450.5 Zig-zag on load side of system grounding connection

Status
Not open for further replies.

Lars Fetzek

New member
Article 450.5 of the 2008 NEC states, in part, "Zig-zag connected transformers shall not be installed on the load side of any system grounding connection..."
The Handbook commentary attributes this restriction to the possibility of ground-fault current of a magnitude damaging to the zig-zag transformer occurring due to ground faults in the larger system located elsewhere than in the circuit in which the zig-zag transformer is connected. I submit, furthermore, that such current over-loading of the zig-zag transformer may even occur in a steady-state unfaulted condition in which large neutral currents manifest elsewhere in the larger system and are inadvertently sourced/sunk in part by the zig-zag transformer. For at least these reasons, the code restriction seems well-founded.
However, the code does not appear to offer an exception which I suspect might be equally reasonable: Suppose that the zig-zag transformer is used to establish a neutral from a 3-wire feeder fed from a grounded system. Suppose that the zig-zag transformer is high-impedance grounded. Would not a properly calculated/specified grounding impedance eliminate such concerns for current magnitudes damaging to the zig-zag transformer originating from faults and/or loads elsewhere in the system while still providing a well-functioning neutral connection from the transformer end of the grounding impedance?
Obviously, this solution would allow for a voltage differential between neutral and ground in the circuit in which the zig-zag transformer is connected. However, the shock/electrocution hazard could be addressed by proper GFI detection to trip that circuit's breaker. Is, then, the only real problem the possibility of exceeding equipment ratings for neutral-to-ground insulation? If so, then can't such a concern be considered then dismissed in cases where the loads are simply lights, etc., which should not be sensitive to neutral voltage rise and in which the conductors are rated for 600 volts, which is well above the theoretical limit of neutral voltage rise in a 480/277 volt high-impedance grounded system that does not involve capacitors? Why not simply consider the so-called "neutral" to be a "floating" fourth hot leg of the circuit and let it be? After all, there is no prohibition of the connection of single phase loads between various pairs of the three hot legs of a three-phase circuit that does not include a neutral at all - let alone at or nearly at ground voltage.
Am I overlooking something - either in my theoretical justification or in my perhaps (hopefully) over-conservative interpretation of the code?
Thank you for any advice.
 
I believe that if you use a zig-zag transformer to derive a neutral, and you _bond_ that derived neutral on an elsewhere grounded system, then you will have the same problems of 'parallel paths' that you get if you bond the neutral at multiple locations in a system. Current which _should_ be flowing on the neutral system will end up flowing on the grounded/bonded metal system.

If, on the other hand, you use a zig-zag transformer to derive a neutral, and then you leave that neutral unbonded as a 4th 'hot' conductor which just happens to have a voltage near zero, then you have a system which in theory could work. However almost all North American equipment that is designed to use line-neutral voltages is designed with the assumption that the neutral is actually grounded. I don't think that you would find any equipment listed for such an application. So you could derive this neutral, not bond it, and have a safe system that nothing is designed to be connected to.

Remember that you would need to switch all 'hot' conductors, including the derived neutral, so you would have some rather non-standard panelboards, double pole switches for everything, etc.

Note that where the NEC permits high impedance grounded wye systems, it doesn't permit line-neutral loads to be served.

I have often thought that high impedance grounded systems, when used with ground fault protection that actually opens circuits in the event of a fault (as opposed to systems designed for continuity of service as currently permitted) should be allowed to serve line-neutral loads; similarly I think that I would be comfortable with a derived neutral being used to serve line-neutral loads if protected by sensitive ground fault protection. Current code does not permit this, however.

-Jon
 
Last edited:
Status
Not open for further replies.
Top