In order to speed up switching times and reduce power (I2R) losses, is it worthwhile to put a fast switching (but high RDSon) FET in parallel with a low RDSon (but slow witching and high Qgs) FET? That way the fast FET handles the switching transition (but only for a few tens of nS) so the high current FET can be soft switched on and off at low loss for the remainder of the switching cycle. The two FETs would have their own gate drive circuits. Does anyone think this is a practical way to switch tens of amps at MHz speeds or should I just bite the bullet and go with the more expensive GaN FETs?

I managed to get a PSpice model of one of the FETs I was thinking of using, if I can push ~6A of gate drive into it, I can switch it in <3nS, I don't need the 1/2nS switching times of GaN FETs nor the headache of dealing with the EMI problems that would cause. The switching losses at 3nS should be manageable provided I can keep \$R_{\theta JA}\$ to less than 10 \$^{\circ}\$C/W.

  • \$\begingroup\$ Good idea, but expensive... Additional driver, additional mosfet, more space. But how much losses do you have? Or do you use that serial resistor on gate? \$\endgroup\$ – Gregory Kornblum Apr 23 '16 at 10:34
  • 1
    \$\begingroup\$ IF the arrangement is DESIGNED and the design shows it will work and it is cost effective for you then it's OK. The trap is that it is easy to implement systems such as this and overlook interactions which make it either harder or of low profit for effort. ie design it well frst. || FWIW as much lower frequencies this can work well with a MOSFET and a high current relay - especialy so at DC where arc breaking is a major factor in current ratings. \$\endgroup\$ – Russell McMahon Apr 23 '16 at 11:21
  • 2
    \$\begingroup\$ At high frequencies, capacitive switching losses dominate in MOSFETs, so lower Rds(on) will not help. This question electronics.stackexchange.com/questions/185721/… (specific to buck regulators but still appropriate) has the equations for losses in that application. \$\endgroup\$ – Peter Smith Apr 23 '16 at 13:09
  • \$\begingroup\$ Interesting question, but I think hard to answer without your mentioning how many tens of amps, what will V_DS be, and which specific MOSFETs and driver you have in mind. My feeling is that it won't be worthwhile (or at least not worth the effort) in most cases, otherwise everyone would be doing it maybe special MOSFETs might exist to fill that niche. Perhaps your application is special in some way and it would be helpful, although based on your question it isn't possible to tell. \$\endgroup\$ – Oleksandr R. Apr 23 '16 at 18:18
  • \$\begingroup\$ Thanks everyone. @PeterSmith I'd totally forgotten about the capacitive loss. That's going to be a real pain to deal with. \$\endgroup\$ – Sam Apr 23 '16 at 22:23

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.