I'm trying to control 12 V or 24 V LED strip (consumption around 5A) with an STM microcontroller. The outputs of the microcontroller are 3.3V PWM.

I'm looking to use the following N-channel MOSFET: PSMN2R8-80BS-1 or FDS6910

I'm a bit worried if the 3.3 V will turn on the MOSFET enough. Because of that, I was thinking to use NPN transistor which will opening the MOSFET with 5 V.

Can you share your opinion which MOSFET to use and if it will work without an additional NPN transistor.

Thank you for your time.


2 Answers 2


You can forget about the PSMN2R8-80BS without an additional stage, at 3V3 it is high impedance (see Fig. 9).

The FDS on the other hand looks promising, it will be in the range of 20 to 30 m\$\Omega\$ (see Fig. 2). At a current of 5 A, the voltage drop will be below 0.15 V.

enter image description here

  • \$\begingroup\$ Thank you for your advice. I will use it. But will be the MOSFET open enough to drive aprox. 5A? Thanks \$\endgroup\$ Mar 22, 2021 at 8:27
  • \$\begingroup\$ With 30mOhm on-resistance the voltage drop at 5A will be 0.15V, with 20mOhm it would be 0.1V. This results in a power dissipation in the FET between 0.5W and 0.75W. \$\endgroup\$
    – jusaca
    Mar 22, 2021 at 8:33
  • \$\begingroup\$ I'm not too keen of driving a MOSFET on the slope of the RDSon chart. I'd use at least 4V for that one, given the spec dispersion. However there is an output curve for the 3.5V, you'll have about 0.2V on the channel in that condition. About 1W dissipation. The SOA for DC is right at the limit and also if he is doing PWM he'll have conduction losses (and a lot of these given the typical GPIO drive strength). IMHO it will run very hot \$\endgroup\$ Mar 22, 2021 at 8:37
  • \$\begingroup\$ It looks more like 0.1V to me in Fig.1 for 3.5V. I assume it would work OK(-ish). But I agree, if this will be for more than a single unit and hobby use, it might be a better choice to get a gate driver. \$\endgroup\$
    – jusaca
    Mar 22, 2021 at 8:49

The concern is correct, since just passing the VGS threshold will not give the MOSFET its full current capability. Also, please refrain to say "open the MOSFET" since it's not clear: in hydraulics open valves pass fluid, in electronics an open MOSFET stops the current. It's better to say "put in conduction". And stay away from the "saturation" word used for BJTs since FETs use it for another thing.

That said, there are MOSFET specified for a gate drive as low as 1.8V, last time I checked. Your FDS is a "logic level" MOSFET, while the PSM is a "standard level" MOSFET. Aside from the name you can see the type looking at the voltages the RDSon is specified:

  • Standard level have it specified at 10V, and are usually driven at 12V
  • Logic level have it specified at 4.5V and are usually driven at 5V

Logic level MOSFETs often have both specified since usually they can handle more current when driven at an higher gate voltage (this is a rule for MOSFET: more gate, more current, until it breaks the oxide!)

Also, depending on the PWM frequency you'll have some non-trivial current floating into the gate: at each cycle the gate is charged and discharged since it is, in fact, a capacitor. The relevant info are in the gate charge section of the datasheet and are useful to compute switching losses (but that's a slightly more advanced topic).

In the meantime I suggest you to use a dedicated gate driver IC. You can actually do one with just two transistor but if you can use an IC is easier and more reliable. Almost every IC maker in the world makes them. If you have 12V available from the supply just use it to power the driver and use either MOSFET. Otherwise you'll need to find a 5V supply somewhere and use a driver rated for 5V operation (and you can't use your PSM part)


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.