I'm going to use a Full Bridge Inverter for my application.

The problem is that the current which flows through the MOSFETs when they are ON can reach 70A RMS and the power dissipation is too high. In fact using a IPW65R019C7 MOSFET (home page, datasheet) one obtains:

Pd = Rdson*Ieff^2 = 0.019*50^2 = 47.5 W

I would like to reduce this power loss by using two MOSFETs in parallel. In this way the current is divided by 2 and one obtains:

Pd = Rdson*Ieff^2 = 0.019*25^2 = 11.8 W

The question is: Are these calculations correct? Do you know the rules and precautions to ensure that the current flows half in one MOSFET and half in the other one?

I thought to use an L6491 (home page, datasheet) as a MOSFET driver. Can I use just one MOSFET driver to drive both of them or do I need one driver for each MOSFET?

Thank you.

  • \$\begingroup\$ How fast must you switch these MOSFETS, to respect the Safe Operating Area self-destruction region? Regarding current sharing, can you afford to insert 10 milliOhms in each Source pin? That would be 20 squares of copper foil, at room temperature. \$\endgroup\$ – analogsystemsrf Sep 5 '18 at 11:38
  • \$\begingroup\$ @analogsystemsrf power MOS have (large) positive temperature coefficient on rds(on), it easily doubles over operating temperature range. No need for extra resistance \$\endgroup\$ – carloc Sep 5 '18 at 12:59
  • \$\begingroup\$ I'm going to work at 100kHz. \$\endgroup\$ – Fabio Sep 5 '18 at 13:27
  • \$\begingroup\$ There are various documents that you can find if you google "paralleling MOSFET". \$\endgroup\$ – Long Pham Sep 5 '18 at 14:41
  • 1
    \$\begingroup\$ @Fabio: "I would like to reduce this power losses making two MOSFETs in parallel. In this way the current is divided by 2 and one obtain ..=11.8 W." Note that what you have calculated is the power dissipation per transistor. I.e. total power dissipation is still 23.6W, because you have now 2 transistors. \$\endgroup\$ – Curd Sep 5 '18 at 15:01

The mosfet paralleling can be done, but it needs some prudence. At least you have to use a gate resistance for each mosfet. The driver is can be the same but each mosfet must have its own gate resistance.

The problem with mosfet paralleling is the gate oscillation which could cause undesired turn-on. The oscillation are due to the parasitic inductance in the layout and the use of a resistor for each mosfet damp the oscillation.


| improve this answer | |
  • 2
    \$\begingroup\$ Why does each have to have a gate resistor. I think people would like to know that. \$\endgroup\$ – Andy aka Sep 5 '18 at 18:14
  • 2
    \$\begingroup\$ answer edited, hope it would be a little bit more complete now. \$\endgroup\$ – RodezIO Sep 6 '18 at 8:10
  • \$\begingroup\$ @Andyaka More specifically, it is to stop MOSFETs gates from oscillating with each other. A single shared gate resistor damps oscillations between the driver and the MOSFETs, but the there are loops between the MOSFETs themselves that do not pass through a shared gate resistor which can give rise to other oscillations. This is show in Figure 3 of the app note. The resistors should also be place as close as possible to each MOSFET gate. \$\endgroup\$ – DKNguyen Nov 22 '19 at 21:30

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.