New Build having issues when gaming

miser44

Hi,

I got a new build not that long ago but now am running into an issue with it. Whenever I play a game for about 10 -30 minutes the computer will completely shut down and the only way to turn it back on is by turnong of the PSU with the switch for a few minutes and then turning it back on. The games this has happened on is Metro 2033 Redux and Rising Storm are examples.

Specs:
Intel i5 4460
Gigabyte GA H97 HD3
Gigabyte GTX 970 G1 Gaming
BeQuiet 500 L8
8gb Crucial Ballistix Sport
Crucial BX 100 500GB
2 X 500GB WD HDD

The computer has no problem for hours if just browsing and watching tv shows /films etc. Only happens when running a game.

Ive been using CPUID HWMonitor to monitor temps to see if its an overheating issue. When idle my CPU is 31,34 degree range, when gaming this increases to the 45 - 48 degrees mark. The GPU when idle is around 28 - 30 degrees and when gaming it increases to the 58 - 62 degree mark. I don't think those temps are bad from what ive read on prior forums but may be wrong.

All my drivers are update from what I can tell. When I check the Event Log Viewer this is what I get:

Critical - Kernal Power - Event ID 41: The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.

Am thinking it may be the PSU, either its faulty or its not getting enough airflow. I have the PSU fan facing to the bottom of the case where there is a vent. The case is located on a wooden cabinet so I don't believe theirs that must dust be sucked in as compared to if it was on carpet.

Sorry for the long post but want to try and include as much in as possible.

Any help would be great.

SterlingArcher

Temps seem fine.

Have you overclocked the gpu?

Need to run some stress tests on the gpu and memory.

http://www.hwinfo.com/download.php

Download this.

It may be the psu,can be anything from gpu/bad memory, to a software/antivirus. So need to start eliminating.

After tests,Can use your igpu to run games see if it happens. Swap out the ram.

miser44

I ran the programs FurMark and Prime95, ran each for an hour and the computer did not crash. Not to sure if an hour is long enough for these tests?

For the GPU on the FurMark test the highest the temp went to was 71 degrees and fro Prime95 the CPU temp highest was 58 degrees.

I went to eliminate the GPU having the issue by replacing the GPU with a XFX GTX 260, in the middle of downloading the drivers for this card so that I could run a game the computer shut down as reported. This is the 1st time it has happened without a game being run.

Am going to run a stress test for the RAM using MEMTEST and will replace the PSU with a XPower 550 W from my old computer, hopefully it has the correct corrections.

Is there anything else I should also? Run the stress tests again and run them longer?

mad turnip

This really sounds like the PSU cannot supply enough power.
After a quick lookup on new egg:
http://www.newegg.com/Product/Product.aspx?Item=N82E16814125685

They recommend 550W, usually if you can match what they say you won't have issues but you have two hard disks an SSD and a pretty powerful card and a strong CPU its not surprising after 40mins when temps get higher it requires more power and your 50 Watts short.

The fact when you plug in your 260X which is even more power hungry it doesn't last long would make me think its the PSU.

HWMonitor gives some good appromiations on power usage which could be worth adding up the motherboard g-card, and hard disks and see how much the wattage comes out at while gaming.

I'd recommend a nice 650W PSU should last you well for a good few years to come.

The fact your benchmarks are one at a time for CPU + GPU would show its right on the line of the power consumption as it does not shutdown when one is "idle".

Digital Solitude

Sounds like a bad PSU to me, an i5 with a 970 should comfortably run on a 400w, so a 500w is miles clear.

Try the RAM test, but its likely the PSU isn't outputting the wattage it should be

miser44

Is there any way I can determine if the PSU is faulty or that the system is not receiving enough power?

I had a quick glance at the XPower 550W before going to work and it doesn't look like it has the connections I need so I do not believe I can test the system using it.

I cant test the beQuiet on the old PC due to having no CPU cooler for the CPU. Was using a Evo 212 and the chip is AMD so cant use the stock cooler that came with the Intel chip.

Only thing I can think of is removing the graphic card from the system and only use the motherboard graphics and run a game. Am pretty sure the PSU could handle that and if it crashes then the PSU would be considered faulty but if it doesn't then its not enough power for the rig. Would that be correct lines of thinking and be suitable enough to determine the PSU or is there any other way?

Digital Solitude

Have you a friend who's PSU you can test? The integrated graphics is very unlikely to draw enough load to crash it.

miser44

I did reach out to someone but it could be 2 - 3 weeks before I can use there PSU to test, he lives in Dublin and am in Donegal. The other PSU does not have the right connections so I couldnt use that to test.

I did come up with an idea and seeing what you think. As Mad Turnip mentioned that the 500 W may not be enough to power due to all the components that I have hooked up.

I can disconnect the 2 HDD and the DVD drive, only connect the GPU, SSD, Motherboard connections and the 3 fans on the case. Would 500 W be good enough for just those components?

If 500 W is enough for just those to run and it doesn't fail then we can nearly saying that the PSU isn't enough power for everything to be connected, if it does fail can nearly say the PSU is faulty. That is if 500W is enough just to run them?

Only thing I can think of as I am limited on how test to this. I will still do the other stress teests just to rule them out also.

Digital Solitude

500w is enough.

What I think the problem is, as I've said is the PSU is faulty and not outputting enough wattage.

This is a hypothetical example.
The PSU should be able to output 500w without tripping. The load wattage for this build is about 350w. Without the 970 its about 190w. So the PSU is cutting when it's asked to draw between 220-360w we'll say, for a little headroom, the GPU will draw a little at idle, so its 220 instead of 190w.

So yeah, between 220 and 350w the PSU in dying, when it should manage over 500w no problem, so removing and testing different components won't test the PSU, as the GPU is the only part capable of bringing the wattage over 200w, which is where the trip lies, imo.

That's my understanding and I've been trying to make this clear to myself as I went along so I don't care if I'm wrong :pac:

The numbers I've quoted for power draws are roughly correct, check a PSU Calculator if you want confirmation

Digital Solitude

That whole speil may be made redundant by he fact it passes a furmark, unless the limit I'm talking about is very close to ~350w

Under clocking/volting the GPU may be able the help a diagnosis, mightn't be able the turn it down enough though. Worth a shot, use MSI Afterburner.

And check all GPU and PSU cables to be doubly sure

mad turnip

After looking at some more reasonable power consumption estimates I'd agree with Digital Solitude the system shouldn't be anywhere close to 500W. Its probably a faulty PSU.

miser44

Thanks guys, am going to return the PSU back to hardwareversand and hopefully get a replacement. I know the company has gone sideways since I last used them so hopefully its a smooth return. Any experience on how long the return process takes?

Thanks a lot for all the help.