Relation Between Stray Capacitance and Ungrounded System

If one of the phases faults to earth/ground, then, self-evidently, the voltage of the other two phases will be 380V to earth/ground now, just as they are 380V to the other phase before the fault occurred. A single phase-to-earth/ground fault does not necessarily operate simple circuit protection devices such as fuses in IT systems, which is why something a little more involved must be used for the best safety.

http://en.wikipedia.org/wiki/Earthing_system - see "IT" systems.

It has very little to do with the stray capacitance, it has everything to do with how ungrounded systems react to the first fault on any leg. The stray capacitance merely provides an artificial floating reference point, and all three legs only read the same voltage when the system is unloaded or perfectly balanced. Under any other conditions the neutral "floats" giving the condition known as "neutral shift". High availability systems take advantage of this phenomena in their protection schemes by monitoring each phase to safety ground voltage. When a fault occurs the voltage to ground on the other phases rise to 1.73 x normal and triggers an alarm and the system continues running normally, but that fault need to be cleared ASAP or the second fault will cause real problems and the protective devices will operate.

First the definition of a capacitor, Two conductors separated by an insulator. In this case the "capacitor" is/can be the current carrying conductor and the conduit, what ever. If the conductors are of sufficient length, you can light a 100 watt bulb, by this, or even electrocuted.

Ships use this system extensively to minimize corrosion, and maintain system operation in the event of a short to ground. As your meter has relatively high input impedance, (this includes the old Simpson 260) you can measure this voltage.

The Simpson 260 has an AC input sensitivity of only 1,000Ω per Volt, as opposed to a modern Fluke Digital Multimeter which has an input impedance of 10MΩ, the same as the old VTVM, which results in far less loading on the AC circuit.

The Simpson will very quickly discharge that loosely coupled capacitor, while the Fluke will read its voltage. That's why Fluke (and others) now have models with a Dual Input Impedance switch to replicate the input impedance of the analog meters, allowing its use on both power and electronic circuits.

Also, that would be one really huge capacitor if it could pass enough current to light a 100 watt bulb! Unlikely to occur with 2 wires running parallel to each other; what does occur is inductive coupling, and that can bite you.

Several thousand feet of cable stuffed in metal conduit routed along steel bulkheads does just that. With these values, even the Simpson 260 would read it.

No doubt at that length. Had a strange case of a 120 Volt relay in a power plant that would not release when the AC was cut, traced it through 1.6 miles of control wiring, did the calculations, and sure enough there was enough capacitance to keep it "energized" despite having no control power feeding it.

That is:

• useful to know and
• a good justification for going DC instead with suppression across the coil.

I have almost run into this in a current project, we need to control a unit 2500 feet away: my tech lead pointed this potential problem out and had a bulletin from Square D about this "long distance signalling" problem. I went to the Schneider Electric website (SE now owns Square D) and got the latest. There are two bulletins: Bulletin No. 8502DB0001 and Bulletin No. M-379F. Good reading, and gives alternative methods of control too. The text in both is pretty much the same, but there are many more tables in the 379F. Do note the tables are pretty specific to Square D, as they are set up for the Square D starter number, size of wire, supply voltage to give a final number of maximum feet distant. In some cases surprisingly short, but the numbers have been worse-cased.

thanks all..
but still no one answer my point,
i want to understand this quotation "When a fault occurs the voltage to ground on the other phases rise to 1.73 x normal"
What is the reason for this rise!!!

Reread post #2, normally phase to neutral voltage is ph-ph V/√3, when you put a phase conductor to ground you have "shifted" the ground and made all ph-ph and ph-n voltages read the same. For more info draw a before and after diagram, then Google on "neutral shift".