2
\$\begingroup\$

I recently worked with a client who experienced a transient on their Ethernet cabling. As you can see, no actual PCs were damaged, only the Ethernet cards of some machines. We don't know where the surge originated.

Can anyone explain how the damage could have occurred as shown? I thought Ethernet had isolation transformers at each connector which would likely stop surges from propagating.

Damaged nodes are highlighted in red. Note that main network stack in the center has AC mains surge protection.

Also note that the main network switch was completely destroyed - i.e. would not even power on.

As a follow-up, would Ethernet surge protection hardware be helpful for preventing this type of damage? If so, where in this diagram should it be installed? main stack or at each PC?

Thank you! enter image description here

\$\endgroup\$
  • \$\begingroup\$ Is it possible the Main switch failed in a spectacular enough fashion that it took out the connected equipment? \$\endgroup\$ – JS. Jan 6 '16 at 20:53
  • \$\begingroup\$ I don't think so... it was a Netgear business class 48-port gigabit switch, and as shown, the firewall/modem/etc weren't damaged nor were most of the computers connected directly. Also this switch was protected with mains AC surge protection, and none of the other equipment attached to that protection was damaged. I popped the cover, but couldn't see any obvious failure indicators. \$\endgroup\$ – Ryan Griggs Jan 6 '16 at 21:03
  • \$\begingroup\$ Was this lightning related? Also, some idea of building size and earth grounds would be in place, as grounding is suspect here. Always remember to think of ground as a good place to get surged unless they are all starred properly. \$\endgroup\$ – Erik Friesen Jan 6 '16 at 21:51
  • \$\begingroup\$ Single story concrete block building. I don't know how the grounding is done. Plugs are 3-prong with earth ground. Don't know if it's lightning related. \$\endgroup\$ – Ryan Griggs Jan 6 '16 at 22:09
  • \$\begingroup\$ are the 5 port switches grounded? \$\endgroup\$ – Jasen Jan 6 '16 at 23:50
1
\$\begingroup\$

Is there a potential difference between the lines supplying the power sockets? In the UK, a large building or a group of buildings in a complex may have three different phases supplying the 240V (different phases to different parts of the building), with a possible 415V potential across different areas. A surge or earth fault could have been to blame.

\$\endgroup\$
  • \$\begingroup\$ That's an interesting thought. The LAN is in a single (large) building with a single mains feed. \$\endgroup\$ – Ryan Griggs Jan 6 '16 at 21:47
  • \$\begingroup\$ twisted pair ethenet ports are good for a couple of thousand volts of isolation \$\endgroup\$ – Jasen Jan 6 '16 at 23:57
  • \$\begingroup\$ That's what I thought, so I'm having a hard time understanding how this surge traveled so far. If it was that many thousands of volts, it seems more damage would have occurred at one location, indicating the source. \$\endgroup\$ – Ryan Griggs Jan 7 '16 at 20:41
1
\$\begingroup\$

I would think the surge entered through in the ground of the big switch, as the other equipment in the rack shares the same ground that equipment was equipotential and was not damaged, the surge then radiated as a common mode signal out through the network cabling destroying two directly connected network cards and hitting two switches.

this sort of damage needs kilovolts, and to get kilovolts on ground it takes a lightning strike. or a shock line, did the did the big netgear switch fail spectacularly?

\$\endgroup\$
  • \$\begingroup\$ No, it just wouldn't power on. I took off the cover to look for blown components but nothing was visible. \$\endgroup\$ – Ryan Griggs Jan 7 '16 at 20:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.