I am very upset. I had the problem on this thread [ http://www.tomshardware.co.uk/forum/page-267280_15_0.ht... ] merely weeks ago. To summarize my 8800 GTX gave up on me and start displaying artifacts all over.
I bought a MSI 4870 1GB with some cash effort. Yesterday after coming home from a short vacation I boot up windows, look around a bit then get some lines and squares on the screen and the system locks. I reboot and the screen seemed okay but didn't even made it more than 3 seconds before some artifacts appeared on the screen and system locked again. Now 2 of 3 boots is a complete non boot (black screen) and the other boot is the bios system info text quickly followed by system lock and black screen. I've tried removing the optical drive, sound card and the tv tuner to check if the psu isn't providing enough power. The problem continues.
What is the mater with all of this people? I want to be 100% it's the card's fault before sending it back and waiting several months for a replacement (IF that ever happens) to be sent back. Thanks for your help.
In neither thread did you tell us what PSU your using. Might want to list out your system specs. It sounds like you have another dead card. To have your old one die on you and then the new one as well must be incredable odds. If it did happen, I'd start looking at your temps and voltages.
You are really having no luck are you, mate?
Take the side off the case and check the CPU fan is working: Boot and enter the BIOS, and set the fan there to maximum or, if that cannot be done, just leave it running in the 'pc health' or monitoring screen so you can check the CPU temps and fan speed.
If the CPU gets over 60C try flicking the fan with a pencil, if that fails, power down and investigate further.
I have an ASUS A-55GA (550 W) PSU;
Full system specs: Asus Commando motherboard, E6600 oc'd @ 3.0GHZ with Tuniq Tower cooler, 4 x 1 GB Kingston DDR2 800 memories, 1 x 400GB Western Digital HDD, Optical drive, old creative sound card.
The 4870 used to run at around 70 *C;
Apparently I only have about 3 seconds after enterning the bios to move around before it locks and bring the black screen. If i try to boot again right after it doesn't boot at all. It appears that waiting for a few minutes before retrying to boot gives me again the 3 second chance.
coozie, yeah I'm completely unlucky. Also The Tuniq Tower is huge and I can't see the fan because of the huge radiator. Should I really get on messing with it? Can the CPU cause the artifacts before the lock as described in my initial post?
LE: Somehow I managed to unscrew some bolts and pull up the cooler from the huge radiator and try a quick boot with the cooler in my hand and it did work.
New development: I've just received a 8600 GT from a friend to try out and my computer boots and works. It's definately the videocard. The question is?
Is it something with my computer that is ruining video cards? Or am I just extremely unlucky?
yes, check that Power Supply, it seems that it is shorting your video cards... maybe it's the PCI-E connector because the two cards that are having issues require the 6-pin connector and that 8600 does not...
coozie, the psu is an ASUS A-55GA
Now I really fear the shop I bought this from online will void my warranty
That means I have to replace the psu when and IF I get a replacement to be on the safe side.. oh crap
I have seen much worse than that in a PSU review!
It might be the powersupply causing the problems, but I am not entirely convinced, still it would only be sensible to get it tested before splashing out on a replacement you may not need. Phone around and see what, if any charges are going to be incurred and offset them against the cost of replacement:
It says it has 2 18amp "rails". I think the 4870 has 1 - 6 pin connector. If you use the "Y" connector you can keep swapping out 4 pin molex connectors until you get 1 from each "rail' and maybe you'll get lucky and it will work. haha. For it to work right I'd look for a power supply that has 34-36amps on a single rail.
Yea, my old 420 watt enermax has 18 amps per rail. Being that overrated brings concerns right away, and a 4870 is too much maybe. Seeing these problems, most likely.
I think you can still return the card, just dont buy your new psu from them heheh
I have also tried my 4870 on the friend's computer before he gave me his spare 8600 GT. He had an OCZ 700 W PSU powering a GTX 260 so if it was power related it should have run well in there. It's dead. I'm just hoping not to have problems with the warranty. It's enough having to pay to ship it back and forth and ordering a new psu.
So it is an almost sure fact, that a new psu will prevent me from having video card problems again?
I have replaced the psu with the corsair. Been using it along with the 4870 replacement for about a month now, working alright.
The interesting thing I am experiencing though lately is that I get at least 1 "video driver has stopped responding and has been recovered...." message everyday. This is a symptom I was experiencing few weeks before my initial 8800 gtx died. :-S
I also occasionally (rarely) had system lockups which required reboot but I always though that is from my unstable overclocking CPU and ram.
Can this be a sign to anything?
I swear to God I'm gonna put a fist through my monitor if my videocard dies again.
Are you still overclocking? What are the temps of the 4870?
Yes I am still overclocking the CPU. I have been running it OC'd ever since I remember. (about 2 years with my first, 8800GTX)
Did a 5 minute test since it's 4am playing WoW and alt-tabbing to GPU-Z and it doesn't seem to go past 65*C
I'll do some further monitoring tomorrow if you consider 5 minute irrelevant.
65c should be fine, it might get a bit hotter but I was looking for more like 90+. General rule says if you have problems, go back to stock and try again. I'd either reset to stock and see if it continues, or start questioning if your 100% sure the OC is stable. This means more then 5min of Prime, more like overnight. I actually prefer all day while your at work seeing as the temp will be hotter.
I've ran this afternoon Prime for about 2 hours and there were no errors reported. When I'll get the chance I'll try running it over night. However even if there's a slight instability with the CPU oc which can only be detected in a 12hour cpu stressing test, do you think this can cause direct video card issues?
Today, the screen turned black and white whilst in-game and the computer locked completely so I had to reboot. I was pretty nervous thinking that the monitor will stay black on my next boot but it didn't.