Right. I said that many / most of their 240 inverters can do corner grounded. I thus assumed that might be the case for their 480 inverters, hence why I asked - but that document shows ungrounded Delta is the only configuration.
Corectamundo, but whether or not 480V corner grounded would be supported in this case is a very minor issue; all the client needs to do is add ground detection to the transformer secondary for the ungrounded delta to be compliant. Most of the problems are on the primary side of the transformer.
Sure. I mean ungrounded is pretty foreign to most people, so down the road someone might be scratching their heads a bit, but the chances of any problems are very unlikely. Hell, the inverters will almost be your better fault detector as presumably they won't like it if a phase grounds out and will fault out.
It's the first time I have dealt with it, that's for sure, but 250.21(B) is very clear; the system must have ground detection.
Yes. What are you going to employ for that? Just the light bulb trick?
Light bulbs are probably the most commonly employed detection scheme at 480V.
Probably any sort of indicator light would get ignored and/or no one would know what it means. Probably some sort of audible alarm would be best.
It cannot be grounded because there isn't one. A 208D to 480/277Y xfmr was specified but a 208/120Y to 480D xfmr was bought and installed. Oops. If you mean the inverter neutral, then in post #29 you are correct.
Right. I mean there is a chance it might work. I am not sure of the internal topology of these inverters, but if they have wye connected power electronics inside, it might sort a ground the system and the inverter might be okay with it. Keyword is "might", I wouldn't try it.
The internal topology is wye, I'm certain. So the grid selection programming is probably just telling it whether to trip with an 'AC voltage out of range' error on the L-N voltages, or to ignore them and only trip based on L-L voltages. So I think grounding the inverter neutral terminal without setting it to delta is just inviting such errors for no benefit whatsoever.
Yes. Standard transformers can only transmit changes in voltage, rather than absolute voltage itself. An ideal transformer takes the derivative of the primary voltage, scales it by the turns ratio, and then integrates it to produce the secondary voltage. As with integration in general, there's an offset constant, which could be anything if not otherwise constrained. This isolates the two sides from one another, requiring a grounding of the secondary system if you want its voltage to ground to be predictable.
The dreaded "+C"!
It might be related to clearances and stress. Having it ungrounded is basically half the voltage stress between internal components and PCB traces. Like everything else, it just money.
I don't know, I think if it's ungrounded you have to assume 480 volts phase to ground in your design. The voltage might float nicely in the middle but you can't count on it.
Maybe the reasoning is really the opposite though?
If the manufacturer's instructions say not to connect a particular inverter to a corner grounded system, I won't do it. It really makes no difference whether it would run or not.
