When grounding of branch circuit equipment was first required, grounding and bonding methods left a lot to be desired. BX (armoured) cable did not have a grounding strip. The armour was used as the ground conductor. After a few years of service, with either aluminum or steel armour, surface corrosion would make the turn to turn resistance from turn to turn higher than the resistance around the loop of armour. The armour became an induction coil with the added resistance of the strip of armour stretched out flat.
EMT or thinwall conduit used indent type couplings. These were a straight sleeve type coupling that was indented into the conduit in four places by a special tool. It just took a little bending for the connections to become a little loose. Mechanically the couplings would hold even when loose but electrically, after a few years and a little surface corrosion, they were a disaster for electronic signals.
Most small commercial buildings housing supermarkets were wired with a combination of thinwall conduit and armoured cable.
Over the years grounding and bonding methods were improved to the point that they were the equal of isolated ground circuits.
The first problems that I was aware of came with the advent of Point Of Sale devices (Cash registers) with a data link to a computer.
The designers assumed that a ground connection was automatically a secure connection to the ground mass of the earth. Despite a clause in the North American codes of the time that no equipment may depend on a ground conductor for its operation except for grounding devices, data designers saved a wire by using the equipment ground as a signal return path.
Through code committees the data people lobbied for changes in the Codes to allow isolated grounds.
Concurrent with this code grounding methods were being upgraded. Although many data installations in older buildings needed upgrades to the grounding and bonding systems to function reliably, bonding methods in new construction was generally suitable for dependable data transfer.
I suspect that many data problems were blamed on grounding when the actual problem was elsewhere. When the problem was eventually located no-one jumped up and said;
"Hey! I made a mistake! The grounding was good all the time!!"
After a long career in construction I have a few anecdotes to support these assertions but no studies.
A good study or two may have saved us a lot of wasted time and money on isolated grounds.
I will be the first to admit that there are a very few exceptions where isolated grounds are a good thing, such as in operating theaters.
There have been studies years ago when explosive anesthetics were used when the combination of an explosive gas and a high frequency scalpel caused a patients lungs to explode on the operating table with concurrent injury and sometimes death of the operating team.
This was surely a case for small transformers with grounded inter-winding shields and isolated grounds.
For most data systems a well installed code ground has been the functional equal to an isolated ground for many years. (like 20 or 30 years).
Bill
--------------------
"Why not the best?"
Jimmy Carter
