- Sep 20, 2016
- 13
- 0
- 1,520
First, introducing my self, I'm a computer scientist, national champion (Brazil) of parallel programming, so I love high performance computing, and decide to build my first gaming rig, lovely, learned a freaking lot and still are! So this a all new gaming rig:
So, first things first, my hardware:
- NZXT H440 Case (3 stock intakes, 1 stock exhaust, 1 Corsair AF140 exhaust)
- Corsair GS800 PSU
- Gigabyte 990-FXA-UD3 R1.1 MB
- AMD FX 8320E (stock BIOS) CPU
- 2x8GB 1866 10-11-10-30 HyperX RAM
- AMD R9 380X GPU
- Corsair H55 Cooler
- 2TB 5400 RPM HDD
- TP-Link PCI Express x1 Wireless board
- Windows 10 fully updated, no anti-virus, only windows defender.
The problem now, the computer randomly (and I mean randomly, videos, downloads, idle, gaming, you name it) hard freezes and reboots (also the uptime can range from 10 minutes to 5 hours).
For my fan setup: I got 3 120mm intakes on the front (plus a H55 in the back as intake) and 2 140mm exhausts on the top, for positive pressure.
What I tried so far (and this a big list)
- Memtest86+, No errors
- Prime95 + Furmark, No reboot so no direct link
- HD health check, No errors
- Mainline Estabilizer 1000 VA, Same thing
- Raising memory voltage (I will attach my voltages below, the memory should be 1.5 and is dropping a little), Same thing
- Raising CPU voltage (didn't cost try within safe temperatures), Same thing
- Scavenging Google, Tried all in earth and nothing
- Changing HDD for isolating software problems, Same thing
- Changing the power cord, Same thing
- Changing the power outlet, Same thing
- Logging temperatures, All normal (I will attach the temperatures after benchmarking)
- Checking Event Viewer for detailed crash information, Just says there was a power surge or crash (thanks Microsft)
And now the BIZARRE PART, I sent it to a respected store within a capital city (Belo Horizonte, I'm from Brazil) and after a week of testing the technicians couldn't reproduce the freeze or reboot!!! WTF!!! This is why I tried a mainline estabilizer as I suspected my power outlet was failing, still same shit and the estabilizer didn't even accused drops or peaks. The technician told me he only run all the possible stress tests and couldn't reproduce the error.
Now I'm logging the PSU voltages so when the problem occurs I can see the last voltages and diagnose accordingly (now I'm waiting for the problem to occur).
And the diagnose I'm most confident is that my PSU is failing because the fan is supposed to respond to thermal or usage sensors, in Idle the fan kicks in sometimes briefly and stops and repeat that, BUT the problem is that when touching the PSU it feels hot, and as far as I know, PSUs are supposed to work below 40 degress celsius. Trying to describe the hotness: it's completely supportable but is easy to feel, one can tell easily that it's hot.
The voltages attachment aren't within benchmarking, but I remember then:
+5V came in as low as 5.000V
+3.3V came in as low as 3.296V
+12.2V came as low as 12.176V
DIMM came as low as 1.472V
VCore uses turbo boost so it's useless info but came as high as 1.424V
SAVE ME PLEASE GUUUYS
Update 1:
Computer rebooted while logging each 2 seconds and the voltages were fine.
Temperatures:
Voltages:
So, first things first, my hardware:
- NZXT H440 Case (3 stock intakes, 1 stock exhaust, 1 Corsair AF140 exhaust)
- Corsair GS800 PSU
- Gigabyte 990-FXA-UD3 R1.1 MB
- AMD FX 8320E (stock BIOS) CPU
- 2x8GB 1866 10-11-10-30 HyperX RAM
- AMD R9 380X GPU
- Corsair H55 Cooler
- 2TB 5400 RPM HDD
- TP-Link PCI Express x1 Wireless board
- Windows 10 fully updated, no anti-virus, only windows defender.
The problem now, the computer randomly (and I mean randomly, videos, downloads, idle, gaming, you name it) hard freezes and reboots (also the uptime can range from 10 minutes to 5 hours).
For my fan setup: I got 3 120mm intakes on the front (plus a H55 in the back as intake) and 2 140mm exhausts on the top, for positive pressure.
What I tried so far (and this a big list)
- Memtest86+, No errors
- Prime95 + Furmark, No reboot so no direct link
- HD health check, No errors
- Mainline Estabilizer 1000 VA, Same thing
- Raising memory voltage (I will attach my voltages below, the memory should be 1.5 and is dropping a little), Same thing
- Raising CPU voltage (didn't cost try within safe temperatures), Same thing
- Scavenging Google, Tried all in earth and nothing
- Changing HDD for isolating software problems, Same thing
- Changing the power cord, Same thing
- Changing the power outlet, Same thing
- Logging temperatures, All normal (I will attach the temperatures after benchmarking)
- Checking Event Viewer for detailed crash information, Just says there was a power surge or crash (thanks Microsft)
And now the BIZARRE PART, I sent it to a respected store within a capital city (Belo Horizonte, I'm from Brazil) and after a week of testing the technicians couldn't reproduce the freeze or reboot!!! WTF!!! This is why I tried a mainline estabilizer as I suspected my power outlet was failing, still same shit and the estabilizer didn't even accused drops or peaks. The technician told me he only run all the possible stress tests and couldn't reproduce the error.
Now I'm logging the PSU voltages so when the problem occurs I can see the last voltages and diagnose accordingly (now I'm waiting for the problem to occur).
And the diagnose I'm most confident is that my PSU is failing because the fan is supposed to respond to thermal or usage sensors, in Idle the fan kicks in sometimes briefly and stops and repeat that, BUT the problem is that when touching the PSU it feels hot, and as far as I know, PSUs are supposed to work below 40 degress celsius. Trying to describe the hotness: it's completely supportable but is easy to feel, one can tell easily that it's hot.
The voltages attachment aren't within benchmarking, but I remember then:
+5V came in as low as 5.000V
+3.3V came in as low as 3.296V
+12.2V came as low as 12.176V
DIMM came as low as 1.472V
VCore uses turbo boost so it's useless info but came as high as 1.424V
SAVE ME PLEASE GUUUYS
Update 1:
Computer rebooted while logging each 2 seconds and the voltages were fine.
Temperatures:
Voltages:
Facebook
Twitter
Instagram