Rui Neves

Distinguished
Sep 20, 2014
1,131
21
19,315
Hi ,

Sone months ago I had build a computer for a friend . Yesterday he told me that the conputer have shutted down from nothing he was oy watching one streamer on twitch , he noticed some fan very loud vefore the pc turned off by itself.

After that the pc turn on but dont give any image the monitor even says no signal and sometimes gpu fan spins very loud and sometimes not .

So I was working on that today . I notice that the gpu fan goes really loud and fast sometimes but not every time . I was able to boot to Windows two or three Times but after some minutes I've got the same problem.

Things that I tested :

Change gpu and with an older one that only have the power from pci express ( same problem )

Clear cmos and tried it again ( same issue )

Change ram to a different slot ( same issue )

Tested PSU with multimeter and seems that everything is fine .



I dont have another ram or psu to test the computer .

The mobo seems to have a led debug that stayed on above one thing labled " dram " can be the ram the issue ? Tbh is the first time that I see that debug thing :/ , i am new when it comes to repair pc's so there's somethings that are new to me xD .



Computer specs :

Msi a320m pro vh plus

Amd rysen 5 2600

Msi itx gtx 1060 3gb

Corsair vs550

G.skill aegis 8gb

Ssd Samsung 860 evo 500gb

Some non branded dvd drive


Ps : image of the motherboard debug led https://ibb.co/X87T0St
The debug led that says on is above the dram label but is the second led so it confuses me a bit since I dont know if it matches the gpu or it matches the ram label .

Removing the Ram make motherboard do a continuous long beep .

I managed to boot Windows some times today and the temps on cpu and gpu was fine .after some minutes it goes black anyway .
I tried to test the ram on memtest and tge computer get black screen and restart right way . I had to test it on safe mode on memtest and is doing the test now ... Have passed 16 minutes so far . I noticed so far that the seconds on the time takes a bit to refresh sometimes it passes from 15:00 to 15:02 for example.
2 pass with no errors until now

So I thinked what if I test my own GPU of my own build that is right here all the time and didin't test it on this conputer ...
For my suprise it worked , turns out my friend didin't have 1 bad gpu but 2 bad gpu's !

I made a stress test of 1,5h on aida 64 , no problems at all , the temps of cpu reach at 60°c , MB at 62°c +- , and mine GPU 75°c .
aida e even said that the last test was ended with the temps of cpu on 54°c , and test time was 1 minute when the conputer had sutdown with the bad GPU .
So much time guessing the problem and turns out that the two gpu's that he got was damaged. ...
 
Last edited:
The DRAM_LED indicates an error with RAM. Conduct a test of the RAM module using Memtest86 or try with different RAM. Google will tell you how to use it.
It could also be a heat related issue with CPU or GPU. Describe the cooling system?
Best is to conduct a stress test to check temperatures under load and at idle. Bios will indicate temperatures at idle. Aida64 for stress test and HWInfo64 for more precise readings.
 

Rui Neves

Distinguished
Sep 20, 2014
1,131
21
19,315
The DRAM_LED indicates an error with RAM. Conduct a test of the RAM module using Memtest86 or try with different RAM. Google will tell you how to use it.
It could also be a heat related issue with CPU or GPU. Describe the cooling system?
Best is to conduct a stress test to check temperatures under load and at idle. Bios will indicate temperatures at idle. Aida64 for stress test and HWInfo64 for more precise readings.
The temps on cpu is at 30°C on bios , I cant even past the post right now so it is impossible to do the ram test on memtest since dont have another pc to test ddr 4 ram's .
I did s stress test on aida when I had boot to Windows one of a few times and the computer had shut off again but the temps was at 40 °c on cpu and could't see the gpu but it was tested two different gpu's already ,if the two was bad it was really bad luck . So I dont think the gpu is the problem or cpu temps .

The cpu colling is a CM hyper 212x
X
Case is a Nox Hummer ZX
 

Rui Neves

Distinguished
Sep 20, 2014
1,131
21
19,315
UPDATE : I tried the gpu of my build on my friend computer .... Guess what it worked , 1,5h on aida 64 , no problems ....
So much time guessing the problem and turns out that the two gpu's that he got was damaged. ...