1
\$\begingroup\$

I have a 18U server rack enclosure in my 4th storey home and want to connect the metal frame to electrical earth. The rails are already bonded to the frame with 2 wires by the manufacturer.

My wall sockets are 3 pin plugs with earth, live and neutral wiring. Would it be safe to use something like an Earth Bonding Plug for grounding the enclosure frame?

\$\endgroup\$
3
  • \$\begingroup\$ @laptop2d - A connection that seems good when subjected to a low-voltage continuity test is not necessarily a robust connection capable of handling fault currents. \$\endgroup\$
    – vofa
    Dec 11, 2016 at 3:57
  • \$\begingroup\$ The rack is likely connected to Earth through multiple paths via the rack mount tabs of the rack mount equipment mounted on it, whose chassis are all connected to ground and all have three pin c14 cables which have earth connections on them. But that's a lot of shouldas and maybes. \$\endgroup\$
    – Passerby
    Dec 11, 2016 at 5:15
  • \$\begingroup\$ From my experience in aerospace, structural attachments are not adequate for handling faults. This is the kind of design that seems adequately bonded, but you'd be hard pressed to prove it without subjecting the assembly to high current testing. The mating surfaces in the screw attachments likely weren't cleaned, and the structure could be anodized and non-conductive. Incidental fay surface bonding is highly unreliable. In the extreme, this can limit fault current to a point where the breaker can't trip. Fault-current-induced hot spots can start fires. \$\endgroup\$
    – vofa
    Dec 11, 2016 at 5:22

2 Answers 2

4
\$\begingroup\$

No!

That plug has a 1MEG series resistance to earth that allows static charge accumulation to bleed away to earth. This prevents you from destroying a component during handling with a static shock from your body to the chip. This resistor will explode if there is a fault to the chassis that causes large fault current to flow. Then the chassis could be live and floating. If you touch it, you are liable to die.

The earth connection is for safety. It must handle large fault currents to ensure that the circuit breaker protecting the circuit has time to trip.

\$\endgroup\$
1
  • \$\begingroup\$ You are absolutely right that grounding for static electricity and grounding for electric line hazards are two completely separate animals and that a 1M resistor will do nothing for the latter. HOWEVER... No, you're not going to blow up a 1M resistor. It won't do any good for faulty wiring or equipment, but it won't be damaged either. Even at 240 V, the amount of power dissipated by a 1M resistor will be 240*240/1e6 = 58 mW. \$\endgroup\$ Apr 5 at 0:18
1
\$\begingroup\$

To bond the frame to ground it is common to bolt an AC socket strip to the frame; the internal grounding of the AC socket strip to its case is reliable from the factory. As long as you use star washers between painted metal surfaces of the socket strip and the metal of the frame, they'll cut through the paint and make a good ground bond, in my experience.

Those strips are (look for an approval label) tested with a high current four-wire ohms tester, to meet safety codes (of which there are many). The intent of such a test, is that any AC loose wire, or damaged insulation, inside the appliance (the socket strip) will not cause harm because current to the case will open the AC breaker or fuse. Since your rack wiring is both insulated and sheathed, this level of protection is not required of the rack frame; for RF suppression and static shock protection, a simple bonding (as is prudent with any architectural metal) fulfills safety requirements.

Electrical safety codes vary somewhat, but for static shock, even the 'earth bonding plug' may be sufficient. It doesn't do RF suppression nor will it trip a circuit breaker if a high-current wiring fault occurs outside your rackmounted grounded cases.

\$\endgroup\$

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