Arc fault Circuit Interupter Breakers

Just go to the site of the inventor:

will find just about everything you want to know. It took the inventor a long time to program a chip to know the difference between a good arc like in a light switch from a bad arc like a hot to neutral arc in the wall.

AFCIs include a 30 ma GFCI - I believe the idea is that if there is a ground wire, an H-N arc will rapidly also become a H-G arc.

You are right that there are a lot of arcs an AFCI should'nt detect. They include operating any switch, which commonly bounce on make producing arcs, opening a switch, light bulb burning out which commonly includes the broken filament remaking at shortened length, motor start, light bulb start. I suspect you might not want an series AFCI for an arc welder.

Arcs can be parallel (H-N, H-G), or series (loose connection). AFCIs only detect parallel (surprised me but note the "fault" in AFCI). Those produced (sold?, installed?) after 1-1-2008 will have to detect both series and parallel. Series arcs are limited by the load (the current will not excede the load current). Parallel arcs are limited by the available fault current at the arc - obviously much higher.

Parallel arcs, because they are not a clean fault, have currents can be somewhat intermittent and are lower than for a fault. The result is that the current is low enough that a breaker may take significant time to trip or not trip at all. 30 seconds may be a very long trip time compared to the time it takes to ignite a fire. IIRC the arc "signature" that AFCIs look for is large current pulses, around 80 amps. The detection is sophisticated to separate from normal arcs. I presume a series arc is much harder to detect and separate from a normal arc.

A good paper from the Consumer Product Safety Commission on AFCIs is at

includes the rationalle for using AFCIs and information on how they work.

bud--

Not every breaker in the panel - just for the bedrooms.

I've never used one

No. The GFCI trips when the current on the white wire and the current on the black wire differ by ~ 5 mA or more. You do not need a ground wire at all in order for a GFCI to trip, and the fault current which causes the GFCI to trip does not have to flow on the ground at all.

Ed

Since a switching power supply or an arc welder

between the white and black it can only be because 5ma is finding some path to ground. I suppose I did not make myself clear. But thanks for clearing that up.

"Bud> A good paper from the Consumer Product Safety Commission on AFCIs is at

Great Paper, lots of good info, thanks -n

The way you originally wrote it implies there is something measuring the ground current. There isn't. Anyway, no ground is necessary to trip a GFCI. A high resistance from the hot (of another circuit or upstream of the GFCI) to either the black or white downstream of a GFCI will trip it. The test button in a GFCI outlet just connects a resistor, and not to ground, it just causes a >5mA difference. Also, I once miswired two adjacent GFCI breakers in a panel by switching the white wires. When I applied a load to one circuit, BOTH breakers tripped. One saw a current flow on the hot with none on the neutral, the other saw current on the neutral and none on the hot. No ground involved.

A shared neutral with only one black going through the GFCI will also trip it, as will any other situation where some current finds a path around the GFCI. True, most of the time it will be a path to ground, but it doesn't need to be. It is better to think in terms of a difference in current, and leave the ground out of it.

Ben Miller

| Arcs can be parallel (H-N, H-G), or series (loose connection). AFCIs | only detect parallel (surprised me but note the "fault" in AFCI). Those | produced (sold?, installed?) after 1-1-2008 will have to detect both | series and parallel. Series arcs are limited by the load (the current | will not excede the load current). Parallel arcs are limited by the | available fault current at the arc - obviously much higher. | | Parallel arcs, because they are not a clean fault, have currents can be | somewhat intermittent and are lower than for a fault. The result is that | the current is low enough that a breaker may take significant time to | trip or not trip at all. 30 seconds may be a very long trip time | compared to the time it takes to ignite a fire. IIRC the arc "signature" | that AFCIs look for is large current pulses, around 80 amps. The | detection is sophisticated to separate from normal arcs. I presume a | series arc is much harder to detect and separate from a normal arc.

I read the target was 70 amps. Minor difference.

But it would need to be amps with some kind of intermittent waveform, else some motors could cause problems, such as air conditioner start.

I notice very little dealing with potentials of radio transmission interference caused to the AFCI device. I know uch an issue exists with GFCI receptacles.