Solar System G-N bond

[jaggedben]

My understanding is if there is a line to ground fault in the AC output circuit of a grid tied inverter, say at the inverter housing, any current output by the inverter will flow back to inverter the shortest path it can take. So if the inverter can output a max of 56 amps and the fault happens in full sun when its putting that out, 56 amps will flow back to the inverter the shortest path it can take.

I described that path above. These are 240V inverters (as most are). So say L1 faults, it's: L1 > EMT/EGC > MBJ > neutral point of utility transformer > transformer secondary winding > L2. Also MBJ > neutral > operating loads on L2 > L2

I suspect modern inverters will detect this condition rather quickly and disconnect themselves.

Yes, these mirco-inverters max fault current would likely be limited to only about 3 cycles. A fault event would almost certainly trigger a 'voltage out of range' error which would keep them offline for 5 mins.

....

As for the DC side if the inverter is putting out 56 amps @ 240VAC then you could have a DC side of say 20Amps @ 700 VDC or some similar high voltage/low current depending how the solar panels are stringed together, and I imagine the wire sizes on the DC side can carry a bolted DC fault continuously.
During a catastrophic bolted inverter AC side fault (say internal to the inverter) I don't think the DC side would see anything other than a large load.
I still think of the DC side of an inverter as a 'one way gate valve' isolated from the AC side, (there is no isolation transformer in a grid tied inverter) but I am no expert on inverter design. I imagine the isolation provided by the electronics of an inverter is equivalent to that of a transformer and what we think of as a 'Separately Derived System' though they are not considered that code wise.
I have never herd of a grid tied inverter failure mode where say fault current from the grid flows thru the inverter or high voltage DC flows out on the AC side, but obviously faults happen during accidents, storms and unforeseen events so anything is possible.

These are microinverters so the DC is something like 35V, max 10A and there are essentially 40 something separate DC systems. The DC side is non-isolated so in theory a DC ground fault could carry AC current however for Enphase I believe the power electronics will effectively prevent that like you suggest. That said, for some string or optimizer inverters that operate at a high DC voltage like you mention I'm not quite so sure. Some (mostly older) inverters do have transformers so they are effectively like an SDS. But nowadays almost all inverter power electronics connect the DC directly to AC as they switch, so non-isolated. My understanding is that only the robustness of the design would prevent utility current from feeding a DC side fault. I saw one or two cases a few years ago where I suspected it happened.

 

[tortuga]

These are microinverters so the DC is something like 35V, max 10A and there are essentially 40 something separate DC systems.

ahhh right

My understanding is that only the robustness of the design would prevent utility current from feeding a DC side fault.

Yeah I'd be curious to learn more about grid tied inverter 'isolation', especially isolating the DC.
You would not want 700V DC leaking thru the system, potentially drawing arcs across 240VAC residential breakers, even at a few amps I imagine it could weld the AC breakers on.