GTX 960 Issues with crashing to black screen (no signal) / randomly won't work at boot up?

HeadphonesOn · Dec 14, 2015

Hey, I was wondering about a problem I've been recently having with my card after installing the latest drivers from GeForce Experience. Whenever I go into a game (like 2-5 minutes of being in a game, tested on League of Legends) , or sometimes when I surf the web, my display will go black, and say that there's no signal being transmitted and permanently stay there until I reboot my computer (no audio either). I tried testing out some stuff while rebooting my computer a lot, and at one point my GPU refused to work after a bootup. Sometimes the computer would just randomly shut down if I tried it too much. Right now there is display on my monitor, but I'm not sure if that will be the case for long. I'm not sure if it's a driver issue, or a PSU issue, or something else entirely. I built this thing around 6 months ago without any problems (except for one driver related issue that hit me about a month ago). Now I can't even go into a game without crashing.

Things I've tried with no success:

1) Rolling back driver version to the previous I was using. (from 359.06 to 359.00)
2) Re-seating the GPU and cleaning the inside of the comp.
3) Trying new HDMI cables.
4) Playing on low and windowed settings in games.

Specs:
Operating System
Windows 8.1 64-bit
CPU
Intel Core i5 4590 @ 3.30GHz 32 °C
Haswell 22nm Technology
RAM
8.00GB Dual-Channel DDR3 @ 799MHz (11-11-11-28)
Motherboard
ASRock B85 Killer (CPUSocket) 32 °C
Graphics
Acer H236HL (1920x1080@60Hz)
2047MB NVIDIA GeForce GTX 960 (Gigabyte) 32 °C
Storage
465GB Western Digital WDC WD5000KS-75MNB0 (SATA) 42 °C
931GB Western Digital WDC WD10EZEX-00BN5A0 (SATA) 32 °C
14GB SMI USB DISK USB Device (USB)
PSU
EVGA 500 B1 80+ BRONZE, 500W Continuous Power 100-B1-0500-KR

Connected Via HDMI

Some of the errors I picked up from Event Viewer are:

- The Service KMSELDI service terminated unexpectedly. It has done this 1 time(s).
- Display driver nvlddmkm stopped responding and has successfully recovered.
- The driver \Driver\WudfRd failed to load for the device SWD\SensorsAndLocationEnum\LPSensorSWDevice.

Eximo · Dec 14, 2015

Might try a memtestx86 run through to verify the system memory is good.

After that I would suspect the entry level power supply. Verify the voltages are within acceptable ranges under load. 11.4-12.6 volt for 12V, 4.75 - 5.25 for 5 volt.

HeadphonesOn · Dec 14, 2015

Eximo :

I've checked the memtest. I let it run for 2 passes and no errors occured. As for the voltage I'm not entirely sure about the 5v reading. These are the voltage readings that I could read from SpeedFan:

3vsb - 3.47V
Vbat - 0V
VCore - 0.89V
12V - 12.04V (would sometimes drop to 11.8 - 11.9 volts under stress).
AVcc - 3.38V
3Vcc - 3.38V
VIN - 0.89V
VIN2 - 1.69V
VIN3 - 0.86V

I've also extensively played games on this PC before for 4-8 hours at a time before this whole crashing situation began if that helps at all.

Eximo · Dec 15, 2015

As long as all the temperatures are good, that 12V should be fine. Sadly most power supplies do not report their internal temperature. It could be the power supply that is overheating.

Not sure what monitoring software you used, 5V is almost always included as it is one of the most basic voltages, particularly for the BIOS and USB. The BIOS should have its own monitoring page to look at if you want to verify that 5V is in range.

Most concerning is that 3.47volt on the 3.3 standby is actually slightly out of spec. 3.65 is the specified maximum. 5V and 3.3 volt are usually output on the same bus. It could be that your 5 volt is a little high as well which could lead to instability. More expensive power supplies put out almost exclusively 12V and then use DC-DC converters to supply 3.3V and 5V, I don't think that is the case here.

HeadphonesOn · Dec 17, 2015

I checked the PSU from the BIOS, and the +5v was good. It was at around 4.902V when I checked it from there (I'm not sure how I would test it under load from the BIOS). I did actually manage to fix the problem (for now at least), and sadly it was the Nvidia Driver and I think me leaving the computer on for a little too long (It's been almost 10 days since I last shut it off.) I rolled back to another stable version (358.5) and there seems to be no issues with my GPU crashing or my comp doing so either. I'm going to keep it on for a while again just to see how the voltage fluctuates and see if there is a problem with my PSU.

Eximo · Dec 17, 2015

Just noticed my typo: 3.65 should be 3.465

That is good news, kind of. Haven't had any issues with the latest Nvidia drivers myself. Might look into latest chipset drivers and audio drivers, might just have a minor conflict going on.

HeadphonesOn · Dec 18, 2015

Ok, so reporting back after investigating for a little bit. I decided to run into the BIOS after I had yet another driver failure with my card. My voltages for my 12V were reading low, around 11.49 to 11.58 and it would fluctuate around there. I've also kept this computer on for the past few days almost, not shutting it down completely. I'm not sure if it's bad practice that lead to it or if my PSU is just bad, but I'm not sure if I should RMA it. I left it off for about 2 hours and now my voltage is reading at 11.8 again. I have no clue what's going on.

Eximo · Dec 18, 2015

That could very well be the temperature of the power supply causing variation which is not a good sign. Many cheaper power supplies have poorly rated 60C or more commonly 80C capacitors. Quality power supplies generally have 105C rated caps from reputable companies. A lot of Chinese knock-off parts that only get good ratings under ideal conditions.

EVGA 500 B1 80+ BRONZE - This is not a highly regarded power supply. Usually the minimum recommendation is at least EVGA's B2 line of products. The power supply itself is only rated for up to 40C, which is way below what the average PC generates, very likely that your long running times are making this evident.

Before wasting time on an RMA/refund, is there anyway you can borrow a power supply from a friend to test? Or perhaps purchase one temporarily (though I still don't like recommending that) Last resort would be some type of computer repair shop that might let you test or test for you for a reasonable fee.

I've not tried out an EVGA return before, though I understand they are very good about GPUs, not so sure on the PSU division (they don't actually make the thing, so they may follow the policy of their OEM)

HeadphonesOn · Dec 18, 2015

I can't really think of anyone who would be comfortable with me taking their PSU out of their build and putting it into my build to test it out. I'm a real novice when it comes to computer building and had to have a friend help me wire up the PSU (which is why I'm so adamant on trying to save the PSU.) Is there any way I could save the PSU to work towards later January? If not I do think I have the receipt and the original boxing somewhere in my home for the PSU unit that I bought from Microcenter which I can attempt to get a refund for. Would you have any recommendations for PSUs around the same wattage (or more) and price range? I was thinking of something like this but I'm not completely sure about what is makes a reliable PSU or not.

Eximo · Dec 18, 2015

When it comes to power supplies there are few resources that you can use. It is mostly people who bother to dissasemble them and run tests to actually verify the claims they make on the packaging:

http://www.tomshardware.com/forum/id-2547993/psu-tier-list.html

http://www.tomshardware.com/reviews/power-supply-psu-brands,3762.html

Generally I also see if hardwarelabs or JohnnyGuru has a review up for a particular product.

To get yours to limp along, I would say try to force more air through it, undervolt/downclock the CPU and GPU to take a little load off (maybe disable boost on both as a minimum)

Power supply installation isn't that complicated, all the cords are keyed and you would just be doing a like for like swap with 4 screws about maybe 10-12 connectors. Plenty of Youtube videos out there to demonstrate.

mjpoulsen · Dec 19, 2015

Before blaming the PSU to much, you should know that the latest GeForce drivers have been known to have these full color screen lockups on Windows 10 atleast.
I have a Asus GTX 770 4GB where i need to force reboot the PC 3 - [?] times a day, because of this issue.
I would try one of the drivers from september, maybe 355.98

HeadphonesOn · Dec 19, 2015

Yeah I tried looking and thinking it was the drivers but the PSU hasn't been working with me much lately. I left it on for a night again and the computer wouldn't respond or wake when I tried to wake it from sleep. Plus the voltages have been really erratic. Either way, I'm gonna check on drivers too cause it might be a double issue. In the meanwhile I'm just gonna use the computer sparingly until I can get myself a new PSU (hopefully a EVGA GS or G2) while I try to figure out how to underclock gpu and cpu safely.

Anyways thanks @Eximo for being patient with me and leading me towards the most likely cause of the problem. The help has been much appreciated and educated me a lot on the stuff. Peace man.

Search

GTX 960 Issues with crashing to black screen (no signal) / randomly won't work at boot up?

HeadphonesOn

Reputable

Eximo

Eximo

Titan

HeadphonesOn

Reputable

Eximo

Titan

HeadphonesOn

Reputable

Eximo

Titan

HeadphonesOn

Reputable

Eximo

Titan

HeadphonesOn

Reputable

Eximo

Titan

mjpoulsen

Reputable

HeadphonesOn

Reputable

TRENDING THREADS

Latest posts

Moderators online

Share this page