GFCI Protect a Whole Sub-Panel from the Main Panel

[WattsaVA]

I have a detached garage that I am rewiring for a customer. Putting in a sub-panel for three circuits. I need GFCI protection for at least one circuit. I could GFCI protect all circuits. My question is if I use a MLO panel in the garage and do not have a main breaker can I install a GFCI breaker in the main panel in the house to feed the entire sub-panel. I need to replace the breaker in the house that is feeding the sub-panel and it would save money to not have a main in the sub-panel since one is not required for only three circuits. But, since I need GFCI protection for one circuit it makes more sense to protect the whole sub-panel from the breaker in the house. Will a GFCI breaker in the house detect a ground-fault anywhere down stream on multiple circuits? The way I understand, it seems like it should. Any thoughts?

 

[david]

For a 120/240 system a multi wire circuit would not be inclusive of three branch circuits

 

[augie47]

Anytime you provide GFCI protection on a feeder other than individual branch circuits you increase the chances of nuisance tripping as it will be looking at ground leakage on all the circuits supplied by the feeder

 

[ramsy]

can I install a GFCI breaker in the main panel in the house to feed the entire sub-panel

As it should be, that ensures qualified persons are required to fix trip issues with inverter appliance, LED lights, and equipment from China, not UL or FCC tested for radio interference.

 

[david]

I have a detached garage that I am rewiring for a customer. Putting in a sub-panel for three circuits. I need GFCI protection for at least one circuit. I could GFCI protect all circuits. My question is if I use a MLO panel in the garage and do not have a main breaker can I install a GFCI breaker in the main panel in the house to feed the entire sub-panel. I need to replace the breaker in the house that is feeding the sub-panel and it would save money to not have a main in the sub-panel since one is not required for only three circuits. But, since I need GFCI protection for one circuit it makes more sense to protect the whole sub-panel from the breaker in the house. Will a GFCI breaker in the house detect a ground-fault anywhere down stream on multiple circuits? The way I understand, it seems like it should. Any thoughts?

Yes a gfci in the service panel in the house will protect the feeder and all he outlets in the garage.

 

[ramsy]

Anytime you provide GFCI protection on a feeder other than individual branch circuits you increase the chances of nuisance tripping

The EU does this with RCD's built for harmonic's, which are better shielded than GFCI's.

The problem with shielded RCD's, is they work with more unlisted junk from Amazon, and tolerate more unqualified persons monkeying with the electrical.

 

[Seven-Delta-FortyOne]

I don’t know what code cycle you’re on, but detached structures are required to have a main disconnect at the structure. The only exception is for a single circuit (MWBC included).

The GFCI protection at the main panel will work, but could be an annoying setup for the owners.

 

[WattsaVA]

You don’t have to have a main unless more than six circuits in current code.

 

[Seven-Delta-FortyOne]

What code are you on?

 

[WattsaVA]

On 2017 but that is still current (2023). Good point about it being an annoyance though. I didn’t think about that. Probably should just run it by customer.

 

[wwhitney]

Square D, for one, says the maximum length of circuit they recommend for a QO-GFI breaker is 250'. https://www.se.com/us/en/faqs/FA117270/

So if the sum of the feeder length to the garage, plus all the cables within the garage, exceeds 250', using a GFI at the supply end of the feeder would be contrary to Square D's recommendation. Now that doesn't necessarily apply to all brands, but the physics is the same (self-capacitance of the cables), so I would expect the limit would be comparable across brands.

Also, if you do have an event in the garage that would trip the GFCI, are you OK with (a) losing power to everything and (b) having to go back to the house to reset the GFCI?

Cheers, Wayne

 

[WattsaVA]

Dang. I didn’t think that someone would break out some physics on me. Really, that answers it. That makes perfect sense. Thank you.

 

[ramsy]

So if the sum of the feeder length to the garage, plus all the cables within the garage, exceeds 250', using a GFI at the supply end of the feeder would be contrary to Square D's recommendation

Not the branch circuits, calculated in parallel resistance, since the series circuit end at the sub-panel.

event in the garage that would trip the GFCI, are you OK with (a) losing power to everything and (b) having to go back to the house to reset the GFCI?

There's nothing wrong with 2-Pole breakers that trip up unqualified persons that tamper with electrical.

 

[wwhitney]

Not the branch circuits, calculated in parallel resistance, since the series circuit end at the sub-panel.

Resistance is immaterial, as it is the capacitance of the wiring that allow for leakage current that can trip the GFCI. And the capacitance stems from the length of ungrounded wire in proximity to the length of the EGC wire. So certainly it will depend on the length of the feeder plus the longest branch circuit. I believe, but am not quite as sure, that any other branch circuits will just add to the capacitance, hence my statement that we just add up the length of all the cables in the garage.

Cheers, Wayne

 

[ramsy]

capacitance stems from the length of ungrounded wire in proximity to the length of the EGC wire.

No reason to push it near 250ft limit.

GFCI outlets at first opening for garage receptacles, accomplish the same thing without risking capacitance from long feeders.

 

[ramsy]

Square D, for one, says the maximum length of circuit they recommend for a QO-GFI breaker is 250'. https://www.se.com/us/en/faqs/FA117270/

Identifying critical factors that break designs can be a thankless effort, if bursting peoples paradigm, or bubble, is not accepted with thanks.

Where 250ft of wire capacitance trips class A GFCI's, and ANSI 101 allows appliance heating elements 3mA of leakage, much shorter raceways would trip that GFCI.

 

[mikeames]

You don’t have to have a main unless more than six circuits in current code.

What reference supports this?


2023 below.

 

[wwhitney]

What reference supports this?

225.33(A), which starts off "The disconnecting means for each supply permitted by 225.30 shall consist of not more than six switches or six circuit breakers . . ."

Cheers, Wayne

 

[mikeames]

225.33(A), which starts off "The disconnecting means for each supply permitted by 225.30 shall consist of not more than six switches or six circuit breakers . . ."

Cheers, Wayne

So where does that say you don't need a disconnect? That says it shouldn't be more than six, it doesn't say you don't need 1. If you have less than six switches in the structure you still are not disconnecting all ungrounded conductors supplying those switches?

 

[wwhitney]

So where does that say you don't need a disconnect?

It doesn't, it says that the disconnect may consists of up to 6 circuit breakers. So if your supply lands in an MLO panelboard (outside or inside nearest the point of entry) with 6 or fewer circuit breakers, that panelboard doesn't need a main. The up to 6 circuit breakers collectively count as the disconnect.

Cheers, Wayne