I want to switch some 12V landscape light LEDs, totaling about 5 watts, using MOSFETs, with the control signal provided by a digital output line from an Arduino microcontroller running at 5V.

I thought I had some logic level N-Channel MOSFETs lying around, but apparently not.

What I have is several 400N80Z and IRLB8721PBF power MOSFETS. I tried using the 400N80Z in m circuit, but the LEDs don't light fully.

I have a variety of low to medium power NPN and PNP switching transistors on hand (BC33716, PN2222A, and 2N4401 NPNs, various others, plus a few TIP120 NPN power darlingtons.

Clearly logic MOSFETs are the right tool for the job and I need to order some. In the meantime, though, can I rig up a circuit that will fully saturate one of my power MOSFETs with a 5V CMOS logic signal from an Arduino?

I'm a software guy, and only passingly familiar with analog circuits.


3 Answers 3


The FCPF400N80Z has a maximum gate threshold voltage (VGS(th)) of 4.5V, which is marginal for logic operation, but the IRLB8721PbF has a maximum threshold voltage of 2.35V. This should work decently well for even 3.3V logic provided you don't need to pass more than about 3A of current.

  • \$\begingroup\$ So with the IRLB8721PbF should I be able to just feed a 5V logic signal into the gate and switch my 12 output? \$\endgroup\$
    – Duncan C
    Commented Aug 13, 2014 at 3:46
  • \$\begingroup\$ At 40A or more, yes. See figures 1 through 3 of the IR datasheet. And figure 8. \$\endgroup\$ Commented Aug 13, 2014 at 3:48
  • \$\begingroup\$ At 40A or more? What does that mean? I thought your original answer said I was limited to about 3A of load. I have trouble understanding what the info in the data sheet means. \$\endgroup\$
    – Duncan C
    Commented Aug 13, 2014 at 10:53
  • \$\begingroup\$ BTW, the IRLB8721PbF is working beautifully in my circuit. It lights the LEDs at full brightness and doesn't get warm at all. Thanks (voted and accepted.) \$\endgroup\$
    – Duncan C
    Commented Aug 13, 2014 at 10:54
  • \$\begingroup\$ @ Ignacio Vazquez-Abrams As link to FCPF400N80Z data,how is Vgs(th) defined? In data, VGS(th) Gate Threshold Voltage is rated 2.5 - 4.5 V. Why is the test condition set as VGS = VDS, ID = 1.1 mA ? How does it related to figure 1, figure 2 in data and actual circuit, in this case, says for example, driving two LED in series by 12V at 350mA? google.com/… \$\endgroup\$
    – EEd
    Commented Aug 13, 2014 at 11:54

When being bitten by the threshold voltage, an option is introducing a simple voltage amplification stage using a NPN BJT. You power your load from 12V, so you can use that voltage to pull the MOSFET's gate up to a level much higher than the logic level.

Basically what happens is the gate is pulled up to 12V unless the BJT is forcing the gate down to ground. The first stage inverts the behavior of the whole power switch, but that should be easy to solve in software of the controller.


simulate this circuit – Schematic created using CircuitLab

Advantage of this circuit is that it barely loads the controller output, only about 0.5mA is drawn from the output pin.

Disadvantage is that you have to route the +12V to the output stage, just for pulling up the gate. When you use a long lead from power supply to the buffer stage, you should decouple the power supply locally with a small (order magnitude) 100nF capacitor.

  • \$\begingroup\$ Cool. I figured I could do something like that. Isn't there a similar circuit using a PNP that would not invert the logic? \$\endgroup\$
    – Duncan C
    Commented Aug 13, 2014 at 10:56
  • \$\begingroup\$ I am using a separate 12V supply just for the LEDs. You're saying I would need to put a decoupling capacitor on the power to my BJT, so it's power doesn't "bobble" as the load switches on and off? And I gather that the specific BJT NPN I use doesn't matter much - any low power NPN will do? \$\endgroup\$
    – Duncan C
    Commented Aug 13, 2014 at 10:58
  • \$\begingroup\$ Any general purpose low power NPN BJT will do fine. You listed the BC337 in your question. The capacitor is to prevent current drawn by the switching transistors from influencing the MOSFET. \$\endgroup\$
    – jippie
    Commented Aug 13, 2014 at 16:24
  • \$\begingroup\$ The cap would go between the supply and ground, physically close to the collector of Q1, correct? \$\endgroup\$
    – Duncan C
    Commented Aug 14, 2014 at 10:51
  • \$\begingroup\$ @DuncanC yes, across the power rails, short leads, near the transistor. \$\endgroup\$
    – jippie
    Commented Aug 15, 2014 at 17:36

Since you already have it on hand, you may use TIP120 NPN power darlingtons which is rated at 5A as Adafruit TIP120 data sheet. Note, in data sheet max. 5A us 'infinite heat sink', refers to case temperature at 25 (or similar) degrees centigrade. Practical heat sink give about half current rating.

1 watt LED is about 300 to 350mA. For 5 watts, one transistors with good size heat sink is enough. Spread to two is better for added safety margin by keeping current to 1A. If individual LED is used, 2 or 3 LED (depends on color, voltage) can be connected in series, so the total current is kept low.

Connect Base with 300 ohms resistor to one of the digital output pin to drive about 10mA into the transistor. E to ground. C to LED, resistor (matching current rating of LED) to positive power. To reduce current, rise efficency by reducing heat loss in current limiting resistor, multiple LED can be connected in series (one LED is 2 to 4V, depends on color). Data sheet,figure 2, Vce Sat is 0.6 volt at 1A Ic, which is good for 2 to 3 LED in parallel (1W LED is 300 to 350mA).

As original poster mention 'right tool', transistor is suitable/usable (have not check price for this one to see if it is 'best solution') and have been working well for decades in application like this one, at a few watts. MOSFET has added advantage of small heat sink requirement, especially more significant in much higher power applications.

  • \$\begingroup\$ Mind the massive VCE(sat) though, which is usual for Darlington transistors. \$\endgroup\$ Commented Aug 13, 2014 at 4:17
  • \$\begingroup\$ Thanks, edited answer. data sheet figure 2, at Ic 2A Vce sat is 0.8V 1A 0.6V \$\endgroup\$
    – EEd
    Commented Aug 13, 2014 at 4:28
  • \$\begingroup\$ @IgnacioVazquez-Abrams Can you explain the "Mind the massive VCE(sat) comment for a non-EE? \$\endgroup\$
    – Duncan C
    Commented Aug 13, 2014 at 11:00
  • \$\begingroup\$ I got the IRLB8721PbF working quite well. It's threshold voltage is low enough that it does the job. I'd rather use a MOSFET with very low Drain-to-source ON resistance, so I don't need a heatsink. \$\endgroup\$
    – Duncan C
    Commented Aug 13, 2014 at 11:01
  • \$\begingroup\$ @DuncanC: Darlington pairs, due to their construction, add both the Vbe of the forward transistor and the Vce of the rear transistor in determining the Vce of the pair. \$\endgroup\$ Commented Aug 13, 2014 at 22:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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