I have the below circuit and when I turn off the MOSFET (Logic 0 at gate vg) the LEDs (LED1 and LED2) still shine with a small intensity. When I replace the LED with an incandescent lamp (car lamp) the lamp does not shine at all. Does anyone know why?

Ciruit 1:

Ciruit 1

Ciruit 2:

Ciruit 2

MCH6445-TL-E -> Q1 Mosfet's datasheet: http://www.mouser.com/ds/2/308/MCH6445-D-108455.pdf

MMBF170LT1 -> Q3 Mosfet's datasheet: http://www.onsemi.com/pub_link/Collateral/MMBF170LT1-D.PDF

Final circuit which works fine: enter image description here

  • \$\begingroup\$ An obvious answer for why the LEDs are still slightly on is Q1 is not off "enough", or all the time. Have you got an oscilloscope to look at the voltage across the two LEDs? \$\endgroup\$
    – gbulmer
    Sep 13, 2014 at 18:02
  • 2
    \$\begingroup\$ @Passerby - I agree with the spirit of your answer, Incandescent power >> LED power to see anything. My experience is LEDs, e.g. 5mm, are visible under 0.5 mA (students sometimes use 10kΩ resistors for current limiting, as orange and red stripes are easy to misread), and duty cycles below 50%. \$\endgroup\$
    – gbulmer
    Sep 13, 2014 at 18:19
  • 3
    \$\begingroup\$ I don't see any current limiting for the LEDs. If you did manage to turn the FET on you would destroy them. Maybe your FETs aren't logic level FETs, so there's not enough voltage on the gate to turn them on. Check the Vth spec on the datasheet. More likely the FET is damaged and the LEDs stay on because of excessive leakage. That would also explain why the lamp doesn't turn on at all. \$\endgroup\$
    – John D
    Sep 13, 2014 at 18:23
  • 1
    \$\begingroup\$ Is it actually at logic 0 or is it high impedance? If it was high impedance, it could take the gate a while to discharge through the 240k resistor. \$\endgroup\$
    – horta
    Sep 13, 2014 at 20:13
  • 1
    \$\begingroup\$ The mosfet is the MCH6445-TL-E. I isolated the gate from the microcontroller and I connected directly to the ground also, I replaced the mosfet with a new but it still has a little shine. \$\endgroup\$
    – DarkKnight
    Sep 13, 2014 at 22:41

2 Answers 2


Looking at your MOSFET's datasheet, the Gate-Source leakage current is 10uA at Vgs=16V. You have almost 12V at the gate of Q3. That's enough current to barely turn on the diodes.

It doesn't look like Q1 would be a culprit. Datasheet says it's only sinking 1uA with 60V Vds. Of course if Q1 is damaged all bets are off. Echoing one of the comments above, you haven't added anything to limit the current through the diode. The on resistance of Q1 is in the tens of milliohms, and it's not rated to take more than 4A. So unless your 12V source is current limited, it's likely that you've damaged either Q1 or the LEDs or both.

The reason you don't see anything in an incandescent bulb is that incandescents take a lot more power to show that they're on. They're about 5% efficient at full power, even less so when dimmed.

  • \$\begingroup\$ The LEDs are car LED lamps and have internal resistors for limit the current. Also, you have right, I remove completely the R7 and LEDs are not shine anymore when Q1 is off. I put an 100K resistor in R7 and then the light idensity was decreased but again LEds shine little and then I put a 200K resistor but the Q3(MMBF170LT1) was not enabled. I will try to put an operational amplifier "OPA172IDBVT" in buffer mode between Q1 and R7 and I will inform you for the results. Thanks all of you very much for your replies. \$\endgroup\$
    – DarkKnight
    Sep 18, 2014 at 14:14
  • \$\begingroup\$ Another option would be to find a more appropriate MOSFET. I'm sure you can find something with lower gate leakage. \$\endgroup\$
    – Dave
    Sep 23, 2014 at 23:45
  • \$\begingroup\$ Finally, I solved the problem with an operational amplifier in buffer mode that as you can see in the third picture in my first post. Thanks all of you for the replies. \$\endgroup\$
    – DarkKnight
    Oct 10, 2014 at 22:51

I would like to point out two things:

  • Add a current limiting resistor in series with the two leds that are in parallel
  • Double check the use of the C3 capacitor because when Q3 goes to ON, C3 will see GND and in that moment C3 is already charged to 5V. So, you are discharging that capacitor too fast and that could be a problem after a while of being using the circuit.

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.