ECN Forum Forums Code Related Discussion NEC & other Code issues Arc Fault Breakers

Paul,
I concur with your thoughts on whole-house AFCI and earth leakage. It would be like trying to find a needle in a stack of needles!

AFCI's retail in the US for about $30, or roughly the cost of a GFCI breaker. The key to getting the costs down is to get the volumes up. Until the contractor community buys in, and starts installing them in installtions that do not require them by code, the cost will never come down. Code right now mandates for bedroom circuits in new construction. There are 1.5 million new homes built a year. With 3 circuits per bedroom, that's only 4.5 million circuits divided among 4 manufacturers. That is not enough volume to drive down the cost of the electronics.

The irony, of course, is that the homes that need it the most (older homes with old wire) will not have to have it by code.

Brenden,
The local hardware store has your AFCIs on the shelf at $79.95 ea.
Don(resqcapt19)

And they may sell one at that price, but I doubt that any electrician will pay that price. Is it a chain, or a mom and pop? I was in the Home Depot here at lunch, and they were $29.99.

The market price will begin to settle down once wholesalers and retailers begin to sell them. How many $90 smoke detectors have you bought lately?

I don't think that the hardware store will sell any at that price. $30 is a much more reasonable price and comparable to GFCIs. I haven't checked any supply houses as I don't do residential work.
Don

Brendan,

Thanks, this is helping.


Can you tell us about the process of the waveform analysis and the nature of the "signature" of the arcing fault?

Al

They may sell one to a Joe Homeowner, but that's about it.

The waveform analysis compares the arc signature (the shape of the waveform anomaly to the hundreds or thousands of arcs that we've mapped in the laboratory. Basically, we analyze the curve and see if it is a non-operational arc...that is, an arc that shouldn't be there. I will paste below a short dissertation from one of our senior engineers on the parallel/series issue. I hope this answers more questions than it generates. :-)

Q. Why was the AFCI designed to sense a short-circuit across two conductors, and not a single open arcing conductor?

A. Arcs between conductors are sporadic (pulsing)in nature. For dwelling unit circuits, the rms available current at an outlet prior to a fault is typically 75A or higher. The arc voltage further reduces the circuit current and, in combination with the intermittent current waveform, the fault current is unlikely to trip the circuit breaker instantaneously (trip level about 200A rms) and possibly not even thermally. The safety enhancement of the AFCI is to recognize the presence of these high energy arcs, and to cause circuit interruption (trip the breaker) far more rapidly than would be the case with a conventional circuit breaker. The UL1699 tests are performed with both NM-B and two wire SPT-2. The focus has been on high current "parallel" arcs rather than low current (load limited) "series" arcs because:
(1) they are higher energy arcing/fire sources
(2) they are far more likely to occur (it is very difficult to maintain a low current series arc between copper wires at 120V)
(3) it is possible to detect parallel arcs without nuisance tripping. Here it is noted that low current safe arcs are an everyday occurrence in dwelling units (switches, thermostat controls, electric motors etc.). Further, many low current electronic-appliances have chopped waveforms (drills, ballasts, power supplies etc.). Thus low current unsafe arcs (at a single open conductor) have to be distinguished from low current safe arcs, and other naturally occurring waveforms, without causing nuisance tripping.

Q. Could the breaker be designed to sense a single arcing conductor? Would it still be called an Arc-Fault Circuit Interrupter?

A. The Branch/Feeder AFCI responds to "series arcs" in NM-B. The test is in UL 1699. In practice, manufacturers utilize the fact that NM-B contains a bare ground conductor. Any break in a single conductor causes current leakage to this ground wire, and the AFCI opens the circuit prior to the arc burning through the NM-B insulation. Since leakage to ground is the response criterion, downstream switching arcs do not cause nuisance tripping. Again it is noted that a series arc in NM-B is highly unlikely.
UL1699 contains an AFCI "Combination" category. This category includes series tests in SPT-2 as well as NM-B. The series test demands that the arc interrupt after 200 to 1000ms (12 to 60 cycles) depending on the current (15A to 5A). In particular this time interval allows discrimination between "operational arcs" (e.g. switching) and "unsafe arcs". There are no commercially available Combination AFCIs.

Q. What was the reason the NEC decided to require AFCI in bedrooms only?

A. NFPA fire statistics show that a high percentage of electrical fires occur in bedrooms. There are all manner of appliance cords in bedrooms (for example, radios, clocks, blankets, air conditioners, heaters, TVs, vacuums) and also lamp cords. All of these cords can be trapped/abused leading to line to neutral arcs. Further, there are long runs of installed wiring (NM-B) between the loadcenter and the bedroom outlets. These runs can abused during installation (e.g. stapling) and after installation (driving nails into the wall etc.).
Summary
The present Branch/Feeder AFCIs provide a significant enhancement in fire safety. They guard against parallel arcing faults in both the installed wiring (NM-B) and the two wire extension wiring (SPT-2). In the unlikely event of a series arc at a break in NM-B, the arc will be interrupted by the AFCI responding to ground current. The overall device is immune to nuisance tripping.

Brenden,
Can you cite examples of how these parallel faults occur in the fixed wiring system? And also why the AFCI manufacturers think that most fires that occur in the fixed wiring system are of the parallel type? This area is the whole basis for my objections to AFCIs. I have been taught and believe that the vast majority of faults that occur in electrical systems are at the points of connection. Faults occurring at the points of connection are of the series type, if they are even arcing faults. Are faulty connections series arcing faults, or just high resistance connections that produce a lot of heat?

By Combo AFCI's is dual AFCI/GFCI function implied?

Does this mean Ul1699 lists AFCI's for NM w/ground only? and older circuits would not be listed for use?

Our wiring methods in England are slightly different of course, but I have to go along with the general consensus that the majority of arcing faults are series, not parallel.

I see evidence of arcing and overheating from wires loose in a terminal, from worn contacts in low-quality switches, etc. -- All series faults.

By contrast, evidence of parallel arcing is quite rare. Most parallel faults I see are where someone drills/nails/screws straight into a cable and the damage is immediately apparent.

The one type of parallel fault that I do see more often is a trapped hot wire behind a fixture, the insulation deformed and cut through to the point that it shorts to the metal box. In most such cases, the direct short just trips a normal breaker or blows the fuse immediately.

I'll concede that in a non-AFCI & non-GFCI circuit such parallel arcing to ground could cause a fire, but it would be interesting to know if any data are available to show whether a sensitive GFCI is any less effective protection in this instance.

Obviously the GFCI wouldn't help against a hot-neutral arc, but again it would be interesting to know if there are any figures to show what proportion of parallel arcs are hot-neutral rather than hot-ground.

A couple of other points strike me from the engineer's report posted by Brendan. It suggests AFCI should be required for bedrooms due to the amount of appliances and cords in use. But with the exception of an electric blanket, surely everything metioned is just as applicable (if not more so) to the rest of the house?

Just one more thing (with apologies to a certain well-known detective!). If we acknowledge that a potential between-outlets fault in NM cable feeding bedroom receptacles needs AFCI protection, then what about other cables in the same wall feeding outlets on the other (non-bedroom) side or just passing through on route to some other part of the house? Shouldn't these be AFCI as well? Just a thought.

I'm all for anything which aids safety, and as someone who has no practical experience of AFCIs I have no axe to grind either way: I'm just trying to look at all the pros and cons with a critical eye.

Paul,
I think that the bedroom was picked as a starting point because more dwelling unit fire fatalities occur in the bedroom then in the other rooms. However, the point of origin of these fatal fires is in fact often in other rooms of the house.
Don(resqcapt19)