Announcement

Collapse
No announcement yet.

VRM shorted, low side shorted ?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    VRM shorted, low side shorted ?

    Hello,

    my motherboard, a Gigabyte b660m-DS3H, broke last night, and I'm trying to fix it.
    The symptoms are as follows:
    - The PC doesn't boot when the CPU power supply is connected (without CPU).
    - When voltage is applied to the CPU connection, the three phase 2 MOSFETs start to heat up. The other MOSFETs do not heat up.

    The phases are made up of one high-side and two low-side MOSFETs.
    My hypothesis is that if the two Low MOSFETs are hot, they must be permanently ON, even when the High is ON, causing a short-circuit.

    Do you know how I can test the driver? Do you have any advice for a case like mine?

    #2
    Test each mosfet. Remove all power. Meter in resistance mode.

    Measure the resistance across:

    source (1-2-3) & drain (5-6-7-8)
    source (1-2-3) & gate (4)
    gate (4) & drain (5-6-7-8)


    Post each measurement. A low resistance often means the mosfet is defective.

    Comment


      #3
      Thank you for your answer

      Values are formatted like this : Mosfet number : [Source->Drain resistance] ; [Source->Gate resistance] ; [Gate->Drain resistance]


      Phase 1:
      1 : [0];[2.9];[2.9]
      2 : [226];[209];[26]
      3 : [225];[210];[27.5]

      Phase 2:
      1 : [0];[6.7k];[6.7k]
      2 : [220];[OL];[OL]
      3 : [220];[OL];[OL]

      Phase 3:
      1 : [0];[6.7k];[6.7k]
      2 : [220];[OL];[OL]
      3 : [220];[OL];[OL]

      Phase 4:
      1 : [0];[6.7k];[6.7k]
      2 : [220];[OL];[OL]
      3 : [220];[OL];[OL]


      The power mosfet in phase 1 is dead ?

      Comment


        #4
        Each is a suspect because a good mosfet should be hundreds of K ohms or higher.

        Suggest to flux and remove the first mosfet, allow it to cool on your work table. Then measure the part again while it is on your bench. If the measurements remain this low, yes, it is defective. Repeat with the others as 6k7 ohms is too low for a good mosfet.

        Comment


          #5
          I found two mosfets short-circuited.
          The others are OK.
          However, when I measure directly on the motherboard, without the mosfets, between the gate and the source, I find 15k ohm (the value rises slowly).
          Attached Files

          Comment


            #6
            Do replace the shorted mosfets and test again the resistance to ground. If higher than the earlier inspection, test the motherboard again with power.

            Comment


              #7
              Thanks ! I will do that.
              I have a some dead motherboards and graphic cards, i can maybe find some compatible mosfets.

              Can i use this QM3056M in replacement for NTMFS4C10N ?
              (QM3056M have lower Rds(on), same Vdss, Vgs and bigger Idmax)
              Attached Files

              Comment


                #8
                Yes.

                Comment


                  #9
                  No. All phases should use the same MOSFETs so that they stay as balanced as possible. Also, QM3056M has much worse switching characteristics, much harder to drive and much slower to switch on/off which can make it less efficient.

                  Also as always the MOSFET driver has to be replaced at the same time as the MOSFETs.
                  OpenBoardView — https://github.com/OpenBoardView/OpenBoardView

                  Comment

                  Working...
                  X