Say, we are building a synchronous buck and we have some beefy 600V 10A MOSFET, but it is not really suited for high frequency application - we need to be aware of dv/dt limits and it is hard to achieve fast switching with FET huge capacitance. Also it may have up to 20V at gate.

What if we add a dirt cheap 20V 20A logic level MOSFET at source of our HV MOSFET.


simulate this circuit – Schematic created using CircuitLab

It sound like this way we will be able to achieve insane switching speeds even with a bad driver in cost of small switching loss of bottom FET. In case of ZVS topology, our switching loss will even decrease because of smaller gate voltage.

I've never seen anything like this in practice however (except GaN and some other scientific stuff). I suppose there is a good reason for it (layout? diode recovery/freewheeling? some other severe problems?)? Why not use two different cascoded FETs as a power switch?

  • \$\begingroup\$ If you leave the 600V FET on the entire time, you'll have your midpoint voltage across the 20V FET all of the time. \$\endgroup\$ – Stiddily Dec 23 '19 at 12:51
  • \$\begingroup\$ @Stiddily But it will turn off as soon as there is at least ~16V across drain-source of bottom fet \$\endgroup\$ – Sergio Dec 23 '19 at 13:06
  • \$\begingroup\$ "But it will turn off as soon as there is at least ~16V across drain-source of bottom fet" Are you certain about that? Because steady state approximations at large time scales aren't the same as transient behaviour as small timescales. Everything has a lag time. If that 20V NMOS turns off faster than the 600V can turn off, it will see the 600V for a short time. \$\endgroup\$ – DKNguyen Dec 23 '19 at 14:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.