Announcement

Collapse
No announcement yet.

Problem MSI 17631 Core power

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

    Problem MSI 17631 Core power

    HI all.
    Got here a GT70 MSI with MSI 17631 v1.1 motherboard. It was already send to MSI under warranty, came back in good order, but broke down after 1 year use, same issue. Dead...
    This time it was out of warranty so i had a look what could cause this problem. Found shortage of the 19 volt rail, caused by PQ11 mosfet in one of the 3 phases of the VCC core power. Replaced it, no shortage anymore, HDD flashes short. Charged the battery again. That's all. The CPU core power comes up 30 ms, with peak level of 1.7 volt. That's a way to much, so i removed the Hashwell to continue. Replaced also the other mosfets PQ8 and PQ9 and again PQ11. No change..
    If i remove the Choke7 from PQ8 the Core VCC will be 0.8 volt (for 30ms). Same to Chose 8 from PQ9 (and connected Choke 7 again).
    Measured the voltages on the ISL95812. 5volt is present, 19 volt is present.
    I see a startpulse on VR_ON. (30ms). VR_HOT will give no pulse.

    Attached the schematics of the corepower. And also a drawing of voltages made by some other guy on this forum. The difference between his values and mine are he got around 1.3 volt output VCC, this board gives 1.7 volt.

    As the processor is removed the ISL isnt programmed to the selected voltages, and should be VID=0, thus 0.8 volt in this configuration.
    Also i made a dummyload by a length of copperwire, but the voltage remains the same: short pulse of 1.7 max, 30ms long.

    Could this be the ISL which is broken, or any other ideas what could cause this problem?
    Attached Files

    #2
    Re: Problem MSI 17631 Core power

    How much resistance (to GND) is on VCC_Core with and without CPU?

    Comment


      #3
      Re: Problem MSI 17631 Core power

      1.7V is perfectly fine for Haswell.
      Also as usual when a MOSFET dies you should replace the MOSFET driver/buck controller at the same time.
      Also it's common for Haswell and newer to die when high-side MOSFET shorts.
      OpenBoardView — https://github.com/OpenBoardView/OpenBoardView

      Comment


        #4
        Re: Problem MSI 17631 Core power

        @caspian. Without Hashwell no resistance at least >100K. With Hashwell i will mesure but low
        @Piernov. Your sure that 1.7 should be correct? Note that i removed the Hashwell, and still the CPU corepower switches off after 30ms. Also with a dummyload (10 meters of wire) the power comes up for 30ms and switches off . Is that behaviour normal without Hashwell?

        Comment


          #5
          Re: Problem MSI 17631 Core power

          Noticed something: the EC_ALLSYSPG signal (start/enable the ISL) is only 30ms pulse i measured. According the sheet this should be much longer. Will investigate this first why this is so short. It's coming from a AND gate which combines the 1.3 Dimm, 1.05RUN and 1.5RUN signals. 1.05V Run i have seen. Will check the others. (Dimm is also removed as well the graphic board)

          Comment


            #6
            Re: Problem MSI 17631 Core power

            How much is the low resistance exactly?
            It is possible that this mainboard has another problem too at another voltage/signal.

            Comment


              #7
              Re: Problem MSI 17631 Core power

              i will measure the resisance tomorrow and let you know. But the mainboard should powerup the Corepower without CPU yes? Not only the 30ms pulse i measured but much longer?


              Now searching for the AND gate 's location to measure why this only gives a short 30ms pulse
              Could be that some other voltage is bad or missing

              Comment


                #8
                Re: Problem MSI 17631 Core power

                The mainboard does not power up the core power without CPU.
                It is possible that some other voltage/signal be missing too.
                It is also possible that this CPU be defective. If you test another CPU on this mainboard, it is possible that the new CPU to get damaged.

                Comment


                  #9
                  Re: Problem MSI 17631 Core power

                  @caspian
                  measured the R when the Hashwell is inserted: my meter is not good enough but something between 200milliohms to 300. (0.2- 0.3 ohm).

                  The issue with the startsignal: As soon the ISL turns down the Power_good signal, all voltages are going done. This works as a loop. Start signal starts the VCC core. When the Powergood from the ISL is going down again, whole powermanagement is switched off.
                  This can easily checked, cause there is a testpoint on U9 AND gate. Ground the testpoint--> Core power will not be enabled--> all voltages remain on--->the startbutton is lit!.

                  Thus conclusion: ISL is defective. Need to replace it. Not a easy job. Dont know the Hashwell is still good. Dont got another Hashwell to test. And buying a new one is too expensive for this machine.

                  Comment


                    #10
                    Re: Problem MSI 17631 Core power

                    You can at the very least get a Celeron 2950M for less than $10 for testing.
                    OpenBoardView — https://github.com/OpenBoardView/OpenBoardView

                    Comment


                      #11
                      Re: Problem MSI 17631 Core power

                      @piernov. Good idea. First i replace the ISL, as the ISL datasheet tells me that a unprogrammed ISL a VID should have of 0, thus 0.8 volt. As there is no processor the voltage should not be 1.7 volts.
                      I got here a good working HP Omen notebook with a Intel Core i7-7700HQ.
                      In the MSI i got a a i7 with i7-4810MQ. Are those 2 interchangeable?

                      Comment


                        #12
                        Re: Problem MSI 17631 Core power

                        i think you are right Piernov.
                        Ordered a Celereon Celeron 2950M for test. The i7 might be gone indeed
                        Thanks for the help

                        Comment


                          #13
                          Re: Problem MSI 17631 Core power

                          Gt70 dominator is working again! With Celeron processor. VCC core is now 0.6-0.8 volts measured with CPU-Z. Without CPU you can expect 1.7 volt.
                          Thanks again for all tips!
                          (Found a new Hashwell processor on Ebay. This will come coming week.)

                          Comment

                          Working...
                          X