Help me out here.
How does the IG stay isolated when it is connected to multiple pieces of equipment on a metal rack?
This is another case where the language and nomenclature used by the NEC, electricians, engineers, laymen, etc. gets confusing.
Of the 14 times that the term "isolated ground(ing)" appears in the 2014 NEC, almost exclusively followed by "receptacle".
Excerpted from the NEC:
250.96(B) Isolated Grounding Circuits
406.3(D) Isolated Grounding Conductor
406.3(D)(2) Exception: Isolated Ground Receptacle
517.16 Isolated Ground Receptacle
250.146(D) Isolated Ground Receptacles
406.3(D) Isolated Ground Receptacles
406.3(D)(2) Isolated Ground Receptacles
517.16 Isolated Ground Receptacles
640.7(C) Isolated Ground Receptacles
647.7(B) Isolated Ground Receptacles
647.7(B) Isolated Ground Receptacles
640.7(C) Isolated Grounding-type Receptacles
645.15 Informational Note No. 2: Isolated Grounding-type Receptacles
250.184(B) Single-Point Grounded Neutral System
Whereas the only time that the term "Isolated Ground" is found
alone in the 2014 NEC is in the Index, and actually refers to section 517.16 which is, you guessed it, a section on "Use of Isolated Ground Receptacles":
Index Isolated Ground 517.16 (
section actually titled) "Use of Isolated Ground Receptacles"
However...
The term
single point ground:
250.184 single point grounded
250.184(B) Single-Point Grounded Neutral System
250.184(B) Single-Point Grounded Neutral System
250.184(B)(1) Single-Point Grounded Neutral System
250.184(B)(1)b. Single-Point Grounded Neutral System
And "single point" where relevant to the discussion:
250.30(A)(1) single point
A
complete Single Point grounding scheme is
not fully (or even
at all...) explained in the NEC.
Learning is for school. The NEC just gives us the minimum standards allowable for a "safe" installation (according to the insurance-industry-driven NFPA). There is a lot more to electrical theory, practice and the resultant installation than what is found in the NEC.
...but I digress...
An isolated ground receptacle (250.146 (D)) means that the ground is not continuous to the mounting of the receptacle. It is literally isolated from the box and is not, when made fast, bonded nor continuous to the box, raceway, and all other items which
are bonded...
at that point. At the point where the grounds come together, the
single point, is where the system is then made electrically safe and meets the intended need of the system in reducing noise or for other purposes as necessary.
A single point ground merely utilizes isolated ground receptacles to achieve the overall system isolation. This allows the boxes, raceways, etc. (exposed and available conductive parts) to be
bonded for safety while allowing for the isolation and protection from EMF and other effects on the grounding conductor.
...again, I digress...
In your example, when connected to multiple pieces of equipment on (in) a metal rack:
The power distribution unit (receptacle), surge suppressor, power strip, etc. is provided an isolated ground receptacle with a dedicated ground which is isolated from any and all bonded raceways, enclosures, building steel, counterpoises, etc. until it reaches either the main service or separately derived service, as applicable. Equipment racks (cabinets, housings, mountings, etc.) used isolated mounting hardware to keep it clear of stray connections to any bonded equipment. A grounding bar, also isolated from the cabinet except at one point (EGC/bonding conductor) is made fast and all EGC's for all rack mounted equipment are attached there. Equipment in the cabinet may also be provided with isolation mounts if necessary (i.e., excessive noise from a power supply at a nuisance frequency), but the EGC must be maintained. All grounding bars and plates for a particular system/installation/cabinet/rack/etc., as applicable must be bonded to the same service where power is derived...this is important for obvious (and NEC derived) reasons.
There is a
lot more that goes into this, but the whole point is to properly understand and then make use of Kirchoff's observations and to keep circulating currents (EMF induced or vice-versa) whether power system or equipment derived from getting into areas which they is not desired (causing problems). There are endless types of "interference" and the resultant safe and legal methods which can be used to mitigate issues are simply the best way (cheap, easy, quick...pick two) of getting the job done.
The IEEE Green (Std 142), Emerald (Std 1100), Standard 3000, and Standard 3003 lay this all out nicely. The NEC and several other documents regularly reference (quoting and interpreting) these engineering standards which can also be directly referenced for exceptional installations of equipment with a very electro-sensitive nature.
This is not hypothesis, theory, conjecture, or some sort of hit-and-miss practice. These use of these (now considered) exceptions have a proper time and place for implementation. I agree that they can and have been abused, used as band-aids, and poorly implemented by many people, but that doesn't mean that each instance is nothing but hogwash.
I have personally been utilizing single-point grounding methods for nearly 30 years, building on knowledge and practice of over 100 years in my field alone.
Works fine...lasts a
long time.
