Video Hardware Error/Power Supply

Yefo · 28 Apr 2013

Another BSOD, temperature was at 82 degrees, not sure about fan speed (although I did hear/see it running).

80 and 82 degrees were the recorded temperatures at the time of crash for my two BSODs today. That seems pretty light for an overheating issue. But then again...

Any comments on the fact that those temperatures are well within safe operating temperatures for the GTX 590? Perhaps the card is hotter than the reading claims it is...? Or, maybe this issue doesn't have anything to do with overheating?

Note: Neither this BSOD or the early BSOD from today logged any files in minidump... not sure why.

Layback Bear · 28 Apr 2013

What case (tower)are you using? 80c is to hot.

Yefo · 28 Apr 2013

I am using the Corsair Graphite 600T Gaming Case (seen here Amazon.com: Corsair Graphite Series Black 600T Mid-Tower Computer Case (CC600TM): Electronics).

According to NVIDIA (source: GeForce GTX 590 | GeForce) the maximum safe temperature for the GTX 590 is 97 degrees. Also, through less scientific sources, many users experience low 80s when under full load on the GTX 590, so it does not sound that uncommon.

While I can't completely rule out overheating, the way I see the evidence, this does not seem to be the issue.

Edit: I'm currently going to try to keep my game settings/framerate at a level that keeps my card temperature lower, around 70 degrees, and see what happens... If I BSOD at 70 degrees, then I'm thoroughly convinced it is not due to overheating.

Yefo · 29 Apr 2013

The issue with GPU-Z not logging fan speed has been resolved. At the bottom of the GPU-Z program is a drop down to select which video card you want it to monitor. While I only have one video card, the GTX 590 has 2 cores, so there are 2 GTX 590 options to choose from. Long story short... one of them is able to retrieve the fan speed and one of them is not since the GTX 590 only has one fan.

x BlueRobot · 29 Apr 2013

The BSOD's which I have seen so far, have been Stop 0x116's, whereby the graphics card becomes "hung" and doesn't recover within the allocated time, here's a explanation to save typing: Timeout Detection and Recovery (TDR) (Windows Drivers)

Here's some of the usual causes for Stop 0x116's from the BSOD Index:

Code:

Usual causes:  Video driver, overheating, bad video card, BIOS, Power to card

Layback Bear · 29 Apr 2013

I would just like to mention the that 97c in my opinion is to hot.
Now I will get out of the way and watch.

Yefo · 29 Apr 2013

Layback Bear said:

I would just like to mention the that 97c in my opinion is to hot.
Now I will get out of the way and watch.

Haha, no need to get out of the way.

I'm just trying to figure out how this could be an overheating issue when I'm so far under the safe operating temperature as indicated by NVIDIA (getting BSODs at 80 degrees vs. the 97 degrees that NVIDIA claims is safe). Granted, 97 does seem pretty high, but I'm almost 20% below that anyway. Perhaps my card is actually hotter than the temperature monitor is indicating...? Or maybe my card simply has a hardware defect that causes it to shut down at lower temperatures than it is supposed to?

Or then again, maybe this is all just a PSU issue in disguise. Personally, I hope this is the case, because a new PSU is about 5x cheaper than a new GPU. Then again, maybe my computer is a complete POS and the PSU AND GPU are malfunctioning at the same time

The fact that these BSODs typically happen when the video card is under stress is concerning and seems to indicate possible overheating... but the temperature monitor just does not seem to indicate that that is the case.

essenbe · 29 Apr 2013

First, you are relying on the monitoring program to get it correctly. Then there is also the sensor on the card itself. Also, the max temp and the safe temp are 2 different things. 97°C is 206°F. I would not call that safe under any circumstances. Just to test it, why not take the side off the case and use a house fan to blow cooler air directly on the card and see if the temps are better and the BSODs improve. I really don't know about GPU monitoring programs, but I do know that CPU temp monitoring programs are rarely correct. They are usually 'in the ball park' but not an exact temp.

As far as the PSU, you can look at the log and see what the voltage was at the time of the BSOD. As I recall, it was pretty normal.

Yefo · 29 Apr 2013

After another BSOD today (at 84 degrees this time), I'm starting to think maybe these BSODs are indeed caused by overheating, regardless of the difference between the temperature monitor and the recorded safe operating temperature by NVIDIA.

One thing I've been noticing in my GPU-Z logs is that my fan is typically only running at ~50 capacity during these crashes. I've used MSI Afterburner to define a more aggressive fan speed curve. At 50 degrees, fan will run at 70%. At 60 degrees, 84%, at 70 degrees, 92%, and at 80 degrees, 100%.

Besides the extra noise, what are the consequences of these changes? Am I overworking my fan (is that even possible)?

I'm going to do some gaming similar to what I've been doing and see how much these fan changes adjust my operating temperature...

Edit: So much for that experiment... I instantly BSODed when trying to start a game. Crashed at 50 degrees this time with the fan at 68% capacity (about 2300 RPM). Now I'm just more confused...

Also, to be more descriptive... these are all black screens of death, and the last 4 or 5 have not even been logged in the minidump files. No idea why.

Edit2:
I tried this again with the same Fan speed curve... and again, nearly insta-BSOD when game started at around 50 degrees. I also tried a flat 85% and, sure enough, BSOD at 52 degrees.

I'm not exactly an expert on how the PSU interacts with the GPU, so maybe this theory does not hold water... But does the GPU draw more power when the fan speed increases? Maybe the card is not overheating at all, but when it reaches a certain temperature and the fan really tries to kick in, the PSU is incapable of supplying the necessary power to the card (and fan), and bam! BSOD. Thoughts on this theory?

essenbe · 29 Apr 2013

Disclaimer: I am not a video card expert, not even close.
Several things you can try:
Try the external fan and see if that makes a difference.

Lower the settings on the game and see if that helps. (I know you shouldn't have to with your card, but maybe)

The black screen sounds more like a driver or the game itself crashed. If either of those happened you should get a message, but not always. Mini dumps are usually only created when you have the typical blue screen with the writing and error code on them. Just make sure you are configured for the mini dumps Dump Files - Configure Windows to Create on BSOD

The definitave test to me: Do you have another computer or a friend that will help? Take your card out and put his card in your rig and your card in his rig. See what happens. If you think it may be the PSU, do the same. I would rather you find out for sure what it is before spending money.

Make sure your drivers are up to date and installed correctly. click on the 'Detailed Guide and read it.It is a work in progress, but very good. Ultimate 2013 Nvidia Driver Guide by MANIAC-VVV-

EDIT: Also make sure to keep an eye on CPU temps too.