Mosfet Smoking within operating limits

Discussion in 'The Projects Forum' started by urb-nurd, Jan 7, 2015.

  1. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    Last edited: Jan 10, 2015
  2. wayneh

    Expert

    Sep 9, 2010
    12,127
    3,049
    You're using an N-channel as a high side switch. Move it to the low side, source to ground. When you reverse it, you send current through the body diode no matter the state of the gate.
     
  3. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    I see - so source to ground and drain to Vcc?
    This stops latch up - but i still get smoking when i drive the gate at 10V
     
  4. BobTPH

    Active Member

    Jun 5, 2013
    782
    114
    Put the load between +24V and the drain.

    Bob
     
  5. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    http://forum.allaboutcircuits.com/attachments/tran21-gif.25142/

    Like this?

    That does work, however i canont seem to make my PWMed optocoupler short the gate voltage to ground.
    If i move the jumper from 10V to ground manually, the mosfet switches on or off.
    With my 500Hz pwm output shorting the gate voltage to ground via an optocoupler - i do not get any control.

    Thank for the assistance! I have made progress!
     
  6. crutschow

    Expert

    Mar 14, 2008
    13,018
    3,235
    We're not clairvoyant.
    Post your schematic.
     
    simo_x likes this.
  7. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    i do apologize!
    Let me get on it!
     
  8. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    [​IMG]
    This awful paint rendition shows what's going on.
    Also - the 25 pulsed dc produces around 34V peak to which the cap charges.
    This dc level is maintained at the top of the v divider for driving the gate.
     
    Last edited: Jan 7, 2015
  9. crutschow

    Expert

    Mar 14, 2008
    13,018
    3,235
    The 4k7Ω resistor provides over 7mA of current from the 34V supply. Can the opto carry that much current to ground?
     
  10. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    yes, the max output is 500mA
    Thanks
     
  11. ScottWang

    Moderator

    Aug 23, 2012
    4,855
    767
    Maybe you should cut off the wire from the common of 4.7K and 2.2K to G of mosfet, you just connecting the wire from the common of 4.7K and 2.2K to the optocouple, and the pwm signal connecting to the G of mosfet.

    There is a question here is that if the vg must >= 10V, then the pwm input also need to increasing the voltage up 10V.

    The current from 4.7K and 2.2K maybe too less for the optocouple, you need to reducing the values of resistor.
     
  12. RamaD

    Active Member

    Dec 4, 2009
    254
    33
    The opto needs a good amount of current for the input diode, to make the output transistor conduct and goto saturation. What is the opto used and how much is the current to the diode? Schematics again!
     
  13. crutschow

    Expert

    Mar 14, 2008
    13,018
    3,235
    But how much input current does that require?
     
    simo_x likes this.
  14. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    The input for the opto is configured correctly. I have already used this configuration to PWM the gate of a IGBT with success.

    As in how much is dropped across the opto output? I am not sure.

    However - i havent got to the point of using the PWM to control the gate as i cant manage to get it working with 10V straight up.
     
  15. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    Your first point relates to passing the gate voltage with the optocoupler rather than sinking it? I did try this but the datasheet shows it being sunk as an example.

    The aim of the PWM is to drive the gate to 10V and then to zero rapidly - rather than to average out a voltage over a time period.

    I think you may be on to something with the gate current !
     
  16. ScottWang

    Moderator

    Aug 23, 2012
    4,855
    767
    I attached the circuit as below, you can check where should to modify.
    24V625mA150W_mosfet_urb-nurd_ScottWang.gif
     
  17. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    First - let me clarify i can understand the circuit - given that it is more than a slight alteration.

    The zener and R1 provide a 12V voltage for the gate - with the capacitor in parallel holding that level.
    R2 limits the current into the opto.
    The PWM output of my microcontroller will short the input to the opto to ground when high via a transistor Q1.
    The 12V gate voltage is passed by the opto when the PWM output is low.
    22k Gate resistor to stop ringing.

    Now i am afraid i have a few questions...
    Why the zener rather than a Vdivider? (i also do not have any zener diodes)
    Why is the additional transistor necessary? The optocoupler i am using (4n25) seems to operate just fine with my 5V output from my microcontroller - and worked in this manner for my high voltage bulb testing.

    I am still struggling to get my mosfets to conduct reliably with a straight 12V applied to the gate - i am not at the point of using the opto and PWM for rapid switching.
     
  18. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    [​IMG]
    This is what i am using at the moment.
    I am simply trying to get the 10+V at the gate to switch the mosfet so that pulsed DC can flow through the bulb.
    EDIT: fixed an issue - with no gate voltage - the mosfet smoked and the bulb was very dim.
    I added a 700K resistor from the gate to ground and now it works as expected when applying the gate voltage and removing it!

    Now to try and PWM through the opto!

    If anyone has an idea of what is going on without that gate resistor to ground - im intrigued to find out.
     
    Last edited: Jan 8, 2015
  19. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    One issue sorted - the optocoupler does not vary from zero to full.
    When the PWM output is at 100% (shorting the gate to ground, therefore should be zero V) i can see that it is only managing to drop to half its max before rising again.

    This has to do with the PWM frequency and the optocoupler maybe.
     
  20. ScottWang

    Moderator

    Aug 23, 2012
    4,855
    767
    If you like then you can still using the resistor.

    I just drawn the normal driver circuit, If you want then you could take the bjt away, and using the port of uC to drive the LED.
     
Loading...