the way a GFCI functions, would it trip
if a bad connection were made to it ?
(if say, a terminal with the hot or
neutral wasn't screwed down tight &
a plugged in load drew current causing
the terminal to warm up - would the
GFCI trip ?)
The load's current use is sensed by the GFCI whether wired dubiously as in
the example you imposed.
The problem child is when the grounding and neutral are swapped.
Arcing, as implied by your example, needs an AFI circuit for detection.
I don't think so. Why would the resistance of a bad connection be any
different from the resistance of a light bulb.
I wouldn't assume there will be arcing just because a connection isn't
tight enough. So I wouldn't assume that even an arc fault circuit
breaker would be tripped by a loose connection. Am I right or wrong?
A GFIC is made to trip on sudden line voltage changes. A slow warm up
would not trip the built in breaker. This very problem cause a house
fire that burned a third of my parents house to the ground. Two wires
in some old Romex touched and caused a slow short but it was not enough
to trip the breakers. Therefore, it warmed up long enough to catch fire.
GFIC's are a nice safety feature but it certainly doesn't catch all problems
There's all kinds of "bad connections".
If the bad connection resulted in a difference
of 5 mA or more between the current in the neutral
versus the current in the hot wire, the GFCI
would trip. The example you gave would not cause
the GFCI to trip.
As others have mentioned, a GFCI would probably not trip. They sense current
differences, and usually have current overload sensing in them also. But a
bad connection causing the terminal to warm up would not trip it unless it
got very hot, and even then the results are indeterminate.
So far there has been a lot of misinformation in replies to this thread.
Only the smoke will indicate a simple HOT connection.
If there is any arcing a AFCI should detect the fault and disconnect.
"The ?AFCI? is an arc fault circuit interrupter. AFCIs are
newly-developed electrical devices designed to protect against fires
caused by arcing faults in the home electrical wiring." ****
A GFCI will only protect from differences in the current flowing in the
two conductors beyond the GFCI. It will cutoff the current when a very
small current difference, 5 or 30 ma., is detected and should do it
quick enough to save a person from electrocution.
Neither device is intended to protect from voltage spikes and neither
device can detect a simple hot connection.
A hot connection alone will not csause a difference in current between
the 2 legs of the circuit, only a ground fault can cause that and a GFCI
should detect that.
thanks everyone for the responses.
the powerpoint link was very informative.
if i understood it correctly, any current differential
if one terminal has contact resistance (loose),
wouldn't the current flow on that leg be reduced
by the amount of resistance ? and therefore trip
the GFCI ? (since it would see less on one leg,
vs the other - or am I misinterpreting, since a
hot/load/neutral circuit is basically in series,
and current is only depended on how much the
Both wires are part of the same series circuit and the current will be
the same whileever there is no other path (ground).
Current in = current out.
The resistance of the contact will be part of the impedance (along with
the real load) that determines the magnitude of the current.
Let me say what I said another way. GFCI outlets and circuit breakers
measure if there is a difference between the amount of current going
throught the hot wire and though the neutral wire. If they are the
same, the gfci sees no problem. Having a loose connection does not
mean that will happen. Using an appliance that shorts to your body
and from there to some other path than the neutral wire *would* cause
it to trip.
OTOH, if you took two all-metal screwdrivers, one in each hand, and
stuck one into each of the slots in an outlett, even a GFCI outlet,
you could burn your heart to a fine grey ash, and the breaker would
not trip. Because the same amount of current was going through the
I agree with others that a GFCi wouldn't trip on a loose connection.
A loose connection can produce enough heat to start a fire without
arcing (a "glowing" connection). Late stages in failure are likely to arc.
AFCIs sense "parallel" arcs - from hot-to-neutral (a "fault", the F in
AFCI). Starting in 2008 the NEC requires them to also detect "series"
arcs, as in a loose connection. As far as I know, no current AFCIs
detect series arcs, so none would detect a loose connection.
GFCIs trip on a difference in current between the hot and neutral, not
line voltage changes. They are primarily for electrocution protection.
The "slow short" you describe, an arc that trips a breaker slowly, if at
all, is exacty what AFCIs are designed to protect against. A more likely
cause is probably an abused extension cord.
AFCIs also include 30mA ground fault protection (GFCIs have 5mA
protection). The idea is, I think, that if a ground wire is adjacent, a
hot-to-neutral arc is likely to also become hot-to-ground.
It sounds like bud has the right answer. I hadn't heard about the
upcoming AFIC's but will certainly pass the info on to everyone at my
office (I'm a construction administrator for a large architectural
firm). Thanks bud for the education.
They call it Ground-Fault Protection of Equipment when it is at the
BTW get used to AFCIs. They will be on all 120v 15 and 20a circuits
in a dwelling in 2008 if the code goes as drafted. Comments are still
open until October.
Yea - it is to detect the arc, not protect people. AFCIs can have a 5 mA
ground fault trip and be used as both AFCI and GFCI. It would require 2
test buttons. Probably would see them if the code change below goes into
To PPS - I presume "upcoming" is the requirement to detect "series" arcs
starting 2008, not 30mA ground fault trip.
Considering the new AFCIs aren't on the market (or are they recently
out?), and it is only about 1.5 years for field experience until they
are required in 2008, it seems like requiring the new AFCIs in all
15/20A dwelling circuits is not a great idea (whatever the wisdom of
requiring the current AFCIs is). Bet there are plenty of comments on
that code change.
An interesting piece on AFCIs, including why normal breakers are
inadequate for arcs and arcs that shouldn't trip them is at: