isolated ground system

[oldsparky]

I keep being asked about isolated grounds as in a data system function.
My thoughts are that you would have to go back to the xo of the transformer,or to the lug in the MDP from the xo which is one and the same.

This is a new area for me and will anxiously await input back from you more experienced guys in this area.

I am still looking in NEC2002 and to no avail.

thank you

 

[bwyllie]

Re: isolated ground system

You are correct in that the IG would need to originate at the Xo of the xfmr.
I feel that IG is a waste of the client's money as devices have evolved today to the point where an IG is useless. If a datacenter has a UPS than that would be sufficient to accomplish the "clean power".

 

[arcandspark]

Re: isolated ground system

According to 250.146(d), and numerous other cross references that are quite vague. It appears that the Isolated ground conductor (see sensitive equipment and electrical noise) is to be run with the circuit conductors back to the point of connection of the main service to ground. It does not have to go back to the xo or beyond the mdp grounding bar.

However, this is complicated if you have separate systems derived by transformers. In which case you have a new service and thus a new mdp for the load side of the transformer. Accordingly, you will be required to acquire a new ground for the new system. If I understand it, this would be the point of attachment for the iso.ground.

I too agree that these systems are redundant with todays technology. Surge arresters that most of us have attached to our computers are equipped to handle this problem sufficiently to render iso grounds obsolete.

See647.6,647.7,250.96(b),406.2

Hope this helps.

 

[dereckbc]

Re: isolated ground system

IGR circuits became popular back in the 80?s when financial institutions started relying heavily on data processing for record keeping and transactions. When the data processing industry started it relied on I/O ports which consisted mostly of RS-232 interconnections and other signal methods that used a common ground to transport signals.

Since this signal ground was in common with the electrical ground at multiple points, minute currents were injected into the electrical ground. These currents produced voltage differential potentials along the length of the ground circuit called common mode noise.

Since the data systems used the electrical ground as a signal path, the common mode noise would sometimes inject errors into the data system causing corruption. So the IGR circuit principles were developed to combat the problem.

But in today?s modern data centers most signal transmission is via optical and balanced protocols. So the need for IGR for the most part has been eliminated by technology, but not completely as some equipment still uses RS-232.

I design data center power and grounding systems. The approach I take is to use large scale UPS systems and distribute power from the UPS via feeders to PDU?s (power distribution units). The PDU?s are located strategically in equipment line-ups to keep branch circuit lengths very short. Each equipment rack has one to four dedicated branch circuits depending on customer requirements. The big difference I take (if budget allows) is the PDU transformers are wound as balance symmetrical output or what the NEC describes as technical power systems covered in 647 (120/60 VAC).

Even if I cannot use balanced power we do not supply IGR as it does not add any value. The SDS created by the UPS and PDU?s clean up any problem created upstream. Sorry to be long winded and hope this helps.

 

[tom baker]

Re: isolated ground system

I don't belive IG systems are worth installing anymore. And, unless you have good supervision, any one can plug any load into the IG circuit, such as a laser printer. No one knew what the orange face receptacle was for, and now with only an orange triange its even harder to tell.