_tom]

The concept is called overstress. Static shock a circuit
while in the shop. Computer fails a month later while at
home. Failure was created in the shop - a condition called
overstress. In your case, an overstress condition to NIC in
old house could have resulted in NIC failure while in the new
house.

Yes, it is possible that other reasons caused NIC failure.
But as those Pulse Engineering data sheets demonstrate, an NIC
can withstand 1500 volts via the Ethernet network. The
destructive surge must first start with a voltage that exceeds
that 1500 volts. And it must have both an incoming and
outgoing path.

The classic destructive surge enters on AC electric (the
typically unprotected utility) seeking earth ground. Incoming
on AC mains. Through NIC and cable router. To earth ground
via the cable. Notice the surge had to have sufficient
voltage to blow through that isolation transformer - to go
through both NIC and cable box router. That surge could have
damaged electronics or simply created overstress. We first
analyze the entire surge circuit. Then replace things that
also may have been overstressed. That means we look at
everything that was in a path from cloud to earth. Learn
what was damaged, connect the dots, and then review stuff
'between those dots' for a potential overstress.

 

[Milleron]

The concept is called overstress. Static shock a circuit
while in the shop. Computer fails a month later while at
home. Failure was created in the shop - a condition called
overstress. In your case, an overstress condition to NIC in
old house could have resulted in NIC failure while in the new
house.

Yes, it is possible that other reasons caused NIC failure.
But as those Pulse Engineering data sheets demonstrate, an NIC
can withstand 1500 volts via the Ethernet network. The
destructive surge must first start with a voltage that exceeds
that 1500 volts. And it must have both an incoming and
outgoing path.

The classic destructive surge enters on AC electric (the
typically unprotected utility) seeking earth ground. Incoming
on AC mains. Through NIC and cable router. To earth ground
via the cable. Notice the surge had to have sufficient
voltage to blow through that isolation transformer - to go
through both NIC and cable box router. That surge could have
damaged electronics or simply created overstress. We first
analyze the entire surge circuit. Then replace things that
also may have been overstressed. That means we look at
everything that was in a path from cloud to earth. Learn
what was damaged, connect the dots, and then review stuff
'between those dots' for a potential overstress.

I see. Yes, the condition you describe could have occurred at the old
house prior to our move. That would explain a lot. Currently, I
think I'm protected because of the grounding of the cable, the
whole-house suppression, and the supplementary surge suppressors at
points of use. Hopefully, this NIC and router will last longer than
the last ones.
Thanks for all the instruction on this difficult subject. Apparently,
according to the pro who installed the whole-house suppressor, some of
the phenomena that occur with lightning and surges just seem to defy
logic. He said he's seen dozens of examples of damaged electrical and
electronic products that did not appear to be nearly the most
vulnerable components in the path of th strike and yet were the only
things damaged. He says the pattern of damage seems to be different
every time..
Ron

 

[w_tom]

Phenomena created by lightning does not defy logic. However
many fail to learn basic concepts. For example, connect a
long wire to an electrical source. The electrical pulse wave
travels down the wire. Where on that wire do we first see the
pulse? Obviously right where the pulse first connects to the
wire.

Wrong. Pulse is first observed at location farthest down
wire. IOW those who think of electricity as an ocean wave
sadly jumped to mistaken conclusions. Electricity is quite
understandable if one first learns basic concepts - such as
where a pulse would first be observed.

Lightning damage is statistically predictable. But when a
human fails to comprehend in advance, then that human must
learn and correct his mistake.

Again, do you know geology beneath your building? One
observed a lightning strike avoid the 40 foot tree to hit bare
earth some 60 feet from that tree. Why? He later learned
more conductive rock was located 60 feet from the tree.
Lightning was not capricious. It struck where expected. A
human did not yet understand all facts.

If lightning causes damage, then a human is reason for
failure. Human must then learn from and correct his mistake.
Lightning is capricious only when a human does not comprehend
principles and facts. All lightning damage is avoidable.