Help diagnosing a (possibly) dead GTX 1080 Notebook GPU

Discussion in 'Hardware Components and Aftermarket Upgrades' started by PrimeTimeAction, Feb 13, 2021.

  1. PrimeTimeAction

    PrimeTimeAction Notebook Evangelist

    Likes Received:
    Trophy Points:
    What is the problem:
    When I press the power button, the button and the lights around it turn on. After few seconds all the three fans start to blow on max. this continues for 30-40 seconds. Then the laptop shuts of. The screen stays dead black during all this time. The green light on the adapter stays on all this time.

    How it happened the first time:
    I was about an hour into a gaming session and suddenly the laptop turned off. I noticed that the green light on the power brick had turned off. I connected the power to a another socket. green light came back on. But as soon as I turned the laptop on, the light went off. The laptop didn't start. The laptop and the brick were very hot. I let these cool down and tried again after two hours. This time when i pressed the power button, the issue I descried in "what is the problem" started happening.

    What i have checked so far:
    • I used the multimeter to check for shorts in the power connector at the laptop. No shorting was found.
    • I did the usual removing battery, resetting the cmos (removing cmos battery), reseating ram, moving ram slots, using a known working ram. reseating hard drives, removing hard drives et.c. Nothing changed the behavior.
    • I tried moving the GPU to the unused slave MXM slot. The issue remained same.
    • I tried turning the laptop on without the heatsinks. During the time when its trying to turn on, the CPU became really hot whereas the GPU remain dead cold. This makes me think that the CPU is ok.
    • I checked the voltage from the powered brick, it is sending 19.4 volts. So this seems to be ok.
    • On the graphics card's MXM pins , I checked for shorts across the PRW_SRC pins and GND pins. There was no short.
    • In the MXM slot of the motherboard, i checked the voltage (while turning the power on) between the PRW_SRC pins and GND pins and it was same 19.4V. The voltage across the GND pin and the 5V pins were 4.8v. Considering margin of error in my multimer. this seems that the slot is sending the correct voltages.
    • On the GPUs, i have tried poking the multimeter at different position, no obvious issue were found,. however i can defiantly use some guidance here as what exactly to check. I have seen some videos on youtube of people diagnosing older desktop GPUs by checking for shortage from ground to different components. But I cannot really use it to test my 1080N because even a no short would result in very small resistance like 0.5ohm so i dont know what the expected values are. Is there any way i can check the vrams or mosfets? Dose no a short between the PRW_SRC pins and GND pins mean that i don't even need to check the mosfets or vram?
    I want to narrow the issue down to faulty part . first at thne main component i,e is it is the motherboard or the GPU or something completely different like fan cable e.t.c. And if it is the GPU then it will be really good to if i can identify the defective part within that component.

    @Darker01 @Khenglish
    Last edited: Feb 13, 2021
  2. Khenglish

    Khenglish Notebook Deity

    Likes Received:
    Trophy Points:
    I'm sorry I didn't respond to this earlier.

    The 19V supply on the MXM slot and the power cable are not the same. The power cable provides the majority of the power. Check the resistance between the 4 pins on the left on the card's power cable jack (19V) and the right side (GND).

    GPU is ok, but never power a CPU without a heatsink. I accidentally did this once for about 5 seconds and it damaged a 3920xm (iGPU would error 43 with Intel driver installed).

Share This Page