Possible that a FACTORY overclocked (GTX950SC) set undervolted??!!!

savorymilkman

Reputable
May 25, 2016
7
0
4,510
No overheating, power supply is perfect, ran diagnostics on every and i mean EVERY piece of hardware, everything is perfect until i start gaming. I just bought a GTX950SC (with acx single fan) and after installing it ive been getting black screens and have to manually restart the computer. Ive done a clean install of WINDOWS (10x64) as i had plenty of time because i wanted my new drivers STERILE forget clean (at the time I thought i was having issues with the old gpu (GT720 installed by cyberpower, turns out its structurally fine aside from being subpar) point being there are no video driver conflicts or anything like that lol. I tried reinstalling AGAIN after having a game crash to see if it would help the black screen. Of course it did nothing.

My temps are totally normal, about 60-65C after gaming a few hours and remain stable. Idle i dont even think was at 40 (39 maybe?) I do have that ACX2.0 cooling thing so im not freezing but I dont think i should be worried about my temperatures. im sure its not overheating, its just not the issue here

Also installed are 2 500 rpm case fans
Power supply is 500w
Motherboard is MSI, processor is AMD FX6300 stock, no overclock
Computer is from Cyberpower and is not even a year old, which is plugged into a surge protected power strip (to which i personally guarantee, is extremely stable)
Ive also got 8GB of DDR5, but my most graphics intensive game (heroes of the storm lol) dosent even hit 2gs of memory on ultra btw worth noting runs perfect on the 64 BIT CLIENT!(for anyone whose familiar) ....before it crashes

I play Heroes of the Storm, Diablo III, Dota2, and HoN mostly, these are not graphics intensive games,however at any graphic setting any game im playing, the display will crash, despite the fact that I could run the computer for days on end without issue. Must hold power button to shut it down afterwards

Could the card simply have incorrect voltage settings from the factory? I didn't overclock this card, I don't know if the voltage is high enough to maintain stability, all I know is +128 base +2 boost vs stock. I would assume this would require extra voltage. Just so nobody has to look it up, my card has the 6x2 (6 pin dual molex) that came with it, i think the stock EVGA is 3 pin (single molex? PCI-E?)
It seems strange to me that so many people should have issues like this and are told to resolve the issue by downclocking, where the voltage and limiter could easily be raised in precision for stability. I really would rather get some input before messing with any of this though, please let me know what you think!

Also, reading a few similar posts I encountered one where someone said 12mV increase caused his card to downclock where 6mV provided no stability for his overclock. Being that precision raises in increments of 6, should I be prepared to overclock it further?
 

Iceman21

Respectable
May 3, 2016
275
1
1,960
You tried to run some benchmarks?
Set power limit +10/20%?
Try OC or underclock your gpu just to see if something will change.
See if other people with the same gpu and same driver have the same problem.
How you tested your psu to think its perfect?
 

savorymilkman

Reputable
May 25, 2016
7
0
4,510
the problem is that it dosent last long enough for a proper test, if im not mistaken 30 minutes - 2 hours under heavy stress (i had OC scanner but did not use it) would be ideal, hence why im suspecting low voltage (ACX dosent kick in until a certain temp and i didnt mess with the fan settings at all, in addiction to the increase in clock speeds theres gotta be a reason my card requires 2 molex while the standard requires 1) It dosent take much extra draw to choke a device if theres not enough potential

Funny enough, there was a "test" i did a while back when I was getting "hard drive errors" confirmed NOT to be hard drive errors (thank you windows 10) and at the time I suspected faulty power. My thinking was that since power hits the drive before the hard drive could actually process and send out a signal was the cause of this (even I know hard drives are notorious). What was happening was that the drive would "fail" and the computer would react by shutting everything down. It turned out to be software conflict, windows 10x64 sucks had the drive checked it was fine solved the software never happened again.

My 60hz TV i was getting as much as 105 FPS in heroes of the storm on ultra if that helps. Had to vsync to get it under control (from game, i took care of the application controlled presets in control panel)

As far as my power supply, im a bit of an electrical hobbyist/enthusiast whose father worked for the local power company 35 years, i learned to use a volt meter before I was ten and have decent experience with both DC and AC systems including my own wiring on guitars (featuring active and passive electronics), fixing electrical outlets, setting up routers, fixing the car radio, fixing my guitar amp after a power surge in texas blew it out of the socket, etc. Im not an electrical engineer but i have enough experience to recognize even in a low volt system whether or not failure in delivery came from the source or not. . bios wasnt reporting any anomalies on any of the rails and theres just an appaling lack of problems anywhere else in the system. the card is powered directly from the power supply leading me to believe that the voltage on the card is simply insufficient. 500w power supply theres certainly enough left for the card, really i could get the volt meter, but its not

Im not sure and theres not enough information about this particular card that ive been able to find (reliable information) what typical volt readings should be for this card, and furthermore I'm not sure what they should read WITH the overclock in place, thats why im here.

As far as limit, I didnt even reinstall Precision X so ill report back with update after trying that.