I'm planning to use an LTC7801 for a step down converter from up to 80V to a selectable voltage of 3.3 to 24V. I find the device interesting because it provides the option of setting the gate drive voltage what probably could allow me to use low voltage enhancement mode GaN FETs for improved efficiency without having an extra GaN-FET-Driver between Buck Controller and FETs. For most buck controllers that is not possible because GaN FETs are rather sensitive at the gate allowing a maximum voltage of 5.5V.
During operation, the bootstrap capacitor of a half bridge driver usually charges to the supply voltage minus the bootstrap diode's voltage drop _plus_ the forward voltage of the lowside MOSFETs body diode (during dead time current flows through it resulting in a slightly negative voltage for a short time at the inductor). GaN-FETs do not have a body diode but show comparable behavior resulting in a voltage drop of about 2 V. When I use a (small) Schottky diode in parallel, this can be reduced to less than 1V, but it still charges the bootstrap capacitor - according to LTSpice - to slightly more than 5.5V when the gate drive voltage is set to 5V. Reducing the LTC8701s gate voltage to about 4.5V solves this effect and seems to be possible within the simulation, but is out of spec according to the datasheet that defines 5V as the lower limit. Is it feasible to set it to 4.5V anyway?
Also it would be nice to parallel two of the devices into a two phase buck controller. Since they implement an accurate current control architecture, connecting their ITH-pins and providing two 180° out of phase external clocks look like a good approach. LTSpice has the same opinion but again this might be out of spec. Is paralleling possible that way?