>> >From what I remember the GFI circuits utilize AC only. One of the
>toroids
>> is double wound: one winding is the hot lead, the other is the
>neutral. The
>> windings are set to cancel out each other: if the same current is
>flowing
>> in each line then there is no magnetic field. A third winding on the
>toroid
>> senses the inbalance,

>> The second toroid is there to detect a I believe a neutral to ground
>short.

I think that is how it works.  Simple GFI's just have one coil, which
detects any line-frequency common-mode current on the line and neutral
wires.  They can't tell if the neutral is shorted to ground unless a
portion of the load current ends up flowing through the short (which it
generally will, unless the load current is zero).

A more robust version would apply a test signal to the neutral wire and
ensure that all of the test current (if any) returns via the line wire.
If it applied the test signal to the line wire as well, then the detector
coil could detect leakage from either wire as a test-frequency signal.
These would work for AC or DC lines (as long as the transformers don't
saturate from DC), in fact they would work with no line voltage at all.
Both line wires need to have a low impedance to ground for the test
frequency.  Loads containing capacitors from the lines to ground could
cause a false fault detection if the test frequency were high enough.


>  Possibility:  Does anything stop you from using 48VDC through a
>couple
>chokes & then addding in an AC Bias, so you can detect the GFI leakage
>of the AC component only?

Unlike utility current, the 48V source doesn't need to be connected to
ground.  It should be loosely connected so that it can't rise high above
ground due to static, etc. and so ground faults could be detected.  If it
is isolated then the first ground fault isn't hazardous, it merely
defines what will be 48V "ground" today.



_____________________________________________________________________
You don't need to buy Internet access to use free Internet e-mail.
Get completely free e-mail from Juno at http://www.juno.com
Or call Juno at (800) 654-JUNO [654-5866]