Just read thru this post again...
The Drivers don't need a heatsink as the Qg of any 'Modern' FET is <10nC.
Forget those silly bolt on things as they're way too slow!
Your gate protection sounds interesting but a resistor is all that's needed!
Str.
Apparently we're back to our breakdown in communication. A resistor is not going to keep the fets alive when the driver has latched them ON because of a drive signal loss to the input of the splitter arrangement, which is based on the circuitry from a Nautel NX series AM transmitter. The drive splitter produces push-pull drive required by the fets in the bridge.
BTW I've tried chip type drivers before and all of them burn up at 7 MHz, the power dissipation is just too high when driving 1000pF gates, and please spare me the rhetoric about the merits of GaN as they are cost prohibitive and not suitable for the power levels I'm heading towards. My only complaint about the SiC devices outside of their cost is that they want +20V/-5V drive for best switching perfomance.
Gaga, the 40% figure come from the speed limitations of the driver/fet combination. The 40% allows the fet to switch off before the other device begins to conduct, preventing a condition known as 'shoot through' in which for a short time there is a dead short across the power supply rail and the devices can be destroyed from excessive current flow.
+-RH