Help with neutral wiring w/critical loads and bypass

[bitbanger]

I'm hoping someone can help save me the embarrassment of a failed inspection (or worse). I just can't figure out a neutral arrangement that makes sense for this setup. I'd like to use a 100A subpanel with interlock kit to act as an inverter bypass for the critical loads. However this seems to be complicating the neutral wiring.

NEC 300.3(B) from the top of my head, requires ungrounded conductors (neutrals) to occupy the same conduit or raceway as the ungrounded conductors. This seems to conflict with how I see critical load panels sometimes wired - just bringing over the hots from the main panel.

Additionally there's an issue with parallel neutral paths in numerous places.

Lastly if the SolArk 12k does an automatic N-G bond under grid-down, that means a double bond with the main panel via the critical load bypass feeder, resulting in objectionable currents.

Is there any way to make the below NEC compliant and functionally correct? I was hoping to avoid the cost of a 3pole 50-60A transfer switch (in place of the 100A panel interlock kit), but I think that would solve some problems.

This is a limited-to-home setup; no intention for supplier agreement/sell-back. Not shown below are CTs and conduit ground bonding.

Thanks for any thoughts folks.

 

[Hedges]

I don't think I'd care about a loop in neutral and ground, so long as each path was high enough ampacity to work if other went open-circuit.

The two (?) black hots and one red hot cross over without neutral or ground. I'd want to route the through same conduit as where neutral and ground tie between those panels.

 

[bitbanger]

I don't think I'd care about a loop in neutral and ground, so long as each path was high enough ampacity to work if other went open-circuit.

The two (?) black hots and one red hot cross over without neutral or ground. I'd want to route the through same conduit as where neutral and ground tie between those panels.

I suppose the issue with neutral loops is there's no guaranteed return path. So if the inverter was for example backfeeding via GRID, and the critical loads also wanted to return via that neutral connection (due to lower impedance), then it would potentially exceed the sizing of the GRID 50A feeder neutral.

I'd have to check sizing but I'm not sure if it's feasible to get the inverter bypass and all branch circuits in one conduit. I guess if it's kept short as a nipple then no derating needed.

Preciate the thoughts.

 

[Hedges]

Have to figure out whether those add or subtract. I'm thinking subtract but I'm not 100% clear where loads and backfeed are. Also consider 120% rule for busbar, and where connections are made.

Yeah, no derating under 2'. Doesn't say anything about a length of wire outside the nipple as heatsink, but my thinking is there should be.

Easier to overstuff if conduit is a straight shot. Curves cramp my style.

I'm planning to look for 2.5" EMT for one elbow, and will have to holesaw the opening. Just a 45 degree into a pull box.

 

[bitbanger]

Thanks - yep if I put GRID connection at bottom of main bus, and derate to 175A (by swapping out 200A MCB) I'm under bus limits w/9kW max backfeed.