Tech Help required

[Mithral_Demon]

Okay, first time I've truly asked others for assistance on this (don't know many who does this stuff).
This is about my computer and gaming issues.

I am experiencing an issue where my computer will shut off during gaming like Borderlands and Black and White 2, it leaves no error and turning back on is like normal with no "Safe Mode" options.
It usually happens within 5-10 minutes of gameplay. I've looked up "solutions" but it didn't help me much except my power supply and CPU/Processor is good, I am guessing my Graphics card.

Heres my specs:
OPERATING SYSTEM:WINDOWS 2.6.1.7600
CPU TYPE:AMD PHENOM(TM) 9550 QUAD-CORE PROCESSOR
CPU SPEED (GHZ):2.22 -2.20-
SYSTEM MEMORY (GB):5.999 -6gb RAM-
GRAPHICS CARD MODEL: ATI RADEON HD 4800 SERIES


The graphics card has memory of 512mb, and my CPU temp stays at 49C (Even tried with Borderlands, the temp doesn't rise much).
But I don't know how to tell of the power supply.
If I missed some things that could help, just say.

[Maes]

Actually, suspecting the PSU is the most logical step, as very few purely software or logic hardware errors can cause an immediate shutoff -usually you get "at least" a BSOD and a reboot.

If you get an immediate shutoff instead, then it's most likely the PSU protecting itself from overheating/overcurrent caused by the GPU sucking up too much current (as testified by the fact that the problem only occurs with graphics-intensive games).

Posting some specs of the PSU (brand, nominal power, actual maximum deliverable power, maximum current on 12V rails etc.) would help. If you have a second one laying around (equal or better in specs than the first one) swap them and run some tests.

Otherwise, the only way to see if your PSU is crapping out is to attact a voltmeter to any 12V output (e.g molex connectors) and see it real-time while you let a game or 3D benchmark run . Don't trust mobo voltmeters on this, they are usually biased/imprecise. Did you change your graphics card recently?

TL;DR version: throw away whatever noname POS PSU you have in there and get a real PSU like an Enermax, Zalman, Heroichi, Thermaltake, Tagan or Coolermaster etc. and be prepared to venture in three-digit $$$ land for a decent one. Don't cheapen out and get the noname $10 "600W" garbage!

[Mithral_Demon]

Maes said:

Actually, suspecting the PSU is the most logical step, as very few purely software or logic hardware errors can cause an immediate shutoff -usually you get "at least" a BSOD and a reboot.

If you get an immediate shutoff instead, then it's most likely the PSU protecting itself from overheating/overcurrent caused by the GPU sucking up too much current (as testified by the fact that the problem only occurs with graphics-intensive games).

It's an immediate shut off, no BSOD, and no reboot, just shut off.
But I guess it might be my cooling system not working well if it's protecting itself.

Did you change your graphics card recently?

Nope, I've had this computer for nearly a year now, same set-up since my birthday in November. It played Black and White just fine back then (after having a mishap with monitor plugins), but then last week this happens.

How can I tell about the PSU? If it's the motherboard, still, how?

[Maes]

Mithral_Demon said:

How can I tell about the PSU? If it's the motherboard, still, how?

Only by monitoring its voltage under a controlled load -for which you'd need PC tech equipment and test benches.

Usually even we didn't bother - we just swapped in another "known good" PSU, fired it up, and it it DIDN'T crash with the "known good one", then the other one became "known bad one" and we had solved the problem :-p

The bad one we either trashed it or used it on a system with known lower power requirements.

It's really a simple "pass" - "no pass" test, but you need a second PSU to test against.

If you have a voltmeter you can keep it hooked to one of the hard disk/cd-rom Molex power connectors (monitor red and black wires) and see the voltage: you PSU is bad if the 12V rail:

1. Is significantly under 12V to begin with. ATX tolerances are 12V +/- 5%.
   
2. Fluctuates. It must NOT fluctuate randomly with an idling system.
   
3. Drops under load.
   
4. ...or even INCREASES under load.
   

Mind ye, you have to be looking at the voltmeter ALL THE TIME, and even then you might not catch the critical moment: it may all happen too fast.

If any of these happens, you have bad components in your PSU that can't regulate voltage properly for a variety of reasons - and whatever condition they cause, it's detected by your PSUs self-protection circuitry which spares you from a costly shower of sparks.

Now if you have a degree in electronics, you can try opening it up and inspecting the PSU's fan, caps, and transistors. But even I'd just get another PSU, especially if it was for my main system.

Ninja edit: it could be the motherboard too, but it's rarer. I still have a semi-fried SOcket 754 motherboard with PCI and AGR slots which one day just shut down for no good reason. From then on, it locked up as soon as "heavy graphics" were being used over the PCI bus. I could still use PCI cards and PCI Express/AGP ones if I just used apps and surfed, but the slightest DirectDraw/Direct3D/OpenGL would lock the system with all AGP and PCIex cards I had. Probably some PCIex rails gor fried...solution? Used my older SOcket 754 mobo. No burned stuff there, and it all went back to normal.

[RestlessRodent]

It could be overheating, the CPU is not the only thing that will cause a shut down when it gets too hot, there's the Northbridge for example.

Depending on the BIOS you can activate an alarm when the temperature gets to a certain point. The temperature only has to rise to the treshold to shut off and you won't notice it when it does.

Is it hotter than usual since November?

[Mithral_Demon]

Yes, and a problem I had that caused some issue was dust, and that solved much of my issue.
So if all else, the game works fine for now, but if it does it again I will need an extra fan for cooling.

Back at Nov it was pretty much new, at like norm temp of 20C-25C, but when I saw issue it was at 40C-55C which is prob why. Even though these are CPU temps, it shows it is over-heating in general.

Thanks for the help, I may know some stuff, but PSU and all that fun stuff I need to know. ^-^;

[Enjay]

Do you live in the Northern hemisphere? If so, your house may have been colder in November than it is now.

[Devalaous]

I have a similar problem with my laptop, which comes and goes. It used to always just turn off, now sometimes it will rapidly turn off then back on or randomly minimise.

[boris]

Maes said:

Only by monitoring its voltage under a controlled load -for which you'd need PC tech equipment and test benches.

Looking at some mother board monitoring software may also help. A couple years ago I had a PSU with a fan that sometimes simply stopped. Then when the PSU slowly overheated the voltage dropped, which I could see at the CPU voltage in the monitoring software.
When that happened I could simply crawl behind my computer and use the piece of paper (don't use something metallic) to push the fan, to get it running again. I eventually changed the PSU, though ;)

[GreyGhost]

Being a cheap bastard I usually open up the PSU and replace the fan.

[Maes]

boris said:

Looking at some mother board monitoring software may also help.

I stopped trusting these a long while ago. They usually report very different values depending on which monitoring software you're using, and may even change ranges between two BIOS updates. The only non-broken way to be sure is a voltmeter.

If you really HAVE to use software monitoring, then the ones built in into the BIOS are the most trustworty, usually. I detected a couple of mobos that had fluctuations in the Vcore voltage due to bad caps this way.

[Mithral_Demon]

Enjay said:

Do you live in the Northern hemisphere? If so, your house may have been colder in November than it is now.

Yes, being in Washington like Danarchy, September to Mid-February it is colder then everywhere else (well... Except this year).
Now it's warm... Feels to much like summer then spring spite our normal rainfall.

[Bank]

Not the same problem per se but whilst on the subject my computer will shut off intermittently whenever I'm watching a video on youtube or playing a game. I'll receive a blue screen for about a second and then the computer will restart, I've been receiving STATUS B.A.D. messages for months now about my hard drive but I haven't had the time or money to fix it because it's causing what is essentially an inconvenience and not a catastrophic failure. Is that likely the problem or could it be something else?

[Maes]

Bank said:

I've been receiving STATUS B.A.D. messages for months now about my hard drive but I haven't had the time or money to fix it because it's causing what is essentially an inconvenience and not a catastrophic failure. Is that likely the problem or could it be something else?

FFS, are you serious? O_o

If you have any data of value on your HD, backup it before it completely dies on you. Also, use this nice little tool to generate an online report about your HD's S.M.A.R.T. status (which is probably somewhere in the area of "your HD is dying" and "Ahhh ya bubba, your HD is rubba!") to convince you that it needs attention.

It's not something that can be "repaired" while the HD is still functioning, unless it's just a controller problem and you can swap it for an undamaged one.

That being said, I have a HD that fails its SMART tests and is unsuitable for use as a boot/OS disk, but works OK in an external enclosure.

[Bank]

Maes said:

FFS, are you serious? O_o

If you have any data of value on your HD, backup it before it completely dies on you. Also, use this nice little tool to generate an online report about your HD's S.M.A.R.T. status (which is probably somewhere in the area of "your HD is dying" and "Ahhh ya bubba, your HD is rubba!") to convince you that it needs attention.

I understand it needs attention, anything of value is always backed up on my external, but I suppose my question was whether or not that was what was causing the shutdowns.

[Maes]

Yeah, a broken controller can cause load surges that the PSU interprets as short circuits and thus shuts downs to prevent damage. Of course, only a double blind test can give full proof (aka swap in a known good hard disk...)

[fraggle]

Perhaps you should upgrade?

[Jodwin]

Mithral_Demon said:

Yes, and a problem I had that caused some issue was dust, and that solved much of my issue...
...Back at Nov it was pretty much new, at like norm temp of 20C-25C, but when I saw issue it was at 40C-55C which is prob why.

If it gets that dusty in such relatively short time, either your room is really dirty, your computer is full of holes or there's something else fishy going on. I almost never clean my computer because I'm too lazy, and it never suffered of drastic dust-induced heating. If your computer case has a lot of holes in the front or side for air intake (likely if it's a cheap one), maybe you can try installing some DIY filters to block some of the dust. But I'd first recommend cleaning more often and putting the computer somewhere else than in the corner of the room.

[Belial]

Or he keeps his case open like I do.

My GPUs huge radiator gathered enough dust in a year to go from 60C to 80+ under stress.

[Mithral_Demon]

Jodwin said:

If it gets that dusty in such relatively short time, either your room is really dirty, your computer is full of holes or there's something else fishy going on. I almost never clean my computer because I'm too lazy, and it never suffered of drastic dust-induced heating. If your computer case has a lot of holes in the front or side for air intake (likely if it's a cheap one), maybe you can try installing some DIY filters to block some of the dust. But I'd first recommend cleaning more often and putting the computer somewhere else than in the corner of the room.

It kinda does have alot of holes, but i'm not worried. And yes, I do need to clean it alittle more, it's going to be 2 times a month (once every two weeks).
But the holes are sides and back, front has "for show" holes, show no purpose. x-x
I'm guess out of everything that makes up my computer the case is cheapest which is probably why too.

[RestlessRodent]

fraggle said:

Perhaps you should upgrade?

I can't afford a 386!

[GreyGhost]

Forget about Windows (it's bound to be just a short-lived fad) and get yourself a copy of GeoWorks, it'll run on everything! :P