I've been looking for a way to get around the saturation voltage
of triacs and have considered FETs and IGBTs for AC switching.  It SEEMS
to me that the saturation voltage on an IGBT is going to be higher than
that of a FET.  I view an IGBT as a sort of darlington with an FET as the
first transistor.  Is this correct?  If so, the input FET can go to a
source-drain short and we still need 700 mV or so on the collector to
turn on the base.  Seems that directly driving the gate of a FET would
give us lower voltage drop (just due to Rds on).
        In AC applications, I've thought of putting two FETs in series
and using the intrinisic diodes to steer the current into the appropriate
FET.  Again, we'll lose about 700 mV across the diode of the FET that is
being bypassed.  It seems that the voltage drop will be approaching that
of a triac, and we have increased circuit conplexity to boot!

Harold