Sign in with
Sign up | Sign in
Your question
Solved

PC Reboots During 3D Games!!!

Last response: in Systems
Share
November 4, 2010 12:52:39 AM

Started about a week ago (after reloading Windows 7... Not so much reboots, it actually powers of completely and then powers back on (I know because I have LED side fans which power off as well.). Before I go any further let me post my PC details:

CPU: C2D E8400 @ 3.00GHz
Ram: 2x 2GB Kingston DDR2 800
Mobo: MSI P43 Neo-F
GFX: MSI GTX 260 (216) TwinFrozr OC
HDDs: 250GB & 320GB Seagate
Optical Drives: 2x LG DVD-Writers
PSU: Aopen 700watt
OS: Windows 7 Professional 64-Bit

These PC cutouts are completely random and only happen during 3D intensive games. Ive taken my entire PC apart and rebuilt it, and nothing helps. It could happen within 5 minutes or it could be as long as an hour before the PC cuts out.

I'm praying it's not my PSU, because it's just out of warranty, in fact I'm praying it's nothing hardware related, since I can't afford to replace anything right now. The thing is, the PC works 100% as long as I don't run any 3D Games.

Heres what I've tried so far:

1. Updating my graphics card drivers
2. Rebuilding and dusting out my PC
3. Loading BIOS defaults
Monitored Temps: All Fine, CPU: Stays Below 40C Graphics Stays Below 65C

I don't have any other components I can try, so my last resort is to reload Windows again.

Any advice that doesn't require additional hardware, or the most likely problem would be
much appreciated. Thanks.

More about : reboots games

a c 103 B Homebuilt system
November 4, 2010 1:25:15 AM

If there is no BSOD involved then it is most likely power issues. 3d games put more load on the PSU.
m
0
l
November 4, 2010 3:57:21 AM

Try opening the side of the case and use a fan to blow air into mother board and see if that works. If computer is overclock, set it to default. Download CCleaner and try to cleanup registry issue.
m
0
l
Related resources
November 4, 2010 9:26:58 AM

It sounds power related try disconnecting both DVDRWs, 1 HDD and any USB devices then see if it still happens
m
0
l
November 4, 2010 10:07:47 AM

Thanks for the replies, I'm also thinking that it's either a power issue or a faulty graphics card. If I could rule out which one that would be great. I'm going to try disconnecting all non-essential components and see what happens.

I'm also hoping it's a software or driver issue, but I doubt that would cause the PC to cut out completely, a software issue is more likely just to cause a soft reset right?

Also, the only OC in my PC is the GTX 260 which is factory OC'd
m
0
l
November 4, 2010 10:35:40 AM

Ok, so I disconnected all non essential components, (4x Fans, 1x Hard Drive & 1x DVD-Writer), had to leave the other one plugged in as my games require the DVD to play. Anyway, I switched off the PSU at the back before disconnecting anything, powered up the PC started a game and it cut out in less than 5 minutes!

The problem is I still can't rule out anything. Signs are pointing to a failing cap in the PSU (eg. the longer the PC has been powered on, the longer the gaming interval before the PC cuts out.), but that still doesn't rule out the possibility of a failing graphics card.

I have an oldish Antec chassis so removing the PSU means removing the entire motherboard first. I want to rule out all other possibilities first.
m
0
l
a c 122 B Homebuilt system
November 4, 2010 11:22:14 AM

You can sort of check the PSU. Try to borrow a known good PSU of around 550 - 600 watts. That will power just about any system with a single GPU. If you cannot do that, use a DMM to measure the voltages. Measure between the colored wires and either chassis ground or the black wires. Yellow wires should be 12 volts. Red wires: +5 volts, orange wires: +3.3 volts, blue wire : -12 volts, violet wire: 5 volts always on. Tolerances are +/- 5% except for the -12 volts which is +/- 10%.

The gray wire is really important. It should go from 0 to +5 volts when you turn the PSU on with the case switch. CPU needs this signal to boot.

The problem, if it is power, will be the 12 volt output. You can check it by measuring between the yellow and black wires on one of the IDE drive power connectors.

You can verify that the video card works by testing it in another nVidia system. You won't need to remove ATI drivers and install nVidia drivers that way.
m
0
l
November 4, 2010 5:19:05 PM

Totempole said:
Signs are pointing to a failing cap in the PSU (eg. the longer the PC has been powered on, the longer the gaming interval before the PC cuts out.) ...
That's not a sign of failing cap. Capacitor is charge and depending on capacity it discharge instantly , unlike batteries.
jsc said:
You can sort of check the PSU. Try to borrow a known good PSU of around 550 - 600 watts. That will power just about any system with a single GPU. If you cannot do that, use a DMM to measure the voltages. Measure between the colored wires and either chassis ground or the black wires. Yellow wires should be 12 volts. Red wires: +5 volts, orange wires: +3.3 volts, blue wire : -12 volts, violet wire: 5 volts always on. Tolerances are +/- 5% except for the -12 volts which is +/- 10%.

That is hard to measure also when power is trip like that, the power level will be ok before it cut off, which come back right away. You'll need an oscilloscope with that record the spectrum data for reviewing.

Things you should also check are the molex terminal pins and the connection joint its connecting. I had the same issue with one of my computer and I had to re-weld the pins of motherboard power connectors. May be it's a cold solder issue, may be the pins are loss. Try wiggling the power wires and see if your pc sut of or reboot.

Another thing to consider is like you said, a software problem causes the game to shut down the PC. Try reinstalling everything from scratch and install game after drivers and updates then play the game. If you still have problem it'll be more likely a hard issue.
m
0
l
November 4, 2010 11:13:59 PM

Ok, so I've tried reformatting my PC, and loading nothing but my Graphics Drivers (260.99). I loaded 3D Mark 2006 Ran a test and "boom" my PC cuts out.
I have no special tools or alternative components I can try.
So, I need to be sure of what to replace, since I gonna have to go out and buy it.


I ran a memory stress test on the card and it seems fine (no errors).
Ran Memtest86 to check my Physical Memory (Passed)
Also loaded Fail-safe defaults in bios (no help).
Disconnected all non-essential components. (nothing)

I'm currently downloading older drivers (197.45), perhaps that's the problem.
Used my brother's monitor, keyboard and mouse (just in case)

So, I'm 99% sure the problem is somewhere between my GFX card and my PSU.
The problem is which one is it?

Also, what are your opinions on furmark? Dangerous or worth a try?
Also, how long can it be safely run for? From what I've read, not longer than 16-minutes. Also, what are the best settings to use, I'm using an LCD TV with a max res of 1280x768, so I'm limited there.
m
0
l
a c 103 B Homebuilt system
November 4, 2010 11:28:30 PM

Have any nice friends that could lend you a GPU or PSU to test?
m
0
l
November 4, 2010 11:40:30 PM

Nope, most of my friends are running cheap generic 450 watt PSU's which would be completely fried by my PC. Also, I could get my hands on a Radeon 4650 Graphics card, which would probably work flawlessly, but won't tell me much since it uses hardly any power.
m
0
l
November 5, 2010 12:25:23 AM

Ok, I thought I'd update you guys (who've been very helpful, thanks.) about what's happening before I get some (much needed) sleep. I uninstalled the 260.99 drivers to load the 197.45 drivers.

Before I loaded the new drivers, I ran 3D Mark (Using just the crappy Microsoft drivers built into Windows 7. Guess what? 1 hour and no reboot!!!

This is very good news, but it doesn't solve the problem. I'm gonna load the 197.45 drivers, try them out tomorrow and see what happens. Wish me luck!

The drivers I was using previously were 258.96 and 260.99 both of which caused my PC to cut out.

The thing is, even if this does solve the problem, why can't the card run properly on the 2XX.XX drivers? Could it be a bug? Could it still be a power issue?

One thing I did notice is that the Microsoft integrated drivers don't have PhysX included with them, maybe that's the reason the PC's not cutting out.

Maybe, maybe, maybe... It's all that's been going through my head today.

Only time will tell.

Tune in tomorrow for an exciting update of............

"THE INTERMITTENT FAULT"
m
0
l
November 5, 2010 2:25:54 PM

So today I tried out the 197.45 drivers. Again 3D ran flawlessly. (I even put it on a 10x loop to be sure). But, gaming is still a problem. I'm only running 2 games at the moment, (Quantum of Solace & Racedriver GRID). Racedriver GRID's Crashes are completely random, but Quantum of Solace now seems to only crash at a specific point in the game.

I think I may have ruled out th PSU as being a problem, because I can put as much strain on the PC as I like, it seems to crash when rendering something specific. (A sore spot on the graphics card, so to speak.)

Anything I could use to diagnose the problem (or confirm that there is a problem) with the GTX 260? I have no additional parts or alternative PC's I could test the Graphics card in, so I'm pretty limited here.
m
0
l
November 5, 2010 2:30:42 PM

Your hardware isn't the issue anymore. Problem rises between games software and drivers. Your OS may have a problem with the install the first time. If you can play one game work for long period... there's no hardware issue. Only hardware issue you'll have if the game you want to play requires more than what hardware you have.
m
0
l
November 5, 2010 11:18:46 PM

Nope, apparently hardware is the issue. I've just realised that the Quantum of Solace's cut outs are also pretty random.
Something I should have tried from the beginning was reduce the graphics settings. I dropped all detail settings to the lowest and my PC no longer cuts out!

I normally run my games with detail and AA etc. V-Sync on maxed out and run my res @ 1280x768 due to the limitation of the LCD TV. However, I've had no problem doing this in the past, so something has definitely gone wrong.

I back to thinking faulty PSU, but is it at all possible it could be the graphics card?
I mean, the games I play are pretty old (2005-2008), so there shouldn't be any reason for the PC to Cut Out at High Graphics settings.
m
0
l
a c 103 B Homebuilt system
November 5, 2010 11:34:16 PM

With out testing I can not say for sure it is your PSU but I think that is more likely than the GPU. If it was the GPU artifacting and other graphical distortion would be more likely.
m
0
l
November 6, 2010 3:50:54 AM

I still think its software than hardware. Try setting the GPU to its factory settings and try playing the games to its original settings and see if it works ok, if not, under clock the GPU to the reference card. If your computer play selected programs without problems then it's not the PC. It doesn't matter that other programs play from it before. The only way you can be sure is to install the OS and drivers and make a clone or back up the image. Then install each game separately and see if the same game gives you the problem.

Could it be your GPU heating up and causing the system failure or a connection problem? This goes back to my first suggestion… home electric fan blowing inside your Pc and wiggling the PSU wires. Right now, I think you can disregard the PSU as a suspect if some games (as you mentioned) paly without problems.
m
0
l
November 11, 2010 7:16:43 PM

Hi again, I know it's been a while, but I still have the same problem. I now know for a fact that it's hardware related since I've now gone back to my old Windows XP and it does exactly the same thing during games. I'm 99% sure the problem is between my PSU and my Graphics card.

Also, I'm about 60% sure it's my Graphics card, and I'll explain why below.

I decided to use GPU-Z in the background to get a full report on temps, usage etc. before my PC cuts out.

First I ran Quantum of Solice, maxed out. My PC lasted about 10 minutes before cutting out. GPU Temp was about 45C and memory usage was at 270MB, GPU Usage was at 41%

Next I ran 3D Mark 06's (3D benchmarks only) on a constant loop, it lasted 45 minutes! GPU Temp was about 48C and memory usage was at 330mb and GPU Usage was at 99%!!!

I'm thinking that if this were a power issue, it would have cut out much sooner on 3D Mark then on Quantum of Solace, wouldn't it?

Temps are most definitely not an issue as my CPU runs at 22C Idle and 36C Load since I applied new thermal paste. I've never seen my GPU exceed 56C either.

I am not an OC'er, but my GTX 260 is factory OC'd as mentioned in the beginning, so my PC is technically running at factory defaults.

I'm taking my parts to test in a friends machine this weekend to try and find the culprit. Any advice in the interim would be much appreciated. Thanks.
m
0
l
a b B Homebuilt system
November 11, 2010 7:42:54 PM

One other software based test you can run it the OCCT - GPU and PSU tester. You can go here to get a copy and use it.
m
0
l
November 13, 2010 8:20:18 PM

Thanks suteck, I've been looking for a program like that one. I ran the PSU test, and my PC lasted about 8 minutes the first try and less than 3 minutes on the second attempt! One thing I did notice, that is cause for concern, is that it shows the 12v as 10.12v :o !

My PSU is supposed to be a quad-rail, which puts out 15A per 12v rail, or a total of 50A combined.

It also says something about total output should not exceed 70A.

In my Bios the 12v shows a stable 12.232v, but OCCT (and speedfan) shows only 10.12v.

I ran another (very quick) PSU test to show the graph of the 12v rail supplied by OCCT.
Everything else is fine, it's just the 12v reading.



So can I be pretty sure that my PSU is failing? If so, I'm just gonna buy a cheaper 600watt. My current one cost me almost 100 GBP! And it's stuffed after 2 years!
m
0
l

Best solution

a b B Homebuilt system
November 14, 2010 12:27:53 AM

Sorry but yes, it sounds like your PSU is causing your problems, When you tax your video cards like that they suck the juice out of a power supply. Even though they have they're own 2-6 pin power connectors it still requires power from the X16 lane and that source is supplying other parts of the motherboard all coming off one rail. And if both OCCT and speedfan agree on the voltage it sounds like a sure thing.

I will gravely warn you though that getting another cheap PSU is gonna put you in the same boat soon enough. Aopen might be ok for the just surfin the net and the facebook crowd but if when you're really gaming the PSU is THE most important part. I would go with a BFG, Corsair, OCZ or coolermaster, something with a big name and proven performance, (not to mention an 80+ rating as a must). Spend the money now or pay the price later. I have been in discussions where people have lost video cards and hard drives due to bad/cheap PSU's so don't just count that as a peripheral, It is a major integral part of a high performance system!
Share
November 14, 2010 10:13:12 AM

By cheaper, I meant Cooler Master. They are really cheap here in South Africa (About Half the price of my 700 watt AOPEN). I have heard complaints that they don't live up to their specifications though. Anyways, the plan is to fix it, sell it and go I5 or I7.
m
0
l
November 15, 2010 12:01:38 PM

Hey guys,

I'm having the same problem with my GTX280. I have a Xilence 800W gaming edition PSU.
After 4 minutes of GPU stress test with OCCT my computer does a hard reboot, the same during heavy game play.
I noticed that when playing games, it reboots only when there's a lot of action going on (GTA4).
On the other hand, when playing Test Drive Unlimited (pretty old, I know) the problem also occurs. But in this game the average frame rate is about 120fps, so my graphics card is definetly overkill for this game, but the problem still occurs.
Like Totempole, I also noticed strange things on my 12V lane. Both OCCT and the BIOS say that +12V=0.35V at all time, wether in idle or under stress.
I must say that I did use a 2 x molex to 8 pin convertor to power the graphics card (the original 8 pin connector from the PSU has a shorter wire), but this setup has always worked for me.
I'm planning to switch the 8 pin connector from the PSU to my graphics card and the 8 pin with the convertor to my motherboard.

Other than this, I have no clue what it could be. Nothing in my system is overclocked. Hopefully someone can help...

My system specs:
MSI P45 Platinum Motherboard
Core 2 Quad Q9300 @ 2.5Ghz
8Gb DDR2 RAM
2 x 250Gb Western Digital in RAID 0
m
0
l
November 16, 2010 12:24:58 PM

Hi JVW, my problem was solved yesterday. I bought a new PSU and... no more cutouts!!! OCCT's 12v readings should be ignored since my new PSU still shows only 10.21v on the 12v rail. My Bios shows 12.408V though. At any rate, If your +12v rail was really only running at 0.35v, your PC wouldn't power on!

I'm pretty sure my problem was that one (or both) of my 2 6-pin connectors was failing (maybe a dry-joint somewhere). My guess is that your PSU also has a problem somewhere, most high-end PSU's have short-circuit protection which will hard reset your machine if something goes wrong.

Try lowering the graphics settings of the games you play, if your PC takes longer to cut out, then it's most likely your PSU.

I know all to well that this is easier said than done, but try bugging a friend who might have a PSU strong enough for you to try (600w Minimum for your machine I would say).

I gave my old PSU to my brother (who is pleased as punch), he has a Radeon 4650 (no 6-pin required) and it works perfectly for him (he needed a PSU with more molex connectors to power 3x Cooler Master Musketeers).

Anyway, I wish you the best of luck, It can get really frustrating at times! You think you've solved the problem, then "boom", your PC cuts out on you again! :fou: 
m
0
l
November 16, 2010 12:33:11 PM

Ok, I'm gonna end off this post by saying thanks to everyone who helped me through this, it's a pity I can only select one best answer, as there are a few good ones. I'm gonna have to go with suteck's answer though, since OCCT's PSU test was the deciding factor in the replacement of my PSU. So an extra big thanks to suteck. It took me almost 2 weeks, but my problem is FINALLY SOLVED.
m
0
l
November 16, 2010 12:35:41 PM

Best answer selected by Totempole.
m
0
l
November 16, 2010 4:08:01 PM

Thanks Totempole, I'm going to lower the settings of my games and see what happens. Sadly that will be a job for tomorow, no time today. I have found someone with a 600W PSU I can use for testing and who I can give my GTX280 for testing, incase its not my PSU. Hope I can play again next week (NFS Hot Persuit is coming out the end of this week).

Thanks for the help!
m
0
l
November 17, 2010 7:29:41 PM

JVW said:
Thanks Totempole, I'm going to lower the settings of my games and see what happens. Sadly that will be a job for tomorow, no time today. I have found someone with a 600W PSU I can use for testing and who I can give my GTX280 for testing, incase its not my PSU. Hope I can play again next week (NFS Hot Persuit is coming out the end of this week).

Thanks for the help!


NFS is already out i think ;) 
m
0
l
!