View single post by Keinokuorma
 Posted: Fri Oct 20th, 2006 02:30 pm
PM Quote Reply Full Topic

Joined: Mon Jun 26th, 2006
Location: Finland
Posts: 1248
The GFI will compare the "in-coming" and "out-going" current on that circuit, or L1, possibly L2, and N wire currents... or L1, L2 and L3 currents to the N current on a 3 phase system. It relies on the fundamental idea that current coming out of a circuit (in an ideal setting) is equal of magnitude and synchronously (but oppositely) phased with the current going into that circuit. 

On a single phase circuit it is simple: L1 current should ideally be equal to N current.

On a split phase circuit, N current should ideally be equal to the difference of L1 and L2 currents.

On a 3 phase "wye" circuit with Neutral in the electrical center point, the N current should ideally be equal to the geometric sum of the three L wire currents.

On a 3 phase "wye" circuit without N wire, or "delta" system, the geometrical sum of the L wire currents should be zero.

Anyway, the GFI looks for a situation where some of the current returning from the circuit is taking a different path than the N wire. This can be anything, like a mouse toasted between the fridge comp wires and chassis, or a sad classical example, a kid putting a knitting pin into the wall outlet while holding the radiator pipe... or something similar. It can also just be airborne "kitchen pollution" that adheres on the insulations, causing a narrow path for some current to stroll and take a longer way home.

The difference in the L and N currents is called "ground fault current". If this current exceeds the predetermined limit, say 20 or 30mA, the breaker will trip. There are GFI's with higher limits like 500mA, but they're no good for personal protection. They are designed to protect the connected machine from fire if an insulation fault occurs.

As it comes to the inner workings of the GFI itself, the older models contain a "comparison transformer" with up to four short primary windings, wired serially to each live and neutral wire of that circuit. The secondary winding has then a lot of thin convolutions, and the idea is that if the incoming and outgoing currents are equal (but polarized oppositely as the N current flows the other way) the magnetic fluxes induced by each primary coil will sum up to zero, and no voltage will be induced at the secondary winding.  If a ground fault occurs, the classical idea is that the secondary winding operates an electromagnetic armature to trip the breaker. The newer models have faster and more precise technology, but the main idea is the same.

One way to try to verify if the fault current in a machine is external or internal is to plug it into a non-GFI outlet and use a precise clamp-on meter over the connection cable... if that reads a steady current (should not) the machine is leaking current out of the circuit, unless it is a very special connection with multiple feed points... if not, there is an internal fault where the ground wire collects all. Not all clamp-on meters are precise enough to record currents this little.

Sometimes there has been GFI retrofitted on circuits that have "Swedish ground" (chassis grounded to Neutral)... in that case the GFI won't notice an internal situation, but you don't need anything else than the machine standing on a concrete floor or touching some plumbing... and that will probably divert enough current. Also, if a Swedish ground is implemented to a machine that is later on wired to a properly grounded circuit, but the loop is not removed, that will draw half the intended N current to the ground wire, which is WAY enough fault current. Consider these points too if you get to troubleshoot a constantly tripping GFI situation.

Last edited on Tue Jul 10th, 2007 12:23 pm by Keinokuorma

"There is no reason anyone would want a computer in their home."
- Ken Olson, Digital Equipment Corporation (1977)