I'm posting this under the PSU forum because i *think* that's what the problem is, but I'm not entirely sure. Was hoping someone else might be able to provide a little insight on a very confusing problem. Here's the story:
I've got a rig ( Core 2 Duo E6600, GT 768MB, 2GB DDR2, 850w Thermaltake PSU, Asus P5N32-E SLI Motherboard), and it was running fine until recently. I made no hardware changes or attempts at Overclocking.
So one day I go to turn it on and the video is... completely messed up -- strange artifacting everywhere, even at the post screens prior to loading into Windows (Vista Home prem). Those of you who remember the NES, remember when the games got a little screwy and you had weird video "bugs" everywhere? Yeah, looked just like that.
Tried initially doing a driver reinstall, no dice (it stopped booting into Windows, oddly). I was able to get into the OS thru CD and reformat the sucker, and it worked fine for about 2 hours, then the screwed up video returned. At this point, I presumed the GPU was shot and got a new one (this being an Radeon 4850 512mb). Uninstalled the nvidia drivers, pulled out the old card, popped in the 4850, installed the ATI drivers, good to go.
...However, here's where things get weird. Machine worked FINE. No problem, I was even able to play Empire: Total War, which is a fairly resource heavy game as a sort of testbed. Game ran fine, for several hours. I shut it off, go to bed.. get up the next day.. and..
Machine boots, goes thru its post, goes to the Windows loading screen and.. reboots. Over and over. I just can't get it to get into the OS. After a couple attempts I eventually try to get into the OS thru the CD again, and then it hangs and crashes while loading from disc. I try again, same. Finally I try to get into safe mode and i get the BSOD. I unfortunately didn't get the exact wording of the message, but it in essence told me that the machine was shutting down because the video card wasn't getting enough power. Really bizzare.
This leads me to my questions: First, obviously, is what the heck is wrong with this machine? Is the power supply bad? Is there a failure in the Mobo? Some greater issue? I'm thinking it's gotta be the PSU, but if anyone's ever seen anything similar to this (I sure haven't), I'd welcome the advice.
And the other questions.. is my initial 8800 dead? I'm thinking it is, but still, it was pricey, so I'm hoping it's might be salvagable. And the one I really can't figure out.. What happened to that 4850? Why would the machine work ok for one night and then fail the next morning? Did whatever's truely wrong with this machine (a short of some sort?) fry this 2nd card too?
Hi, right......disclaimer to all nvidiots (I waited in line and bought the nv3 so F off for this comment as its true) BUT.....
The entire 8 series of cards as well as some 9 series nvidia gpus are made faulty - the bump issue - google it its quite real and is causing a lot of users real issues now when its really starting to surface outside notebooks into desktops. Several friends stupid enough to buy nvidias rubbish are now replacing a lot of cards HOWEVER...
It is almost certainly your psu causing this problem now.....but it was almost certainly your nv gpu that caused it. When the gpu's fail on nv cards they draw a LOT more power (measurable easily) and subsequently can damage the motherboards 12v supply and definitly the psu's 12v rail.
Basically they overheat drawing a lot more power and premeturely age your components related to power - only about 30% of your gpu's power comes from the slot its plugged into nowadays with the rest being supplied by the rear power connector - this is the problem.
Bottom line - get a new psu before you possibly damage the new (reliable) ati card and pray the nv pile of crap hasnt damaged your board - if in doubt and you can afford it replace both. DO NOT buy a cheap psu though as you will require a hefty supply to competently drive that system for a few years so please make sure you do some research.
I will point out that Nvidia hid this manufacturing defect as long as they could and even when it became public said screw you to all their customers choosing to take a one time hit on profits and deal with "some" oems only on reimbursment. Their idea was to force bios / firmware updates to run fans faster to aid in cooling when the power draw increased hoping to hell their crap would make it past the warranty period.
Well there you go.....you got caught out in a nasty hole (**** happens) but there is always a way out - at least you now own an ATI card so can rest assured they dont cheap out on manufacture (cant afford to in their current position) so get a good psu and pray your mobo is ok.
Shame on Nv I will say as this is a possibly (accruraly) their last nail to drive them out of business. Not a lot of customers will replace their failing gpus with the same make now will they.....
**unless of course your psu was total crap to start with - but - the nv gpu issue has claimed a lot of pc's in this area alone so just replace the psu and hope**
oh for christ sake - you have an Asus made (argh) Nvidia based motherboard too - bin it - asus mobos are for top end tweakers only that can deal with the problems they cause and as for Nv they cant make a chipset to save their lives and each and every one has caused user problems especially when they overheat and or are overclocked.....jesus you bought the load on that build.
Just replace the board and psu now and look forward.....
Nvidia really ought to publicly apologise for their manufacturing issues (unless of course its 3rd party).
Anyway night and GET A GOOD PSU and replace that damn MOBO too.
Gelde.. I'm gonna agree with you on the Nvidia rant. That's the 2nd machine I've had over the years that's had one of their "high end" (for the time I bought it) cards, I believe the earlier was a 4400...Ti, maybe? I don't quite remember, that crapped out on me (though that one was the CPU cooking, if memory serves). Going from here on, I'm throwing my luck in with ATI. But what you explained makes a lot of sense on the time line of how my system went about failing.
For what it's worth I noticed while I was troubleshooting the thing, the 8800 was *really* hot, even just during the earlier described troubleshooting. I know those things get pretty warm, but to be hot enough to be uncomfortable holding after being powered on only a few minutes struck me as very concerning.
I had already returned the 4850 from where I bought it as it didn't fix the problem, but I've decided to just build a new machine from the ground up around an ATI card.
I'm gonna write both the mobo and PSU off on this machine just for safety's sake, and see if I can't salvage the RAM/HD/Optical drives from it to turn into a secondary machine with an inexpensive integrated graphics board and a decent PSU to replace the fairly ancient 2nd machine we have lying around now.
On a bit of a side note, vern, you might be onto something. Quite awhile ago (maybe 6 months or a year), we had reason to suspect either the house, or a powerline near by got strick by lightning, as over the span of a month of so after a particulary severe storm, a bunch of electronic devices started failing around here (a dishwasher, a radio, a tv). While I do have my PC plugged into a surge protector, now that you mention power surges, it *IS* plugged into the same socket as that tv was (though the tv was just directly plugged into the socket). I wonder if perhaps maybe the initial damage was done then and it just took awhile to give up the goat...
Hi again, yes ditch the mobo - nvidia chipsets are complete garbage - I didnt spot that in my original rant but please PLEASE get an intel chipset next - it might not support SLI but no one even remotly sensible cares about it anyway (not being nasty but).....as for the power line thing - get a good line cinditioner extention gang or maybe a ups - that will sort that out - but seriously - nvidia ALWAYS overvolt their entire line of products - I currently am responsible for a few nv based systems that have ALL required active cooling to remain STABLE in operation.
The people who boght then wont be replacing with nforce chipets I can assure you......as for the now totalling 20+ customers with failing nvidia gpu's - same / similar symptoms as your own - will again NOT be buying nvidia again. Its a shame as 10yrs ago I was championing nvidia as the next big thing.....nv3 (riva128) then the tnt - gf2 (dont count the beta gf1 that required mega cooling to run reliably) etc etc were all good for the time.....
But that said its painted a picture of a good company with excellent design team that cheap out at manufacture to boost profits. That is not a long term good bet.
Shame but what can you say - please do as you say and get a GOOD psu with at least 50amps on the 12v rail (combined or total) and for god sakes DO NOT BUY nv based motherboards.......