Sign in with
Sign up | Sign in
Your question
Solved

Nvidia gtx 280 crashes computer while playing games

Last response: in Graphics & Displays
Share
August 28, 2012 3:25:21 PM

Alright, so this just started happening 2 days ago. Whenever I go into a game of any kind, the game will crash eventually with a screen of weird colored repeating pixels everywhere. The mouse turns into a square version of that about 1 inch wide, sometimes I can hear the sound and sometimes I cannot. With graphic intensive games the game crashes right upon logging into the world (I may see character for 1 or 2 seconds). This is Guild Wars 2.

Here are my specs.

AMD PHenom(tm) 9850 Quad-Core Processor 2.50 GHz
4GB of RAM
Windows 64 bit operating system
GTX 280 Nvidia
500W PSU

My temperatures overall seem fine (never goes above 70C). I always keep my fan on 100% no matter the card because I like keeping my card from getting anywhere near dangerous temperatures.
a c 289 U Graphics card
August 28, 2012 3:27:54 PM

Can you take a picture with your cell phone of those repeating pixels?
August 28, 2012 3:28:33 PM

I took one but have no idea how to upload it xD. I hooked up my phone but dunno how to get the pics.
Related resources
August 28, 2012 3:29:31 PM

I also just recently before making this post put my Core Voltage on 1075 just incase. I run 600 Core Clock (MHz) and my Memory Clock (MHz) at 1107. Right now with fan at 100% my GPU is at 42-44C. Usually when gaming with 100% fan it goes up to 60C, I have never seen 80C.
August 28, 2012 3:38:30 PM

It did look a little bit like that except with more pixels, doing GPU-Z now with Furmark.
a c 164 U Graphics card
August 28, 2012 3:38:58 PM

First off, restore the entire system to stock setings to give yourself a leval playing field.
Do the pixels ook like any of these?:

http://www.playtool.com/pages/artifacts/artifacts.html

You're right to suspect temperatures, if you have not done so already use GPU-Z to monitor the GPU and Realtemp for the CPU-both will show maximum and minimums (cilck on the reading boxes in GPU-Z to cycle through min, max and average).
a c 289 U Graphics card
August 28, 2012 3:41:04 PM

Do you have the pictures on your PC? You can upload them through www.imageshack.us
August 28, 2012 3:47:17 PM

I have my phone hooked to my computer via charger but when I go to the phone it only has a file named DevLogo on it. But this GPU-Z stuff is confusing me.
August 28, 2012 4:10:12 PM

For some reason lowering the voltage did it, I can now play without crashing, don't delete forum topic, might have problem again.

Best solution

a c 172 U Graphics card
August 28, 2012 8:12:32 PM
Share

You need to rebuild your card, the inductors used on these cards are known to go bad after regular use due to their design. These cards were not produced to last but mainly for the casual but very big buyers that usually only use their cards for a few months to maybe a year before upgrading.

Nvidia went with the ferrite shielded inductors are physically small while being able to handle high loads but not for sustained use and certainly not for years. The problem is that they are not able to handle the high current loads from the power vrm to the gpu so they will start failing. At first it is only some annoying coil sequel or buzz. Later but usually noticed by those who overclock is that the max stable clocks gradually decline and for cards that have a lot of use it even goes below stock clocks. Once the insulation inside begins to fail the inductors become very weak until they short out completely. When they do they break the circuit so no current can pass through to the gpu or memory. It increases the load on the remaining vrm phases and quicken their failure until the card is no longer usable.

You can tell if the inductor is bad because when they short they sometimes break up and only hanging by only the leads soldered on the board. You can shake the card and if you hear any thing like a baby rattler or a rattle snake's tail then one or more is already bad. Those that haven't burned out yet but are weak they are harder to spot.

The easy way to fix the card is not to reflow unlike older cards but to replace the old inductors. The standard coil type with plastic shell is very common and the card will work with R45 up to R60 rated inductors. You can use R50 inductors from salvaged board if you want but they won't fit under the stock cooler so you will have to run wires.
September 4, 2012 10:39:38 PM

Best answer selected by infinity467.
January 7, 2013 4:04:51 AM

infinity467 said:
It did look a little bit like that except with more pixels, doing GPU-Z now with Furmark.

I know this is a bit old, but i'm having the same exact problem as you did. I wanted to know how you lowered the current to fix it.
October 19, 2013 12:34:31 PM

nforce4max said:
You need to rebuild your card, the inductors used on these cards are known to go bad after regular use due to their design. These cards were not produced to last but mainly for the casual but very big buyers that usually only use their cards for a few months to maybe a year before upgrading.

Nvidia went with the ferrite shielded inductors are physically small while being able to handle high loads but not for sustained use and certainly not for years. The problem is that they are not able to handle the high current loads from the power vrm to the gpu so they will start failing. At first it is only some annoying coil sequel or buzz. Later but usually noticed by those who overclock is that the max stable clocks gradually decline and for cards that have a lot of use it even goes below stock clocks. Once the insulation inside begins to fail the inductors become very weak until they short out completely. When they do they break the circuit so no current can pass through to the gpu or memory. It increases the load on the remaining vrm phases and quicken their failure until the card is no longer usable.

You can tell if the inductor is bad because when they short they sometimes break up and only hanging by only the leads soldered on the board. You can shake the card and if you hear any thing like a baby rattler or a rattle snake's tail then one or more is already bad. Those that haven't burned out yet but are weak they are harder to spot.

The easy way to fix the card is not to reflow unlike older cards but to replace the old inductors. The standard coil type with plastic shell is very common and the card will work with R45 up to R60 rated inductors. You can use R50 inductors from salvaged board if you want but they won't fit under the stock cooler so you will have to run wires.



I´m having the exact same problem with my gtx 280.

Now i can´t even enter on windows.

I read nforce4max answer and i would like if possible to obtain more information about those inductors.

Thanks you
!