ECN Forum Forums General Topics of the Trade General Discussion Area afci parallel series

We were anticipating a less-than-sanguine perspective on the protection provided by circuit breaker AFCIs, supported by a live demonstration. What we got, which kept us talking far longer than many another program, was more than we had expected. The presenter was consultant Bernard Schwartz, P.E., a Fire Protection Engineer formerly of the United States Consumer Product Safety Commission. He was ably assisted, and opposed, by Clive W. Kimblin, Cutler-Hammer's Manager for Applications and Standards. It was not a catfight, but it certainly was lively.

Rather than present all the back-and-forth-ing, I will try to lay out what we learned. There was much more information on which the two experts agreed than on which they disagreed.

Here are the main points that Schwartz raised:


AFCIs interrupt parallel arcing faults, and do so much more quickly than would normal circuit breakers operating in their instantaneous range. Bring a hot and a return conductor together and there's will be a BANG! from the arc as current flowing through the gap trips a standard circuit breaker. Bring the same two wires together and there will be a mere pop from the arc as current flowing through trips an AFCI circuit breaker. Unfortunately, there is no predicting how much life and property protection they can provide, even though they work exactly as they are designed to work. Many people, including, as was demonstrated, many people attending this meeting, presumed that they are designed to interrupt all sorts of arcing and even glowing faults, and do so right quickly. Not so. They are designed to interrupt parallel arcing faults that reach and sustain a flow of 75 amperes. Glowing faults and series arcing faults may last indefinitely without tripping an AFCI, as Schwartz demonstrated, even while sparking repeatedly, and even when sustaining sufficiently high current flow to ignite nonmetallic conduit.


Many of those who are unhappy with the AFCI requirement, in the meeting and out, are concerned about this fact. Montgomery College instructor Cleveland Tyler worried that homeowners might be outraged to have been misled, by the devices' name, into thinking they are receiving more protection than AFCIs provide. Larry Griffith, who teaches Code classes nationwide, was concerned that those of his students who don't study the Code and product standards carefully can end up with the same false idea. He is going to start making the distinction very explicit. Contractor and Electric League Officer Buddy Friedeman was relieved when Kimblin assured him that AFCIs will interrupt faults at aluminum terminations. Buddy estimated that 75% of aluminum conductor terminations at refrigerators are dangerously defective, especially when nonmetallic boxes have been used. He has measured more than 200 degrees Fahrenheit at these outlets. David challenged Kimblin on this assertion, confirming that AFCIs will interrupt these arcs only after they melt or burn enough to cause ground faults. People then seemed less reassured.


So AFCIs only interrupt a certain subset of arcing faults. Why, then, the did U.S. Consumer Product Safety Commission encourage the development of this technology, and its adoption into the NEC? Woefully incomplete information seems to be responsible. Whenever a fire truck is dispatched to a fire, Schwartz told us, upon its return the driver has to check off the cause of the fire it attended on a "run sheet." Among the causes from which this fire officer, who is not trained as an investigator, must choose, there are ten electrical options: one is an overheating electric-discharge ballast; the other nine are all phrased in terms of arcing. This is the basis for the national data reports attributing nearly nine percent of home fires to arcs, and this is why arcs are the problem that was proposed for solution.


How about capturing data on the real causes? There's no money. In Montgomery County, Maryland, for instance, a fire investigator goes out only if a fire causes $10,000 worth of damage or loss of life. Other jurisdictions have similar restrictions, due to resource limitations. Therefore, using actual fire investigators' data would bias results-in some unknowable direction.


Schwartz concludes that there are no good data to indicate whether a significant proportion of fires would be prevented, because they can be attributed to the arcing faults that AFCIs are designed to interrupt. Therefore, their worth falls, in his view, to a judgement call based on personal experience and opinion. He strongly suspects, based on his decades of fire investigation, that the vast majority of electric fires are the result of the heating caused by a lower level of current flow, through resistance; not the result of high-current parallel arcs. Fire in the permanent wiring occurs, in his experience, due to overheating at the termination points. He believes that it is extraordinarily rare for a staple to be driven through NM-B cable hard enough, and at just the right angle, to cause a parallel arc from the hot to the return conductor. AFCIs are not designed to interrupt overheating, and he suspects that AFCIs will not help much to prevent such fires through indirect means. Therefore, while not questioning the ability of AFCIs to do their job, he questions the degree to which they are worth the investment. He believes the manufacturers that urged that AFCIs be required do not spend enough time talking about their limitations.


Here is where Kimblin differs. A couple of Code cycles ago, representatives of the Electronics Industry Association and the CPSC urged that the instantaneous-trip settings of circuit breakers be lowered, to interrupt large faults more quickly. Manufacturers fought this, to avoid nuisance tripping. AFCIs, he believes, provide a superior solution.


He focused on the fact that as much as a billion dollars of property damage annually is attributed to electrical fires, nearly 400 deaths, and many more injuries. However much (or, one is polite enough not to say, little) of these costs could be reduced by utilizing AFCIs, he believes, doing so will be well worthwhile. They interrupt sputtering faults, both hot to grounded conductor and, he asserts, hot to grounding means, that would never trip time-delay circuit breakers, and other low-level AC arcs tend to self-extinguish at the zero-crossing point, making them highly unlikely to start fires. Furthermore, he emphasized the fact that the AFCIs presently on the market or planned for marketing incorporate 30 mA ground fault circuit interrupters. He talked repeatedly of AFCIs as protecting against other than parallel faults. His logic is that if a series or glowing fault does not burn through enough insulation to become a high-current parallel fault, in modern wiring it will burn through enough insulation-even, he asserts, at receptacle terminations--for 30 mA to flow to ground.


That is the meat of what we heard. Here is some more background. The Consumer Product Safety Commission got this project rolling originally because it is concerned about the dangers associated with older wiring, defined as 40 years old or older. (Unfortunately, as Kimblin acknowledged, AFCI requirements are not retroactive, and furthermore electrical systems going back to the 1960s and earlier may not have grounding means available to trigger the 30 mA protection.) They named 14 fire precursors, and entertained a variety of proposals from people and organizations offering means to identify them and interrupt power. None of the means proposed could do the complete job, but the best, AFCIs, caught six.


Kimblin made a number of points in favor of AFCIs' worth:

First, he appealed to authority. The Code-Making Panel (CMP) was not blindsided with the AFCI concept; they have considered it for two Code cycles. Furthermore, word on field performance has been available. AFCIs have been sold commercially for more than three years. Schwartz has had the opportunity to bring his arguments up with the CMP, and they voted down his objections eleven-to-one.
Kimblin pointed out several advantages of AFCIs over circuit breakers in protecting against arcing faults, even faults somewhat above the circuit breakers' instantaneous-trip range. First, as Schwartz had noted, they are faster. Bring two wires together and there will be significant arcing as current flowing through the gap trips a standard circuit breaker. Bring two wires together and there will be far less arcing as current flowing through trips an AFCI breaker. Second, many arcs sputter, making and breaking: high current, then off; high current again, then off. The intervening "off" time will prevent a standard circuit breaker from tripping, but not an AFCI.
He also argued for the efficacy of AFCIs in stopping fires that occur beyond the fixed wiring. Two-wire zip cord also can sustain arcs from one conductor to the other, and they can carry well over 75 amps. This will trip an AFCI.


What about series arcs, which AFCIs cannot stop? Both Schwartz and Kimblin agreed that there is no commercially-available technology that hopes to identify them and distinguishes them from legitimate loads. (Schwartz said that, to his knowledge, there is no such technology even in the works.) If there are dangers associated with arcs occurring at wiring terminations on devices or at splices, their characteristics may be indistinguishable from the arcs drawn by thermostats or switches. But Kimblin downplayed their danger, which Schwartz's demonstration might have caused the audience to see as major. He said,"It is essentially impossible to get a series arc by drawing two pieces of copper apart up to100, even 200 amps, past the half-arc [the zero-crossing point]. It will be very difficult for it to start a fire." In Europe, Kimblin noted, with standard utilization voltage of 220 or 240 volts, the risk that a series arc, whose current is limited by the load being fed, will not self-extinguish is greater than it is here. He named UL 1699 as a standard that requires includes protection against possible series arcs in NM-B (but not apparently, other cables such as ACHH, SEC or SER). Any series arc in a conductor that is broken inside its insulation must self-extinguish before a flame can leave the cable sheath. This is his basis for arguing that, in any grounded cable, the 30 mA ground fault protection will interrupt power, thanks to contact between the damaged conductor and the grounding conductor, before a fire can be started. Further, the ground fault interrupting characteristics provide some "glowing connection" protection at receptacles.


Larry Griffith asked Schwartz and Kimblin where AFCIs are likely to be required in the future. The speculations included additional residential circuits beyond bedrooms, and higher voltage circuits. Later, Kimblin explained that inherent in all our discussions was the expectation that
most residential dwelling unit circuits will eventually be covered. He added that the impact on the fire statistics (dealing with whole house) will be limited if only bedroom outlets are protected.


Larry asked whether AFCIs are likely to be useful in hazardous locations, to protect against arcs around explosives. Kimblin was uncertain on this issue. After all, an explosion may take just one spark, and AFCIs interrupt arcs, they don't prevent them. However, he asserted that a hazardous circuit location would be safer with AFCI protection than without. On the other hand, he talked of aircraft fire or explosion prevention as one application that is being investigated seriously for AFCIs. Even though AFCIs are not a reliable means of preventing explosion, and jet fuel is highly explosive, there are thousands of feet of wiring per aircraft, including lots in the fuselage, which is not a hazardous location. It is so expensive to upgrade the wiring that it can be cheaper to build a new aircraft.


Larry also asked about the requirement, new as of January 2002, that room air conditioners be supplied with AFCIs. Kimblin explained that while functionally they are being supplied with such protection, technically the mechanism is different from that used in AFCI circuit breakers. The equipment's cable is covered with a grounded sheath, and provided with inherent ground-fault protection.


Membership Co-coordinator Ed Holt asked Kimblin for speculation on the cost-benefit analysis of installing thousands and thousands of AFCIs, noting that while life safety is important to all of us, the reality is that cost must be considered. Kimblin demurred. Buddy suggested waiting to see whether insurers lower the cost of insurance on a house with AFCI protection. He pointed out that Allstate already asks for an electrical contractor's certification of the safety of any electrical system more than 40 years old.


David pursued the cost-effectiveness question, asking whether adding a few smoke detectors to a home, for the price of an AFCI, might not increase life safety as much or more. Both Schwartz and Kimblin responded that while smoke detectors are valuable, there is a world of difference between preventing a fire and detecting a consequence of fire. Schwartz warned that you can be left with only a minute and a half to get out of the house once a smoke detector goes off. He said that a smoke detector that is properly installed, and functioning in accordance with the UL standard, can wait two to four minutes after the fire begins to go off.


David asked why the 75 amp trip level was set for arcing faults, and the 30 mA trip level for ground faults. Since the arcing faults are unique waveforms, wouldn't lower be safer, making suppression come earlier? Since their ground fault interruption is designed to step in after major breaches of insulation, wouldn't it be better for the 4-6 mA Class A GFCI level be incorporated?


With regard to the AFCI setting, Kimblin responded that at least 75 amps of fault current is available at all branch circuit outlets tested, with average available fault current more like 200 amperes. Therefore, the present setting will-not should, but will--result in tripping. A lower setting would require identifying the signatures of many more benign loads that may be present on branch circuits. These are loads that do not need to be differentiated from apparent arcs when using the 75 amp trip level. A 10 amp motorized appliance can draw 100-150 amps as inrush current; a fluorescent ballast can draw 75-100 amps.


With regard to ground fault protection, Kimblin speculated that the Class A GFCI setting probably would not significantly increase safety from arcs over the present, GFPE, setting. Schwartz speculated that combination AFCI/Class A GFCI equipment will be chosen increasingly in the future, explicitly for the added shock protection.


Larry expressed concern about false tripping of Class A GFCIs, listing various circuits that he prefers not to protect from ground faults. Schwartz talked about the research based on home inspectors' reports indicating that many GFCIs die over time, especially in high-lightning areas. Kimblin mentioned that the GFCI standard, UL 943, is undergoing change right now. Past President Jim Wooten expressed concern over the possibility of false tripping of AFCIs. Kimblin responded that while there are waveforms similar to those associated with parallel arcs, for example inrush currents and the spikes associated with incandescent lamp burnout, these can be distinguished from parallel arcs.


Finally, David asked why the 2002 NEC requires protection for that part of the circuit upstream of the bedroom outlets themselves. Kimblin's response was that this bit of protection is "free," by which he meant this it is both important and a bonus. Schwartz's was that a wide range of knowledge is represented on any Code-Making Panel. Some of those present are there to vote for their organizations' primary interests, and do not necessarily familiarize themselves with the details of some proposals. It takes three to five thousand dollars a year to attend committee meetings, so pretty much everyone on a CMP has a sponsor with an ax to grind. Kimblin disagreed with this cynical analysis.

Subsequent to the meeting, Square D Engineer Alan Manche offered a number of comments, condensed to the following:


"As far as the fire data, we gathered the historical information available to the industry, from numerous sources, in order to understand whether AFCI would provide an answer to concerns voiced and documented by the CPSC, UL and EIA. A number of manufacturers, alongside the CPSC, UL, and EIA, arrived at the conclusion that AFCI technology was the answer to reducing a large portion of the potential ignition sources from the electrical arc hazard. Reporting forms can always be enhanced when you're looking for very specific information, and a debate over the statistical significance of the data will always arise. . . .

"NAHB held the only negative position on AFCI as a participant on CMP-2 during the 2002 NEC process.

"I noticed that Larry asked about AFCI being used in hazardous locations. The AFCI product and product standards that are currently available have not included a review of hazardous location issues, and therefore it should be made clear in your article that AFCI is not an answer for a hazardous location and should not be applied or relied upon as a safety device for such a location. . . Clive may have turned his engineering hat on with a spirited intent that maybe this application needs to be reviewed and might help. My stance is that no such evaluation has been performed and I did not want anyone to carry away that AFCI is an answer to protecting circuits in a hazardous location.

"As far as the LCDI and AFCI for window air conditioners, the LCDI led the way and the introduction of AFCI technology into the NEC provided another means to protect the cord from the document incidents so Clive's comment obviously reflect that he is an AFCI manufacturer and that LCDI is an alternative. Either will get the job accomplished using different technology.

"You indicate that 'both Schwartz and Kimblin agree that there is no commercially-available technology' ([for direct] series detection) - but the series detection discussion is where the industry is currently engaged.

"Note that UL listed a receptacle that was evaluated at the 5A series condition, but as you say it is not yet commercially available nor will it meet NEC 210.12 "to deenergize the circuit," but it does demonstrate a move toward the 5A series level. Square D listening to the pulse of the industry discussion thought that industry was demanding 5A series protection in AFCI technology, so Square D submitted a proposal to the UL 1699 product standard that would require AFCI technology to detect the 5A series condition. Industry, the UL STP (balanced committee) in this case, rejected the 5A series performance requirement to the AFCI technology. So, as an AFCI manufacturer, I'm wearing the white hat here, we attempted to give the industry the 5A series protection that we thought was being demanded from the discussion as found in your meeting. . . .The 5A detection is a series evaluation. How should I say. . . the industry is divided on this issue, not even all of the manufacturers are on the same side of the fence here. . . I would recommend you go directly to the horse's mouth such as UL, or an inspector member of the STP, to get the bulletin of proposed changes to UL 1699, voting and comments from the STP members on the AFCI topics."

The reasons for not putting Arc Detection in the main are simple.
1. If you have an arc, you would trip the main leaving the house without power.
2. Finding the circuit with the problem would take far longer if you had to troubleshoot the entire home.