Your objective is to extend the zero volt reference to the neutral-ground input at the equipment location.
Specifications often require an isolated grounding (IG) system when there's a concern for electrical noise on the equipment grounding system. This is because noise can cause operational problems for electronic equipment. By using an IG system, in lieu of a solidly grounded (SG) system, we may think we've eliminated potential problems. This isn't necessarily true.
IG systems can create operational problems, even when you install them according to NEC requirements and IEEE recommended practices. How so? Read on.
Opening Pandora's box. In the early '80s, with the unveiling of a line of high-speed data processing products, a large computer equipment manufacturer began specifying a grounding design called an isolated ground. It required you to run an insulated equipment-grounding conductor to the grounding terminal of a special receptacle. The objective of this design was to extend the zero volt reference (created by the neutral-equipment ground bond at the electrical service entrance) to the neutral-ground input at the equipment location.
Unlike standard receptacles in SG systems, these receptacles had the grounding terminal insulated from the mounting strap. Thus, their notation as IG receptacles.
In cases where technicians tried to save the customer money, they altered existing standard receptacles, often creating unsafe installations. Over the next few years, other equipment manufacturers specified this type of grounding system.
Most were lax in the exact requirements of an IG installation. Some technicians relied on an electrician's interpretation of how to install it. To make matters worse, the NEC and IEEE didn't have requirements or recommended practices for installation of such a grounding system. Many sites developed serious safety and equipment performance issues.
How you should install an IG system. In the 1987 NEC, Sec. 250-74, Exception No. 4 and Sec. 250-75, Exception provided installation requirements of the IG system for receptacles and hardwired equipment, respectively.
Today, you'll find those requirements in Sec. 250-146(d) and Sec. 250-96(b) of the 1999 NEC. They require you to run an insulated equipment-grounding conductor with circuit conductors from the equipment ground terminal at the receptacle to the equipment-grounding terminal at the derived system or service. (See Figs. 3, 4, and 5, on page 18, original article.)
For safety purposes, you can't run an IG equipment-grounding conductor in its own conduit or outside the branch circuit or feeder conduit. You also can't terminate the IG equipment-grounding conductor to a lone ground rod.
For equipment performance purposes, IEEE recommends you contain the IG conductor and the circuit conductors in metallic conduit to protect against radiated electromagnetic interference (EMI)/radio frequency interference (RFI). IEEE recommends you provide a separate hot, neutral, and equipment-grounding conductor for each branch circuit. However, most contractors view this recommended wiring practice as cost-prohibitive.
Sources of common-mode voltage. Most equipment, with the exception of incandescent lighting, is a source of common-mode voltage. Devices containing single-phase motors direct-couple noise to grounding planes of the AC power system. Examples of common-mode voltage include vending machines, copiers, laser printers, refrigeration/chiller units, and space heaters. Loose connections on the grounding circuit, subject to vibration, may also cause mid-level electrical noise.
Other sources can induce a voltage on the grounding circuit via EMI or RFI. Radio/TV antennas, motion detectors, two-way radios, cellular phones, pagers, and fluorescent lighting bring about these types of disturbances.
Regardless of the source of the disturbance, any voltage greater than 1V between neutral and ground generally will cause equipment malfunction.
Effect of common-mode voltage on equipment. Electronic equipment communicates internally through a digital pulse known as a bit, or as a string of bits known as bytes. A typical bit resembles a square wave pulse. The amplitude of this pulse varies with equipment design and application. Typically, transistor-transistor logic (TTL) and complementary metal oxide semiconductor (CMOS) logic operate at 5V. When active, the pulse is at a "logic 1" or high state. When not active, it's at "logic 0" or a low state.
At the leading and trailing edge of the pulse are transition points. Here, for a short time, the pulse is neither at logic 1 or logic 0. If a spurious common-mode voltage occurs between the neutral and AC equipment ground at the same time transition points occur, it's possible the intended signal could be reversed. As a result, there is one less bit in the stream of information travelling to circuitry within the device or to an external device. When this occurs, the internal circuitry won't function as a result of parity error to the bit structure. Results of parity error include keyboard lockup, read/write errors to disk drives, and system "reboots."
In part two of this series, we'll discuss benefits and drawbacks of the IG system, and present rules for testing and installation.