NVIDIA MSI GTX660 Twin Frozr 3 Crashing

Just got my card back from MSI because I had to RMA it after 2years. Thankfully the card was covered under warranty for 3. The card I got back is horrible!!!!

Anytime I try to do anything that requires the gfx card to operate it crashes... (Current driver is 352.86)

I don't know much about overclocking but I have been using EVGA Precision X and MSI Afterburner to monitor the card. When game(s) aren't running the GPU clock stays at a steady 324MHz. As soon as I try to play games or do anything that needs the card it jumps to 1019MHz, and if I max out game graphics it shoots up to over 1149MHz! I didn't think that was possible with this card? My screens go black and the computer reboots. Sometimes it doesn't reboot. It wipes out my wireless internet adapter too.

I have to reboot, unplug the wireless internet adapter, plug it back in, and wait before I can use my pc again and surf the web.

Stats without games running: Screenshot by Lightshot Note the GPU clock at 324.

Stats with games running: Screenshot by Lightshot Note the GPU clock at 1149.7!

According to this webpage Graphics card - The world leader in display performance the core GPU Clock is 1033MHz and the Boost is 1098MHz. Not sure how I'm getting up to and maybe even over 1149.7MHz.

I never monitored any of my hardware before I started having problems with this card so I don't know how the other card was operating compared to this one...

I can work my way around a computer on a more advanced level than your average user so I'm willing to try some things... but messing with flashing bios, or editing the registry I'd rather not do... =)

I just got this card back after 2 weeks of being without it... I don't want to have to send it back AGAIN! I just find it odd that the card I RMA'ed ran perfectly and this being the same one does not!?!? IDK...

Thanks in advance. =)

Here are my PC Specs:

OS - Windows 7 Home Premium 64 Bit
CPU - Intel Core i5-2500K 3.30GHz Quad Core
RAM - 16GB DDR3 SDRAM Patriot
Motherboard - MSI H61M-P23
Chipset - Intel H61 (Cougar Point)
Graphics Card - NVIDIA GeForce GTX 660 2GB of GDDR5 on board RAM
Hard Drive - Seagate ST500DM002-1BD142
Power Supply/Case - Thermaltake V2 Mid Tower 450W PSU (purchased in 2013)
Monitors - 3x 26'' LG LCD Tv's

If you need any other info just let me know.

Downclocking when idle is normal.

You can try manually downclocking the GPU to first 1000Mhz and see if it will show same issues. Increase to where it should be later.

I'm sorry I don't understand? If I downclock the GPU won't that make it crash more?

No, it is the overlocking that makes the hardware to be unstable requiring extra voltage etc.

Underclocking - Wikipedia, the free encyclopedia

Underclocking, also known as downclocking, is modifying a computer or electronic circuit's timing settings to run at a lower clock rate than is specified. It is analogous to driving a car below the speed limit. Underclocking is used to reduce a computer's power consumption, increase battery life, reduce heat emission, and it may also increase the system's stability and compatibility. Underclocking may be implemented by the factory, but many computers and components may be underclocked by the end user.

You can use MSI Afterburner for this task.

So what should I set the GPU Core Clock at?

Edit: So as stated above I have 3x 26'' LG LCD Tv's hooked up to the gfx card... I swapped from one monitor to another and the card stabilized. I am currently testing now by playing Final Fantasy 14 on maximum graphics settings and it has been 20-30 mins since last problem.... On the other monitor the card would have crashed by now. Card is running consistently between 810 and 1000MHz. I also dropped the in game frame rate from No Limit to 30FPS. Does this seem normal that the monitor (LCD TV) would be causing the problem? Seems odd to me.

I guess everything is possible when electronics are involved. However, 1000Mhz isn't pushing that card, is it?

Try with FurMark. It should push it to the limit. However, don't leave it unattended, Furmark also produces a lot of heat!

Yeah 1000MHz seems a bit high to me seeing as it's factory set core is 1033 and its boost is 1098... It still spikes to 1150MHz+ but it depends on where I am in the game.... but it seems to go back down and it doesn't or I should say hasn't crashed yet...Do you think I should still downclock it anyways? If so to what value?

Considering it was the card and not the monitor! Disregard if it turns out to be the monitor as the culprit.

I meant the card is not pushed to it's limit with that game you are playing. Push it harder to see if it will destabilize. If it does, then downclock with MSI Afterburner to 1100Mhz. And work your way down in 10Mhz intervals until it is stabilized.

If you can find the spot, you may later request a custom BIOS at some other forum (or perhaps from MSI) so you wouldn't need to use software to downclock.

I have never used FurMark before... How long does the stress test take and how do I know if something is going wrong other than a BSOD or black screen(s)? What is a good operating temp for a gfx card? Mine fluctuates between 39C-49C. Is that a safe operating temperature?

FurMark will make it burn (which is expected). It increased my GTX970's temp by 10C in 15 seconds

Hmm, maybe you should start with 3D Mark(1.7GB) or Unigine Valley benchmark (300-400MB).

If you are not experiencing anomalies then it should be alright.