Tl;dr version: when the target temp(factory default) is 75C the card dies, when I manually force it to 70c (and unlink power target leaving it at 100% in precisionx) the card runs fine.
 
I'll try to keep it short.
Background: I work at a pc store and we just built a new pc with a 970 and video on it keeps dying.
 
Here is how it dies, Running unigine Heaven benchmark or furmark for more than 5minutes causes the video signal to the monitor to completely go away and monitor goes to standby. 
I still hear sound from the machine and can shut it down with the keyboard. After a shutdown the video is restored and working, doing a reset does NOT restore video.
 
I tested the card in another computer with the exact same result, at first I thought it was a defective card and we ordered a second identical one with the same result.
After this I tried setting the target temp down and now both cards work fine in both computers with the lower target temp. If i set the target temp back up to 75 both cards fail in both computers again. I can even do it back to back, run the benchmark 3 times at 70C with no issue, alt-tab and set temp to 75 and the card dies on the very next run in less than 5mins.
 
Can anyone else try this and let me know what they see?
I find it 100% impossible that both brand new cards do the exact same things in two completely different computers and the "fix" is as simple as setting the target temp down to 70c. 
75c is far from really hot in the first place but I have the evidence right here and I can reproduce it with 100% reliability.
Also why do I not see any other posts online with similar issues.. I'm stumped.
 
Also our supplier refuses to even look into the issue and tells me to contact EVGA directly.
 
Specs & stuff:
Using the latest Nvidia drivers: 347.88
 
New computer:
Operating System: Windows 8.1 64-bit
CPU:Intel Core i7 4790K @ 4.00GHz 33 °C
RAM: 16.0GB Dual-Channel DDR3 @ 934MHz (10-11-10-30) (kingston hyperX 1866)
Motherboard: ASUSTeK COMPUTER INC. MAXIMUS VII HERO (SOCKET 1150) 28 °C
Graphics: NVIDIA GeForce GTX 970 (EVGA) 35 °C (model: 04G-P4-2972-b6)
Monitor VE228 (1920x1080@60Hz)
Storage: 465GB Samsung SSD 850 EVO 500GB (SSD) 27 °C
Optical Drives: HL-DT-ST BDDVDRW CH12NS30
EVGA 850 bronze PSU
 
My other test machine thats about 1 year old and in use every day all day long:
Operating System: Windows 8.1 64-bit
CPU: Intel Core i7 3770 @ 3.40GHz 43 °C
RAM: 16.0GB Dual-Channel DDR3 @ 665MHz (9-9-9-24)
Motherboard: Gigabyte Technology Co., Ltd. Z77M-D3H-MVP (Intel Core i7-3770 CPU @ 3.40GHz) 28 °C
Graphics: 4095MB NVIDIA GeForce GTX 970 (EVGA) 40 °C
Storage: 931GB Western Digital WDC WD1002FAEX-00Z3A0 (SATA) 32 °C
Optical Drives: HL-DT-ST DVDRAM GH24NS90
Thermaltake TR2 600w PSU
 
*edit* pictures cause.. why not?

You can register the card now and whoever buys it can also register it with the warranty transfering over. But the 3 year warranty would start immediately when you registered the product possibly being an issue if the card stays in your stock. (Whoever holds the card at the time has the warranty as long as the serial number is on the card still)

To get to your issue, you need to obviously keep your BIOS updated if it is not. Secondly, seeing as both cards export emcee the same issue, it may be a BIOS setting which may make it ideal to reset the BIOS/CMOS. Lastly, when the card heats up, this means the card is using more watts leading to a possible PSU issue. You mention your work at a computer store so you could easily use or buy a PSU error reader (plugs into back and gives a report back). Other than that, if you have not tried calling EVGA, I would try that and you do not need to have the card registered to do that. However, you usually still need the cards serial number if you do call.

Good luck, hopefully you find the issue or maybe someone knows more on the issue.

@bsmegreg:
BIOS in both machines are on the latest version.
Two different motherboards, different processor generations, different PSU's same result.
 
@Sajin:
The new machine was a build from scratch with no OS or drivers installed. I did not install afterburner or precisionX or anything similar for the initial test, using only the driver windows update downloaded it died.
Then downloaded 347.88 drivers(as stated in my OP) and installed only the drivers no other software and same issue.
Only *after* I installed PrecisionX and lowered the temp target out of desperation did the issue go away.
 
So no one else can reproduce this? I just can't believe that I'd have two cards on two machines with the exact same behavior and no one else has it happening.
Also I did play a game on the machine (Cities: Skylines) for about 15 minutes with no crashes with the temp target at the default 75, it only really happens under load - which still isn't supposed to happen.
 
I was hoping to uncover a similar situation for others that they were no aware of but I guess I'll have to contact EVGA support directly.
 
Ps. I also had a reference design 970 that I tested in the new machine without issue even though it's target temp is WAY hotter than the 2 fan design ones.

You don't have a choice, its done in seconds after the OS is installed and no it doesn't interfere with anything because windows installs the last WHQL version of the nvidia drivers not some random one.
 
Also everyone keeps ignoring the obvious facts that simply by making a temp change everything works, how can that be the drivers fault or the OS or the BIOS or the rest of the computer in any way shape or form (other than the psu, which is brand new and way more powerful than required and another separate computer all together.)

My end results was after benching the card like a maniac for 2 days the black screen issue just went away by itself.
 
Still the temp target thing did seem to help until I got to the point when the card just worked.

I can't speak to your experiences but the two cards I was was insanely quiet and cool. The fans didn't even turn on until about 1minute into benchmarking and when they did come on the fan speed was at like 40% to keep the cards at target temp.
So calling the acx cards loud and hot is just nonsense, also I did have two cards but their serial numbers was only 1 apart so I figured it must be from the same batch.
 
Anyway I'm just kinda glad to see there is others with similar issues, maybe there will be a driver/firmware fix for it. When I initially encountered the problem I couldn't find a single post anywhere with the same issue.

so this issue is commonplace? i've recently bought the 970 (ftw version) and i keep getting blackscreens and restarts when my card heats up for extended amounts of time. What exactly did you do to solve it?

Issue still persists, and although i'm unsure i think it took one minute longer than usual for screen to black out.

lipe123
 
Install precision X and set temp target to 70C, see if that helps.

I was having similar issues with my GTX 970 SSC ACX 2.0 but only during overclocking. Unlinking the Target Temp and setting to 70C helped but didn't solve the issue completely. After hours of benching Unigine Valley and slowing incrementing the GPU clock only, I got to the point where anything over+70 would crash the driver during Unigine Valley bench. It was't until I enabled K-Boost that I could extend well beyond +70. After researching K-boost, it made perfect sense this would solve the driver crashing issue.
 
Enabling K-Boost allows the GPU's clock speeds and GPU voltage to run at their highest set frequencies all of the time. K-Boost disables all of the power saving features that current video cards use to aid with lower temperatures and lower power consumption. The major benefit of K-Boost is to help remove the sudden drops in voltage and clock speeds that happen dynamically when stressing video cards. Removing these can reduce performance inconsistencies and instabilities when overclocking.
 
So in my case, when K-boost was disabled the voltage and core would spike when starting Unigine Valley. This large swing from zero or base clock speeds would cause the driver to crash. By enabling K-boost the swing is much less allowing the driver and GPU to "ease" into the OC settings.
 
Final OC settings (Stable):
110% Power Target
72C GPU Temp Target (Thanks lipe123)
+88 GPU Clock Offset (I was able to push +100 but found it to be very unstable)
+455 Memory Clock Offset (I was able to push to +610 but lots of artifacts)
+37mv OverVoltage (I found raising the voltage would help stabilize the OC if the driver crashed)
K-Boost: enabled
 
I hope this helps.
 
 

Is their not a different support chain for resellers?
Also for a reseller that is putting together a computer with a ASUSTeK COMPUTER INC. MAXIMUS VII HERO Motherboard would you not also install the ASUS GTX 970 like the STRIX-GTX970-DC2OC-4GD5 Graphics Card?
I think mixing different brand components in a custom computer for a reseller is odd.
Unless your customs can ask for the components by Make and Model of what they want installed.
Still if your company needs to come here for support I hope you find your solution.
lipe123, Being that the Motherboard is from ASUS have you asked them the same question?

Hi Everyone,
 
I would first like to say welcome to the forums to some of the newer posters in this thread.
 
I can confirm that the screen blacking out can be caused by a number of different causes from system stability to heat (over 98C for the 970) and we are here to help. Please call or email our support using the information included below and we will do our best to assist you with finding a solution.
 
 Please keep in mind that you may email EVGA support at Support@EVGA.com or call support 24 hours at 1-888-881-3842 option 1, if support is needed.