Lightning and grounding... what else can I do.

Regulator

n3wb
Dec 2, 2018
16
13
Atlanta
Ive had my system installed for around 1 year and have replaced the Longspan extenders twice and the gate controller twice. Im not sure what else I need to do. These were not direct hits either.

The system starts coming off the hub to a Longspan into Ubiquity ETH-SP-G2 and then exits the house on a 1,100 ft CAT6e shielded run through the woods to another Ubiquity ETH-SP-G2 and then the second Longspan. From there it enters the outdoor box where its distributed through a hub and POE injectors. There is no direct grounding from either end of the run even though there is a copper grounding line in the CAT6E line. I was not sure if that needs to be grounded. Also the ETH-SP-G2 is not grounded. After typing this I'm realizing there really is no grounding anywhere but also am not sure where to most effectively do this.

Ive also lost two cameras to lightning. both cameras were in trees with separate CAT6e POE lines running from the basement out to the cameras. On one occasion the 4 Port POE injector was also blown.
 
Last edited:
Since you apparently have power on both ends of the 1,000 ft. CAT6e I'd replace it with a Ubiquiti Layer 2 Transparent bridge like below using 2 Nanostation M5's.

Ubiquiti_layer2_bridge-cams.jpg

I would install the UBNT surge protectors as directed by their publication EXCEPT I would advise against using a STP (Shielded Twisted Pair) cable going into the house; I would use UNshielded:

Ubiquiti_ESD_Protection_no-STP-to-LAN.jpg
 
Last edited:
  • Like
Reactions: looney2ns
1,100 ft CAT6e shielded run
Base-T Ethernet is only rated for 328'. By the time you add devices to boost the range of Base-T Ethernet and surge damaged devices, it would have probably been cheaper to run fiber in your trench.

Tree mounted cameras are generally more likely to be damaged by lightning. Ethernet SPDs won't do much if they aren't grounded.
Putting a lightning rod in a tree is always an option, but doing this correctly would be complicated in this setup (if it's even possible).
 
  • Like
Reactions: TonyR
Great, thank you Tony. Do you know if the airmax can punch through 1,000 ft of foliage. My understanding is even 900mhz would struggle at that distance. I had considered fiber but the pricetag was significantly higher than the 6e.
 
I had considered fiber but the pricetag was significantly higher than the 6e.
I just checked and you can get 1000' of outdoor single mode fiber for as little as $250 for non-armored cable and around $600 for armored. You would need a few other things to make it work and likely to pay someone to terminate it.
Great, thank you Tony. Do you know if the airmax can punch through 1,000 ft of foliage. My understanding is even 900mhz would struggle at that distance. I had considered fiber but the pricetag was significantly higher than the 6e.
Can you go below the foliage?
 
Great, thank you Tony. Do you know if the airmax can punch through 1,000 ft of foliage. My understanding is even 900mhz would struggle at that distance. I had considered fiber but the pricetag was significantly higher than the 6e.
Yikes. No, it cannot. Bad assumption on my part that you had clear LOS.

EDIT: You'd think I'd know better...was born there in ATL, grew up in East Point...'49 to '69.
 
Last edited:
The thing that most people do not visualize when they're imagining how lightning strikes cause damage is this:

The earth in most places is a crummy conductor of electricity. And in the other places, it's an even worse conductor of electricity!

Try to envision the earth in any given area as a 3 dimensional volume of varying resistivity. Then imagine a lighting strike inserting a pulse of, perhaps 100,000 amperes of current into this mass of varying resistance, trying to find its way to "earth ground" (whatever that is).

We know that current times resistance = voltage. As this enormous pulse of current spreads out through the earth from the point where it made contact, voltage gradients appear from any point to any other point within this volume of random resistances.

So let's say we've driven a ground rod in at one building, and another at a second building. And let's imagine lightning striking a tree or whatever anywhere near these two buildings. As the current spreads out from that point of impact, very high voltage gradients suddenly appear between any two points, and certainly between our two ground rods. Suddenly "ground" at one building can be thousands, even tens of thousands of volts different than "ground" at the other building.

It is this voltage DIFFERENCE that causes damage.

And this happens even within a building. Grounds are never perfect. And even if we achieve a low resistance grounding scheme, that doesn't overcome the inductance of the ground wiring. So when we present this ground system with a very fast-rising current waveform, we can still end up with momentary voltage differences from any point to any other because these high currents are prevented from flowing, for a moment, by the inductance of the wiring itself.

The concept of "a good ground" is mostly a fantasy. You can go to heroic lengths to achieve a better ground, but that's about it. And within a large building, or certainly from one building to another, you're just going to have to deal with high voltage differences appearing when lightning strikes nearby. Let alone a direct hit!

Further, high current pulses, like we see with lightning, even in the lightning stroke itself, will radiate an EMP (electromagnetic pulse) that can be quite sizable. And nearby conductors will act as receiving antennas and high voltages can be induced in those nearby conductors as well.

So for connections between buildings, since we know that their "grounds" will often be tens of thousands of volts different from each other, the best policy is to NOT connect the buildings directly with signal wires. Radio links or optical links are the way to go. And even for shorter runs, it's good to have surge protection at both ends of any direct electrical signal connections.


When setting up radio sites on mountaintops, we would drive an array of ground rods into the earth surrounding the radio shack, often 25 or more rods, connect them in a ring using enormous copper cable exothermically welded at every point. You know what was considered a "good" ground resistance for such an array? 25 Ohms! All of that, and 25 Ohms to "earth" was considered good!

Anyhow, this "good ground" was then connected to a heavy (perhaps 1/4" thick) copper plate that serves as the "entrance" for all electrical connections entering the building. This plate is, then, our "ground" reference.

The tower was grounded to that same plate, and all equipment grounds went directly and independently (isolated grounds) to that plate (no daisy-chaining allowed). Every signal or power connection into the building had a surge protector mounted to that plate as the connection passed through a hole in said plate.

This is the concept of a "single point" or "star" ground.

Let's imagine now, lightning striking our radio tower on the top of that mountain. And this happens hundreds of times or more every year for one of these sites because they're always the best point for a discharge in the area. As the lighting strikes the tower, the current pulse goes down the tower and its ground cable to meet our "entrance plate" and the huge ground system for the building and tower. At that moment, let's say this strike is a big one, and has 100,000 Amperes of peak current. What's 100,000 times our 25 Ohms to true "earth ground"? It's 2.5 million Volts! So our "ground" is suddenly at 2.5 Megavolts away from "earth ground"!

So why doesn't that radio station get blown up and be off the air constantly? Because of the single point or "star" ground system. Everything in our radio shack jumps to that same 2.5 Megavolts potential at the same time. So there is no DIFFERENCE in potential between any of our pieces of equipment. So the gear doesn't see any harmful voltage across it or through it. Properly set up, a site like this shrugs off thousands and thousands of direct lightning strikes over its lifetime.

We actually do the same thing when laying out printed circuit boards for analog applications.

The idea is that even if a high current passes through one path to our "single point ground", the voltage drop in that particular ground path will not impose a voltage, with respect to that ground, in any other device's "ground". That's why we cannot abide "daisy chaining" our grounds. To do so is to invite ground-referred noise.

But unless you're going to set up your buildings to each have this sort of lightning-tolerant single-point grounding system, you won't ever find much safety from lightning-induced surges even within a large building. And you'll never avoid high voltage potentials between different buildings.

So you just have to envision the way the current from a lighting strike spreads out in the crummy "ground" of the earth, and realize that you must simply accept that voltage gradients will be induced from point to point on the ground.

This is why they tell you to squat down, making your body into a tight ball, with your feet placed as close together as possible when you feel a lightning strike is imminent and you're caught outdoors. With your feet right against each other, the hope is that the voltage potential between them will be relatively small as the current spreads out in the earth under your feet, so there will be no reason for it to choose passing through you as one of the "preferred" current paths.

Go with the radio or fiber-optic links. You'll never truly "ground" two buildings together in any lightning-proof way.
 
I don't recommend getting struck by lightning :blankstare:
It's pretty hard to do much between buildings that are 1100 feet apart.

The system starts coming off the hub to a Longspan into Ubiquity ETH-SP-G2 and then exits the house on a 1,100 ft CAT6e shielded run through the woods to another Ubiquity ETH-SP-G2 and then the second Longspan. From there it enters the outdoor box where its distributed through a hub and POE injectors. There is no direct grounding from either end of the run even though there is a copper grounding line in the CAT6E line.
Just so we're clear, do you have electricity on the building in the woods?
If yes: is the power fed from the house or does it have it's own utility connection and electric meter? What type of cable was used to supply power to the building in the woods?
 
Well stated @J Sigmo. One thing I learned early in my career about grounding and shielding is that, as simple as it sounds, it is one of the most complex things to really achieve in reality. Great explanation of this problem! The use of Single Point Ground (SPG) is the best design practice for anything from a simple single bay equipment rack, to the large radio towers you speak about.

Edit: Another good point to bring up is if using shielded cable for EMI/RFI. Only ground the shield at one end, never both ends. Once you ground at both ends, you now have a conductor for current flow and defeat the purpose of the shield. Typically ground shield at signal source end, and of course tie it to your SPG.
 
Last edited:
  • Like
Reactions: c hris527 and Ckb3
A little off topic, but to illustrate the point of potential differences. Power substations have a grid of wire buried just below the surface, maybe a foot down. That grid is the single point of ground for the whole station. At each intersection there's a ground rod. Some thieves caught on to this and decided to steal the copper wire of that grid from a 500KV substation, or maybe it was just a switch yard. Anyhow, they were working on it and when they cut the grid wire at one point, they got the jolt of death due to the potential difference in the ground system, itself.

Buildings that far away demand a non-conductive link, fiber optic or RF. Worst case there is only the equipment on one end gets fried in a strike.
 
  • Like
Reactions: c hris527
The thing that most people do not visualize when they're imagining how lightning strikes cause damage is this:

The earth in most places is a crummy conductor of electricity. And in the other places, it's an even worse conductor of electricity!

Try to envision the earth in any given area as a 3 dimensional volume of varying resistivity. Then imagine a lighting strike inserting a pulse of, perhaps 100,000 amperes of current into this mass of varying resistance, trying to find its way to "earth ground" (whatever that is).

We know that current times resistance = voltage. As this enormous pulse of current spreads out through the earth from the point where it made contact, voltage gradients appear from any point to any other point within this volume of random resistances.

So let's say we've driven a ground rod in at one building, and another at a second building. And let's imagine lightning striking a tree or whatever anywhere near these two buildings. As the current spreads out from that point of impact, very high voltage gradients suddenly appear between any two points, and certainly between our two ground rods. Suddenly "ground" at one building can be thousands, even tens of thousands of volts different than "ground" at the other building.

It is this voltage DIFFERENCE that causes damage.

And this happens even within a building. Grounds are never perfect. And even if we achieve a low resistance grounding scheme, that doesn't overcome the inductance of the ground wiring. So when we present this ground system with a very fast-rising current waveform, we can still end up with momentary voltage differences from any point to any other because these high currents are prevented from flowing, for a moment, by the inductance of the wiring itself.

The concept of "a good ground" is mostly a fantasy. You can go to heroic lengths to achieve a better ground, but that's about it. And within a large building, or certainly from one building to another, you're just going to have to deal with high voltage differences appearing when lightning strikes nearby. Let alone a direct hit!

Further, high current pulses, like we see with lightning, even in the lightning stroke itself, will radiate an EMP (electromagnetic pulse) that can be quite sizable. And nearby conductors will act as receiving antennas and high voltages can be induced in those nearby conductors as well.

So for connections between buildings, since we know that their "grounds" will often be tens of thousands of volts different from each other, the best policy is to NOT connect the buildings directly with signal wires. Radio links or optical links are the way to go. And even for shorter runs, it's good to have surge protection at both ends of any direct electrical signal connections.


When setting up radio sites on mountaintops, we would drive an array of ground rods into the earth surrounding the radio shack, often 25 or more rods, connect them in a ring using enormous copper cable exothermically welded at every point. You know what was considered a "good" ground resistance for such an array? 25 Ohms! All of that, and 25 Ohms to "earth" was considered good!

Anyhow, this "good ground" was then connected to a heavy (perhaps 1/4" thick) copper plate that serves as the "entrance" for all electrical connections entering the building. This plate is, then, our "ground" reference.

The tower was grounded to that same plate, and all equipment grounds went directly and independently (isolated grounds) to that plate (no daisy-chaining allowed). Every signal or power connection into the building had a surge protector mounted to that plate as the connection passed through a hole in said plate.

This is the concept of a "single point" or "star" ground.

Let's imagine now, lightning striking our radio tower on the top of that mountain. And this happens hundreds of times or more every year for one of these sites because they're always the best point for a discharge in the area. As the lighting strikes the tower, the current pulse goes down the tower and its ground cable to meet our "entrance plate" and the huge ground system for the building and tower. At that moment, let's say this strike is a big one, and has 100,000 Amperes of peak current. What's 100,000 times our 25 Ohms to true "earth ground"? It's 2.5 million Volts! So our "ground" is suddenly at 2.5 Megavolts away from "earth ground"!

So why doesn't that radio station get blown up and be off the air constantly? Because of the single point or "star" ground system. Everything in our radio shack jumps to that same 2.5 Megavolts potential at the same time. So there is no DIFFERENCE in potential between any of our pieces of equipment. So the gear doesn't see any harmful voltage across it or through it. Properly set up, a site like this shrugs off thousands and thousands of direct lightning strikes over its lifetime.

We actually do the same thing when laying out printed circuit boards for analog applications.

The idea is that even if a high current passes through one path to our "single point ground", the voltage drop in that particular ground path will not impose a voltage, with respect to that ground, in any other device's "ground". That's why we cannot abide "daisy chaining" our grounds. To do so is to invite ground-referred noise.

But unless you're going to set up your buildings to each have this sort of lightning-tolerant single-point grounding system, you won't ever find much safety from lightning-induced surges even within a large building. And you'll never avoid high voltage potentials between different buildings.

So you just have to envision the way the current from a lighting strike spreads out in the crummy "ground" of the earth, and realize that you must simply accept that voltage gradients will be induced from point to point on the ground.

This is why they tell you to squat down, making your body into a tight ball, with your feet placed as close together as possible when you feel a lightning strike is imminent and you're caught outdoors. With your feet right against each other, the hope is that the voltage potential between them will be relatively small as the current spreads out in the earth under your feet, so there will be no reason for it to choose passing through you as one of the "preferred" current paths.

Go with the radio or fiber-optic links. You'll never truly "ground" two buildings together in any lightning-proof way.
I recently had to mount Cameras and a wireless bridge on a Radio Tower and had a HELL of a time figuring out a Grounding Game for them, The Grounding of the Tower, Shelter, Propane Tank ,generator and Security fence did not seem Right and IMOHO left a LOT of Ground Potential between them. I did have a conversation with the Tech for Motorola and he told me about a year and a half back they had a hit and did a half a million damage to the equipment inside that shelter but could never figure out where it came in. I did finally did contact a company called DITEK and suggested a setup on the Inside and Outside. Its always on the back of my mind about my grounding scheme at the site. I will include a pic any suggestion would be helpful. The Cams are Mounted of the perimeter fence and the wireless bridge is about 60' up the tower.

shelter.jpg
 
  • Like
Reactions: iseeker
Well, not a whole lot to see from 1 pic, but it seems that you have 2 Single Point Gnds. You do have a heavy gage wire from the terminal block to the SPG copper buss bar, so it will be pretty negligible for the purpose. I think I know what you were thinking w/ mounting everything on the board for easy assembly and mounting. Ideally, you want your term block mounted directly on copper buss SPG, since it looks like its conductive. Ditec recommends that the Surge Suppressor enet patch cable from/to equipment be at least 3' long. Don't really understand why, except they maybe playing with the cable inductance to help w/ surge. Make sure that power strip safety gnd it tied to that copper SPG somewhere near. Best of luck with your install, and figuring out what the previous tower installers did.

https://www.diteksurgeprotection.co...ion-guides/688-dtk-mrjexts-installation-guide
 
  • Like
Reactions: c hris527
Well, not a whole lot to see from 1 pic, but it seems that you have 2 Single Point Gnds. You do have a heavy gage wire from the terminal block to the SPG copper buss bar, so it will be pretty negligible for the purpose. I think I know what you were thinking w/ mounting everything on the board for easy assembly and mounting. Ideally, you want your term block mounted directly on copper buss SPG, since it looks like its conductive. Ditec recommends that the Surge Suppressor enet patch cable from/to equipment be at least 3' long. Don't really understand why, except they maybe playing with the cable inductance to help w/ surge. Make sure that power strip safety gnd it tied to that copper SPG somewhere near. Best of luck with your install, and figuring out what the previous tower installers did.

https://www.diteksurgeprotection.co...ion-guides/688-dtk-mrjexts-installation-guide
Thanks for the reply, When I called Motorola, They pretty much told me to put in a system like the native cameras guarding the outside of the shelter, They have 4 Analog cams guarding the outside of the shelter so I saw the Ditek stuff and talked to them. That is where I copied the bus bar from.
shelter1.jpg

Re-Mounting the bus bar should not be a issue and grounding the power strip either. If My equipment gets destroyed I really do not care, I worry about my equipment being the path of destruction for the Radio gear. It has been up and running since last fall with No issues but storm season is approaching soon. I have had a few discussions with higher up on the grounding issues and they kind of fell on def ears. Motorola did not put in the site pads or grounding system and shelter BLDG and they told me they have a team that can come out and analyze it at some point. I also have outside protection for the 2 cams and wireless bridge, basically what you see on the inside but for outside use.
 
You can always look at it this way. Everything is located on a tower (lighting rod). Direct strike and those little surge suppressors will be popping like popcorn. Probably everything else too. Then no one can blame no one, start over and you have another tower job to do.
 
  • Like
Reactions: c hris527
You can always look at it this way. Everything is located on a tower (lighting rod). Direct strike and those little surge suppressors will be popping like popcorn. Probably everything else too. Then no one can blame no one, start over and you have another tower job to do.
I will as time permits make those changes, thanks for the input, The setup is not permanent and is guarding a nearby construction project and also taking snapshots for a time lapse movie of the new facility, that end of it has been working out better than expected. Again thanks for the pointers.
 
Well stated @J Sigmo. One thing I learned early in my career about grounding and shielding is that, as simple as it sounds, it is one of the most complex things to really achieve in reality. Great explanation of this problem! The use of Single Point Ground (SPG) is the best design practice for anything from a simple single bay equipment rack, to the large radio towers you speak about.

Edit: Another good point to bring up is if using shielded cable for EMI/RFI. Only ground the shield at one end, never both ends. Once you ground at both ends, you now have a conductor for current flow and defeat the purpose of the shield. Typically ground shield at signal source end, and of course tie it to your SPG.

The point of grounding the shield of any shielded twisted pair type cable is excellent. The shield needs to not end up providing a ground loop.

My experience for analog signal wiring has shown that I always want to tie the shield at the "receiving" end of the run. The reason for this is that you want the signals that the shield "picks up" to "drive" the signal ground or "signal low" terminal of the receiving device. That way, when there is induction from the shield to the internal wires in the shielded twisted pair cable, the interference or noise becomes common mode to both the signal hi and signal low conductors and will not appear as a differential mode signal at the receiver. After all, noise is "seen", and will only be a problem at the receiving end.

A great way to prove this is to play with shielded twisted pair microphone or guitar cabling, or runs between audio pre amps and power amps. If you have cables built with the shield tied at only one end (as we are recommending), you can try connecting them either way. With the shield tied at the receiving end, the hum disappears. With it tied at the transmitting end, you will likely get hum.

This is also how I connect all 4-20mA devices, and it has never failed to produce the lowest noise pick-up.

But I have seen instrument manufacturers' equipment manuals advise tying the shield at the transmitter end, or worse, at both ends! So you have to even take "expert" advice with a grain of salt in this area. Trial and error has always resulted in lowest noise for me with the shield tied at the signal receiver end.