[BC] LAN Weirdness

[Jerry Mathis]

A situation with the neutral bonded to ground anywhere outside the distribution panel or service entrance is a serious problem. Even more serious is the fact that your computers do not have a safety ground connection. Chances are good that the computers do not tie neutral and (chassis) ground together internally, or with anything more than a small capacitor or an MOV, so static charges could easily build up on the chassis and find a discharge path through the NIC card.
 
I would recommend that you correct the house's wiring problems. Also, I would try moving the switch out of the rack, away from the STL and see if that helps.
 
--
Jerry Mathis

On Wed, Feb 10, 2010 at 9:15 PM, Tom Taggart <tpt at literock93r.com> wrote:

>The station is in a house, which I suspect was a DIY (Do It yourself)
>project for the original and sole owner back in '85.  So there is a
>fair amount of weirdness in the electrical wiring, such as not
>connecting the third wire (safety) ground at various outlets.

>The switch, however, is mounted in a metal rack. The rack is tied to a
>2" copper strap, which in turn runs to several ground rods outside the
>building.  Neutral and safety ground for the rack power is tied to the
>strap. The electrical neutral and safety grounds for the house are
>tied to the ground strap by a 15 foot piece of #4 (since we couldn't
>find an easy way to run the strap to the box). There is strap run into
>the control room, which is connected to the console, and to the
>neutral and safety ground of the branch circuit for the control room
>equipment. This includes the 4 computers in the control room. The PD's
>computer, in his office, and my computer (in the rack room) aren't
>tied to anything in particular, except through the neutral. I know
>there is no safety ground on the outlet for my computer--but that
>computer was installed after the first time we had a lock-up of the
>switch.

>The little router and cable modem (two separate units) sit right above
>the switch. They are in plastic cases, run by wall warts, but are all
>plugged into the same UPS. The cable modem, of course, is grounded to
>the copper strap by a feed-through "F" connector ground.  (Had a
>switch burned up at another location when I relied on the cable
>company's "ground" at the service entry box).

>The computers on the LAN are connected by CAT5;(solid wire CAT5
>between rooms to jacks, jumpers to the computer, Leviton rack mount
>jack field in the rack, then jumpers into the switch). I presume there
>would be some way that a ground loop could develop depending on the
>design or failure of the NIC's in each computer, but one would assume
>the switch would account for this in the design.

>RF: There's an Armstrong STL in the same rack--which drives my little
>weather radio nuts, but I would doubt that would cause this
>problem--or it would be a more continuous problem. Not one that would
>pop up when there is no-one moving around to reflect RF.