The common base configuration has been well documented for the BJT .Small signal analysis shows better high frequency capability then common emmitter for the same transistor.I have seen common gate used on JFETs at VHF where the low input impedence would be a better match to 50 ohm. Would a common gate circuit arrangement which would be more complex on an orthodox SMPS give lower switching losses ? Is there a definative statement about how much lower they would be ? I did this on a small ZVs royer where the circuit complexity penalty for common gate was minimal but the actual switching losses of ZVs are so low anyway when input volts are high and input currents are low .Is it worthwhile to look at common gate ?
If the gate is set at maybe 10V (fixed) and the source is hit with a 20Vp-p square wave (0V to 20V) then I can see good reasons to trial it - the MOSFET would turn off very quickly because the gate would be taken negative with respect to the source and to turn the device on the source would be at 0V.
However, the source signal is having to conduct the same current as the load so this is its chief disadvantage I reckon. If the drain current is 10A the signal at the source has to cope with 10A when it is low falling to virtually nothing when the source is high.
I'm not saying that there is a big power loss in the drive signal, just that the drive signal has to cope with a rapidly changing load.
Of course on the use of a common gate, any CMOS op-amp using MOSFETs as the differential pair uses a common gate by virtue that the sources are connected.
It's an interesting idea.