Hello Usaghi,
Q1) VDD2 is shorthand for EITHER (VDDA to GNDA) or (VDDB to GNDB).
The UVLO for each channel block works independently, so if VDDA to GNDA falls below the UVLO level, but VDDB to GNDB doesn't fall below the UVLO, VOB will still output, while VOA will be tied low.
Q2) Yes, it is possible to supply VDDA to GNDA with a different voltagefrom VDDB to GNDB. This is due to the fact that each output channel is on a separate silicon die. This is what allows them to float with respect to each other so well. There is no concern about latch-up due to the grounds bouncing with respect to each other, like there is in high voltage level shifter gate drivers.
Additionally, there is no reason you need to have GNDA sit at a higher potential then GNDB. You could even use VOA as your lowside output on a half-bridge if it would help with layout. The two channels are essentially symmetric.
Q3) VDD2 would be meaning whichever channel block you are talking about. VDD2 is equivalent to saying (VDDA or VDDB)
Q4) Yes, both output channels can be operated independently. The most common use for a two channel chip like this would be to run a half bridge with thetwo outputs mutually exclusively with a small amount of deadtime (usually around 150-500 ns).
For more exotic operations, I've used one channel as an on/off, and the other as a PWM. If you are using a bootstrap configuration, just make sure that the on/off channel has sufficient duty cycle and frequency to keep the bootstrap capacitor charged.
Q5) Yes, you can operate only one side if you desire. As an example, if you only wanted to use VOA, you could just not connect VIB, VDDB, VOB, and GNDB. A more common method would be to short VIB to GND1, and VDDB to GNDB, just in case. A floating VIB has the potential to increase power on the primary side, but usually wouldn't.
RSchnell