wnjj wrote:
Maybe it is semantics but you did say that it "does not use the current imbalance detector" and I disagree.
I did say that, and upon looking deeper than I originally intended to I see that you are correct. For more FYI, here's what I suspect is an early implementation of grounded neutral detection that used 120 Hz signal injection via a separate transformer:
http://www.masterworksservices.com/files/GFCI_by_Sam_Goldwasser.pdfAnd here is a chip approach to 120 Hz. signal injection that I believe "could" use a separate transformer, but in practice likely just adds a coil to the current sense transformer:
http://media.digikey.com/pdf/Data%20Sheets/Texas%20Instruments%20PDFs/LM1851.pdfI have no clue what devices use the LM1851 or FAN4147, if one is obsolete, and so forth. Intuitively the FAN4147 sounds like a more advanced approach to me, and a single 2-coil transformer is probably required for the oscillator to work. I suspect somebody did a good job thinking out of the box to come up with this approach.