when did the nec begin to require grounding conductors to be installed with service feeders?

Per the "History and Mystery of Grounding" in the 8th Edition of 'SOARES Book on Grounding'
the 1913 code required mandatory grounding of certain alternating-current secondary systems.

Some of the wording closely matches todays' (2002) NEC 250.20(B)(1).

I'm interested mainly in the last 20 years or so. My wife's office building is this age and has no gec in any of the 3 panels serving her tenent space. the tenent is a computer server company and I'm concerned for their computers in case of a gf. I'm wondering if i should recommend retrofitting a gec to their panels.

Brian,
There shouldn't be a Grounding Electrode Conductor (GEC) connected at these panels, if they are sub-panels. The GEC connection to the grounded (neutral) conductor would be made at the main service equipment.

Maybe you meant Equipment Grounding Conductor (EGC). If the panels are supplied with metal conduit, you would not neccessarily see an additional conductor for equipment grounding. The metal conduit is often permitted to perform that function.

Ed

there are 3 separate panels fed from 3 separate meters, each with its own disco. should there not be a ground wire from the disco to these panels? Although the original conduit runs were emt there have been additions in the form of mc and flex run into these panels. the egc from these new circuits are bonded to the can with lugs. I might also tell you that the panels in question are residential panels and not commercial panels. Don't know if this was standard when the building was built or not.
I'm in commercial elec. construction and we always (!!!!) pull in a ground. It seems to me that the mc and flex is not rated for gf protection and therefore circuits are not grounded.

The panels are grounded by the EMT feeder and the loads served by the MC are grounded as you said by lugging the grounding conductor to the panel.

Code compliant in my area.

If its in and working I would not change it, if I was to do it new I would pull a ground.

Fairly sure IAEI used to sell a brown-and-white booklet on history of the NEC, with some very controversial early times with respect to system and equipment grounding. In bryan’s initial post—if the service is ungrounded 480V [or possibly even 240V] 3ø 3-wire, there used to be some “liberties” permitted with these circuits, before the nowadays well-understood need for limiting potential difference between enclosures and raceways [where first-fault overcurrent- device operation is not an issue.] I may be nuts talking about it, but that’s nothing new—the nuts part anyways.