AFCI - Why I Have a Problem With It

By Bob Huddleston, rhud@eastman.com

 

Arc fault circuit interrupters (AFCIs) are required by the 2002 edition of the National Electrical Code to be installed to protect all 15 and 20 ampere dwelling unit bedroom outlets (NEC 210.12B).  According to current literature, these devices are designed to protect against fires caused by arcing faults in the home electrical wiring.  The AFCI circuitry continuously monitors current flow through the device.  AFCIs use unique current sensing circuitry to discriminate between normal and unwanted arcing conditions.  Once an unwanted arcing condition is detected, the control circuitry in the AFCI trips the internal contacts, thus de-energizing the circuit and reducing the potential for a fire to occur.  An AFCI should not trip during normal arcing conditions, which can occur when a switch is opened or a plug is pulled from a receptacle.(1)

Manufacturer ‘A’ advertises their Arc Fault Circuit Interrupter as a residential circuit breaker with an integrated processor which recognizes the unique current and/or voltage signatures associated with arcing faults, and acts to interrupt the circuit to reduce the likelihood of an electrical fire  They go on to say that the most familiar and dramatic arcs in residential wiring are associated with short circuit arcs between two conductors.  A parallel arc can often sputter on and off for extended periods of time without detection, increasing the risk of fire.  Series arcs can also occur at a break in a single conductor.  These series arcs are relatively low current since they are in series with the load.  In practice it is difficult to sustain such arcs since copper-to-copper arcing is unsustainable at 125 VAC.  However, such arcing can occur in the presence of carbonized insulation.  The arcs are typically sputtering and intermittent in nature.

Manufacturer ‘B’ also makes an AFCI device.  Their literature states that typical conditions where an arc may be generated (include):

 

 

Manufacturer ‘C’ makes an AFCI device.  Their information states: Worn insulation or an out-of-the-way extension cord, wiring in the wall that's been nicked by a nail, or any sort of exposed contact or conductor can lead to arcing - a free, uncontrolled flow of electricity - which can, in turn, cause an electrical fire.  An AFCI uses electronics to recognize the current and voltage characteristics of arcing faults, and interrupts the circuit when the fault occurs.

The Plant Protection Electrical Safety Team at a large chemical manufacturing company, in cooperation with their resident electrical contractor, designed and built a test set-up to demonstrate AFCI actual arc fault tripping.  The intention in building this rig was to provide a demonstration for instructional purposes at the local Area Safety Council Workshop, held annually.  The following circuit diagram shows the device that was built. 

 

 

The set-up consisted of a 2-circuit distribution panel, with a standard 20-ampere circuit breaker in one slot and an AFCI in the other slot.  The panel was fed by a standard 115 VAC outlet through a cord and plug arrangement.  A switch (SW-1) was placed (three-pole / double-throw) on the input and output from the distribution panel so that either the standard circuit breaker or the AFCI breaker was feeding the load(s).  The other pole was used to feed the current either directly in to the neutral bus of the distribution center, or through the AFCI neutral connection.  A set of adjustable carbon-arc rods (AR-1) was mounted at the load side of the distribution panel, so that current to the load(s) was forced to flow through them.  By adjusting the spacing between the rods, an arc could be established and maintained.  Loads were attached to the set-up either by the hard-wired lamp holder, or the duplex receptacle.  Loads consisted of a heat gun (very similar to a hair dryer) rated at 1500 Watts, an electric drill which had a 2.4 ampere current draw, and a 205 Watt incandescent light bulb.  Various combinations of these loads were tried.

Two Manufacturer ‘B’ AFCIs and one Manufacturer ‘A’ AFCI were used in the test set-up.

Testing consisted of the following:

 

Carbon-arc rods were adjusted so that they made contact with each other.  Switches were placed in a position such that current flow went through the AFCI device.  Loads were activated (heat gun or heater turned on; drill started; bulb inserted), and the carbon rods were adjusted so that a sustainable arc was established.  AT NO TIME DURING THIS PROCEDURE DID THE AFCIs EVER TRIP OFF. 

We became concerned when we could not manage to trip the AFCI no matter how the carbon rods were adjusted.  A second Manufacturer ‘B’  AFCI was tried, with exactly the same results the AFCI simply would not trip, despite the fact that we had a beautiful, sputtering, arc between the carbon-arc rods.  A Manufacturer ‘A’ AFCI was tried same results.  For all AFCIs, when the blue test button was pressed, the device did indeed trip; it appeared that the only way to make these AFCI devices trip was with the test button!

 

Carbon-Arc Rods with series arc occurring

 

               

Close-up of series arc, with small heater providing load current

 

        

Next, we changed our strategy somewhat in an attempt to make the devices trip.  Manufacturer ‘B’ says their device will trip on loose connections, so the hot terminal screw on the side of the duplex receptacle was loosened, and while the load was attached and operating, the wire was moved around the screw with insulated needle-nosed pliers in an attempt to simulate the effect of a loose connection.  Amazingly, we could not get any of the AFCIs to trip.  Next, we removed the wire from the terminal screw and, with the load attached, touched the wire to the screw, simulating a broken wire or an extreme loose connection.  Beautiful arcs were produced using this method, yet the AFCIs DID NOT TRIP.

 

At this point, we thought that there was something wrong with the test set-up, and a wire-by-wire check was conducted.  No problems were found with the assembly.  Polarity of the input-powering receptacle was checked and found to be correct.

 

It is my opinion that the three AFCI devices that we tested (two Manufacturer ‘B’ and one Manufacturer ‘A’) did not perform in the manner in which they are advertised, and would not prevent a house fire if a series arc fault were to actually occur.


Manufacturer ’B’, when contacted about our test results, stated that their AFCI device met all UL testing requirements, and then proceeded to describe them:

1)     Carbonized path arc ignition test insulation is cut on a conductor of NM-B cable.  The resulting arc creates an arc to the UNINSULATED GROUND CONDUCTOR (author’s note this is a ground fault, regardless of whether it is arcing or not.  GFCIs trip at 5 mA of ground fault current, and will detect this same event).  The test arc current level is 5 amps for tripping.

2)     Carbonized path arc interruption test insulation is cut on both conductors and ground.  The resulting parallel arc between the conductors and/or conductors and ground causes the AFCI to trip provided the available current is 75 amps or more (author’s note 75 amps of fault current will trip ANY circuit breaker rated at 15 or 20 amps).

3)     Point contact arc test conductors are cut through with an insulated steel guillotine.  As soon as a second conductor’s insulation is cut enough for an arc to start, the AFCI will trip, provided the fault current is 75 amps or more (author’s note again, this is short circuit current (phase-to-neutral or phase-to-ground, and ANY circuit breaker rated 15 or 20 amps will trip with 75 amps of fault current).

There is another test that Manufacturer ‘B’ referred to the carbonized path arc clearing time test.  This is a load-sustained series type of arc, such as that produced by a broken wire or a loose connection (what our test set-up produced).  This test requires that the insulation eventually be burned away so that it becomes like #1 above a fault to ground.  According to Manufacturer ‘B’, there are no commercially available devices that are required to meet this test.

My question is this why does the National Electrical Code require these expensive devices to be installed in bedrooms of new houses, when they WILL NOT TRIP ON A SERIES ARC FAULT?  The authors believe that broken wires and loose connections are a common cause of electrical failure which results in fire, and Manufacturer ‘B’ says that their AFCI will trip if these events occur (and we have proven that they will not trip). 

We believe that the electrical community in general (authors included) has been duped into thinking that these devices are a significant safety improvement, while in reality they will only do what GFCIs and regular circuit breakers do already.  We are very concerned that people are spending their money on these devices, and they are truly not getting their money’s worth.

We respectfully request that Code Panel 2 eliminate the requirement for AFCIs until these devices are capable of detecting and tripping on a series arc fault.

(1)   U.S. Consumer Product Safety Commission website

Mike Holt’s Comment: If you have any comments with the above, contact the Author:

Bob Huddleston rhud@eastman.com

Copyright © 2002 Mike Holt Enterprises,Inc.
1-888-NEC-CODE (1-888-632-2633)