220V in twisted pair

A little randomly obtained extreme testing of network equipment.

Initial configuration: between a computer and server approximately 40 meters of twisted pair. On the one hand, a computer with a curved power supply supplying 220 V (relative to the ground) to all contacts. (Ie, the voltage between the contacts of the pair was correct, but each of them had a potential of ~ 200V relative to the ground). Until the moment when an enikeyshchik, who dealt with a stupid computer, used this method “oh b # $ on # & @ to% & @ #!”, The history of periodic network loss lasted more than a year, during which time quite a lot of hardware was tried. Everything was complicated by the fact that the BP did not apply 220 directly to the case, so the primary measurements (somewhere on the third complaint about the loss of communication) did not show anything seditious.

So, forcedly tested equipment:
')
SOHO:

• Intel Business Hub (4 ports) - stopped working after an hour or two, and did not work until it was turned on / off
• Zyxel ES-2108 - burned port
• Dlink DES-1005D - burned out with a complete set (all ports stopped working).
• Cisco 1750 - silently made the port down (as if there is no signal there) before the hardware restarts

19 "equipment :

• 3COM 4200 - after a couple of days, turned off the port until the switch was completely de-energized (after turning it off and on, the port came to life again)
• 3COM 4200G - similarly turned off, but the port "came to life" within a day after disconnecting the cable
• Allied Telesis Rapier AT24i - ignored and worked
• Cisco 1841 with a card with switchports - ignored and worked
• Edimax ES-3124RL - the only one of the rack, burnt his port

I tried to find information about the “endurance” of the devices, but apparently this is an abnormal mode of operation of the devices ...