Hello,
Today I have received from RMA (old was GTX-560Ti/448 cores) EVGA GeForce GTX-660 Ti / 2GB (02G-P4-3660-KR), but fan seems to produce rattling noise probably from bearing or vibration. I got few questions about that:
 
- Is there something which I can do myself to avoid new 2 weeks with non working PC while RMA is processed and traveling from Bulgaria to Germany?
Fan is producing noise like something is stuck in fan blades or vibration, but there is nothing inside the fan - no dust nor fan is not touching case of cooler and there is no visible reason for that sound. Sound can be heard on idle (@40%) and changes it's frequency as temperatures are changing with RPM change. It is more irritating during system idle because it is louder than all other 5 fans inside the case (4 x 120mm and one Intel LGA 775 boxed cooler).
 
- Am I allowed to open cooler case cover WITHOUT removing whole cooler (as far I can see there are several screws on plastic) and if I disassemble fan to change lubricating grease will this void warranty of whole card?
- If it is from vibration and not from bearings can I add piece of foam from old laptop HDD ( nearly 2mm thick)? Will this stop vibration, and will this void warranty of card (of course with assumption that fan won't touch and stop on housing)?
- Eventually if above tries are not successful is it possible to get new fan and send old one from/to EVGA support team or to order new one and on what price?
Again - I can do myself almost everything, but with proper advises. Eventually fan model number will be very appreciated - mostly to look for schematics and documentation.
 

Thread can be closed. Fan became almost silent and sound is hardly audible now. Looks like fan needed some time to work ...

Yes, of course. Fan now is almost silent and because I had some free time, discovered that card is freezing on random. Even with disabled HDMI audio and all verclock related settings in BIOS set to safe defaults. Looks like there is always new "surpirses" with NVIDIA latest GPUs....
Now I got some bitter feelings that this card should be RMA'ed again and again in no more than month I will stay without computer for new 2 weeks period...

Above fixes were tried already. Interesting part is that card is crashing only on DirectX based games - furmark, GPU-Z stability tester (looks like furmark based) MSI kombustor tests does not crash. But ... there is no more than hour in Bioshock Infinite, Mortal Kombat X is crashing on 4th maximum 5th demo combat. There is no change if fan is forced @80%, Set to "Curve mode", or Auto - when card goes in Boost frequency mode and hits 61-62 degrees it's crashing. Old drivers (310.90, 314.22) in most cases can produce "Nvidia kernel mode stopped to responding and was restarted", newer (355.XX) are ended either with solid colored screen or just message from monitor "Going to standby".
Tried everything - Integrated LAN was disabled and exchanged with old Intel Pro100/PCI (used same IRQ as VGA), Realtek sound was disabled in BIOS and drivers were removed, NV HDMI was disabled in all of the time of tests. Hooked one old SB AudigySE in exchange of RTL888 Sound. Played with enable/disable/force use of HPET and there is no change. Only disabling HPET maked card to last maybe 10-15 minutes more before crashing. PSU is out of scope because voltages according to multimeter and HWMonitor 1.28 are over 12.1V in all of the time, CPU and RAM are on their stock settings... Even clean Windows install does not change situation. Now I just think that got next faulty EVGA card or just Nvidia are actively working on losing all of their customers with these crappy drivers. Tomorrow will drop a line (again) to EVGA support either for suggestions, BIOS update or RMA start. Again I am out of jokers in the pocket and I am not sure about the root cause for crashes.
Now I will go to sleep as here is almost 1:00 am.
 
 

Hello mapamop, 
 
I apologize for any inconveniences you are having with your replacement card. Please contact our EU support team directly 
Tel: +49.89.189.049.11English: supportEU@evga.com so we can further assist you in this matter. 

Ok, I have more light on situation:
 
-- card is crashing on GPU frequencies over 1100MHz in boost mode.
-- latest driver 355.98 does not fix issue. Playing with power profile, Single / Multi display performance does not change situation.
-- PSU is not the case because voltage in idle state on desktop is 12.34-12.35V, just before crash was 12.31-12.30V measured with 2 multimeters hooked on both connectors. Moving between PSU's 2 12V rails does not change situation - tried to connect card on every one rail and to mix them. Tried to clean PCI-E power connectors of PSU with dry contact cleaning spray - without change.
Mail to EVGA EU support was pointless - they suggested me to play with drivers from download section and check for VBIOS update, which does not exist there.
Current VBIOS is 80.04.4B.00.60, VRAM is Hynix
At this point could someone from EVGA give me new VBIOS (if available) or tool which I can use to modify frequencies?
This afternoon will try to change PSU with new one and if there is no change I will open RMA, because all possible solutions will be exhausted.
 

With the testing you've done, it certainly appears to be an issue with the card. You'll need to contact our EU office for assistance, as mentioned previously. You can do that by calling +49.89.189.049.11 or emailing supporteu@evga.com and they should be able to assist once they re-open for business on Monday morning.

Just a small update:
Today I have received third faulty card in row - artifacts on BIOS splash, crashes on desktop, freezes on Youtube SD videos, GPU-Z bus test (furmark based render test) crashed it in no more than 5 minutes. Now 3 months from 12 months warranty are gone in ping-pong game with EVGA RMA team and I am really begin to feel frustrated from endless testing of faulty cards...
 
Seems that nobody tests hardware before releasing to user...

Yeah,
They asked me to check PSU again ... I am really became frustrated by this purchase - it was a big mistake.

EVGATech_JaesonW
mapamop,
 
I'm sorry to hear you're still having trouble with your system. I can assure you that all video cards are tested thoroughly as part of the re-certification process. We can definitely work with you to either get you another replacement, or if you would prefer a refund we can do that instead. Please contact me directly at jwilda@evga.com and I'll work with you and our EU team to get this resolved.
 
In the mean time, it sounds like you may have more than just a video card issue with your system. The first card you sent back to us did fail testing, so there was definitely an issue there. The second card though did pass testing, which would indicate another component as well as the original video card is faulty. If at all possible at this point, I'd highly recommend seeing if you can test a known-working video card in your system and see if the if the problems go away?
 
I also see you mentioned this latest card artifacts in the BIOS. That's a sign that this card may indeed be faulty. If there is any way you could test the card in another computer to see if the artifacting and other issues follow the card, this would confirm is this card is indeed faulty.
 
Please email me at your earliest convenience so we can get this resolved for you.

Hello,
 
Yes, second card is working flawlessly until race condition is met (I have stated it several times):
- card needs to go in boost mode and frequency should go over 1072Mhz (1072 is last known frequency before crash)
- RAM usage needs to go over 1300MB - with usage around 1200 - 1250MB card is stable
 
- I am not sure but temperatures in my opinion wasn't issue, because CPU does not go more over 46-48C and north bridge does not go over 55C, usually with heavy load previous card did heat in most cases more than 63-67C, because in room ambient temperature is around 23-25C. Also I have a 120mm fan which is blowing cold air on front south bridge, VGA and PCI slots. I suspected that there is slightly unsoldered ram chip which looses contact when it is heated.
Few days ago I get old ATI 4650/1GB DDRII which survived all the stress tests I can run on it. When tested first card (GTX560Ti/448) I was borrowed my friend's Sapphire Radeon 7770 Gigahertz edition and put it on stress testing inside my PC - it does not crashed for a 3 hours stress tests. I am not sure if these facts have some meanings because 7770 is 128-bit and has TDP at 80W or 100W while GTX-660Ti have 130W-150W as I have seen in different sites.
 
Now on current case - Card was received today @9:30 A.M. packed in your usual carton packaging. At 10:30 after I have put it in PC and checked it if it seats securely on slot. When powered on for a first time PC there was on first top half 4 lines with incorrectly colored bold dots. After cleanup of ATI and Nvidia drivers (with uninstall before for AMD drivers) with DDU pixels do not appeared on restart from safe mode, but appeared on next reboot after install of 355.98 driver. Also I noticed slight hiss sound from card's power circuit. Shortly after my son opened video on youtube from Firefox and on 5th minute of video got driver crash warning with black screen freezed PC. When I come back after work in next 2 hours I will send you mail with pictures to your mail as advised.
 
I still have one questions without sensible answers:
- Why card boosts over 1000 Mhz, while on Nvidia.com and on your site it is stated that boost frequency is 980Mhz max?
- Is there a reliable (without BIOS mods, soldering additional elements, up/down voltaging card)way to prevent boosting over let's say 1000Mhz without constant running PrecisionX or similar tools (setting something in Nvidia CP is OK)?
 
 

vBios modification is the only way to disable boost. My titan x's run at a constant speed on every application without any fluctuations in clocks whatsoever.