Those of us who have worked on house wiring in the USA understand that the "ground" wire and the "neutral" wire connect together in the circuit breaker box. The wires from the breaker box common connection point go two ways, first to the transfer feed incoming transformer (generally to one of three or four incoming wires depending on the particulars of the connection), and second to a stake in the ground (the "true" earth ground potential, at least until some current is involved). The problem with this arrangement is that for single-phase power in the USA is a common point for neutral and ground pins from the wall outlet. The use of complete isolation (isolation transformer with a grounded shield between primary and secondary windings), or high isolation (low capacitance coupling between two separate bobbins for primary and secondary), or even high frequency plus galvanic isolation (as can be gotten with use of a toroidal non-conductive ferrite core with separated windings and the use of a flyback converter), is still not perfect if the "ground" connection is passed around the isolation boundary. Consider what happens with high frequency noise or a short-circuit from ground to neutral on other devices on the same circuit as the "grounded" device. Potentials have been measured as high as 30V for the simple neutral-to-ground fault condition with a heavy load on 120 VAC in residential applications. The problem is much worse where RF, particularly microwave, frequency is involved.
As a result of these types of issues, I have long personally lobbied for fully isolated and floating safety schemes for lower voltage and RF systems, and for multi-point and multi-path earth grounding for higher voltage systems. I have been on the receiving end of too many electrical "surprises" ranging from old 25KV CRT discharges, to aluminum housed "saber-saw" power to case shorts (with my sneakers in grass glistening with morning dew no less), to getting an unplanned "buzz" from the "case ground" of a lab isolation transformer during testing, and many others. None were pleasant, and a few could have been fatal. Fully balanced and floating potentials with low inductive and capacitve coupling to earth ground generally help here, but at high potentials the only hope is to shield and isolate flesh from wires.
Good article that more people need to read and understand. I have two comments, and the first is that I do have, and use, an isolation transformer on my work bench when I need to isolate an item for either experimental reasons or safety. (Yes, I am a confirmed "boat anchor" and vintage electronics fan, so an isolation transformer is required when working with radios and similar consumner electronics from the 1930s thru the 1970s that have "hot" chassis or a "common" bus that was referenced against one side of the incominmg AC power and did not use a polarized power plug.)
My second comment has to do with my work as an EMC engineer. My emplyer, a maker of RF shielded enclsoures, had a customer who was having trouble trying to achieve a quiet environment for his new enclosure in the range of 10 kHz to 1 MHz or so. I was sent to investigate. Long story short, the enclosure had been properly installed and grounded to the customer's rather massive building ground system. I was fortunate to have brought along an EMC receiver that could operate using batteries, thuis freeing it from connection to an AC power source and safety ground. After 2 or 3 days of intense checking, the conclusion was that the better the enclosure was grounded, the worse the noise was.
Reason? This enclosure was located in a very rocky area with poor soil conductivity. Using the EMC receiver connected to multiple grounds, and even a steel well head, provided clear reception of signals from high power Navy transmitters used to communicate with submarines to AM broadcast stations located over 100 miles away. Since the enclosure could not be safely used without grounding (due to the on-board electrical power filters), the solution was to relocate the enclosure to another facility (owned by the same customer) in an area with better soil conductivity and "quiet" grounds.
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...