Question Surge Damage Over Networks

Aikouka

Lifer
Nov 27, 2001
30,383
912
126
We had some bad storms roll through the other day, and a lightning strike close to my house caused some damage to my electronics. From what I can tell, the surge entered in through the coaxial line, went through my modem, and hit my Ethernet network. This leads me to a few different questions/thoughts...

Coaxial Grounding
At the coax's entry point to the house, there is a grounding wire that is attached to the coax run. I checked the continuity between the two, and it definitely is active. However, as we know, this ground run did not stop the surge at all. The ground run goes from the communications box about 8-10 feet to my electric meter where it is simply clamped to the box itself.

So, I guess I'm wondering... is this a good grounding method? I'm not an expert on this, but during my reading, one thing that was noted was that just because something is grounded doesn't mean a surge will take that path. This made me wonder if my ground connection would have been better if it was hooked to a nearby grounding stake rather than the electric meter.

Coaxial Surge Protection
One thing that I didn't do was protect my network lines beyond the original coax grounding line. I had always heard negative things about the coax protection in power strips, but I knew my APC UPS had the same functionality on it. I did a little reading on it, and APC states that it only degrades the signal by about 2dB. I did reroute my coax to run through the UPS before it goes to the modem; however, I was wondering how useful this would be in the end.

Ethernet Surge Protection
I also heard about Ethernet surge protection that is supposed to have minimal impact on performance. I picked up one of these to try out, and I figured I could attach it to my UPS's grounding screw. Although, I guess I'm not sure about the worth of these devices. So, I'm curious if anyone has had some experience with them.

Accepting Wireless
I think one thing that lead to me losing some equipment was likely my philosophy of "if it stays in one place, it goes on Ethernet". For example, my Denon A/V receiver shut off immediately after the lightning strike and entered Protect mode. Fortunately, there seemed to be no lasting damage that I saw. (I didn't check to see if the Ethernet port was ruined by the surge though.) I was thinking about it, and I realized that while the receiver does meet my aforementioned philosophy, it doesn't really need a faster connection due to it mostly using the network for firmware updates.

Unfortunately, while my AVR hasn't had any issues, I can't really say the same for my NUC. I used it on the back of a touch-capable monitor in the kitchen for recipes and such. While I did use the NUC for some multimedia (e.g. watching YouTube while cooking/baking), it never required a lot of bandwidth and would've been fine on WiFi. Although, this brings up another question... at least from my understanding, network surges don't often destroy computers due to Ethernet PHYs having fuses. The Ethernet port tends to die as a result, but the rest of the hardware is (usually) still good. (I did have a computer lose its Ethernet port too.) That makes me wonder... in reducing the size of the electronics on the NUC, did Intel eschew surge protection for the Ethernet port?

In the end, I was probably pretty lucky. The only dead device was my NUC8i5, which has unfortunately gone up in price $100 since I bought it. My modem did still mostly function after the surge; however, its LAN port would only operate at 100Mbps and it would sometimes just stop working correctly for a bit. I also had to replace a 16-port D-Link switch, but it had been giving me problems even before the lightning strike. (Ports would just stop working after power outages.) I was probably pretty lucky that my Ubiquiti network gear didn't kick the bucket. Although, I was a bit worried when my APs were acting up in the beginning. However, resetting power to them fixed any issues that I've seen. Although, the "WiFi Experience" listed in the UniFi Controller was in the 80% range for quite a while yesterday, and I noticed a number of retries. However, last I checked, it has been steadily back in the high 90's.
 

QuietDad

Senior member
Dec 18, 2005
523
79
91
10 feet is a long grounding line and it depends how the electric bo is grounded. You'd be better off hammering in a 4-5 ft piece of rebar right under the cable box and going straight down. It might still not stop everything but would be better. When I did cable installation, the preferred method was to install the box next to the power meter and share the ground with the power meter, not just attach to it.
 

SamirD

Golden Member
Jun 12, 2019
1,489
276
126
www.huntsvillecarscene.com
When I did cable installation, the preferred method was to install the box next to the power meter and share the ground with the power meter, not just attach to it.
That's what we have as the coax ground is attached to the electrical ground.

I dealt with a similar issue a while back when here was a nearby lightning strike. In our case, it was concluded that due to the height of the cabling (the attic where everything is, is probably 50 ft off the ground) that current was probably induced through the air. Many things were already running through ups units, surge suppressors and the like, including ethernet and coax protection built into these units.

The conclusion that we came to was that you can't stop something as strong as lightning and even ethernet has a limited ability to resist high voltages and currents. It's time like this that the equipment repair policy on your ups units comes into play, if not, homeowners insurance (if the damage is significant).

If you can use a usb nic on the nuc, I think you can salvage it. We have this one thin client that lost its built-in nic years ago, and then its parallel port, and a usb nic still keeps it going.
 

Aikouka

Lifer
Nov 27, 2001
30,383
912
126
10 feet is a long grounding line and it depends how the electric bo is grounded. You'd be better off hammering in a 4-5 ft piece of rebar right under the cable box and going straight down. It might still not stop everything but would be better. When I did cable installation, the preferred method was to install the box next to the power meter and share the ground with the power meter, not just attach to it.

Hm, I can always give that a try. It sounds like a simple, cheap solution to improve results in the event that this occurs again.

Also, I went out and measured it instead of just guessing, and the media box is about 13 feet from the meter, and the ground line wasn't run straight. It runs to the meter and straight up to the mounting point at the top, which adds another two feet, and from the media box down about 20 inches at about a 30 degree angle. So, the run itself is probably about 15-16 feet.

Although, this makes me wonder... what liability does the cable company have for an ineffective grounding implementation? Not that I think I'd have a leg to stand on in getting them to replace the damage components nor would it be worth the legal hassle for about $500 worth of stuff, but rather, maybe I could get them to come out and properly ground it?

I dealt with a similar issue a while back when here was a nearby lightning strike. In our case, it was concluded that due to the height of the cabling (the attic where everything is, is probably 50 ft off the ground) that current was probably induced through the air. Many things were already running through ups units, surge suppressors and the like, including ethernet and coax protection built into these units.

Out of curiosity, how much did those protective units actually help? Do you think they actually managed to save any devices?

I was a bit curious about the path of the surge, but also about the lack of destruction in some areas. Theoretically, it seems like it would've likely hopped to most devices -- if not all -- on my network, but quite a few appeared to come out unscathed. Although, now that I look closer, it appears that Port 1 on my main switch is currently limited to 100Mbps for some reason. I might have to jump in there and test this out by swapping ports and seeing what happens with another device in the same port. (A few of my runs go from the main switch to a patch panel to a wall jack.)

If you can use a usb nic on the nuc, I think you can salvage it. We have this one thin client that lost its built-in nic years ago, and then its parallel port, and a usb nic still keeps it going.

Unfortunately, the entire NUC appears to be dead. When I plug in power, the internal standby LED lights up for a second and goes away. Pressing the power button results in nothing. For kicks, I tried removing the battery to see if that would help, and nothing. I tried its power supply on another NUC (Haswell i3), which does have a smaller power requirement, and it worked fine. While I could assume that the NUC's power supply was damaged and can't provide enough amperage to boot the faster NUC, I just don't think that's likely.
 

SamirD

Golden Member
Jun 12, 2019
1,489
276
126
www.huntsvillecarscene.com
Out of curiosity, how much did those protective units actually help? Do you think they actually managed to save any devices?

I was a bit curious about the path of the surge, but also about the lack of destruction in some areas. Theoretically, it seems like it would've likely hopped to most devices -- if not all -- on my network, but quite a few appeared to come out unscathed. Although, now that I look closer, it appears that Port 1 on my main switch is currently limited to 100Mbps for some reason. I might have to jump in there and test this out by swapping ports and seeing what happens with another device in the same port. (A few of my runs go from the main switch to a patch panel to a wall jack.)
I think they did a fair amount of protection. The only unprotected route was from our cable modems to the main router, and it seems that route might or might not have caused damage. The unit has some flakiness going on and it got much worse after--looking like dead ports, but it seems like it may be a firmware issue that just requires a reload to get back up and running 100%. This is our core router so we got a replacement and put it into production. I still have to try the firmware load on the older one. The reason I think it is just a software corruption is that 'dead' ports would come back alive after a few reboots and vice versa, which wouldn't happen if they were hardware dead.

Prior to this event, we had one other event which trashed actually a large amount of equipment--televisions, almost all the ports on a 16-port 10/100 switch, our main router at the time--and this was before we had so much surge and power protection in place, so I think for this event, there was a lot of protection from the devices.

I would definitely check that port--we've seen a port here and there die in cases like this.
 

Aikouka

Lifer
Nov 27, 2001
30,383
912
126
I think they did a fair amount of protection. The only unprotected route was from our cable modems to the main router, and it seems that route might or might not have caused damage. The unit has some flakiness going on and it got much worse after--looking like dead ports, but it seems like it may be a firmware issue that just requires a reload to get back up and running 100%. This is our core router so we got a replacement and put it into production. I still have to try the firmware load on the older one. The reason I think it is just a software corruption is that 'dead' ports would come back alive after a few reboots and vice versa, which wouldn't happen if they were hardware dead.

Prior to this event, we had one other event which trashed actually a large amount of equipment--televisions, almost all the ports on a 16-port 10/100 switch, our main router at the time--and this was before we had so much surge and power protection in place, so I think for this event, there was a lot of protection from the devices.

That's good to hear. I was a little worried about overreacting and trying to put in some protection that just wouldn't do much in the end.

I would definitely check that port--we've seen a port here and there die in cases like this.

Yeah, it looks like Port 2 on my Unifi US-24-250W is bad. At best, it will visibly connect at 10/100, but it won't do anything. (I hooked the old NUC up, and it was stuck at identifying.) That port was where my kitchen connected through, so I guess it makes sense as the NUC, which was connected there, also died. Fortunately, I never tore apart my old NUC, and was able to just hook it back up, perform some updates, and it's like nothing ever happened!