I'm having some difficulty narrowing down the exact cause of my problem and am hoping to get some help here rather than simply replacing the various components of my system to figure out what's wrong.
Recently I had upgraded my video card from a GTX 285 to a GTX 580. The PSU I had for the 285 didn't have enough power going to the 12v rail and so I had also picked up a new PSU to go along with this, the HX850 which from the specs was ample power. I hadn't thrown any particular demanding games at my new setup for about the first month I owned the new card and PSU but recently while attempting to play both Deus Ex: Human Revolution and Dead Island, my system would black screen and be irrecoverable unless I restarted the system. The flashing EVGA logo on my graphics card indicated that it was having a power related problem, as if it simply stopped working due to not having enough power. When this happens, my monitors will cycle through both analog and digital displays looking for a signal but it won't find one until I restart the system.
At this time, I've RMA'd my video card back to EVGA for them to look at after talking with their tech support and trying several of their suggested steps with no luck:
-Manually setting memory timings
-Moving the card to a different PCI slot
-Checking video card temperatures
-Checking the voltage on the 12v rail in the BIOS
They narrowed down my symptoms to either faultiness with the video card or PSU. I don't have a spare PSU to test my system with so I've opted to have the card RMA'd and looked at for now since I do have a backup 8800 GTS so I'm not completely without a computer.
The interesting thing that happened just last night though, the same black screening problem occurred where the monitors would cycle between analog and digital to find a signal but once again couldn't. The only difference in this occurrence was that the system remained running in the background and I was able to continue with my Skype call even though I couldn't see anything. When this happened with the GTX 580, the whole system would hang as the sound would distort and repeat the last millisecond of sound that was being played at the time. I continued my Skype call for a time only to have it eventually lead to the same demise of the system crashing in the background as well, but it took a good 5 or 10 minutes for this to happen rather than instantly with the 8800 GTS.
This is leading me to believe that there may be something else that's problematic. I suppose this could be the PSU or I even got to thinking if there was potentially something wrong with my motherboard and it's PCI slots perhaps?
Does anyone have any ideas as to how I can go through a process of elimination to figure out exactly what could be the issue?
ok tell me, this happens only when you play games ? and when you owned the PSU and GPU for the first month? or it just happened few days ago with both 8800 and GTX 580 ? do you have 6 × 2 or 3 × 4 RAM ? is the CPU overclocked ?
someone here posted a week ago a similar problem but with HD 6990, he gets no signal on screen and computer freezez when he plugs everything, when he removes some HDD powers and unplug all fans he gets to windows environment and everything seems to be fine and he does have the same PSU.
i suggest you do a RAM test with MEMTEST 86 (test RAMs all in once) then test them individually, if the problem still presets i you can only put one stick and remove any add-in card in your pC, simply the CPU,PSU, GPU & RAM and see what happens.
is your PSU fan gets louder? or do you hear any voice coming out from it ?
it sounds like a PSU problem for me, also check that the CPU is good and not overheating
Not all games, just ones that are more graphic intensive like the ones I mentioned above. I was able to play League of Legends and Team Fortress 2 without any trouble for several hours. The problem has only sprung up a few days ago since I just recently tried playing Dead Island and Deus Ex.
The ram is in 3 sticks, 4GB each. I've run Memtest with no errors found and the CPU is currently at stock settings.
If it is the PSU, it seems to be working "good enough" for it to boot to Windows and such just fine. I have no trouble with that at all, it just seems like it runs into trouble when more power is needed when the GPU is drawing more power for the more demanding games.
The PSU fan seems to stay consistent from what I can tell, its usually the GPU's fan that starts kicking in as expected since the heat starts to go up immediately.
I've done some temperature monitoring while recreating the crashes and as far as I can tell, they were all within reasonable ranges, nothing absurdly hot.
First off, I see that the GFX card is "superclocked". Be advised that I had been fighting with EVGA for over a year with my son's box over the same thing. Despite 4 RMA's, the "Superclocked" GTX 295 card simply refused to run at the "factory overclocked speed. I couldn't even get it to run 10% of the way between the "reference" card speed and the factory OC'd speed. They finally ..... and begrudgingly ......sent a GTX480 as a replacement which is in the box now.
The extra specs are of less interest than a few test results would. Download OCCT and run the GPU test. It will monitor both temps and voltages. Even if the machine crashes, the graphs which the program produces will be able to be viewed after you recover / reboot. Look for hi temperatures of any voltage anomalies on these graphs. The ATX standard allows 5% voltage variation but I find anything over 2% can impact just how high your OC goes.
I would ask if you have run memtest86+, especially that your 2nd post lists 12 GB of Gskills. Is this a 6 x 2 GB set or 4 x $GB .....if all 6 slots are filled, I'd expect you may have to drop the timings a bit. If ya got 6 slots filled, 1st do it w/ all 6..... then try 3 at a time ..... MAKE SURE when running 3 that you have them in the right slots ... don't assume...... open the MoBo Manual and check. If one batch fails and one don't, send the 2nd batch back for an RMA......I test them all one at a time and mark the bad one for the manufacturer's convenience but I still want matched set back not a single module.
Thanks for the reply Jack, I'll give OCCT a try when I get a chance tonight. It'll have to be with the 8800 GTS though since the GTX 580 is on it's way to EVGA at the moment.
My RAM configuration is 4GB x 3 and I've made sure to double check with the manual to ensure they're in their correct slots for triple channeling.
The OC in relation to voltage is an interesting theory though I'm not quite sure why it would happen on the 8800 GTS in that case which isn't factory OC'd at all unless the voltage variation is going over the normal allowance of 5% and just driving it way up.
But again, I'll give OCCT a try and see if there is anything unusual happening.
So something interesting I've spotted while running OCCT with my 8800 GTS is that I notice the 12v rail isn't actually running at 12v like I saw when I went into the BIOS to check the voltages.
Not knowing enough about PSU's I'm not sure if this is supposed to be a steady ~12v all the time or if it only ramps up when necessary. At the moment its running at any where from 4.2 to 6.7 I've seen. This is quite a considerable amount of fluctuation which I'm thinking is now the problem.
Based on this, it would make sense that at certain load, this just wouldn't be enough power for the GTX 580 which would cause it to lose power and shut off.
Let me know if the way I'm reading this is accurate or not. Thanks for the help so far!