Sign in with
Sign up | Sign in
Your question

BSOD or CTD after several minutes of gaming....

Tags:
  • Homebuilt
  • Blue Screen
  • Systems
Last response: in Systems
Share
October 11, 2007 7:42:54 PM

This post will probably be pretty long as the problem has "evolved" and it's necassary to understand this system's history to get a better idea of what the problem is. So here goes.

I built this system about 11 months ago. It started with these specs....

C2Duo +6400 w/ stock HS
2x1gb crucial DDR2
ASUS P5B-E Mobo
Gigabyte X1950 Pro PCI-E 256mb
Maxtor 80gb IDE
SB Audiby LS
Antec TruePower 430W & Sonata 2 Case
Windows XP Pro

The system ran great for several months with no problems whatsoever. One day I was messing around in the case, although I can't remember what I was doing....possibly installing or uninstalling my sound card. I hit the button to start the system and I realized that the HS wire got stuck in the HS fan which kept it from spinning. Knowing I had to shut the system down ASAP, I just pulled the power cord from the PSU.

After that, I had to hit the button 50-200 times while turning the PSU switch off and on for 15-30 minutes to get it to boot again. When it did boot, everything worked fine until we had a power outage and the PC lost power without shutting down. I would have to repeat hitting the button and turning the switch off and on for 15-30 minutes to get it to boot again. If I shut the system down normally however, it would have no problem rebooting.

A month or two went by and my ethernet card stopped working. It wasn't being seen by the system and I could not install the driver. At first, I thought I damaged my mobo so I RMA'd it and they sent me a new one (I know it was new and not the same one) with an updated BIOS. I installed it and I had the exact same problems as before.

Next thing I did was bought a new power supply (Antec TruePower Trio 430W, 16A on each of 3 rails) which fixed my boot up problem and my dead ethernet problem. However, a new, more annoying problem took it's place. After playing IL-2 or Oblivion for 5-30 minutes, the game would lockup with repeating sound. I was able to alt-ctrl-del to get back to desktop each time. STALKER would just crash to desktop which seems to be par for that game and GPU combination.

I figured the PSU wasn't enough so I upgraded to an Antec TruePower Trio 650W with 19A on each of 3 rails. I still had the same problem. However, if I opened the case and pointed my house fan inside, I could play about twice the amount of time before it locked up. My CPU temps were around 55 at full load so I figured my GPU was overheating. So I upgraded to a 8800 GTS 320 Mb which stays around 60 under full load.

The problem is, I'm still having lockups. So far IL-2 is the only game that runs perfectly, but it's not a very demanding engine. Oblivion will often cause an error when I exit out of it from in-game and it stutters a lot in-game especially when I'm running and originally it would C2D, but I played around with the settings and it hasn't been crashing since. World in Conflict BSOD's, locks up, or crashes to desktop with or without error. STALKER "still" randomly crashes to desktop, which seems to happen more or maybe exclusively when I'm running (as opposed to walking). I've also noticed AA doesn't seem to be working in STALKER regardless of where I have it set. I've seen two different BSOD errors from WIC, the first was concerning the NV4_disp.dll, the second was a PFN_LIST_CORRUPT.

Just for kicks, I tried overclocking my CPU to 2.20 (default is 2.13) and it BSOD from desktop shortly after bootup . This is the first time I've ever oc'd anything. Upon restart, I set it back to default.

I can run Prime95 with no problems although I need to try again since I installed the 8800. I can make it through the latest 3dmark benchmark with an 8700 score and no problems or stuttering. I'm using the latest drivers off the nvidia website.

I've ran checkdisk and driver verifier with no errors. I've ran memtest with no problems.

Things still on my list to try:

Play game without soundcard installed.
Run Prime95 for 30+ minutes while monitoring temps.
Reseat RAM.
Jump off roof.

Anyone have any other ideas?

The problem is that I have no idea which problems are related and which aren't. I don't know if my CPU could be damaged and still be capable of playing a high-end game for a period of time. I don't know if my RAM could be damaged and memtest86 still give it a green light.

Help!!

More about : bsod ctd minutes gaming

October 11, 2007 8:52:03 PM

UPDATE: Uh oh....

I just ran Prime95 again using the In-Place Large FFT's and Blend and now I'm getting errors. What gives?

Guess my next move is to try one stick of ram at a time?
October 11, 2007 9:39:54 PM

You seem to be on the right track. Try one stick of ram at a time in different slots. Also, try playing with the memory voltage a bit. Trying it in different slot could narrow it down to the motherboard of the memory. I have also found before where memory wanted more voltage than the motherboard was giving it.
Related resources
October 11, 2007 10:06:20 PM

I've read somewhere about possibly needing to up the voltage a bit but the mobo doesn't say what the default is so I don't know where to start. I know nothing about voltages and timings.... :( 

However, I am trying one stick at a time right now. One stick failed within 3 minutes by itself and I'm testing the 2nd right now in the same slot. It went through the Large test for about 20 minutes with no error and I'm running the Blend right now and still no error. I guess it could turn out to be the first stick and memtest just didn't catch it? I hope that's the problem anyway as Ram is easy and cheaper to replace than mobo, plus it gives me an excuse to buy another gig. :) 

I guess the final test will be to play wic with the good stick to see if it is stable.
October 11, 2007 10:19:01 PM

You are on the right track. After you narrow it down to what you think it is, leave that bad stick out and play the game. It will likely have low performance, but if it doesn't CTD or BSOD you found the problem.
October 11, 2007 11:01:33 PM

So much for that theory. I thought that stick had tested long enough and I went to play and it crashed after about 30 minutes. I re-ran prime95 on it and got an error within 3 minutes.

Is it possible that my cpu is causing these problems even though everything seems to be working ok for quite some time....and some games I don't have problems with?
October 12, 2007 12:39:29 AM

UPDATE: I switched the 1 stick to channel B and prime has been running for about an hour with no errors. Maybe it's channel A or just the first slot....more testing will tell the tale, but I'm out of time for today. :(  I'll keep updating.

I'm wondering though, if memtest86 completes and finds no errors, does that guarantee that the ram is ok? Or is there still a way for some bad sectors to "slip through the cracks" ?
October 12, 2007 2:16:56 AM

I know you say that you've run chkdsk with no errors but if you could somehow get your hands on another HDD then you might want to give that a shot.

At the very least you may want to move from IDE to a SATA2 drive but you may want to only do that if you can verify whether or not your 80GB is at fault.
October 12, 2007 12:51:58 PM

It is a relatively old HDD and has been a good one. I've had it for 3 or 4 years. However, my next HD will be a SATA because my mobo only has one IDE port. I have some reading up to do though because I don't know the difference between SATA and IDE. Doh!

Unfortunately I have no spare parts to play with. :(  But my HD wouldn't be causing an error in Prime95 would it?
October 12, 2007 1:36:02 PM

Stranger things have happened.

Without spare parts to try out different scenarios it will be difficult to say for sure which component might be at fault.
October 12, 2007 2:14:52 PM

What are the voltage specs for the ram? The DDR2 default is 1.8V but some more aggressively clocked rams require more voltage, even up to 2.2V. Try increasing it a bit and see if that helps

MEMTEST isn't really 100% accurate if you only run it for one pass, better to leave it running for several hours and something like 10-20 passes. If it doesnt find any errors after that then the ram is probably fine.

edit: maybe the cpu isnt getting enough voltage either, but start with the rams first.
October 12, 2007 2:52:59 PM

Ok Kari, I'll try that. I'll have to find the ram default first because my mobo only says default and doesn't give me a value so I don't know where to start. The custom values start at 1.8 and go upward from there. I tried 1.8 just for giggles but no improvement, and I didn't want to risk going higher than that before I knew what I was doing.

I'll also run memtest overnight tonight if I don't have any answers by then.

Thanks for your replies, I'll keep the post updated until I solve the problem.
October 12, 2007 3:41:58 PM

Only advice I have, after reading the PSU problems, is prolly to redo windows. I'd back up saved games and such, format the drive for a new fresh install. So that would rule out any data that may have been corrupted during the crashes you described.

I saw that the Ballistix memory does need 2.2v, but if your running the reg stuff, 1.8 should be enough. Running it at 2 volts should not really hurt it at all.

Also, try post CPU-Z link of your system. That can help others see some of the MB settings you got going.

Try using CoreTemp as well, and if you can run the game in a window so you can see the temps change as you play the game or let it run. RivaTuner's hardware monitor also is good in keeping track of GPU temps and logs it as well.

Anyways.. Good Luck on Trouble Shooting it. :D 
October 12, 2007 3:47:49 PM

Thanks Grimmy, what is CPU-Z?

I formatted my HD after I got the new mobo but not since the new PSU's or GPU.

I use riva to monitor GPU core and both CPU cores. Temps look great while in game. CPU maxes at 54, GPU maxes at 63'ish.

Thanks for the post!
October 12, 2007 3:50:31 PM

CPU-Z Program

Its a free app. You can validate it on their site, which they provide a link like this:

My E4400
a b B Homebuilt system
October 12, 2007 4:15:42 PM

CPU-Z v1.41 under the SPD tab will give you the memory voltage you are looking for.
October 12, 2007 4:43:55 PM

therschbach said:
Thanks Grimmy, what is CPU-Z?

I formatted my HD after I got the new mobo but not since the new PSU's or GPU.

I use riva to monitor GPU core and both CPU cores. Temps look great while in game. CPU maxes at 54, GPU maxes at 63'ish.

Thanks for the post!


Does Riva really show the CPU temps? Can't seem to find that option on my 2.05 setup. I can only get GPU and Ambient Temp for the video card.

Can you try using CoreTemp when running your games? It can do a txt log in the same folder you put it in.
October 12, 2007 9:19:55 PM

Yeah Grim, but you have to get the Core 2 Duo plugin. It'll then show CPU0 and CPU1 so you can see each core separately. Google will point you to it.

UPDATE: I let memtest86 chew on one stick for almost 2 hours. The test completed 4 times with no errors, but the game still crashes on this stick so I've ruled out ram and dimm slots, unless both dimm channels are bad, in which case I'd think I'd be having more problems than I do right now.

My next step is to get CPU-Z to find out my current cpu and ram voltages and upping them a bit to see if I increase stability. Thanks for the link and info Grimm and Evon!
October 12, 2007 11:42:23 PM

Well, I bumped up my voltages to ram and cpu a bit and so far it seems to be stable, but I haven't quite done enough testing to be 100% sure. I'll keep the thread updated. Thanks for the help guys.
October 12, 2007 11:44:52 PM

Does it fail with both sticks in? I once had a glitchy motherboard that caused similar problems. Also you might want to try a fresh install of GPU drivers. It's possible they became corrupt from the improper shut downs. Also, if it is crashing after stressing the system over a period of time, you might want to check your temps after playing before a crash.

Other than that, I would try a fresh install of the OS and drivers. Weird problem my friend, I have been in a similar situation before and know how you feel.
October 13, 2007 4:44:26 AM

therschbach said:
Yeah Grim, but you have to get the Core 2 Duo plugin. It'll then show CPU0 and CPU1 so you can see each core separately. Google will point you to it.

UPDATE: I let memtest86 chew on one stick for almost 2 hours. The test completed 4 times with no errors, but the game still crashes on this stick so I've ruled out ram and dimm slots, unless both dimm channels are bad, in which case I'd think I'd be having more problems than I do right now.

My next step is to get CPU-Z to find out my current cpu and ram voltages and upping them a bit to see if I increase stability. Thanks for the link and info Grimm and Evon!


Hmmmm. I found the plug-in for Riva. Not too sure if those temp reading should be trusted. The core temp readings off that plug-in for Riva actually about the same as Speed Fan (4.33). Usually Speed fan Core temps are 15C cooler then Core Temp and TAT (Thermal Analysis Tool).

Right now my bedroom temp is around 80F, readings as follows:

Riva - Core0 27C - Core1 30C
Speed Fan - Core0 27C - Core1 27C
CoreTemp - Core0 38C - Core1 42C

I usually have speed fan put the 8800 gpu temps on the taskbar, and CoreTemp for both core temps, to keep things organized. I usually hate using Riva's hardware Monitor since it won't minimize. :lol: 

Anyways... about Speedfan and Coretemp. The older CoreTemp actually had the same temps as SpeedFan. The newer version of it, started matching the TAT temps. I remember reading the fix log they had on their website, stating it was a driver conflict.

So what ever temps you had, if off the core sensors I'd say might have been off (54C + 15C= 69F). There is a good C2D Temp guide which explains more in detail, which you may have seen or not. There are some confusing things about the sensors for C2D systems. There is actually 3 sensors. One on each core, and one in between the 2 cores.

So... in my conclusion of thought, if your were at 69F, there's a chance it could have gone higher (when your sytem froze). 70C is consider hot, but should still run. The thing is, throttling may kick in, which may causing studdering. Not sure, but the thermal trip should be 85C, and the throttling will occur in the 70's. Not exactly sure what 70's number. :lol: 

Hope those thoughts may help your trouble shooting.

Edit:

Oh that C2D guide on temps is:

Core 2 Quad and Duo Temperature Guide

Not sure if you ever seen it. Looks like its been updated for Quad temps. Been awhile since I last saw it. :lol: 
October 13, 2007 8:21:41 PM

Thanks Grimm. The temps from riva seem to coincide with bios temps as far as idle goes....unfortunately, I can't go from full load to bios quick enough to check that temp. :( 

I still haven't had time to do a thourough test with WIC to see if it's still crashing since I upped the voltages, however I did notice my CPU running a degree or two cooler for some reason. I'll check back as soon as I find out if I'm stable.
October 14, 2007 3:51:30 AM

No problem. Just check out that link for the Temp Guide. More then likely, the temp you see in the bios uses the same sensor that is in between the cores, as well as riva, and perhaps speedfan.

CoreTemp and TAT use the sensors that are above each core.
October 15, 2007 2:16:44 AM

Fixed! I just played WiC for 3 or 4 hours with not even a stutter. SUCCESS!! Turns out it was just the voltages for either cpu or ram were too low. I don't know which but I left them both up as they probably needed to be higher anyway. THANKS EVERYONE!!!
!