Newsgroups: sci.electronics.design,alt.binaries.schematics.electronic
Subject: Re: GFCI Failures + Gadgets
References:  <[email protected]>  <[email protected]> 
buck rojerz  writes:


> The first part of the input of the sensor for the GFCI was a toroid 
> transformer with three windings.  The first two (I will call the main) 
> windings were wound as current opposing, around the toroid.  Such that if 
> the two main windings(Hot & Neutral Lines) had the exact same amount of 
> current passing through them, they would cancel each other and no output 
> would be present at the third winding.  The third winding was connected 
> to the differential inputs of an op-amp.  If there was any imbalance of 
> current between the hot and neutral lines, the imbalance would negate the 
> cancellation and a current would be induced in the third winding of the 
> toroid, biasing the op-amp and in turn triggering the relay to break the 
> power circuit.  The imbalance would be indicative of current taking an 
> inappropriate path to earth, through a path other than the neutral line 
> of that same outlet.  

True. Good explanation of GFCI operation.
 
> The statement that no safety ground line is required for the GFCI to 
> operate the way it was meant to, is true. 

I can agree on this.

> However the test button would 
> not operate, because the test button put a resistor from the hot line, to 
> the safety ground line as a test, to cause the imbalance in the hot and 
> neutral lines.  

I can't agree as general case. 
There is another way to create imbalance in the hot and
neutral lines without ground connection: Wire the resistor 
that causes the imbalance current so that it is wired 
from GFCI output live connector to GFCI input neutral 
connector. In this way when it is connected, it creates 
imbalance current through GFCI. I have seen GFCI device 
testy button implemented in this way. So no need for 
safety ground connection for this reason.

> Even if no safety ground line was connected to the given 
> outlet, any amount of leakage above 2 or 3 ma, to another path to earth, 
> such as a waterpipe, would be measured, as an imbalance in the toroid 
> transformer, flip the output of the op-amp and in turn, trip the relay.  
> A safety ground is not necessary for a GFCI to operate normally, but of 
> course it would be stupid not to have a safety ground in any event.
> (this post was read in alt.binaries.schematics.electronics)

-- 
Tomi Engdahl (http://www.iki.fi/then/)
Take a look at my electronics web links and documents at 
http://www.epanorama.net/