Computer crashing while under heavy stress

juomm

Reputable
Jul 31, 2015
4
0
4,510
Hey all. I recently upgraded some parts of my PC and since then I've only had problems. I bought and MSI R9 390 and a new PSU to support it: an Evga Supernova 650 GS which seemed to be more than enough for it. With that my specs are:

Motherboard: Asrock z68 extreme4 Gen 3 (latest BIOS installed)
CPU: Intel i5 2500 3.3 Ghz
Memory: 8 GB Corsair DD3 1600Hz (2x4 GB)
GPU: MSI R9 390
PSU: Evga Supernova 650GS
One 100GB SSD and one 1TB HDD.
Running Windows 7

After I installed everything seemed fine until I tried gaming for a bit and my PC shut down by itself a few times after playing for around 20-30 minutes. I started monitoring temperatures and noticed my CPU was a bit high so I changed my cooling unit and put new thermal paste and it seems fine now (It doesn't go over 78C in stress tests).
I thought everything would be fine still but it's still having the same issues. The GPU temperature is fine, it doesnt go over 70.
So I started running more stress tests and found that it shut down first during a FurMark while the GPU temp was at 68.
Then I decided to test the PSU even if I had never done it and scared me a bit so I loaded OCCT and the PSU test lasted for about 3-4 minutes before it all shut down again. The temperatures seemed normal and I sadly have no ideas how the voltages should read.

Another thing that happens is that usually after the pc shuts down it tries to restart but it gets stuck on a black screen not loading anything until I power it down and wait around 3-5 minutes until turning it back on, after that it will boot up normally.

So I don't know what to do. I contacted the store that sold me these parts and they say that the PSU should handle my PC perfectly (which I also thought) but it seems it doesn't. Or could it be something else that I haven't thought about?

Thanks in advance for your help.
 
One of the PSU's protection circuits is being triggered. Resetting the protection circuit requires cycling the PSU's AC power switch.

Your vendor telling you "the PSU should handle my PC perfectly" means that they really don't know.

AMD recommends a 750W PSU for that card and I can see why in some cases (e.g. when overclocked to 1.22GHz GPU and 1.65GHz Memory the power draw of the card can reach over 600 Watts).
 

juomm

Reputable
Jul 31, 2015
4
0
4,510
So I should return the PSU and get a 750W one?
By the way now the shut downs are happening even while the pc is on but not under stress like when browsing or reading something.
 
Check your voltages to see if they are within spec:

+12V rail should not drop below +11.40 Volts or rise above +12.60 Volts.

+5V rail should not drop below +4.75 Volts or rise above +5.25 Volts.

+3.3V rail should not drop below +3.14 Volts or rise above +3.47 Volts.
 

juomm

Reputable
Jul 31, 2015
4
0
4,510
Went into the BIOS (I don't know if I can check them out of it) and found them inside the numbers you said:

Vcore: +1.080V
+12: +11.985
+5: +5.040
+3.3: +3.312

Does that mean the PSU is fine? I also did 2 runs of memtest just to be sure it wasn't the RAM and they went without any errors.
 


It only means that the voltages were in spec at the time you checked them. You don't know what they are when the crashing happened. That requires using a utility that can monitor and log the sensor readings frequently enough over time and hopefully retain a copy of the log file even when the system crashes.
 

juomm

Reputable
Jul 31, 2015
4
0
4,510
Hey again,
So I used HWiNFO64 to log my sensor readings while doing some stress testing and gaming to check if there was something different when it crashed. I crashed while playing a game (not during the 10 minutes Furmark test or the 15 minutes Prime95 one, no idea why).

I can't find anything specific in the logs but I don't really know what half the things registered mean so I uploaded the logs to see if anyone can find anything.

https://docs.google.com/spreadsheets/d/1MRCHCrUmyRz7FBtR5gdVW4upBkpJTbRpxJ0Nih-LSkQ/edit?usp=sharing