grounding for a data rack

Status
Not open for further replies.
Hey all,

I have a specific question about grounding for a metal rack that houses servers. In my case, I am dealing with 4 server rooms in 4 different locations. At each location, a conductor is being used to bond the rack to a water pipe that is nearby each location. Is this allowed by the code, or do the racks need to be bonded to the service ground directly?
 

erickench

Senior Member
Location
Brooklyn, NY
The rack itself is sort of a structure and would therefore be under the heading of electrically conductive materials as outlined in NEC 250.4(A)(4). The water pipe would either be connected to the service ground or be bonded to another grounding electrode which is in turn connected to the service ground.
 

dereckbc

Moderator
Staff member
Location
Plano, TX
It is allowed. However if each rack has a Server in it and powered by an AC branch circuit, it should already be grounded by the EGC ran with the phase conductors. The supplemental ground does not serve much of a purpose other than to serve a BICSI requirement. The single conductor impedance is way to high to do much of anything. In fact it can do more harm than good.
 

jumper

Senior Member
It is allowed. However if each rack has a Server in it and powered by an AC branch circuit, it should already be grounded by the EGC ran with the phase conductors.

This I understand.

The supplemental ground does not serve much of a purpose other than to serve a BICSI requirement.The single conductor impedance is way to high to do much of anything.

This also.

In fact it can do more harm than good.

Okay, as usual, I am clueless. Why?

What am I missing?
 

dereckbc

Moderator
Staff member
Location
Plano, TX
Okay, as usual, I am clueless. Why?
OK this is a good question.

Let's say the equipment rack servers are supplied from a transformer in the same room some given distance away from the building service entrance. The transformer is bonded to building steel a few feet away. The N-G bond of the transformer creates a new ground reference point which is a good thing and one of the main purposes of using an isolation transformer. Isolation transformers removes virtually all common mode noise sources.

Now you go off and run another supplemental ground to the equipment racks and you reference it back to the service entrance GES. You have just now created a loop injecting unwanted currents which adds common mode noise which you were trying to avoid in the first place. It is not dangerous in any way, it just defeated the purpose of the isolation transformer to begin with. You just put back in the Common Mode Noise you tried to get rid of.

There are other scenarios, but this is the easy one to grasp. Some scenarios can be very dangerous in high rise buildings, or very large slab on grade buildings like manufacturing environments using isolated ground planes.

EDIT NOTE:

Now let's say you have a customer with the above situation and insist on a supplemental ground. What do you reference it too? Simple the ground buss in the isolation transformer. It is the exact same potential as the EGC's and forms no loops, just a parallel ground conductor. It does no harm, but not much good either other than lower the over all impedance a bit. Well it does one positive thing for you, mo money.
 
Last edited:
Status
Not open for further replies.
Top