MOSFET Problem...

Discussion in 'The Projects Forum' started by dksoba, May 2, 2010.

  1. dksoba

    Thread Starter New Member

    Jan 8, 2010
    25
    0
    Hi, I'm using this MOSFET.

    http://www.vishay.com/docs/68998/si4174dy.pdf

    My circuit is as follows:

    Drain: +12V
    Source: [MOSFET] -> [8 ohm resistor] -> [Ground]
    Gate: +5V

    My power supply is supposed to be rated to 22A, but that resistor should only be pulling 1.5A. Here is the problem, when I apply +5V to the gate of the MOSFET (I measure about 4.93V on the gate, relative to ground), the drop over the 8 ohm resistor is only 3V. If I put a 1k ohm resistor, I get the same thing. I'm not sure what's going on! Shouldn't the voltage over the resistor be nearly 12V+?

    A picture of my circuit is attached.

    Please advise,
    Thanks,
    Matt
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    The problem is that the resistance on the source terminal causes Vgs to drop to the threshold of the MOSFET.

    Then your MOSFET dissipates a lot of power.

    If you want to limit current, then put the resistor between +V and the load. Use the MOSFET to sink current from the load. If you want to know what kind of current the load is drawing, use a small resistor (0.1 Ohms to 0.5 Ohms) between the drain terminal of the MOSFET and ground.
     
  3. dksoba

    Thread Starter New Member

    Jan 8, 2010
    25
    0
    Damn. I completely understand. I read the source and it's +3V, and my gate is at +5V, then my Vgs is +2V, not +5V...

    I'm not sure what you mean by limit current... I need each MOSFET to source 1A to the resistor. It will be controlled by PWM. The final design will have +24V power supply, and 8 ohm resistors, so 1 amp should be feasible.

    So I should have, as far as drain and source go, this order:

    [V+] -> [Load] -> [MOSFET Drain] -> [MOSFET] -> [MOSFET Source] -> [GND], that way Vgs is always referenced to GND, not the voltage over a resistor.

    So the only way to fix my circuit, I believe, is to glue my MOSFETS upside down, then solder wires from it to the PCB... :(

    There will be another iteration eventually.

    Matt
     
  4. dksoba

    Thread Starter New Member

    Jan 8, 2010
    25
    0
    Forgot to ask you, SgtWookie, know of any better ways to fix this mistake?

    Matt
     
  5. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    If for some reason you want the resistor to be between the source and GND, then you will need a MOSFET gate driver IC to drive the MOSFET.
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You can't turn those MOSFETs upside-down; they are N-channel and have a body diode.

    Since you have your load on the source terminal, you need a high-side driver that will reference your gate voltage to the source voltage instead of gate to board ground.

    If you went to a P-channel MOSFET, you would then need a driver circuit that can reach your 24v V+ to turn the gate off, and from 24v-5v to 24v-10v to turn it on.

    Without seeing your schematic and board layout, it's hard to say where you could splice in such fixes.
     
  7. rjenkins

    AAC Fanatic!

    Nov 6, 2005
    1,015
    69
    I think he means to attach the FETs to the PCB 'dead bug' style, so he can get around the PCB trace that hard-wires the positive supply to the drain.

    It's then possible to add link wires to configure the FETs with grounded source and load in the drain circuit.
     
  8. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    No, he can't "dead bug" them. The source and drain traces provide the only cooling that the MOSFET gets. Trying to dead bug them will result in little blobs of burnt plastic.

    Might be able to make daughterboards for the MOSFETs and high-side driver IC's, but it would add a lot of extra fooling around and extra connections to make. The best way to go is to just re-design the board with the proper high-side drivers.
     
  9. retched

    AAC Fanatic!

    Dec 5, 2009
    5,201
    312
    I agree with the re-design. but, the daughter board approach will be helpful.

    You can add pins to your existing traces, and route them to the daughter board. This is an easy way to provide access to the fets, and thermally isolate them from any other sensitive components.
     
  10. John Luciani

    Active Member

    Apr 3, 2007
    477
    0
    For testing you may be able to take a TO-220 and solder the pins to the
    existing pads. You could add a heatsink if necessary.

    I am not sure how much power you are dissipating statically or switching
    but there does not seem to be a lot of copper under the FETs. For SMD power
    devices the cooling comes from the surface area of the drain pad. There is a
    fair amount of area required which varies with the copper thickness.
    For the device you are using the Rja is 50degC/W (max) for 1in^2 of
    copper.

    (* jcl *)
     
    Last edited: May 3, 2010
Loading...