This is part of the project


I designed this circuit from the original ZVS Mazzilli driver. I had to build a driver capable of working at 110kHz, but the original circuit was very slow and my Mosfets burnt because they spent too much time in the ohmic region.

So I had the idea to use a Mosfet driver, I chose the TPS2814P for this task and it has proven to work very well. This driver is now working at 110kHz, with IRFP250N Mosfets, 48DC and around 148V at primary, it can provide more than 120W and the Mosfets stay cold.


If you're looking for high voltage, this is your driver. You can feed it with higher DC voltage, and it will work well with less primary turns (to prevent saturation) because it's much faster than the original circuit. Also because you can make the primary with less turns, your secondary voltage will be higher.

The benefits you get using this are:

-Higher DC voltage input

-Higher AC voltage output

-Colder Mosfets, because the switching is very fast

-Higher frequency

-Lower primary resonant current

-Negligible power lost through the UF4007

This is the full schematic:


The two inductors that forms the primary (L) and the capacitor next to them (C) don't have values because it's supposed that you give them any value you want, just remember that the ZVS will work at resonant frequency given by:

I used 136nF for the capacitor and 3+3 turns around the ferrite core you can find in the Components list.

Note: if you are wondering why I put 100ohm resistors before the UF4007, I saw with the oscilloscope that placing those resistors reduced the noise coming to the inputs of the TPS2814P. This is because those resistors delay a bit the turning off process of a Mosfet. This prevents the scenario where the Mosfets are switching and we can't consider that one of them is fully turn on, so the 100uH inductor creates a peak of voltage to both drains. If you make the turning off process a bit slower, the other Mosfet will have time to turn on.