GTX 660 ti OC PE Crash on all drivers?

fonzi0243

Honorable
Aug 27, 2013
2
0
10,510
So i have had this card since the start of the year could not play crysis 2 because it would crash within a few seconds of playing...

i eventually worked out that it was the PSU causing the problem and replaced it with my corsair CX600... that worked but my card always crashed at some point in the game, it would be a black screen and a bubble would pop up on the desktop saying my card had recovered...

i thought it might be overheating... but this problem has become more and more frequent and now i cant even play with the graphics turn up... i have a feeling its the PSU again but i just want to be sure, in OCCT the 12 volt rail says its dipping to about 11.4? not sure if i believe that but i have ASUS suit 2 and that says its hanging around 11.7... i dont have a voltage tester...

The problem how i see it seems to be involving the boost clock, could it be possible that the cards trying to draw more power but obviously cant because the power supplys too weak?... i have taken both my ram sticks out and tested them separately, this makes no difference so i know my ram is fine...

I have ran furmark to test my card and i tested the memory on the card it seems to be fine... i cannot run OCCT testing the CPU and furmark at the same time because the GPU and CPU fight for power tuning off and on constantly, this tells me my system needs more power????

Please let me know what you think?





CPU: AMD Bulldozer FX-8120 Black Edition
MB: asus m5A99X EVO
Ram: Corsair 8GB Blue vengeance
PSU: Corsair CX600
Graphics: MSI GTX 660 ti Power Edition
Storage: OCz 60GB SSD - Seagate Barracuda 7200 6GBs
Cooling: Noctua NH-D14
 
You can run CPU and GPU test at same time in OOCT but the "PSU test" is likely to fry ya PSU.... the CX series is not very strong. I have had similar issues with EVGA SC series cards but not w/ MSI PE ....then again, never with 11.4 volts which BTW fits within the 5% allowable under ATX standard.

Next time, if ya can't do HX, try and stay w/ the TX V2 series. Might try complaining to Corsair and see if they let ya trade in for a TC series.
 

fonzi0243

Honorable
Aug 27, 2013
2
0
10,510
I ran the PSU test and the 12v rail dropped to about 11.33v the CPU and ram was running on full load the whole time but the GPU was dropping from 99% to 60% then 80% just keeps doing that the whole time :/ im really not sure what this problem is? i go into MSI afterburner and drop the core clock by -100 and im able to play games... but soon as i reset it to stock it crashes still?
 
i had same problem. The trick was to force a constant clock speed with nvidia inspector, fluctuating boost clock seems to cause issues with this game. All other games seemed to work fine except C2 and C3. Once the game crashed it would also revert to 500mhz and could not recover without reboot. Google, "nvidia inspector batch file", you can set up a batch file on your desktop for normal, idle, and force constant clock speed. The first couple links that come up have detailed instructions with templates, you just insert your cards clock speed for ram and core clock. This was a 100% fix in my case.
 
I had this problkem with afactory OCd EVGA card....they gave me the run around for 18 months .... finally they sent a stock next gen card.

Was an issue about over volting some MSI PE cards some time back......

10 minutes later I found it

http://www.tomshardware.com/news/MSI-GTX-660-670-overvolting-PowerEdition,18013.html

Still....the PSU is running outta spec...they should swap it ....or at least let you trade in for a TX
 


monitoring voltage with software is taking it from the motherboard, so its not necesarily the psu running out of spec. only way to check is with a multimeter on the 12v connectors of the psu.
 
Whatever the voltage might actually be is of less concern...if it's off 0.1 volts doesn't really doesn't concern me. The fact that OCCT is showing it drop 0.7 volts is scary though and this is something weve seen before with the CX series.

With the CX and a possible "cheating" PE GFX card, he could have a host oif problems.
 


a similar thing used to happen when i had a core 2 duo and nforce 650i board, was getting voltage drops on the 12v under load. went from a thermaltake 470w to an antec 650w psu and the voltage drop was still reported by hwmonitor. So it wasnt the psu at all in that case.