Ground Fault Breaker - advice needed...

| I am to build an electrical panel for a mobile unit which is to be powered | by 110VAC / 60 Hz (1 phase, 1 neutral and ground). | The unit is equipped with medium heating elements and have a max. current of | 20A which is connected through a Hubble connector. | | To protect from ground fault I wish to build the panel with a ground fault | breaker included. | | According to European rules both the phase and the neutral have to be | disconnected in case of a ground fault. | Also the trig limit for protecting personell is 30mA current. | | - What Ground Fault Breaker should I use. I would prefer one from Allen | Bradley but can't seem to find one that | disconnects the neutral line also. | - Is it mandatory that the neutral have to be diconnected also according to | US rules ?
US rules do not require disconnection of the neutral. IMHO, they should, but in actual fact they don't. Typical GFCI breakers do not have this capability, but most GFCI receptacles these days do.
The cheapest solution would be to use a 20-amp rated GFCI receptacle device that does have neutral interruption, and use that in a panel. You can wire your load to the "load" terminals. There are such devices made with no outlets, called "dead front".
Since your case involves only one phase wire, you could do this with a 2-pole GFCI breaker. One of the phase inputs will power the circuitry inside, and you'll have to figure out which one. Then wire the other to neutral. Load will be connected to the breaker terminals as usual, and the white pigtail wire will be connected to the neutral bus (not to the neutral terminal on the breaker). You would not want to do this in a distribution panel, just in a single circuit panel.
Another more expensive option, which I am looking at to deal with 2-pole circuits where I want the neutral interrupted, is to follow a GFCI breaker (that doesn't interrupt neutral) with a contactor that interrupts all conductors together.
| - Also as far as I know US rules say that the trig limit for protecting | personell have to be 5mA - is this true ?
Somewhere between 2mA and 6mA, with 5mA being the preferred level.

I am not sure about the "rules" but I know for a fact that the GFCI breakers for residential DO open the neutral as well as the hot.
I found this page.
I was aware of AB making medium voltage stuff, never molded case circuit breakers... Learned something new today...

The Great Outdorrs. Sound like a nice job to do. Build Up the panel with the GFI Breaker of your choice, if memory serves me right the neutral and hot are wired into the Breaker you have nothing to worry about but to terminate it with nice workmanship using every terminating item & the proper tools for a job well done, with no scratches or marrs on the plastics & metals and Raintight Enclosures, Fittings & Covers. spend,spend,spend };-) I do when I have to. =AEoy

The 120 volt units that stand alone (with or without duplex outlet included in the unit) definitely do disconnect the neutral (likely because they want to "protect" against wiring errors whereby neutral and hot are reversed. If you don't believe me, you can take one apart and see for yourself. If you have one in service, you can "test" trip it and see if there is continuity between neutral and ground.
There is nothing to stop you from separating the overload protection from the ground fault protection. The total cost may be less and since GFCI units sometimes fail, it will be cheaper in the long run to separate the functions.

| |> |> There's no textual indication that these breakers interrupt the neutral. | | The 120 volt units that stand alone (with or without duplex outlet included | in the unit) definitely do disconnect the neutral (likely because they want | to "protect" against wiring errors whereby neutral and hot are reversed. | If you don't believe me, you can take one apart and see for yourself. If | you have one in service, you can "test" trip it and see if there is | continuity between neutral and ground.
Are you referring to the dead front GFCI receptacles? Yes, I do know those will disconnect the neutral. But these do not also do overcurrent protection. I do not refer to these devices as "breaker".
When someone says "breaker", I think of typically a molded case circuit breaker ranging from small 3/4 inch distribution panel breakers to big 15 inch industrial main breakers. These breakers all have separate "contact chamber" for each pole in which there is a mechanism intended to be capable of interrupting high fault currents. In order to be able to nterrupt the neutral, and additional such "contact chamber" is needed. The illustrated breaker was one with 2 such sections. In theory that could be used to interrupt the neutral with hot 1 pole. But there may be issues in the design that would preclude this unintended usage.
| There is nothing to stop you from separating the overload protection from | the ground fault protection. The total cost may be less and since GFCI | units sometimes fail, it will be cheaper in the long run to separate the | functions.
So how would you do this for a 3-wire shared-neutral circuit with the requirement that the entire circuit be interrupted on any cause that would interrupt any part of it? A 2-pole panel breaker won't do it. A dead-front receptacle-style GFCI won't do it. The "hack" I came up with is to use a 2-pole GFCI breaker followed by a 3-pole N/O contactor with a 240 volt coil powered from the 2 hots.

Nor do I. But combined with a "real" beaker they will offer protection. If the guy wants to make his design truly idiot proof and relatively cheap, he can combine a cheap two pole breaker and a cheap two pole GFCI. A high current ground fault would likely destroy the GFCI but maybe the breaker would clear it in time. Well, using a "real" two pole breaker (including the type that has screw terminals for both line and load and intended for stand alone use) and a GFCI would provide near idiot proof safety.
That wasn't the problem the OP states. He said 120 volts. If you "hack" you have to take apart the GFCI and pass the neutral conductor through the sensing/injection toroids.
> > -- > --------------------------------------------------------------------------

|> Are you referring to the dead front GFCI receptacles? Yes, I do know | those |> will disconnect the neutral. But these do not also do overcurrent | protection. |> I do not refer to these devices as "breaker". | | Nor do I. But combined with a "real" beaker they will offer protection. | If the guy wants to make his design truly idiot proof and relatively cheap, | he can combine a cheap two pole breaker and a cheap two pole GFCI. A high | current ground fault would likely destroy the GFCI but maybe the breaker | would clear it in time.
I've never seen a two pole (when not counting the neutral as a pole) GFCI device, other than the ones that are also breakers and don't disconnect the neutral.
|> When someone says "breaker", I think of typically a molded case circuit |> breaker ranging from small 3/4 inch distribution panel breakers to big |> 15 inch industrial main breakers. These breakers all have separate |> "contact chamber" for each pole in which there is a mechanism intended |> to be capable of interrupting high fault currents. In order to be able |> to nterrupt the neutral, and additional such "contact chamber" is needed. |> The illustrated breaker was one with 2 such sections. In theory that |> could be used to interrupt the neutral with hot 1 pole. But there may |> be issues in the design that would preclude this unintended usage. | | Well, using a "real" two pole breaker (including the type that has screw | terminals for both line and load and intended for stand alone use) and a | GFCI would provide near idiot proof safety.
Near. But the neutral isn't disconnected. While it is not as dangerous as the phase lines, it can cause problems to have it not disconnected. For one thing, a continuing ground fault current over the neutral could cause the GFCI tripping mechanism to continously operate. If the neutral is sufficiently distant from the ground bond, and has a sufficient voltage due to imbalance, and has a fault to ground, enough current could flow to trigger a GFCI. It is probably not rated for continuous operation of the internal solenoid tripping mechanism. I have reproduced this kind of failure using radio instead of neutral to ground fault and have found that GFCI devices do indeed try to continue to trip even after they have tripped, making a continuous buzzing noise. I have not (yet) attempted to see if I can burn one up and show it to cause a fire by doing this for a long period of time. I think that exercise should be left to a safety laboratory.
| That wasn't the problem the OP states. He said 120 volts. If you "hack" | you have to take apart the GFCI and pass the neutral conductor through the | sensing/injection toroids.
Or put the neutral on the distribution panel bus. A single phase panel then becomes a single-ended 120-volt-only panel. A three phase panel can then handle 120/240 single phase. Significant risk and likely non-compliance would be issues with doing this.