tiger025

Often when I play a 3D game for ~30 mins or longer, The game image just freezes on the screen, and I have to restart my PC. The amount of time after which it freezes is not consistant. Sometiems I can play for 4 hours, sometimes after 30 mins it freezes. Recently I have been getting some "artifacts" (video) after playing HL2 and Guild Wars, for a while. So the next played a game, I had a program open to monitor the temp of my mobo, and I had nvidias control panel open to monitor the temp of my Vid Card. I played for a few hours, and the cpu temp was fine, 48C most of the time, and the GPU was fine as well, 65C most of the time. The only thing that was was troubling that my "NB fan speed" slowed down to like 900rpm at one point, and some alarm went off. However its been fine for a few days now, and its gone back to ~2500 rpm.

I just need some advice as to what is the most likely to be wrong here.

I have a 2.8 P4 [Northwood] 800mhz FSB, No OC
4 X 256MB ddr400 some budget ram.
evga 6800nu 128MB [all stock hsf etc, i dont know much about chageing them]

I know a bit about computers, but even so when my computer was overheating a while ago, I dinnt feel comfortable installing a new hsf over my processor, so I took it to Comp usa and had them do it. ( I just dont know how to put on the hsf and compund or something) And it Has worked fine since. Anyway, if I need to get a new vid card to fix this, thats no problem for me. I'd be happy getting a 6800GT or ultra if I have the $$. However, if it looks like it maybe a Mobo/CPu Fan HSF problem as well, I think I will want CompUsa to try and fix it or something.

please let me know what you think. Also i have thing CD i made to check my RAM. So I think I will do that again just to rule out the ram factor. But I think that the fact Im having artifacts, means my Vid card is shot or something?

Sarkast

Well you dont have the artifacts all the time so i am not sure if the card is really damaged.
Take a close look at the cards heatsink and make sure it touches the GPU and all the memory chips. I know the temperature reported look fine but there obviously arent sensor everywhere on the card - and they often cant be trusted anyways. Maybe one of the memory chips doesnt have contact and overheats - then produces the artifacts.

Dont change (or take off) the heatsink though if you still have warranty on the card - as this usually voids it.

Just to see if there might be an overheating problem somewhere take a big housefan and let it blow into the open case - play your games and see if the situation changes.

Have you tried slowing down your AGP to 4x and turning fastwrites off in the BIOS ? Unlikely this will help but its worth a try.

Last - is your power supply sufficient enough for your components ? Maybe after a certain ammount of time of running under full load it doesnt supply enough or clean enough power to your components anymore - and creates these problems.

tiger025

OK, ill take a look at the card, and maybe post a pic. What are am I looking at? The side with the fan on it, or the other side?

In terms of my PSU, its a 420 Watt PSU, but its just some random probably budget psu, so It might be part of the problem. I posted a screenshot of my Mobo Monitor and the some said the "rails" (which are varying between 11.84v -11.90v) may mean the PSU is having problems.

Sarkast

When it comes to the heatsink just make sure its sitting flat on the card and there is no visible gap between the memory chips or the GPU. You should be able to see this if you look from the top/ side and bottom of the card. Dont try to wiggle it or pull it off. I dont think i'd be able to see much on a picture because the gap could be 1mm - so pretty impossible to see unless you have the card right in your hands and 10cm away from your eyes.

What is the supposed output in Ampere (A) on the +12V rail ? You find this information on the PSU label.

tiger025

i took out my card, and there doesnt seem to be a gap between the hsf and the actual GPU chip its self, the main chip in the middle of the card. However the card is a little bent, and there are some little gaps on the sides, but I dont know if that matters. The Ampere is suppose to be 13.0A on the +12V rail.

Also since I reseated my vid card, because I was checking it out like you said, I also updated my drivers, and I havent had a freeze or artifact problem since, After 3 hours or playing or so. But that dosent mean it still wont mess up. I'll let you know if it messes up again. Do you think the driver update could have fixed it?

Sarkast

Yes keep me updated. I am pretty sure eventually this problem will return. I tell you why - 13Amps is WAY low for your systems specs anda 6800 series card. Recommended minimum for stabil operation is at least 18 - better more.

Driver updates are there to fix problems - so it would sound like a possible reason for the problems to go away - but i am quite sure your main problem here is the insufficient power supply.

tiger025

ah, i dinnt know that. Well in that case, why dont u recommend a good psu for me to get. I just have a stardard ATX Raidmax668 case. Is a PSU easy to install? Can I just unscrew old on, take it out, put in new one, and re attach all the power wires? Ill look inside my case later on, but It think I should be able to do it. Ill probably get the PSU off newegg. Gimme a good brand or a link to a good one. 500W or something i guess. Enough for this to not be a problem.

Sarkast

These are some candidates i'd recommend. I dont know your budget but i wouldnt go any lower (wattage) - of course you can always get a model with more output.

http://www.newegg.com/Product/Produc...82E16817103936

http://www.newegg.com/Product/Produc...82E16817103927

http://www.newegg.com/Product/Produc...82E16817103459

These have the advantage of 2x 12V rails - so one can power your CPU (Nr.1 major power consumer on the +12V rail) and the other your video card (Nr.2). This gives you exceptional stability.

PSUs are very easy to install. You unplug all connections, unscrew the old one - take it out, put the new one in and plug everything back in.

tiger025

$ isnt a issue. I just dont want to waste money on something better than what I need. However It's possible I may get a 6800 Ultra that need the 2 molex connectors. So I want something that could handel that.

How about this one?

http://www.newegg.com/Product/Produc...82E16817103928

tiger025

also, I have a average amount of stuff running off it, expect I do have 3 optical drives.

tiger025

Forget abou that one one new egg. I arftifacts have returned and im ordering it from amazon.com for the good shipping.

How is this one:


http://www.amazon.com/exec/obidos/AS...nnorswebguidec


I thinks its ok but I cant tell if It has the 2x rails.

This is the manual for it though, its on the page. or this link may work

http://www.amazon.com/exec/obidos/re...&append-uid=no

Sarkast

Unfortunalety this one doesnt have seperate 12V rails. Aslong as it has enough output power on the +12V rail it'll still work but you just get better stability and more upgrade options with the seperate rails.
Considering the ammount you're gonna spend you should probably try to get one with 2x12V rails - but its obviously your decission not mine. Either way this is also a very good quality PSU.
What you need to look for is ATX 2.01 standard - that means the 2 seperate rails (and a 24 pin connector for boards that need it).

tiger025

I can't seem to find one on amazon that I can tell has 2x rails. And all the antec ones they offer seem to use that same instrution manual, so I guess they dont have the dual rails either. Are there certain brands that have the 2x rails, or a certain model or something that I could search for? Please Let know if you find one on amazon.com that has the 2x rails. Also, I'm just trying to find once that I can get here fast, one that uses the amazon prime thing. I may just use that one I posted before, but if amazon offers one with amazon prime I will buy that one.

Sarkast

Here is a list of all Antec PSUs:

http://www.antec.com/us/pro_powerSupply.html

The one with the 2.0 are the ones that support dual 12V rails.

I'll take a look at amazon now - but have never used their page for anything besides books so it might take a while. lol

tiger025

well if you want to take a look at amazon its probably easist to use:

http://www.amazon.com/exec/obidos/re...f=pd_ir_pc_nav

I checked and they dont sell any of the 2.0's by antec though. And since ive already orderd that one I just posted i'm probably gonna just use that one, if it gets here fast. I mean it should be adquete right? I dont really attach that much stuff to it.

Anyway, Thanks for all of your help, and Ill let you know if my computer still messes up with the new PSU.

Sarkast

Sorry i couldnt get back to you sooner but i was kinda busy but also spent quite some time searching amazon for the "PSU of your dreams". lol Well you are right they dont sell the dual rail ones - except one Enermax model but even that wasnt in stock.

Either way - the one you ordered is a nice model and is defenitely strong enough to power your systems specs. Please let me know how things go once you have it installed.

tiger025

no probm, with any luck, i should have it tomarow. And thanks for all the help.

tiger025

Well I got the PSU and installed it, and my games dont seem to have the problem, so far, which is after a few hours of testing. BUT... I had a very weird message come up when I was capturing some DV video useing adobe Premire...






- i know I installed a PSU, but i mean there is no drivers, or software or bios settings that need to be changed is there?

The first time, this error appeared out of now where, and I had to restart. Then 3 times, within about 10 minutes, adobe premire just restarted out of the blue... and I was like

So far this whole random restarting thing is limited to adobe premire, and it seems to random within adobe premire, BUT it always restarts when I export a small movie 40MB>save it>and say I dont want to imped project info. As soon as I hit dont impeed project info... Comp restarts...
I dont know if that is important but I thought you should know.

Anyway do you have any idea what might be causing this? Is there some wire I may have connected the wrong way or something?

tiger025

Well, this may be ulrelated... But my core center alarm went off again telling me the "nbspeed" is in the red zone i guess.




And then i check it a few minutes later and it was at 0rpm



Maybe the sensors are off or something... I really have no idea what is going now now....

Sarkast

Well you are right - this cant have to do with the PSU as there are no drivers for it. Keep monitoring and of course do more testing to see if you have any errors or crashes in other applications.

This error is (like so many other windows errors) dreaded cause it can have so many reasons and can be hard to track down. This page is pretty comprehensive:

http://www.updatexp.com/stop-messages.html

I'll read my way through some of the stuff when i have some more time later.

Give this program a try though please:

www.memtest86.com

Let it run for 10 loops or overnight. I am wondering if there might be a RAM issue.

EDIT: Didnt see the post of the NB untill i posted. Well if it reports 0 then take a look at the board and see if it actually spins. Its not unusual for sensors to report wrong readings or no readings at all sometimes.

EDIT2: Was there a certain file reported together with the error ? If not look into your event viewer to see if there is a certain driver file associated with it that would help to figure out whats going on.