AFCI Recall Notice - look for the blue test button on SQ D AFCI's

[Jerry]

CPSC, Schneider Electric North American Division Announce Recall of AFCIs WASHINGTON, D.C. - The U.S. Consumer Product Safety Commission announces the following recall in voluntary cooperation with the firms below. Consumers should stop using recalled products immediately unless otherwise instructed.

Name of product: Arc Fault Circuit Interrupters (AFCI)

Units: About 700,000

Manufacturer: Schneider Electric North American Division, of Palatine, Ill.

Hazard: An AFCI is an electrical circuit protection device (circuit breaker) that detects electrical arcs from cracked, broken or damaged electrical insulation and shuts off power to the circuit before the arcing leads to a fire. An electronic component failure inside the AFCIs can cause the devices to not detect an electrical arc. Although the AFCIs will function as regular circuit breakers, they may not detect an arc fault, posing a safety risk to consumers. Clarifying Statement

Incidents/Injuries: Schneider Electric is investigating one reported fire during a new home construction that may be related to this problem. No injuries have been reported.

Description: The recalled Square D QO=AE and Homeline=AE Arc Fault Interrupter circuit breakers are used with 15- and 20-amp branch circuits. They are required to be installed in bedroom circuits in accordance with the 2002 National Electrical Code. The recalled units were manufactured after March 1, 2004, and have a blue test button. The AFCI circuit breakers have one of the following date codes - CN, DN, EN, FN, GN, HN, or JN - stamped in red on the breaker label located just above the wiring terminal. The recalled units also have one of the following catalog numbers printed on a label on the front of the breaker: QO115AFI, QO115AFIC, QO120AFI, QO120AFIC, QOB115AFI, QOB120AFI, HOM115AFI, HOM115AFIC, HOM120AFI, HOM120AFIC, QO115VHAFI, QO120VHAFI, QOB115VHAFI, or QOB120VHAFI.

Sold at: Electrical distributors and retailers sold the AFCIs between March 2004 and September 2004 for between $30 and $130.

Manufactured in: Mexico

Remedy: Installed AFCIs will be replaced free of charge through electrical contractors. Consumers can return uninstalled AFCIs to the retailers or distributor from whom the unit was purchased for a free replacement unit.

Consumer Contact: Consumer should call Schneider Electric toll-free at [phone] between 7:30 a.m. and 5 p.m. ET Monday through Friday or log on to the company's Website at

[Long Ranger]

Most of these have been rounded up already. This news is about 4 years old.

[gfretwell]

This is 3 years old but I do wonder how many they have actually recovered. If it is 40% I would be "shocked".

[Gerald Newton]

I was unfamilar with this recall until you brought it up in a previous post. Absolutely nothing was said about this at the NW IAEI Section meeting that I attended in Anchorage during the first week of September. With 700,000 units out there, it is a real problem as you explained in your post. I am creating my 2008 NEC Change course and have been on this AFCI thing for three days. It is interesting to note that at the Sq D site they explain that they are working with vacuum cleaner manufacturers to insure their products do not trip afci's. Can anyone imagine how many new products are going to have to go through this same evaluation. I suppose that we can expect increased costs for the evaluation of an electrical product to insure that it will not trip an AFCI. Is this going to part of the listing requirement for new products or does anyone have a handle on this at all? It looks like a bucket of worms to me. I just talked to an electrician in Hawaii that told me that some types of computer monitors are tripping afci's.

[Long Ranger]

I had a "home inspector" call me and get me involved in this when he thought he had found them in a house. Turned out he had the wrong vintage, and they were O.K. I spoke to my Sq-D rep and they said they would be surprised if many were left "out there" and that Sq-D had really done a thorough job on it. Anyway, the info is about 3-4 years old. That is why it wasn't a topic at your meeting.

[gfretwell]

Look at the archives in the magazine. I think it was in U/L corner or one of the other submitted articles.

[gfretwell]

I would expect this uis really a ground fault problem. I know my TV trips a GFCI when it is connected to my PC and it has a 2 wire plug. I looked and there is no voltage on the A/V ports but I bet they are coupled to the neutral somehow. I suspect they are creating a current carrying ground path with the PC that bonds the DC ground to the EGC.

[Gerald Newton]

I live in Alaska where we have about 50,000 people living in 200 remote villages some accessible by water but mostly accessible by air. So we don't get the information out for 2 or 3 years sometimes. Anyway the IAEI NW Section meeting only comes here every 7 years and this was our year. So I would consider the AFCI recall significant for the meeting. Square D says they will pay for the rplacement. Hmmm, does anyone know what it costs to fly out to the Yukon River to replace a AFCI? Most of these villages do not have an electrician.

[gfretwell]

I still would like to see real numbers on this. I can't imagine the builders were really pro-active it going out looking for them (I know the ones around here do not go out and stir up "closed" customers) and I bet SqD doesn't really have a clue where these things went.

[Gerald Newton]

What is needed is a new label for cord and plug connected electrical equipment used in dwelling units: "Evaluated for connection to an AFCI protected circuit." Perhaps this can be put in the NEC some place.

[gfretwell]

Since the main reason something trips an AFCI is a ground fault I think this is adequately handled with the leakage standards in the listing requirement. I am not sure how you could certify against a false "arc fault" trip since this is a proprietary technology and the various companies don't use the same algorithm. This is most apparent when you look at the spotty performance of "testers". It is this type of varying standard that gives the AFCI the reputation of being more voodoo and snake oil than science. The intent is great, the implimentation is suspect. In the current state of the technology I believe you would get virtually the same protection from a GFCI.

[Gerald Newton]

I think the afci is a device that can help prevent fires. However, it is an application of high tech to an industry that has a history of not willing to pay the price for the strick quality controls that high tech requires. For instance, the AFCI was designed to differentiate between bad arcs and good arcs such as created by a common light switch. But doesn't this assume that the light switch is built within some design parameters. What if the contacts' material are changed from one metal to another, or what if an electronic switch is introduced. It seems to me that the creators of the AFCI circuitry have got an ongoing tripping crisis on their hands. I did my original research by going to the Zlan site. Evidently these are the people that originally invented the AFCI microprocessor. The waveforms are very interesting at:

ref:

[Gerald Newton]

The more I research the AFCI circuit breakers the more evident it becomes that more R&R should have been done.

Quest 1: Did the inventors test every electrical apparatus that is used on 15- and 20-ampere 120 volt circuits in dwelling units that are on the market today? Answer: NO

Quest 2: Did the inventors test every electrical apparatus that is used on 15- and 20-ampere 120 volt circuits in dwelling units that are in use but no longer on the market? Answer: NO

Quest 2: Did the inventors test every electrical apparatus that is used on 15- and 20-ampere 120 volt circuits in dwelling units that will be available on the market? Answer: NO

Apparently, these devices are being forced into homes by new rules in the 2008 NEC while they have not been fully evaluated. The future decisions now require that either the manufacturers develop only those products that have waveforms in the set of waveforms accepted by AFCI's or the AFCI's be reprogrammed to any new waveform signatures that new products may have

I am seriously thinking about writing a letter to the State of Alaska suggesting to not adopt the 2008 NEC without and exception to

210.12(B).

[bud--]

In addition to the expansion of locations where AFCIs are required in the 2008 NEC, the 2005 NEC requires "Combination AFCIs" to be used after

1-1-08. AFCIs are now required detect arcs at a 75A level and will detect parallel arcs. The new "Combination AFCIs" are required to detect arcs at a 5A level and will detect series arcs. As far as I know there are no "Combination AFCIs" on the market now only 2.5 months before they are required to be used. So a far more sensitive AFCI will be required in many more locations without field experience. If you think it is bad now....

IMHO, the expansion of AFCIs to many more locations in the 2008 NEC should have been based on the track record of the AFICs previously installed for bedrooms. It wasn't.

[Gerald Newton]

There seems to be some confusion with the term "combination-type AFCI." At the 2007 IAEI NW section meeting a combination-type afci was defined as one that provides series and parallel protection, the type that you referred to in your post. However, some people think they are afci's that provide AFCI protection and GFCI protection. But the article in the IAEI News article on aci from the July-August 2003 issue puts a new twist on this. In this article written by George Gregory and Alan Manche both of Square D (ref:

"The name "combination" here means that it combines protection of fixed wiring with protection of cords." The 2008 NEC Section 210.12(B) does not define combination-type afcis, but requires that listed combination-type be installed. Perhaps a definition should be added to Article 100.

Your statement that combination-type AFCI's are not on the market yet is interesting. You mean to tell me that Code Making Panel 2 let a combination-type AFCI requirement be added to the NEC when there are no combination-type AFCI's on the market yet? Houston, we have a problem here!

The series of articles in the IAEI news provides convincing evidence that AFCI's do provide substantial protection and have detected many problems in existing old wiring. A program to test dwelling units' old wiring with a type of AFCI tester is warranted. However, I doubt that consumers are willing to pay for the test and had rather take their chances. ref:

[Gerald Newton]

(ref:

)state, "The name "combination" here means that it combines protection

ref:

quoted text -

There appears to be an overwhelming amount of support for AFCI's at the National Association Fire Marshal's web site and videos too. Also, the Siemen's site has numerous references supporting the installation of AFCI's. It is obvious that Code Making Panel 2 was overwhelmed by the support and had little choice about adding requirements to install AFCI protect circuits in all living areas in dwelling units. The tripping problems appear worthwhile for the degree of safety that is achieved. ref:

[gfretwell]

Firemen are not electrical professionals and their investigations of accidents may be anle to pinpoint the origin of the fire but they can't make a professional judgement about whether an AFCI would have prevented it.

Siemens sells AFCIs Of course they like the idea of selling a $50 part when they used to sell a $5 part.

The mere fact that these things are being rushed into the code before they actually exist should be a red flag to anyone who debugged something else in their home, at their own expense. The title of this thread should make that concern perfectly clear. Who knows exactly WHAT this so called AFCI really does. The industry can't even come up with a reliable way to test them other than a button that might be a shunt trip for all we know. I always hear the old saw that we resisted GFCIs too but, at least, they could define a measurable performance standard (5ma imbalance in current) when they sold them to us. Even with that there was still a big learning curve in the industry and the first several generations were crap. The Central Florida electrical inspectors in IAEI determined that up to half of the installed GFCIs were defective. That is also in the magazine if you want to look it up.

[Charles Perry]

I agree with your points. AFCI's are being rushed into the market without proper study, imo. On the plus side, it likely means more work for the labs I manage in the long run as we help people chase down problems with tripping AFCI's ;-)

Charles Perry P.E.

[bud--]

(ref:

> )state, "The name "combination" here means that it combines protection

"Combination AFCIs" combine the requirements of

- "branch circuit AFCIs" - now required - that protect fixed wiring (and plug-in cords) and

- "outlet circuit AFCIs" that protect cords plugged into outlets (also protect supply wiring from series arcs).

The magazine article at:

very good. It is the only place I have seen that explains the current detection levels. "Branch circuit AFCIs" detect arcs at a max 75A level - good for parallel arcs. "Outlet circuit AFCIs" detect arcs at a max 5A level - they will detect series arcs (and parallel arcs). All 3 types have max 50mA (usually 30mA) ground fault detection. This is not detection at the 5mA level required by GFCIs.

The code panel wants the definition in the UL standard.

IMHO it is a major stupidity. IMHO a second major stupidity was vastly expanding the locations where the nonexistent AFCIs had to be installed. It may have made sense when the 2005 NEC was written to think "Combination AFCIs" would be available 1-1-08. But there were 2 proposals from Eaton/Cutler Hammer to delete the requirement for "Combination AFCIs" in the 2008 NEC because the devices did not exist. Now, even later, the devices still apparently don't exist.

IMHO a third stupidity was expanding the locations required for AFCIs without basing it on data from the AFCIs already installed. Far as I remember, all the 'data' was anecdotal or suppositions. (Anecdotal evidence proves burning witches was effective.) AFCIs may be a great idea, but can't someone generate some data based on experience?

The fire marshals specifically:

- Brush off the lack of series arc protection in the existing AFCIs. Presumably series arcs are much less dangerous than parallel. (I haven't seen data that indicates most fires are caused by parallel arcs.)

- Brush off nuisance tripping - annecdoatal comments that nuisance trips were caused by actual problems (*some* no doubt were).

- Does not address expansion past bedroom circuits. A cost analysis was based on 2 likely circuits (how many will be required with the new rules) and the cost of the original AFCIs (will the much more sensitive "Combination AFCIs" have a lot higher cost?)

Another issue that annoyed me was smoke detectors. There were several proposals to exempt smoke detectors installed in bedrooms from being required to be on AFCI circuits. Losing the power, particularly when a fire may be starting, is a larger danger than fire caused by detector arcing. Many jurisdictions apparently feel this way. The response of the code panel was the proposals had no data that "indicate that AFCI devices are not compatible with smoke alarms". This is totally irrelevant to the argument. Does the code panel have data that indicates smoke detectors are a significant source of fires from arcs?

Incidentally, a couple proposals indicate NFPA 72 (smoke alarms?) has been changed to require a "secondary power source for smoke alarms that are installed on AFCI protected circuits" (battery backup - or maybe a smoke detector emergency generator?). [Is NFPA 72 adopted by anyone? Is it adopted by reference in fire codes?]

Rumor is the locations required to have AFCIs was significantly limited from the original language. (Basically not required where a GFCI is now required?)

It will be interesting what jurisdictions do with all of this. Greg?

[gfretwell]

The problem with depending on Nationally Recognized Testing Labs to provide code language is there are over a dozen NRTLs that compete with each other for testing business. There is no gurantee that these will use consistant terminology.

The current plan is for the state of Florida to skip the 2008 NEC and pick it up at the 2011, probably some time around 2012 if past history is any guide. They did cherry pick the softening of 680.26(C) requirements and put that in the building code.