Possible Power Supply Failure?

[barky502]

Recently (for over a month now) I have been experiencing issues with my computer which I built back in November of 2015. I built my computer with an EVGA SuperNOVA 550w Powersupply, an ASRock Fatal1ty H170 Mobo, Intel i5 6600, NVidia/MSI Gtx 970, 2x8GB Corsair Vengeance LPX 2133 DDR4 Memory, a Crucial 256 GB SSD, and a Seagate 1TB HDD.
I have tried many things to solve my issue including: Reinstalling/Updating BIOS, Reinstalling/Updating Graphics Drivers, checking if all cables were fit in properly, cleaning the little dust out of the system, doing stress tests on the CPU (no problems,) fiddling with my overclock on my 970 and doing stress tests with it using Furmark, and doing a disc scan or two. Note: This only happens when my GPU is under stress; I ran AIDA 64's CPU stress test for a while and was stable, but when I ran Furmark on a usually stable OC my computer would just restart (not blue screen, just instantly off.)
I believe this is a powersupply issue because my entire computer shuts off (case LEDs included,) and not a hard drive or video card problem, which would likely results in a BSOD, instead of the entire system shutting off at once, and just turning back on per usual. If anybody else knows has had this happen, or knows anything about power supplies (I'm sure there is somebody who fits the latter criteria) can help me out here. Sorry for posting here, there was no power supply category, and I didn't know whether to post it in here or components. Thank you!

 

[R_1]

reboots like that are usually PSU, RAM, or a software setting, and usually in that order.
the good news is if it is the PSU it has a 7 yr warranty. to test the only real way is to replace the unit and test.
you could try to clear the RAM as the issue by running memtest for a pass or two.
you can try to eliminate the software by running in linux (bootable USB/DVD) and running some benchmarks under the [strike]clean[/strike] fresh environment

 

[MeanMachine41]

Hi barky502

To check your PSU under load, Have both HWMonitor and AIDA64 on your desktop together side by side.
Run the stress test for CPU, FPU and Cache and observe your Voltages on all rails in HWMonitor.
It will indicate if the PSU is doing its job under load. You have +/- 5% tolerance on all rails to be in spec.
You can also run a test on the GPU separately and your RAM.
Also try with one RAM module in first slot and try switching them around in case one is failing.
You can also run Memtest86 from a USB stick to eliminate your RAM as a possible cause.

 

[barky502]

Hi barky502

To check your PSU under load, Have both HWMonitor and AIDA64 on your desktop together side by side.
Run the stress test for CPU, FPU and Cache and observe your Voltages on all rails in HWMonitor.
It will indicate if the PSU is doing its job under load. You have +/- 5% tolerance on all rails to be in spec.
You can also run a test on the GPU separately and your RAM.
Also try with one RAM module in first slot and try switching them around in case one is failing.
You can also run Memtest86 from a USB stick to eliminate your RAM as a possible cause.

Thank you for replying so quickly, I will try these and report back if I encounter anything. Thanks!
Edit: Since I'm not quite sure what I should be looking for, what should I be looking for? In AIDA64 It displays my 3.3V, 5V, and 12V rails, but only my 3.3V rail is showing up in HWMonitor. Also, where would it display the deviation in the rails, or would I just have to judge that by myself? Thanks again.

 

[R_1]

the 12V rail +/-5% should be between 11.4V-12.6V the closer to the rail the better.
5V should be very close to 5V 4.75V-5.25V. the 3.3 V should be 3.17V-3.47V

 

[barky502]

the 12V rail +/-5% should be between 11.4V-12.6V the closer to the rail the better.
5V should be very close to 5V 4.75V-5.25V. the 3.3 V should be 3.17V-3.47V

After 10 minutes of testing my average on each rail was 3.323V for the 3.3V rail, 5.064V for the 5V rail, and 12.191 for the 12V rail. This is very close to what they should be, so I'm not concerned really. I'll put the statistics for the test right here.

 

[MeanMachine41]

the 12V rail +/-5% should be between 11.4V-12.6V the closer to the rail the better.
5V should be very close to 5V 4.75V-5.25V. the 3.3 V should be 3.17V-3.47V

After 10 minutes of testing my average on each rail was 3.323V for the 3.3V rail, 5.064V for the 5V rail, and 12.191 for the 12V rail. This is very close to what they should be, so I'm not concerned really. I'll put the statistics for the test right here.

Yes your rail voltages are within spec.
Now check all your RAM in AIDA64 under load with all your DIMMs for 30mins.
Then run a test on your GPU for 30mins. Keep an eye on temps.
Furmark in EVGA Precision X 16 Stress test for NVidia cards is a more demanding tester than AIDA64 and if you ran it for 10mins and temps are OK then I think that rules out the GPU.
As I don't know what changes you have made in Bios, it might be worth clearing CMOS back to default. Did you make any changes in Bios without knowing the consequence.?

You cannot overclock your CPU as it is none "K" version. You can however change RAM Primary Timings and Voltage if your RAM proves unstable.
Have you set your XMP profile in Bios.?

 

[barky502]

the 12V rail +/-5% should be between 11.4V-12.6V the closer to the rail the better.
5V should be very close to 5V 4.75V-5.25V. the 3.3 V should be 3.17V-3.47V

After 10 minutes of testing my average on each rail was 3.323V for the 3.3V rail, 5.064V for the 5V rail, and 12.191 for the 12V rail. This is very close to what they should be, so I'm not concerned really. I'll put the statistics for the test right here.

Yes your rail voltages are within spec.
Now check all your RAM in AIDA64 under load with all your DIMMs for 30mins.
Then run a test on your GPU for 30mins. Keep an eye on temps.
Furmark in EVGA Precision X 16 Stress test for NVidia cards is a more demanding tester than AIDA64 and if you ran it for 10mins and temps are OK then I think that rules out the GPU.
As I don't know what changes you have made in Bios, it might be worth clearing CMOS back to default. Did you make any changes in Bios without knowing the consequence.?

You cannot overclock your CPU as it is none "K" version. You can however change RAM Primary Timings and Voltage if your RAM proves unstable.
Have you set your XMP profile in Bios.?

I am quite positive that I did not change anything in the BIOS aside from my boot order. I use MSI Afterburner instead of EVGA Precision X because my card is an MSI, and it offers more support, but hopefully the built in Furmark for that is the same as EVGA's. TLDR: I ran from +50MHz core and +100MHz Memory up to +15mV, 110% Power Limit, +125MHz Core and +350MHz Memory before the computer did the usual turn off suddenly and turn back on. The last stable test was +15mV, 110% Power Limit, +100MHz Core and +300MHz Memory.
The memory test was also fine for 30 minutes. Since you said that Furmark should be good if it's stable for 10 minutes I didn't bother with the 30m of Aida's test on the GPU as I had ran over 30m of Furmark working on the OC. I just reset my CMOS as well, so I'll go play some games and see if I crash anymore.
Edit: Just ran a Memtest 86, no errors, ran fine.

 

[MeanMachine41]

So is your system still shutting down randomly.?
What is your current Bios version.?