My company have an automatic outdoor carpark barrier system installed on three exit points. This carpark barrier system are interconnected to a master network system which include booth payment counter, autopay console and proximity card access system. Since the carpark barriers are installed outdoor, the comm cable are buried underground. Sometimes during a thunderstorm, the NIC card of various system were damaged which we think is the surge from the lightning to the underground network cable. we have installed a Furse ESP 06D on all the NIC however the recent thunderstorm kill 3 of the NIC card. Any suggestion to help to improve the situation is highly appreciated.
Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations JAE on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
Lightning surge protection on underground network cable 1
- Thread starter nicole10g
- Start date
"Surge Protector" devices are available for Ethernet cables. They typically accept the 8P8C ("RJ45") connectors and you use a jumper to complete the connection. The protector then has a heavy ground wire. Here's an example. I think that one is sold out, so it's just an example.
You should also look into correct grounding techniques.
And since there's never a 100% assurance, keep spare NIC on the shelf - if not an entire server since they're so cheap.
You should also look into correct grounding techniques.
And since there's never a 100% assurance, keep spare NIC on the shelf - if not an entire server since they're so cheap.
- Thread starter
- #3
Your situation is no different than what every telco switching station must solve. First some basic concepts. Your NIC already has significant protection typically rated about 2000 volts. It is damaged because a transient current is incoming (ie via a buried wire) and outgoing via some other path (ie AC electric). Obviously it must be a serious current to increase voltages well above 2000 volts.
Furthermore, a protector that tries to stop or absorb that current is bogus. Voltage increases as necessary to blow through. No protector has numbers sufficient to absorb a destructive surge.
Second, the solution. Never try to stop or absorb that current. Connect that current to what it wants. On a path that does not pass through and stays away from the NIC. Telcos do this by routing wires underground before entering the facility. By earthing each wire in every cable. By making that connection to earth as low impedance (ie short) as possible. No protector does protection. Earth ground does the protection.
Some protection systems have no protector. A wire (no protector) makes that connection to earth. Every protection has and carefully installs the only item that does all protection. The item that absorbs hundreds of thousands of joules - single point earth ground.
Best protection is also distant from electronics. Telcos want up to 50 meter separation. But your location sounds like that is impossible. Therefore an earth ground must be even better. For example, surrounding the building with a buried loop would help. Any wire (AC electric, ethernet) that enters must connect at a common point. A low impedance connection to earth means no sharp wire bends, no splices, is not inside metallic conduit, and routes separated from any non-grounding wires.
How many AC electric wires enter? Three? A neutral wire might already be earthed. Other AC wires must connect to the same earthing electrode just as short. But make that connection via a 'whole house' protector.
Same rule applies to each ethernet wire. Each cable has eight wires. Each must also connect to that same earthing electrode; make the same low impedance connection. Obviously, ethernet cannot connect directly. That is what a typically green wire from the ethernet protector does. Connect each wire to earth via a protector.
The resulting surge could be incomcing on ethernet and outgoing via AC electric. Or incoming via AC electric. Choosing a best ethernet wire that connects to earthborne charges miles away. Either way, damage means an incoming current path and a completely different outgoing current path exists. Makes no difference if wire is overhead or underground. It must make that low impedance connection to single point earth ground where entering NIC building. Otherwise current will find other paths to earth destructively via NICs.
Appreciate how major that current must be with each storm to cause damage. Approaching 2000 volts for no damage; or greater than 2000 volts to have damage. Serious transient problems even when damage does not occur. IOW no protection other than what is already inside electronics. Does the Furse ESP 06D have a dedicated ground wire? If that wire is connected to a wall receptacle safety ground, then not earthed. Receptacle safety ground is electrically different from earth ground. The Furse must connect low impedance (ie 'less than 10 feet', no sharp bends, not inside metallic conduit) to single point 'earth' ground. Only then does it become part of a protection system.
Third, every layer of protection is only defined by the earth ground. Above defines the 'secondary' protection layer. Also inspect a 'primary' protection layer. A picture demonstrates what must be inspected:
Fourth, some facilities use metallic conduit on buried incoming wires to increase protection. Buried conduit is typically up to 50 feet long. And all conduit join together at the single point ground before wires enter.
Fifth, above only defines protection for the building. Treat each gate as a separate structure. Therefore each gate also has its own single point ground. Any wire that enters a gate controller must also be earthed to that gate's earth ground. The rule is this simple. Any incoming wire (single or power) must first connect to earth before entering. Either by a direct connection or via a protector. Only then will a surge not connect to distant earthborne charges destructively through electronics.
Furthermore, a protector that tries to stop or absorb that current is bogus. Voltage increases as necessary to blow through. No protector has numbers sufficient to absorb a destructive surge.
Second, the solution. Never try to stop or absorb that current. Connect that current to what it wants. On a path that does not pass through and stays away from the NIC. Telcos do this by routing wires underground before entering the facility. By earthing each wire in every cable. By making that connection to earth as low impedance (ie short) as possible. No protector does protection. Earth ground does the protection.
Some protection systems have no protector. A wire (no protector) makes that connection to earth. Every protection has and carefully installs the only item that does all protection. The item that absorbs hundreds of thousands of joules - single point earth ground.
Best protection is also distant from electronics. Telcos want up to 50 meter separation. But your location sounds like that is impossible. Therefore an earth ground must be even better. For example, surrounding the building with a buried loop would help. Any wire (AC electric, ethernet) that enters must connect at a common point. A low impedance connection to earth means no sharp wire bends, no splices, is not inside metallic conduit, and routes separated from any non-grounding wires.
How many AC electric wires enter? Three? A neutral wire might already be earthed. Other AC wires must connect to the same earthing electrode just as short. But make that connection via a 'whole house' protector.
Same rule applies to each ethernet wire. Each cable has eight wires. Each must also connect to that same earthing electrode; make the same low impedance connection. Obviously, ethernet cannot connect directly. That is what a typically green wire from the ethernet protector does. Connect each wire to earth via a protector.
The resulting surge could be incomcing on ethernet and outgoing via AC electric. Or incoming via AC electric. Choosing a best ethernet wire that connects to earthborne charges miles away. Either way, damage means an incoming current path and a completely different outgoing current path exists. Makes no difference if wire is overhead or underground. It must make that low impedance connection to single point earth ground where entering NIC building. Otherwise current will find other paths to earth destructively via NICs.
Appreciate how major that current must be with each storm to cause damage. Approaching 2000 volts for no damage; or greater than 2000 volts to have damage. Serious transient problems even when damage does not occur. IOW no protection other than what is already inside electronics. Does the Furse ESP 06D have a dedicated ground wire? If that wire is connected to a wall receptacle safety ground, then not earthed. Receptacle safety ground is electrically different from earth ground. The Furse must connect low impedance (ie 'less than 10 feet', no sharp bends, not inside metallic conduit) to single point 'earth' ground. Only then does it become part of a protection system.
Third, every layer of protection is only defined by the earth ground. Above defines the 'secondary' protection layer. Also inspect a 'primary' protection layer. A picture demonstrates what must be inspected:
Fourth, some facilities use metallic conduit on buried incoming wires to increase protection. Buried conduit is typically up to 50 feet long. And all conduit join together at the single point ground before wires enter.
Fifth, above only defines protection for the building. Treat each gate as a separate structure. Therefore each gate also has its own single point ground. Any wire that enters a gate controller must also be earthed to that gate's earth ground. The rule is this simple. Any incoming wire (single or power) must first connect to earth before entering. Either by a direct connection or via a protector. Only then will a surge not connect to distant earthborne charges destructively through electronics.
"...a protector that tries to stop or absorb that current is bogus."
It's not clear exactly what you're claiming here. Yes, there are many bogus protection devices; but not all protection devices are bogus. You're correct that a larger system view is a far superior approach; but it may not be cost effective.
"By earthing each wire in every cable."
That approach obviously requires clarification.
These are minor clarifications. Much of your lengthy post is exactly correct and there's a ton of good info within it.
It's not clear exactly what you're claiming here. Yes, there are many bogus protection devices; but not all protection devices are bogus. You're correct that a larger system view is a far superior approach; but it may not be cost effective.
"By earthing each wire in every cable."
That approach obviously requires clarification.
These are minor clarifications. Much of your lengthy post is exactly correct and there's a ton of good info within it.
Some examples of ethernet protectors:
Ethernet has eight wires. Each wire connects to a ground stud or ground wire via a protector. The protector only works when the ground stud connects low impedance (ie 'less than 10 feet') to single point earth ground. Then all eight ethernet wires are earthed.
Telcos typically have a 'whole house' protector installed for free inside the subscriber interface. Again, two or four telephone wires enter. If the NID is properly earthed, then every wire is connected to earth via that protector. Such protectors were installed for free long than you or I have even existed.
Cable TV needs no protector. The wire that connects cable directly to earth (as also required by the National Electrical Code and other standards) does that protection. But again, the wire must be low impedance (ie 'no sharp bends').
If the protector does not make that always required low impedance connection, then it only claims to protect from surges made irrelevant by protection already inside every appiance. Key to any effective protector is that earth ground connection. Because protection is always about where energy dissipates.
Ethernet has eight wires. Each wire connects to a ground stud or ground wire via a protector. The protector only works when the ground stud connects low impedance (ie 'less than 10 feet') to single point earth ground. Then all eight ethernet wires are earthed.
Telcos typically have a 'whole house' protector installed for free inside the subscriber interface. Again, two or four telephone wires enter. If the NID is properly earthed, then every wire is connected to earth via that protector. Such protectors were installed for free long than you or I have even existed.
Cable TV needs no protector. The wire that connects cable directly to earth (as also required by the National Electrical Code and other standards) does that protection. But again, the wire must be low impedance (ie 'no sharp bends').
If the protector does not make that always required low impedance connection, then it only claims to protect from surges made irrelevant by protection already inside every appiance. Key to any effective protector is that earth ground connection. Because protection is always about where energy dissipates.
It's not clear that the Tripplite product has a grounding connection. Perhaps it does, but it's not explicitly shown nor mentioned.
Here are two that clearly show a grounding lead: Example One and Example Two. I think we probably both agree that a grounding lead is an essential part of such a device.
Here are two that clearly show a grounding lead: Example One and Example Two. I think we probably both agree that a grounding lead is an essential part of such a device.
-
1
- #9
Surge protection is a very inaccurate science. All surge suppression devices have limits. And none will with stand a direct hit. What they protect against are magnetically or capacitively coupled induced voltages and currents. Also there is a third type of surge which is ground potential differences during a strike event. During a strike (maybe somewhat distant) there is a potential difference between the ground at the two ends of the cable. The further apart, the higher the voltage.
The best way to eliminate the third type of surge is by using fiber optic cable. There are now available NIC cards which directly accept fiber, no need for converter boxes. In addition to that I would also recommend placing surge suppression on all wires entering or leaving the local box housing the computer componets.
The best way to eliminate the third type of surge is by using fiber optic cable. There are now available NIC cards which directly accept fiber, no need for converter boxes. In addition to that I would also recommend placing surge suppression on all wires entering or leaving the local box housing the computer componets.
Because surge protection cannot work for direct strikes, then your town is without phone service for four days while they replace that $multi-million switching computer? Since a CO will suffer about 100 surges with each storm, this damage must happen frequently.
Or read papers even in 1950s Bell System Technical Journal. Researchers confirmed existing protection from direct lightning strikes would also protect transistors.
The 'inaccurate' science of protection has been proven by well over 100 years of science and experience. What has been known that long is not found in a typical power strip or UPS that most buy in retail stores. Those devices do not even claim to protect from destructive surges. Read its spec numbers.
Attached to an ethernet protector is what any effective protector always must have. A low impedance connection to earth ground. Many will recommend a protector that has no earthing. Experience damage. Then assume "none will withstand a direct hit." Nonsense. Already implemented in telco switching centers, muntions dumps, electronics atop the Empire State Building, rocket launch facilities, cell towers, radio stations ... anywhere that damage cannot happen. WTC electronics would suffer 40 direct strikes annually without damage.
Homeowners can purchase an equivalent solution even in Lowes and Home Depot for about $1 per protected appliance. A typical lightning strike is maybe 20,000 amps. So a minimal 'whole house' protector is 50,000 amps. So that a protector remains functional after each direct lightning strike.
What makes that Ethernet protector effective? What makes a 'whole house' protector effective? In every case, a protector makes a low impedance (ie 'less than 10 foot') connection to 'single point earth ground'. All four words are significant. To even make ground potential difference irrelevant.
Protectors are simple and easily understood. Protectors do not do protection. Earthing is where mistakes are easily made. Then many assume nothing can protect from a direct strikes. The art of protection is earth ground; not a protector.
Or read papers even in 1950s Bell System Technical Journal. Researchers confirmed existing protection from direct lightning strikes would also protect transistors.
The 'inaccurate' science of protection has been proven by well over 100 years of science and experience. What has been known that long is not found in a typical power strip or UPS that most buy in retail stores. Those devices do not even claim to protect from destructive surges. Read its spec numbers.
Attached to an ethernet protector is what any effective protector always must have. A low impedance connection to earth ground. Many will recommend a protector that has no earthing. Experience damage. Then assume "none will withstand a direct hit." Nonsense. Already implemented in telco switching centers, muntions dumps, electronics atop the Empire State Building, rocket launch facilities, cell towers, radio stations ... anywhere that damage cannot happen. WTC electronics would suffer 40 direct strikes annually without damage.
Homeowners can purchase an equivalent solution even in Lowes and Home Depot for about $1 per protected appliance. A typical lightning strike is maybe 20,000 amps. So a minimal 'whole house' protector is 50,000 amps. So that a protector remains functional after each direct lightning strike.
What makes that Ethernet protector effective? What makes a 'whole house' protector effective? In every case, a protector makes a low impedance (ie 'less than 10 foot') connection to 'single point earth ground'. All four words are significant. To even make ground potential difference irrelevant.
Protectors are simple and easily understood. Protectors do not do protection. Earthing is where mistakes are easily made. Then many assume nothing can protect from a direct strikes. The art of protection is earth ground; not a protector.
As usual, an ounce of prevention is worth a pound of cure. The best approach in this type of situation is to avoid exposure, and it would seem prudent to put the cables in some sort of metal or armored conduit. You could still have a whole-house protector as a backup, but the fact of the matter is that your cables need not be exposed to direct or indirect strikes at all. Dissipating the energy before it gets to your cables would put less stress over time on your cables and circuitry, regardless of how well the whole-house protector does its job.
If this is supposed to be a reliable system, then it would make even more sense to not rely on a single-point of failure circuit.
TTFN
faq731-376
If this is supposed to be a reliable system, then it would make even more sense to not rely on a single-point of failure circuit.
TTFN
faq731-376
- Thread starter
- #12
> Do you think that I should have a separate dedicated earthing for the surge protector?
An important expression is 'single point earth ground'. Grounds for every incoming wire must be to a common earthing system. All incoming wires or their surge protectors must connect low impedance to one ground. Multiple grounds may even make damage easier.
A utility demonstrates good, bad, and ugly grounds:
Earth a surge current at one point so that current need not find earth destructively via some other ground.
An important expression is 'single point earth ground'. Grounds for every incoming wire must be to a common earthing system. All incoming wires or their surge protectors must connect low impedance to one ground. Multiple grounds may even make damage easier.
A utility demonstrates good, bad, and ugly grounds:
Earth a surge current at one point so that current need not find earth destructively via some other ground.
- Thread starter
- #15
> My main electrical earth ... is far to the point of surge protector.
Again, a protector does not do protection. Grasp that critically important fact. The protector does not do protection. By wire or by protector, what is the connection to earth?
A surge is incoming on AC utility wires. AC hot wires connect to earth destructively via appliances. Why? If not earthed low impedance (ie 'less than 10 foot') where it enters the building, then a surge is connected to earth destructively via appliances - no matter how many protectors exist. An expression 'less than 10 feet - or shorter' cannot be ignored. Nothing stops that destructive hunt. Protection is about a short connection to earth; not about a protector.
Excessive impedance: ie 'more than 10 feet'.
Again, a protector does not do protection. Grasp that critically important fact. The protector does not do protection. By wire or by protector, what is the connection to earth?
A surge is incoming on AC utility wires. AC hot wires connect to earth destructively via appliances. Why? If not earthed low impedance (ie 'less than 10 foot') where it enters the building, then a surge is connected to earth destructively via appliances - no matter how many protectors exist. An expression 'less than 10 feet - or shorter' cannot be ignored. Nothing stops that destructive hunt. Protection is about a short connection to earth; not about a protector.
Excessive impedance: ie 'more than 10 feet'.
nicole10g (Electrical) said:
Westom said:
Not all lightning induced surges arrive on the AC utility lines. I've seen them with my own two eyes arriving on a coaxial antenna cable.
Although we obviously can't see through the Internet to be perfectly certain of what exactly happened in this case, the working assumption for this thread (at least for most of us) is that the lightning induced surge entered the system through the Ethernet cables. If this assumption is true (and it almost certainly is true), then providing a surge protector (i.e. Whole House protection device) on the AC service entrance would not address the root cause in this case.
As a general rule: If something is damaged and that something is connected to very long cabling (for example, several hundred feet of Ethernet), then one must consider the possibility that the surge was coupled into and arrived on those long cables (almost certainly in common mode).
Anyone claiming that a 'Whole House' protection system (on the AC service panel) is the be all and end all of lightning protection is not providing complete advice.
The best solution for this case, if it is feasible, is to switch to fiber optic cable for the main length of Ethernet cable.
> ... the working assumption for this thread (at least for most of us) is
> that the lightning induced surge entered the system through the Ethernet cables.
So a surge enters on a phone line that already has surge protection installed (for free)? Or enters on a cable that already has surge protection installed? Not likely.
Electricity must have an incoming path and another outgoing path. Often damage happens on the outgoing path (ie ethernet cable). Many then forget how electricity works. See damage on the outgoing path. Then 'assume' that was an incoming path.
AC electric is a most common incoming path that almost never has protection. Cable or telephone are ideal outgoing paths due to already installed protection. Cable and telephone are often further protected by AC electric wires.
A lightning strike to AC wires far down the street is a direct strike incoming to every household appliance. A typical incoming path.
> that the lightning induced surge entered the system through the Ethernet cables.
So a surge enters on a phone line that already has surge protection installed (for free)? Or enters on a cable that already has surge protection installed? Not likely.
Electricity must have an incoming path and another outgoing path. Often damage happens on the outgoing path (ie ethernet cable). Many then forget how electricity works. See damage on the outgoing path. Then 'assume' that was an incoming path.
AC electric is a most common incoming path that almost never has protection. Cable or telephone are ideal outgoing paths due to already installed protection. Cable and telephone are often further protected by AC electric wires.
A lightning strike to AC wires far down the street is a direct strike incoming to every household appliance. A typical incoming path.
Westom said:
"Typical" yes. 100% of all lightning damage - no. "Typical" does not imply "all". You are confusing one aspect of lightning protection (AC) for the complete solution in all cases as is your long standing habit. You do your readers a disservice with your single-minded focus on "Whole House" protection devices.
The OP reported damage to the "NIC card". It's unlikely in the extreme that a common mode surge on the AC lines would be sufficient to cause damage on the NICs and cause no other damage. It is obviously far more likely that the surge arrived on the long, partially outdoor, Ethernet lines and sought ground via the computer. This is common sense.
It must also be pointed out that AC power systems in commercial properties are typically fed through indoor or pad mounted transformers. Those transformers will not pass a lightning surge in common mode. The AC voltage will go to zero; there will not be a common mode surge in the typical case. Suggesting that a $50 "Whole House" gadget is applicable to a commercial property is probably an over simplification.
- Thread starter
- #20
As per what VEiBLL said, the AC power system is fed through indoor transformer (stepdown 11kv/415v.) The recent case of lightning strike, beside the 3 NIC, on of the network repeater/router switch also gone. We are now in process to get quote for fiber optic link.
- Status
Similar threads
- Locked
- Question
- Locked
- Question
- Locked
- Question