There's a whole working group on this issue you can join.
That's called "breaker masking" in the world of plug in solar.
GFCI receptacles and likey GFCI circuit breakers in general all have just one CT with all the "protected conductors" passing through it.There's a whole working group on this issue you can join.
Send me a PM if you're interested in checking that out.
---
All these topics are active discussion, and there is a lot of regulatory and advocacy work on this topic now. Recent press:
Forget Rooftops — Bay Area Residents Are Plugging Solar Into the Wall | KQED
A movement is growing to bring small, portable, affordable solar to a balcony or backyard near you. But before you see them everywhere, advocates must break through significant barriers.www.kqed.org
On the GFCI the issue is that some legacy GFCI products measure current with current transformers, and the designers never considered that power might flow backwards through the wires. CT's are directional, and this means negative current perhaps at a 50:1 or whatever ratio flows into the measuring circuit. The claim is that this negative current can fry the GFCI and/or render the
Breaker masking is a greater theoretical concern, which is more tangled.
----
Millions of these units are in use in Europe where it's been accepted at all levels from the power companies, to regulators to the checkout gal at Ikea.
That's called "breaker masking" in the world of plug in solar.
There are hundreds of GFCI designs over the decades. Cataloging which are safe for bidirectional flow, and which are not, is part of the quandry.
Is any of the traditional outlets (GFCI or not) tested/listed for back feed (i.e. inlet)? It would seem odd if they were as that's contrary to their original intended use case.
The electrons can't tell which direction the receptacle outlet is oriented. Current is current.
The older style receptacles could feed loads side power to their 'slots' even when the line side power had been opened by the GF action or by loss of the .
The inverter would be grid-tied, and the only N-G bond would be at the service. So if you have a ground fault, current will be flowing through that N-G bond, and that current will show up as a net current through the GFCI CT, regardless of whether that current originated from the utility ungrounded conductor or the inverter ungrounded conductor.
I think tortuga chose his resistor example so that the L-N imbalance at the GFCI would be, in his calculation, 4ma on a 8ma fault, and therefore not trip the GFCI on a fault that's supposed to trip it. But he assumes the fault current is divided among the sources, and I'm not convinced this is correct. I think that (almost) the entire fault current would be supplied by the grid tied inverter, if it's capable of backfeeding the receptacle. I believe in that case the fault current travels back to the inverter on the neutral, and trips the GFCI that way. That is, unless there is other load on the circuit that somehow results in some very unlikely equilibrium between the sources at the fault. Which, worse case, doubles the minimum trip current of the GFCI. Unless there's a scenario I've missed.
Maybe I'm missing something, but as I understand it: the service has an N-G bond. The grid tied inverter doesn't. The grid tied inverter is on the load side of a GFCI. We have a high impedance ground fault L-G on the load side of the GFCI.
An important point that is worth stating is that these two different sources of imbalance current at the GFCI are additive with each other, rather than potentially cancelling. With the convention that current flows from L to N, the usual balanced case is that the same current travels on L from the line side of the GFCI to the load side, and then travels back on N from the load side to the line side.
