Troubleshooting pc switching off

[ErgoPandemonium]

Good morning/afternoon/evening to you all

 

In the last week my pc has shown me a bunch of wonderful new way to crashes, see attached images. Those crashes see to be due to nvidia drivers, but that seems to be fixed after a format. Along with those crashes my pc has switched off with no warning. In the one case I was just reading something on my browser and on the other I was playing Supreme commander, spanning 2 screens.

 

Current configuration:

• CPU: i5 2500 OC to 4.0GHz (been running like this for 2 years)
• RAM: 2 x 8 GB Corsair vengeance CL9 1600MHz + 2 x 4 GB Adata CL9 1600MHz (Not the best combination, but it has been working)
• GPU: MSI 1050ti 4GB (No 6pin connector)
• SSD: 128GB Samsung 840evo
• PSU: Corsair TX 750
• Motherboard: GA-P67A-UD3 (http://www.gigabyte.com/products/product-page.aspx?pid=3698&dl=1#ov)
• Windows 10 Pro

What I have tried:

1. Format windows
   1. Got rid of the coloured crash screen
2. Removed all oc and xmp.
   1. Ran valley benchmark 3 times and was stable
   2. Evaluated Event viewer and found no errors during the benchmark
3. Restored CPU OC
   1. Ran LinX with a 16GB problem set along with Kombuster's frame buffer filling stress test
      1. PC ran at 100% cpu load @4GHz at 79C
      2. Ram was filled to 20/20 GB
      3. GPU used the full 4GB frame buffer
      4. Test period 20min
      5. System was stable

2017.02.04

1. Removed the 1050ti and installed my old 660
   1. PC powered down when starting supreme commander FA
   2. Struggling to replicate the behavior

 

2017.02.05 May today be the end of this crap

1. Stripped the pc in its entirety and took a few hours to clean and inspect every nook and cranny. All parts were photographed during the process.
   1. The motherboard
      1. Inspected the are under the slots to remove the dust that might have settled in ther
         1. Found a little dust bunny lodged under my PCI-E lane between the pins
      2. Inspected all slots for dust partials and damaged pins
      3. Inspected all capacitors for ovality, buldges, leakage and superficial damage
      4. Inspected all power connectors for damage or burn spots
      5. Clleaned and replaced the thermal paste as the paste was more of a putty than a goo.
   2. PSU: The warranty was up so decided to remove the lid and see if all was still well in there
      1. Cleaned the fan
      2. Cleaned 6years worth of dust
      3. Inspected the capacitors for signs of damage
      4. Resembled the psu and tested it before using it again
   3. Removed some redundant case lighting
   4. Redid cable management to remove any possible pinch points
2. Testing
   1. Ran Aida 64 for 30min passed
   2. Included a picture of a shorter run after the temps had stabilized

 

2017.02.13

1. The Pc randomly powered down again.
   1. As a next step I was investigating my psu's different voltages (3.3V, 5V and 12V)  for deviations exceeding 5%.
   2. For the evaluation I used a multi-meter rather than software.
   3. Measurements were taken at molex and sata connectors
   4. All the voltages are within tolerance and did not shift regardless of idle or load conditions.
       
2. I think my initial gpu crash were due to windows driver issues. After allowing windows to do all its updates, the problem went away.

[LogicalDrm]

It looks like GPU is toasted. Try with iGPU or another card to verify.

[marshmallowbam]

reseat your RAM and make sure all of your power cords are connected properly. Mine would randomly switch off on me too, then i relized that I had a new 4pin cpu connector above my northbridge (it was my first gaming board cut me some slack :P) As for the discoloration... I could be an issue with your RAM try reseating it, if that does not work then try running just with the 2 x 8 solution if the problem still occurs try testing each ram stick individually. either the problem with clear up if it is the ram, but if it is not. Then it is your GPU and youre going to have to contact who you bought it from and get a new one.