My 4850 asplode

krameriffic

2[H]4U
Joined
Mar 1, 2006
Messages
3,214
So I was playing a little bit of TF2 earlier today like I normally do on my Visiontek 4850 @ 700/1100 with an Accelero S1 cooler, the sinks on all of the RAM modules and sinks on the VRMs, and suddenly my system turns off. When I tried to turn it back on, everything powered on just fine, but the red warning LEDs were flashing on the back of the 4850 and there was no video output to my monitor. I figured "I'm fucked, I hacksawed my stock heatsink off for the VRM coolers, they'll never accept this RMA."

I proceeded to take it out and investigate to see if there were any telltale signs of electrical shorting or any kind of burns anywhere on the card; there were none. Nothing else in my system seemed to be messed up, and I booted fine with video working on another card. So I removed the S1 to take a closer look at everything, especially the core and I still noticed nothing obvious that would have bricked the card. There appeared to be no shorts, and it certainly wasn't overheating (none of my temps break 60C in GPUZ while running Furmark).

I ended up cleaning off the stock TIM that came on the S1 and reapplying with AS5 like I normally would. I put the card back in and booted up; the warning LEDs went away and I am getting video just fine out of it. Temperatures are roughly what they were before with the stock MX-2, and I didn't change anything about the way I have anything mounted. I played about an hour of TF2 since then without incident.

My question is this: what do you think happened? It seemed like a 100% electrical breakdown rather than an issue with thermals, but nothing I did had any effect at all on the electrical aspects of the card. It seems to be fixed, but I'm concerned about it happening again. Please post your thoughts.
 
Power glitch maybe? Got a good power surge protector?
 
A bump in the grid man. Take it for whats it worth - might be better to bring the card back down the stock settings.
 
Surge protector is good, PSU is a brand new Corsair HX520W and I'm not running a very power hungry system. The concern I had was the flashing red warning lights. I'm not sure if different flashing patterns indicate different problems or not though.

The overclock seems totally stable, no issues with it since I set it about a week ago. I've played hours and hours of TF2, all the way through Crysis at max settings. It's been totally stable except this.
 
well you are overclocking so that could have caused it oh and an RMA wouldn't be right would it if the card did die from you adventure beyond the warranty.

 
If the card still tests stable then I would just forget it happened and go on with my business unless you have reason to doubt your PSU.
 
If the card still tests stable then I would just forget it happened and go on with my business unless you have reason to doubt your PSU.

This is what I'm thinking. It's been stable for about an hour on Furmark, and as I said I played all the way through Crysis using high settings with the CCC config at this overclock/cooling configuration with no incidents. I have an 8800GT sitting here that I can use if this thing bites the dust again anyway.
 
I think a few people have told you about all your going to know without it happening again. power surge or just an overheat of the card.

1. Make sure you have good ventilation in the case, this can cause more problems then most people outside of [H] really know.
2. If your really concerned about heat, lap the heatsink on the gpu. I have spent hours lapping heatsinks for both CPU and GPUs. Just a hint, womens nail files work really well on final polishing.
3. If you can get ahold of a PSU tester, test it. Stability issues normally come down to dirty power and heat. Especially with overclocking.
4. Last but not least, make sure your getting good airflow acrossed your south bridge. This air will not only cool off your south bridge heatsink but will also provide nice cool air going out the back of the case and normally give your vid card some extra cooling.
 
Back
Top