Hello everyone.
I got a P775DM3-G that I have been troubleshooting. Didn't get anywhere, so I thought I should post here. Trying out my luck I suppose.
Background: The laptop itself came from CyberpowerPC. Previous owners are not tech-savvy, so it's most likely that the stock thermal pad configuration resulted in abysmal heatsink contact with the GPU core, VRM, and memory. [Picture]. Nevermind the heatpipes for now, they were bent after the failure during disassembly. Before failure, laptop was found to hard crash/shut down from benchmarks when plugged in, but it ran benchmarks just fine on battery. This was probably a thermal problem (more power available on AC -> things can run hotter, not so much on battery), not a mobo or charger one, though I couldn't really run any benchmark on it in its current state to confirm.
The problem: With GPU installed, laptop boots to black screen, nothing happens for 30 seconds, then backlight turns on, turn off briefly, and then on again, but nothing is displayed. Laptop's stuck at this stage until manually shut off. Keyboard lock keys and FN + 1 work as per normal (indicator lights toggled/fans hit max RPM). Keyboard shortcut to Sleep and Airplane Mode do sort of make the indicator light change. Disk-read activity indicator light is lit most of the time. USB flash drives are powered when plugged in. Nothing comes through the HDMI and mDP ports though. Nothing comes out of the laptop speaker either. Laptop draws 110W when powered with adapter. Normally battery charges at around 55W or so. Hot air do come out from both vents.
Reseating/removing RAM, reseating CPU/GPU, removing hard drives, clearing BIOS via CMOS battery removal (hold FN + D then power on), cleaning contacts did not not work. At this point I think only the GPU is dead while the 7700K and mobo are still alive.
[Card front]
[Card back]
[Card side]
I'm guessing top row + far off to side cluster are the core VRM components, and the ones near the MXM connector pads are for memory. Card doesn't look bent.
I went through the troubleshooting steps that @Khenglish suggested while I was trying to fix the 980M. I measured some resistances. The results are as follows:
_MXM PWR (big) Pad -- MXM GND Pad: Open circuit.
_MXM PWR Pad -- Power cable PWR pins: Open circuit.
_MXM 5V PAD -- MXM GND Pad: 16 kOhm.
_Both sides of all black capacitors/all MOSFET Vsw -- MXM GND Pad: 0.2 Ohm.
_Across all black capacitors: All except the cluster near the MXM connector measured 0.2 Ohm. The cluster itself measured 1.0 Ohm.
_Power cable PWR pins/all MOSFET Vin -- Power cable GND pins/MXM GND Pad: 14 kOhm.
I already contacted Khenglish for his opinions. He stated that Pascal cards typically read 0 Ohms across the core when off (consistent with the resistance measurements across the capacitors). The system does power on, so there is no core short. The card may have solder damage which progressed over time. He suggested that the core be removed and resoldered.
What's next: All that said and done, I guess I'm looking for advice about whether or not the card is worth fixing. The alternative would be selling the 1080 for parts and buy a new card + heatsink. Parting out the entire laptop would also be an option that I'm considering as well.
-
-
Dr. AMK likes this.
-
Flash the right BIOS and vBIOS with a programmer, and resoldering the GPU's (not reballing).
Broken GTX 1080
Discussion in 'Sager/Clevo Reviews & Owners' Lounges' started by Darker01, Sep 1, 2019.